The Commissioners for HM Revenue and Customs v SDM European Transport Limited: [2015] UKUT 0625 (TCC) [2015] UKUT 0625 (TCC)

UPPER TRIBUNAL
TAX AND CHANCERY CHAMBER
[2015] UKUT 0625 (TCC)Case No UT/2014/0068
The Commissioners for HM Revenue and CustomsAppellantSDM European Transport LimitedRespondent
Judge Colin BishoppJudge Jonathan CannanDate 19 November 2015Category: Tax
[1]The respondent to this appeal, SDM European Transport Ltd (“SDM”), as its name implies, is a haulier undertaking the transport 5 of goods between the United Kingdom and continental Europe. It seems that it is a modest concern, with two directors, David Cranny and Martin Hodgkins, and is run from a small office in the garden of Mr Cranny’s house and from a yard in Bicester, where it keeps a small fleet of lorries. It also employs some drivers. When it is unable to 10 meet demand from its own resources it arranges for others—individuals trading on their own account or, in at least one case, another company employing drivers—to undertake the work on its behalf. In some cases the person or company so engaged sub-contracts the work. Some of the drivers of the consignments with which we are concerned in this appeal, although not employed 15 by SDM, worked exclusively, or almost exclusively, on its contracts, and in some cases they drove their own tractors towing SDM’s trailers.[2]We were told that the bulk of SDM’s work consists of the collection of frozen goods from warehouses on the continent and their transport to warehouses within the UK. Its business is not, however, confined to such work, and SDM 20 actively seeks contracts for the carriage of goods to the continent since otherwise the lorries sent to collect goods would travel there empty. The contracts with which we are concerned in this appeal are in that category. They differed from SDM’s principal work, however, in that they were contracts for the transport of goods liable to excise duty but in respect of which the obligation to pay that duty 25 had been suspended. Special rules, to which I shall come, apply to the transport of such goods.[3]Between July and November 2006 SDM agreed to undertake the transport of 65 consignments of spirits which were to be taken, by duty-suspended movements, from authorised warehouses in the UK to authorised warehouses in 30 other Member States of the European Union. One (identified as movement 12) was consigned, according to the relevant documents, to Unistock SA in Riga, Latvia, another (movement 65) to a warehouse at Neuss, Germany, owned by Dialog Logistik, and the remaining 63 consignments to a warehouse at Vaux-sur- Sûre in Belgium belonging to the well-known supermarket chain, Aldi. In each 35 case the load was carried by a driver who was not employed by SDM. SDM guaranteed the payment of excise duty on the goods comprised in all of the movements, a factor to whose significance I shall return.[4]HM Revenue and Customs, or HMRC, who are the appellants before this tribunal, concluded that none of the consignments reached their destinations and 40 that SDM, as guarantor, was liable to pay UK duty amounting to £6,306,137 on the goods. SDM accepts that the goods have all been diverted—that is, they have been removed from duty suspension but no duty has been paid on them in any Member State of the European Union and there is no evidence that they were exported to a country outside the European Union. SDM maintains that the 45 diversion took place after the movements for which it provided the guarantee had 3 been discharged and that in consequence its guarantee is not engaged. There is no dispute about the amount of the duty if SDM is liable to pay it at all.[5]It is common ground that SDM has not been able to produce the customary documentary evidence of discharge, that is a properly receipted and stamped accompanying administrative document (“AAD”) in 5 respect of any of the consignments. At the relevant time paper AADs were in use. The European Directive dealing with the imposition of excise duties and, among other matters of detail, duty-suspended movements of excise goods in force in 2006 was Council Directive 92/12/EEC which, by arts 18 and 19, provided that AADs were to be 10 produced in quadruplicate: the top copy was retained by the consignor, while the remaining three copies accompanied the goods (see Commission Regulation 2719/92/EEC). Following their arrival, copy 2 was to be retained by the consignee; copy 3 was to be returned by the consignee to the consignor endorsed by the consignee and, if national rules so required, stamped by the fiscal authority 15 in the destination Member State, as evidence of receipt and discharge of the movement; and copy 4 was to be made available to the fiscal authority.[6]No copy 3 AAD has been produced in respect of most of the consignments, and, as SDM accepts, those which have been produced (all relating to consignments said to be destined for the Aldi warehouse) bear stamps and receipts 20 which are not genuine. HMRC obtained evidence, which SDM does not dispute, that a corrupt Belgian excise officer (since convicted and imprisoned) had applied false stamps to some of the AADs relating to deliveries to the Aldi warehouse. HMRC obtained further evidence that the signatures, purportedly of Aldi staff, on some international cargo transport documents, known by their French acronym of 25 CMR, and relating to the relevant movements had also been forged, and of the activities of some continental criminals; I shall return to the significance of this evidence.[7]Although a properly receipted and stamped AAD is the usual means of establishing the discharge of a duty suspended movement it is open to HMRC to 30 accept alternative evidence. The copy 2 AADs held by the consignee would, in most cases, be acceptable alternative evidence but SDM has not been able to produce a copy 2 AAD for any of the consignments. It has, however, produced some other evidence, which HMRC say is incomplete. Moreover, some of it, they say, far from establishing that the movement to which it relates was properly 35 discharged, shows that it could not have been.[8]The appeal has a complicated procedural history which I shall explain in more detail below. For present purposes it is sufficient to record that the First-tier Tribunal from whose decision HMRC now appeal concluded that save in two cases—movement 29 to the Aldi warehouse and movement 65 to Germany—the 40 alternative evidence, coupled with the written and oral testimony of various witnesses, was sufficient to establish that the goods did arrive at their declared destinations, and that SDM is not liable to any duty on those goods, though it is liable for the duty due on the goods comprised in movements 29 and 65. HMRC appeal against its findings save in respect of those two movements. In a 45 respondent’s notice SDM indicated that it wished to challenge the decision relating to movement 29; it accepts that it cannot challenge the decision in respect of movement 65. 4 The relevant law[9]There was no disagreement between the parties about the relevant law, but it is helpful for understanding to explain it briefly. The Directive provides that various goods, among them alcoholic drinks, become chargeable to duty on manufacture within, or on import into, the European Union. 5 The obligation to pay the duty may, however, be suspended as long as the goods are held in a suspension arrangement—either a holding arrangement, within a warehouse approved for the purpose by the fiscal authorities of the Member State in which the warehouse is situated, or a movement arrangement, which meets various 10 conditions, by which the goods are transported from one approved warehouse to another, in the same or a different Member State. The duty becomes payable when the goods leave the suspension arrangement (are “released for consumption”), and it is payable in, and at the rate specified by, the Member State in which the release occurs. 15 10. Thus art 6 of the Directive provided that:
“1. Excise duty shall become chargeable at the time of release for consumption … Release for consumption of products subject to excise duty shall mean: (a) any departure, including irregular departure, from a suspension 20 arrangement ….”
[11]Article 4(c) defined a “suspension arrangement” as “a tax arrangement applied to the production, processing, holding and movement of products, excise duty being suspended”. It is undisputed that the movements undertaken by SDM were subject to suspension arrangements. If SDM is right, the goods reached their 25 intended warehouses of destination and, as HMRC accept in that eventuality, they were taken (or, as I shall explain below, are to be treated as having been taken) into a duty-suspended holding arrangement, with the consequence that the movement arrangement was discharged. Thus the goods remained within a suspension arrangement, but no longer one for which SDM bore any 30 responsibility. If that is correct it is not liable for any duty. On the other hand, as SDM accepts, if the goods did not reach their intended warehouses of destination there was an irregular departure from the movement with the consequence that excise duty became payable.[12]In some cases of irregular departure from a duty suspended movement 35 between one Member State and another there may be evidence of the place where the departure occurred (and correspondingly the Member State in which duty is payable) but when, as HMRC assert is the case here, the goods have been fraudulently diverted the place of diversion will often be unknown. The Directive provided for that difficulty by art 20.2 and 20.3: 40 “2 When, in the course of movement, an offence or irregularity has been detected without it being possible to determine where it was committed, it shall be deemed to have been committed in the Member State where it was detected. 3 … when products subject to excise duty do not arrive at their 45 destination and it is not possible to determine where the offence or irregularity was committed, that offence or irregularity shall be deemed to 5 have been committed in the Member State of departure, which shall collect the excise duties at the rate in force on the date when the products were dispatched unless within a period of four months from the date of dispatch of the products evidence is produced to the satisfaction of the competent authorities of the correctness of the transaction or of 5 the place where the offence or irregularity was actually committed ….”[13]HMRC’s case is based on art 20.3: they say that the goods did not arrive, it is not possible to determine where they were diverted—they may never have left the UK or may have been diverted on the continent—and therefore, as the 10 movements began in the UK, it is UK duty which is payable. It is art 20.3 which permits a person said to be liable for the payment of duty for this reason to produce satisfactory evidence, other than a properly receipted and stamped AAD, of the arrival of the goods (ie “the correctness of the transaction”), or to show that the diversion occurred in one Member State rather than another. It also makes it 15 plain (though SDM does not argue otherwise) that it is for the person said to be liable for the duty to produce evidence of arrival or place of diversion rather than for the competent authorities to produce evidence of non-arrival, or to establish where the goods were diverted.[14]The Directive left certain details, mainly of administration, to the Member 20 States. In the United Kingdom, at the time with which we are concerned, the relevant domestic provisions (which in part implemented and in part supplemented the Directive) were contained in the Excise Duty Points (Duty Suspended Movements of Excise Goods) Regulations 2001 (SI 2001/3022). The regulations which are material in this case were as follows: 25 “3 Irregularity occurring or detected in the United Kingdom(1) This regulation applies where: (a) excise goods are (i) subject to a duty suspended movement that started in the United Kingdom … and 30 (b) in relation to those goods and that movement, there is an irregularity which occurs or is detected in the United Kingdom.(2) Where the Commissioners are satisfied that the irregularity occurred in the United Kingdom, the excise duty point shall be the time of the occurrence of the irregularity or, where it is not possible to establish when 35 the irregularity occurred, the time when the irregularity first comes to the attention of the Commissioners.(3) Where it is not possible to establish in which Member State the irregularity occurred, the excise duty point shall be the time of the detection of the irregularity or, where it is not possible to establish when the 40 irregularity was detected, the time when the irregularity first comes to the attention of the Commissioners.…(4) For the purposes of this regulation, detection has the same meaning as in article 20(2) of the Directive. 4 Failure of excise goods to arrive at their destination 45 (1) This regulation applies where: 6 (a) there is a duty suspended movement that started in the United Kingdom; and (b) within four months of the date of removal, the duty suspended movement is not discharged by the arrival of the excise goods at their 5 destination; and (c) there is no excise duty point as prescribed by regulation 3 above; and (d) there has been an irregularity. (2) Where this regulation applies and subject to paragraph (3) below, the 10 excise duty point shall be the time when the goods were removed from the tax warehouse in the United Kingdom. (3) The excise duty point as prescribed by paragraph (2) above shall not apply where, within four months of the date of removal, the authorised warehousekeeper accounts for the excise goods to the satisfaction of the 15 Commissioners. 7 Payment (1) Subject to paragraph (2) below, where there is an excise duty point as prescribed by regulation 3 or 4 above, the person liable to pay the excise duty on the occurrence of that excise duty point shall be the person shown as 20 the consignor on the accompanying administrative document or, if someone other than the consignor is shown in Box 10 of that document as having arranged for the guarantee, that other person. (2) Any other person who causes or has caused the occurrence of an excise duty point as prescribed by regulation 3 or 4 above, shall be jointly 25 and severally liable to pay the duty with the person specified in paragraph (1) above.”[15]The effect of reg 3 was that when an irregularity occurred in the UK or, in the case of a consignment whose journey began in the UK, an irregularity was detected in the UK, a duty point arose. Regulation 3(4) stated that the word 30 “detection” bore the same meaning as in art 20.2 of the Directive. That article simply used the term without offering a definition; but it is accepted that in this case the detection in respect of each consignment was in the UK (see also Anglo Overseas Ltd v HMRC (VAT and Duties Tribunal decision E01090) at [69]). Regulation 4 applied when the goods did not arrive at their destination and it was, 35 by virtue of reg 4(1)(c), a mutually exclusive alternative to reg 3. However, as long as either reg 3 or reg 4 was engaged, so too was reg 7 (see Anglo Overseas at [57]). HMRC assessed SDM, as guarantor, for the duty in accordance with reg 7(1).[16]Article 13(a) of the Directive required a despatching warehousekeeper to 40 guarantee payment of the duty on any consignment of excise goods, should the consignment be released for consumption before arrival at another authorised warehouse, but art 15(3) permitted Member States to accept a guarantee provided by the transporter, either jointly and severally with the warehousekeeper or alone. It is not in dispute that SDM was the guarantor of all 65 consignments within the 45 meaning of art 15(3) and it seems that it was the sole guarantor. The liability of a despatching warehousekeeper and, by parity of reasoning, a guarantor is not dependent upon causation of the duty point or any other kind of fault: see 7 Greenalls Management Limited v Customs and Excise Commissioners [2005] 1 WLR 1754 at [17]. HMRC therefore do not need to show that SDM was complicit in, or caused, any diversion which took place, or that it was in any other way blameworthy.[17]HMRC have made further assessments against some of 5 the drivers (others could not be identified) for the duty on the consignments which they had undertaken to deliver, and against the vendors of some of the goods, in each case alleging against them that they caused the diversion, and relying on the imposition by reg 7(2) on those others of joint and several liability for the duty. I understand 10 that the appeals which they, or some of them, have brought against those assessments have been stood over by the First-tier Tribunal pending the outcome of this appeal. I understand too that no assessments have been made upon the despatching warehousekeepers. The history of SDM’s appeal 1518. SDM appealed to the First-tier Tribunal against the assessment made against it. As I have said, it did not dispute HMRC’s case that in most instances the copy 3 AADs had not been produced and that those which were produced bore forged receipts and customs stamps. It also did not disagree that the evidence that the goods had been diverted at some point was compelling. Thus the only real issue 20 was whether the goods had arrived at their stated destinations and had been discharged from the movements, or they had not. Part of SDM’s case was that the evidence of criminal activity which I have mentioned supported the proposition that the diversion occurred after delivery, while HMRC argued that it supported the contrary proposition. 2519. It was and remains common ground that the relevant authorities in Belgium, Germany and Latvia have concluded, whether rightly or wrongly, that the goods consigned to their respective countries did not arrive, and no assessment for Belgian, German or Latvian duty on the consignments has been made against SDM or any other person. Although HMRC rely on what was said by those 30 authorities and on the absence of overseas assessments in support of their case, they do not argue that these are determinative factors.[20]SDM’s appeal was heard by a First-tier Tribunal panel of Judge Wallace and Mr Coles (“F-tT 1”) who released their decision in March 2011. They decided that SDM had not shown that one of the consignments, movement 65 destined for 35 the warehouse in Germany, had reached its destination, and that SDM was liable for the duty on that consignment. They decided that SDM had shown that all of the other consignments, movement 12 due to be taken to the warehouse in Latvia and the remaining 63 due to be taken to the Aldi warehouse, had reached their intended destinations, and that SDM was not liable for duty on any of those 40 consignments. They reached that conclusion despite some doubts about ten of the consignments, referred to by others as “the allegedly impossible journeys”, a term I shall also adopt. The doubts arose from the perceived difficulty of reconciling the drivers’ evidence of delivery at the Aldi warehouse with the times shown on various documents SDM produced in support of its case, a difficulty which is 45 central to the present appeal and with which I shall deal in detail later. 8[21]HMRC sought permission to appeal against that decision to this tribunal. Permission was refused by Judge Wallace, but granted by Sir Stephen Oliver QC in the Upper Tribunal. SDM did not seek permission to appeal in respect of the German consignment (movement 65), though it emphasises that F-tT 1 dismissed its appeal in respect of that consignment because of a lack 5 of evidence that the goods had been delivered or taken into a legitimate holding arrangement, rather than because they were satisfied that they had not. F-tT 1’s conclusion in respect of that consignment has therefore, SDM says, no bearing on the conclusions which should be drawn in respect of the remaining consignments. 1022. HMRC’s appeal to this tribunal came before Judges Sinfield and Hellier (“UT 1”) who, in a decision released in May 2013, allowed the appeal (though on only one of the four grounds advanced and even then with some modification) and set aside F-tT 1’s decision. I shall deal with the detail of the ground which UT 1 accepted (and, so far as they are still relevant, the grounds they rejected) later. 1523. UT 1 decided to remit the appeal to a differently-constituted panel of the First-tier Tribunal. They explained in their decision that they would have preferred to remit to F-tT 1, as originally constituted, because of the difficulty facing a differently-constituted panel of determining an appeal by reference to evidence it had not itself heard, but that they could not do so as Judge Wallace 20 had by then retired (and, although this was not stated, had attained the age of 75 and was no longer able to hear it). At the end of their decision they invited the parties to make submissions about the terms of the directions to be given to the new panel.[24]The parties could not agree on the directions which should be made and a 25 further hearing took place before UT 1 in November 2013. HMRC’s position, at that stage, was that the appeal should be re-heard while SDM preferred reconsideration without a re-hearing of the evidence, not least because one of the witnesses on whose evidence it relied had died. UT 1 came to the conclusion that a complete re-hearing should be avoided (a conclusion in which HMRC now 30 concur), and that the appeal should be remitted to a differently-constituted panel which should reconsider the conclusions to be drawn from the evidence heard by F-tT 1 and in so far as necessary itself hear evidence, in accordance with the detailed directions UT 1 made. I shall come to those directions after I have examined the substance of the earlier decisions in more detail. 35 25. In June 2014 the remitted appeal came before Judge Berner, sitting alone (“Judge Berner” or “F-tT 2”). He recognised, like F-tT 1, that movement 29 posed particular difficulties, but he rejected the explanation overcoming those difficulties which F-tT 1 had accepted, and concluded at [142] that this consignment had not reached its destination but had been diverted at an 40 unidentifiable place. He found, however, that the remaining nine of the allegedly impossible journeys were in fact possible. He then applied UT 1’s direction, as he understood it, to that finding and reached the conclusion that SDM had discharged the burden of showing that the remaining consignments had arrived. Thus SDM was found to be liable for the duty on the goods comprised in movement 29, as 45 well as those in movement 65, but its appeal was otherwise allowed.[26]HMRC again sought permission to appeal, which was refused by Judge Berner but granted, albeit on only two of the three grounds advanced, by me in 9 this tribunal. I shall make some comments about the ground which was rejected later. The grounds on which permission was granted, in very brief summary, are that Judge Berner (HMRC say) failed to do what was required of him by the direction issued by UT 1, and that he had in any event applied an impermissible approach to the task before him. That is the appeal now 5 before us. In the respondent’s notice to which I have referred SDM challenges the manner in which Judge Berner approached his examination of the evidence given to F-tT 1 about movement 29 and his conclusion in respect of it.[27]HMRC were represented before us, as they were before F-tT 2, by Miss 10 Jessica Simor QC, leading Miss Isabel McArdle, and SDM by Mr Richard Barlow. We both express particular gratitude to Mr Barlow, who appeared before F-tT 1 as a practising barrister but has continued to represent SDM before UT 1, F-tT 2 and now us, we understand without reward, following his retirement from the bar. Our task would have been significantly more difficult without the benefit 15 of his submissions.[28]I should also mention at this stage that we have been asked by both parties, should we allow HMRC’s appeal in whole or in part, to re-make the decision, or so much of it as is necessary, rather than remit the appeal yet again. For that reason we heard argument about the manner in which the matter should be 20 resolved if we were to allow HMRC’s appeal. Unfortunately, we are divided about the outcome.[29]Ordinarily my conclusion would prevail, by exercise of the casting vote conferred by art 8 of the First-tier Tribunal and Upper Tribunal (Composition of Tribunal Order) 2008 (SI 2008/2835), but I have considered whether I should 25 exercise that casting vote, in view of what was said on the topic by the Court of Appeal in PF (Nigeria) v Secretary of State for the Home Department [2015] EWCA Civ 251. I have reached the conclusion that, even if this were a case in which the presiding judge of a first-instance tribunal should refrain from its exercise, it is not appropriate for me to do so, sitting as I am on a second appeal to 30 the Upper Tribunal and when it is at best doubtful what purpose would be served by remitting the appeal yet again.[30]I will, therefore, exercise my casting vote so far as necessary and allow HMRC’s appeal. I also think it is appropriate, notwithstanding the disagreement within the panel, to re-make the decision: again, remitting the appeal to a different 35 panel would not advance matters. For those reasons I have dealt, in what follows, much more extensively with the evidence than I would have felt necessary or desirable in the context of a simple appeal, and I have attempted moreover to make this decision comprehensible without recourse to the decisions which precede it. In doing so I have drawn in part from F-tT 1’s decision, occasionally 40 from that of UT 1, and have taken a good deal from F-tT 2’s decision, as Judge Berner helpfully distilled and summarised much of the evidence. I have also found it convenient to set out some of the parties’ arguments as I deal with the earlier decisions. F-tT 1’s decision 4531. F-tT 1 heard extensive evidence and submissions over 13 days in September and October 2010. They had the witness statements and oral evidence of SDM’s 10 directors, of the directors and staff of some of the vendors of the goods and of the consigning warehouses, and of the drivers (named Bunce, Waters, Blunsden, Woods, Francis and Parnham) of 30 or 31 of the consignments; the exact number is not clear. Another driver (Wild), of 14 consignments, had made a statement which F-tT 1 admitted; his statement was not agreed by 5 HMRC, but he did not give oral evidence because, SDM said, he had emigrated to Canada (he is one of the drivers whom HMRC have assessed). As SDM had delegated the transport of the goods to others who, in some cases, further delegated the work, it had not been possible after so long an interval to identify all of the drivers. Mr Bunce was one 10 who had arranged for others to drive the goods. He had been in business on his own account until August 2006, but then became the transport manager of the UK operation of Connie International BV (“Connie”), a Dutch company, which had a base in Kent with some employed drivers. It too engaged self-employed drivers, of whom Mr Waters was one. 1532. Mr Cranny told F-tT 1 that in 2006 SDM had decided that it should begin to undertake the duty-suspended transport of alcohol, for which purpose it obtained a movement guarantee of £90,000: that is, HMRC approved SDM as the guarantor of duty-suspended movements up to a limit of £90,000 of UK duty per movement. SDM then received requests from a number of customers to transport 20 consignments, including Doktor Czech UK Ltd (“Dr Czech”) and Liquid Marketing Ltd (“LML”); SDM had not previously dealt with either company. They both wished to have goods transported to Belgium, which suited SDM as one of the warehouses from which it regularly collected goods was also in Belgium (I shall come to the German and Latvian consignments later). The goods 25 were to be collected from the authorised warehouses of Edwards Beers and Minerals Ltd (“EBM”) at Leighton Buzzard or Checkprice UK Ltd (“Checkprice”) at Norwich, where they were held under duty suspension.[33]Mr Andrew Airlie, a director of Dr Czech, told F-tT 1 that the company was in the business of selling duty-suspended alcoholic goods, mainly to cash and 30 carry outlets in France and Belgium, and he related the circumstances in which he was introduced to two new and apparently legitimate customers, Cyber Comp and Tele Audio Group, or TAG; it is not clear whether they were French or Belgian companies but it is clear that their respective directors or managers knew each other, and that Dr Czech was aware of that connection. Dr Czech sold 55 35 consignments to Cyber Comp and two consignments to TAG, of which ten were despatched by Checkprice and 47 by EBM, all to the Aldi warehouse. TAG had paid late and it was for this reason that Dr Czech had sold it only two consignments. Cyber Comp had initially paid promptly, though in cash which Mr Airlie received from a courier, but had later fallen behind and at the time of the 40 hearing owed Dr Czech £221,000. Mr Airlie said that SDM had been recommended to him as a reliable haulier; it does not seem that SDM had carried any goods for Dr Czech before the consignments with which we are concerned. Dr Czech has been assessed by HMRC for duty on those consignments of which it was the seller on the basis that it was jointly and severally responsible for their 45 diversion.[34]There was no live evidence from LML as its director, Mr Devinder Chahal, was ill, but he had made a statement to the effect that LML had sold six consignments to TAG and Cyber Comp, all despatched from Checkprice to the 11 Aldi warehouse. He too had been recently introduced to TAG and Cyber Comp and knew of the connection between them; it appears that LML came to use SDM as its haulier on the recommendation of another customer. It seems that LML, too, experienced difficulties in collecting payments from its customers. It has been assessed for duty on its consignments in the same 5 way as Dr Czech.[35]Mr Barlow laid some emphasis before us on the manner in which the transport of the consignments was arranged. Typically, as F-tT 1 recorded at [53], SDM received a telephone call from Dr Czech or LML in which it was given instructions to collect and deliver a consignment. Mr Hodgkins, the director 10 responsible for operations, or the transport manager, Mr Young, then arranged for one of the drivers or Connie to undertake the work; it seems that they aimed, whenever possible, to combine the outward journey with a collection from the warehouse of a continental customer.[36]Mr Cranny, whose evidence on this topic was not challenged, told F-tT 1 15 that the instructions to SDM often came with little notice, and that in consequence drivers were selected shortly before a journey took place. In some cases, as I have said, the selected driver engaged a sub-contractor whose identity was not known to SDM, and it was not told by Mr Bunce or Connie who would be the driver of those movements for which Connie was engaged. The warehousekeepers were 20 given the numbers of the vehicle and trailer that would collect the load, but not the name of the driver, even in those cases in which SDM knew who he would be. Although F-tT 1 did not say so in terms, it appears that the witnesses from the despatching warehouses confirmed that they did not know, and were not interested in, the names of the drivers. The receiving warehouse, too, would not 25 know the name of the driver and would know only approximately when a load it was expecting might arrive. If HMRC are right, of course, the stated receiving warehouses would not be expecting any goods to arrive, although if Aldi had one or more dishonest members of staff, as SDM accepts must be the case if the goods were diverted after arrival at its warehouse, those employees would no doubt be 30 aware of an imminent delivery even if not of the exact time of arrival.[37]Although there was some evidence that a few of the vehicles might have travelled by ferry, it seems that the great majority crossed the Channel by Eurotunnel. SDM had an account with Eurotunnel which enabled the drivers to use the service at SDM’s expense. The vehicle registration number had to be 35 supplied in advance, so that it could be identified and related to SDM’s account by Eurotunnel’s number plate recognition equipment, but the vehicle was not required to cross at any particular time on either the outward or inward journey; it could travel on the first available train following its arrival at the terminal.[38]F-tT 1 also had the written and oral evidence of six HMRC officers who had 40 made enquiries, including investigations into what had become of the goods. They produced various items of information, some obtained from the Belgian authorities about their own investigation and the subsequent prosecution of the corrupt official and others, and the witness statements of some Aldi employees to the effect that the warehouse did not take in goods belonging to third parties, that 45 Aldi did not deal in goods of the kind comprised in the 63 consignments said to have been received at the warehouse, and that it did not have any dealings with EBM, Checkprice, Dr Czech or LML in the UK and did not know of Cyber Comp 12 or TAG. None of the Belgian officials or Aldi employees gave oral evidence to FtT 1, but with a limited exception their evidence was not in issue. The exception is that, as Mr Barlow pointed out, the Belgian authorities had at first confirmed that some of the loads had arrived, later changing their minds, but even then accepting that one had done so. Mr Barlow drew attention to this point 5 not as a means of showing that the consignment had been taken into Aldi’s stores but in order to show that the Belgian investigation into the eventual disposal of the goods had been somewhat casual and could not be relied on as a clear indicator that the goods had not arrived at the warehouse. 10 39. Some features of the investigation are nevertheless important. The corrupt Belgian customs officer, named Schmit, was arrested on 30 November 2006. He had been the officer responsible for the Aldi warehouse until about 2002 or 2003 when he was replaced by another officer. He admitted before an examining magistrate that between about July and September 2006 (although F-tT 1 found 15 that forgeries continued into November) he had applied false customs and Aldi stamps to an undetermined number of AADs (the decision records that 18 were found) in exchange for about €50,000 paid to him by a person he claimed not to know.[40]The consignees declared on the falsely stamped AADs, consistently with the 20 evidence of Mr Airlie and Mr Chahal, were Cyber Comp (whose general manager was one Kucuker) and TAG (general manager one Motheu), whose agent was said to be called Renaix. The evidence also showed that a telephone belonging to the former managing director of TAG, named Vrielynck, had been tapped. Mr Barlow emphasised that the tapping did not reveal any calls to the United Kingdom, 25 suggesting no evidence of any connection between the criminals and the drivers. Schmit, Motheu (who it seems was probably using Renaix as an alias), Vrielynck and Kucuker were all charged with various offences. The decision records that Schmit was convicted, but one is left to assume that the others were convicted as well. 30 41. I think it appropriate to mention in passing F-tT 1’s expression of surprise that the Belgian authorities did not carry out a more thorough investigation into what became of the goods. I do not share that surprise since it is evident from what else F-tT 1 said that they considered that HMRC had led the Belgian authorities to believe that the goods had not left the UK, despite the evidence that 35 TAG and Cyber were involved. Given the nature of their investigation, described by Mr Barlow himself as casual and in any event not focused on these movements, I do not attach any significance to the fact that the Belgian authorities recorded that one load arrived. I do not, in particular, accept it as reliable evidence that the load did arrive. F-tT 1 also expressed surprise at [469] that HMRC had 40 not made their own enquiries of Aldi; but as it was for SDM to show that the goods had arrived, and its case remained that they had, one might think that it was SDM which should have made those enquiries. In fact, as we learnt (though F-tT 1 may not have been aware) SDM’s representatives did make some enquiries at Aldi, after the assessment had been served. We did not learn the results of those 45 enquiries.[42]The parties disagree about what the evidence of conspiracy tends to show. F-tT 1 dealt with that disagreement by comparing what they described as two 13 competing scenarios: that the goods were diverted before arrival at Aldi’s warehouse, and that they were diverted after arrival. They pointed out at [440], as Mr Barlow did before us, that although SDM’s case—which it maintains—was that the 63 consignments had duly arrived at Aldi’s warehouse, it was necessary for it to do no more than show that they had arrived in Belgium 5 since, if that was the case, Belgian duty was payable irrespective of non-arrival at the warehouse, and the present assessments must be discharged. HMRC accept that this would be correct if arrival of the goods in Belgium before irregular departure could be established; but they argue that there can be no certainty that the goods left the 10 UK or that, if they did, they were diverted in Belgium rather than France, through which they would have to pass on their way to Belgium, and it is conceivable though perhaps unlikely that they were diverted elsewhere.[43]Nevertheless SDM’s case before us remains as it has always been, that the 63 consignments were duly delivered to Aldi’s warehouse. Even though it may 15 not be necessary to go further Mr Barlow does not argue, nor realistically could he as SDM relies on the drivers’ evidence of delivery at the stated destination, that the goods were diverted in Belgium or France, and that arrival on the continent is enough. He also accepts, with a possible caveat about movement 29, that the only reasonable conclusion is either that all of the 63 consignments were duly delivered 20 at Aldi’s warehouse, or that none of them was. Whichever of those conclusions is correct the Latvian consignment needs to be considered separately.[44]HMRC accept that if it is shown in respect of the Belgian consignments that the goods arrived at the warehouse SDM should be taken to have discharged the burden of showing regular completion of the movement, and that there is no 25 obligation on it to go further and show what happened to the goods thereafter—in particular, they accept that SDM does not need to show that the goods were taken in to Aldi’s stock, so as to enter a duty-suspended holding arrangement, which FtT 1, unsurprisingly, found they were not. HMRC too accept that the only realistic conclusion is that all of the goods, or none of them, arrived, again with a caveat 30 about movement 29.[45]Although they did not say so in terms, it is implicit in the manner in which F-tT 1 described the competing scenarios that, whichever of them was to be preferred, TAG and Cyber, or those running them, were involved in the diversion. If the goods were diverted before they arrived at the warehouse (whether in the 35 UK, or Belgium or France) it also necessarily follows that the drivers were implicated, and that the evidence of those of the drivers who appeared before F-tT 1 to the effect that they had delivered the goods to the warehouse was false.[46]At [443] F-tT 1 addressed the significance in the context of the competing scenarios of the evidence that the identity of the drivers would not have been 40 made known to the despatching warehouse or, indeed, anyone else in advance: “The documentary evidence showed SDM as merely giving the numbers of the vehicles and trailers to the warehouses. No explanation or even theory was advanced by Miss Simor as to how the ringmasters could have known the identity of the drivers of each movement if the goods were never 45 delivered at Aldi. We observe that if a driver had become involved in one diversion, he could have told the ringmasters about further movements by him. That would not explain how such driver could have become involved initially if his identity was not known.” 14[47]At [445] they addressed the corresponding difficulty inherent in SDM’s proposition that the goods were diverted after arrival:
“Diversions at Aldi after the goods were unloaded would have involved at least one dishonest employee at Aldi and very probably more than one. Although such insiders would no doubt have been 5 informed by TAG and Cyber that the individual consignments had been ordered from England, the insiders could not know when the goods would arrive. All of the drivers gave evidence that they went to the office on arrival. The insider would have needed to ensure that the AADs and CMRs were not processed by anyone 10 not involved in the conspiracy and that the consignment was not entered into Aldi’s stock records.”
[48]They added elsewhere that for the conspiracy to work diversion in this manner would have to occur on as many as 63 occasions without detection by the management staff present at the warehouse. It would also be necessary to take 15 care to ensure that the drivers did not realise that there was something amiss in the manner in which the goods were received and the paperwork was handled.[49]F-tT 1 then considered which of the two scenarios was the more likely. They began with some further background material with which I too need to deal. At [452] they referred to EC Regulation 3820/85, which limits the time for which 20 the driver of a heavy goods vehicle may drive, and imposes rest time requirements:
“[452] … the maximum driving time per day was 9 hours which could be extended to 10 hours on no more than two days in a week. A break of 45 minutes was required every 4½ hours, but this could be split into three 25 breaks of 15 minutes each; time spent on a train counted as a break provided a bed or couchette was available. Time spent waiting to load or unload counted as a break if no work was performed. One break of at least 9 hours was required each day or of at least 8 hours if there were 3 separate breaks. [453] Thus, for example, if a driver took a 9 hour break before boarding a 30 train at Folkestone, he could legally drive 4½ hours from Coquelles to Vaux without a break, spend 15 minutes waiting and ½ hour being unloaded and drive straight back to Coquelles giving a return time Coquelles to Coquelles of under 10 hours, however he would then need another long break.”
[50]At [454] they mentioned the requirement derived from the same Regulation 35 that each lorry should be fitted with a functioning tachograph, and that the discs should be retained for a minimum of a year. The discs would reveal the speed at which the vehicle had travelled—and, on analysis of the journey to which the disc related, whether it exceeded the speed limit—and whether the driver had taken the requisite breaks. It is noteworthy that (as F-tT 1 recorded) none of the tachograph 40 discs for the vehicles covering the times at which the journeys were made, or were said to be made, were produced. Some explanations for their non-production were offered to F-tT 1; I shall return to those explanations at a later stage.[51]F-tT 1 then turned to an analysis of the drivers’ evidence, making the point that it had been given four years after the relevant events, and that most (though 45 not all) of the drivers had not been interviewed—and therefore made aware of any doubts about the deliveries—until two years after the events. I shall undertake a similar exercise later in this decision, and it is not necessary to set out the detail of F-tT 1’s analysis; it is sufficient to record for the present that the drivers who gave 15 evidence maintained, despite the secondary evidence to which I come shortly, that every load for which they were responsible had been correctly delivered to its stated destination. Mr Bunce gave evidence to the same effect about the consignment destined for Latvia, which he had driven; he also spoke about the movements for which Connie had provided a driver. The driver 5 of movement 65, destined for Germany, did not give evidence.[52]The secondary evidence of the arrival of the goods produced to F-tT 1 by SDM consisted of documents identifying the trailers on which the goods had been loaded and the tractors towing the trailers; documents emanating from Eurotunnel 10 showing that the tractors and trailers had travelled through the Channel tunnel on the appropriate days; copies of the vignettes which the driver of a heavy goods vehicle is required to buy before he is permitted to travel on Belgian motorways; and some fuel purchase receipts. The purpose of its submitting that evidence was to show that the relevant vehicles had travelled through the tunnel and on from the 15 French end to Belgium on days corresponding to the collection of the goods and the stipulated delivery date, thus establishing that the goods had reached Belgium and, inferentially, the Aldi warehouse. This evidence related to only some of the movements, and it was often incomplete. In other cases there was no secondary evidence of delivery at all, apart from the driver’s statement that he had taken the 20 load to its intended destination. In some but not all cases the documents produced included some relating to the collection of the goods from the despatching warehouse.[53]Many of the documents bore indications of the time of day at which the event for which they were the evidence had occurred. I will need to deal later in 25 more detail with the information which could be gleaned from the documents, but for the present I need merely to record that the Eurotunnel records showed the times of check-in at the relevant terminal, and the vignettes could be interpreted to determine not only the period for which they were valid but also the date, time and place of purchase: some were bought after the driver arrived on the continent, 30 but others were purchased at the UK tunnel terminal at Folkestone while the driver was waiting to board the train. The fuel receipts, too, showed where and when the purchase had been made.[54]HMRC’s argument before F-tT 1 (and which they pursue now) is that an analysis of the documents from which it can be determined where a particular 35 load was at an identifiable moment shows in ten (later reduced to nine) cases, and if the documents can be taken at face value, that the journeys could not have been completed in the time available: hence the description “allegedly impossible journeys”. The assessment of the time required assumed a journey from the French end of the Channel tunnel at Coquelles to the Aldi warehouse at Vaux-sur- 40 Sûre, the discharge of the load there, and the return to Coquelles including the collection of a return load on the way. There was some evidence from the drivers about the extent to which they respected the applicable speed limits and rest hours requirements, to the detail of which I shall come later.[55]F-tT 1 were therefore required to resolve a conflict between what the 45 available documents showed, or appeared to show, on the one hand, and the oral testimony of the drivers to the effect that they had delivered the goods to the warehouse on the other. There was some further evidence, particularly that in 16 most cases the goods had been paid for, which was said to support SDM’s case, but it seems F-tT 1 attached little or no weight to it. It may have contributed to their finding, at [493(c)], that “[t]here was a conspiracy or a series of conspiracies involving TAG and Cyber, but in which SDM, EBM, Checkprice and Mr Airlie (of Dr Czech) were not involved”. It was not disputed that 5 SDM had earned no more than an ordinary commercial charge for undertaking and guaranteeing the movements. There was also some evidence which HMRC had extracted from the drivers’ mobile telephone records, showing that they had spoken to each other on a number of occasions, and also showing that they were in France or Belgium at 10 identifiable times. F-tT 1 referred to those records while describing the drivers’ evidence, but they did not indicate what, if any, significance they attached to them.[56]F-tT 1 summarised the evidence and submissions relating to the conflict between the drivers and the documents in this way, at [467]: 15 “Miss Simor contended that many movements were impossible within the timescales indicated by the documents. We heard evidence from the drivers that they variously timed their arrivals on the continent for the early part of the day when the roads were quiet, chose routes which were known to have few traffic police and did not strictly observe legal speed limitations. Under 20 these circumstances they were able to cruise at speeds of up to 80mph. The Tribunal analysed the timings of all of the movements identified by Miss Simor in the light of the drivers’ evidence, and concluded that only one movement, Movement 29, was impossible. However, the evidence indicates that this movement was originally intended to be part of a three vehicle 25 delivery, subsequently amended to two vehicles, and which included collection of a backload from a customer in France. It was not clear which vehicle or vehicles carried out which part or parts of the journey, and it may well be that the actual journeys performed by the individual vehicles were different from those originally planned. Overall, we conclude that the 30 journey timings offer no support to Customs’ case that the goods could not have reached Aldi.”[57]They then dealt with some matters of detail, before setting out their conclusion about those movements for which they had the evidence of the drivers at [475]: 35 “Faced with the difficulty of the two competing scenarios, the burden of proof which rests on SDM is important. Having heard the evidence of Mr Waters, Mr Blunsden, Mr Parnham and Mr Francis, we are satisfied on the balance of probabilities that they did deliver to Aldi; their evidence tips the balance in respect of those deliveries. Similarly we are satisfied on the 40 balance of probabilities that Mr Bunce delivered to Latvia.”[58]Mr Bunce’s role, both when trading on his own account and when managing Connie’s UK operation, had been to provide a driver and, in most cases, a vehicle in order to transport a load for SDM. The only consignment he drove himself was that destined for Latvia. It differed from those sent to Aldi because the UK vendor 45 was Tradium Ltd, which had sold the goods to Unistock SA. The goods were collected, Mr Bunce told F-tT 1, at Calais. No copy 3 AAD or any alternative documentary evidence of arrival was produced, save for a CMR, but equally HMRC had no more than a statement by the Latvian authorities that the goods had not arrived. Mr Barlow contended that there was some confusion about the AAD 17 relating to this load, and that it could not be certain that the Latvian authorities had been asked about the correct load.[59]F-tT 1 addressed the 14 movements for which Mr Wild was responsible in this way: “[476] A detailed comparison of the evidence relating 5 to Mr Wild’s movements and those performed by the other drivers does not indicate any material differences between his movements and the movements of the drivers whose evidence we accept. We do not consider it appropriate to disregard Mr Wild’s statement on the ground that he has gone to Canada 10 having received an assessment of £1.3 million. There was nothing exceptional about any of his fourteen movements. In fact he exhibited fuel receipts from Veurne which is on the road from Coquelles to Vaux about one hour from Coquelles; these were timed and dated and fit well with a journey to Vaux. Twelve of the fourteen movements have evidence of return loads of 15 food. Although Mr Wild did not send Eurotunnel paperwork to Mr Simpson [the HMRC officer who made the assessment], SDM did. There were stamped CMRs for all his movements except Movement 62. It is clear that he did not send his tachograph discs to Mr Simpson despite repeated requests, but we accept the evidence of the other drivers that tachograph 20 discs were commonly not retained by drivers. Although the fact that he was not available for cross-examination is relevant (since we had no opportunity to assess his credibility) the evidence produced did not indicate any qualitative difference between Mr Wild’s movements and the movements of the drivers whose evidence we accept.” 25 60. They resolved the apparent difficulty in respect of movement 62, that is the absence of a stamped CMR, in SDM’s favour, and concluded at [477] that Mr Wild delivered all of his 14 consignments to Aldi.[61]There was no driver evidence at all for 22 of the movements, which F-tT 1 discussed from [479] to [488]. Although I do not need to deal with one of them, 30 the movement destined for Germany, since there is no appeal against F-tT 1’s conclusion that SDM is liable for the duty on it, I do need to record something of it, since it was another for which Mr Bunce provided the driver. He was asked about the movement as he gave evidence, but said he could remember very little, including the name of the driver. The UK seller was Pierhead Purchasing Ltd, and 35 the continental buyer was Intermédiaire Europe Eurl which, despite its French name, appears to have carried on business in Germany (Dialog Logistik was stated to be the receiving warehouse). HMRC produced documents showing that the vehicle claimed to be carrying the goods to Germany was found to be empty in a Eurotunnel security scan while travelling outbound, and a letter from a 40 director of Pierhead Purchasing to the effect that when the copy 3 AAD did not arrive as expected he contacted Dialog Logistik and was told that the goods had not arrived there. However, Intermédiaire later told him that they had been delivered, on Intermédiaire’s instructions, to a different warehouse, also in Germany, at Halle. 45 62. F-tT 1 rejected SDM’s appeal in respect of this consignment because the correct procedure for a change of consignee had not been followed and because they were not satisfied that the diversion occurred in Germany. They did not make a clear finding that the goods had been diverted (rather than merely sent to a different receiving warehouse) but it must follow from what they did say that they 18 were satisfied the goods had been diverted. SDM had produced no evidence of actual arrival at the Halle warehouse.[63]There was some documentary evidence before F-tT 1 relating to the 21 remaining consignments for which there was no driver evidence, mostly revealing unexplained delays rather than shortages of time. HMRC 5 argued that the delays were in fact explicable: they represented the periods during which the goods were illicitly removed from the vehicles before they travelled, empty, to the continent if indeed they travelled there at all. F-tT 1 did not address that argument directly, but did deal with the difficulty in which they were placed by the absence of adequate 10 evidence and explanation, and then set out their conclusions: “[489] We have had considerable difficulty in reaching a conclusion in relation to the Aldi consignments for which there was no driver evidence. [490] The mere fact that there was no direct evidence of delivery is not decisive in spite of the burden of proof. In legal proceedings of any nature 15 facts can be established by inference. We have to decide whether the inference of a consistent system of diversion or concealment of diversion by the ringmasters is sufficient to satisfy us that it is more probable than not that all the drivers delivered their consignments to Aldi being given receipt stamps on the CMRs and that they were not personally involved in the 20 conspiracy. [491] We have concluded that on the balance of probabilities all of the Aldi consignments were delivered by the drivers to Aldi and that the irregularities occurred thereafter. The appeal therefore succeeds in relation to all movements except Movement 65 to Germany.” 25 UT 1’s decision[64]As I have said, HMRC secured permission to appeal against F-tT 1’s decision on four grounds. As recorded by UT 1, they were that F-tT 1 incorrectly considered that they were precluded from reaching any decision that SDM or its employees had acted dishonestly; that they had incorrectly imposed the burden of 30 proof on HMRC; that F-tT 1 reached findings of fact that were not open to them on the evidence; and that their approach to reaching their decision and their assessment of the witnesses’ evidence were erroneous. UT 1 did not, as I understand their decision, allow the appeal on any of those grounds, as so expressed. However, before coming to their conclusion I need to explain more of 35 the reasoning.[65]Much of UT 1’s decision relates to the grounds they rejected, and I do not need to deal with what they said in respect of them. The part which is relevant now focuses on the allegedly impossible journeys, that is the ten (or now nine) journeys in respect of which SDM produced Eurotunnel documents, vignettes or 40 fuel receipts, or some combination of them, which suggested that the time available was too short for the journey to which they related to have been undertaken. Documentary evidence was not produced for all 63 movements to Aldi and, as I have said, in some cases the documents relating to a journey suggested an inexplicable delay, or at least a delay for which there was no evident 45 explanation, between collection of the load and the driver’s arrival at Folkestone, although some explanations were offered by the drivers as they gave evidence. HMRC continued to maintain, at least in those cases in which the delay occurred 19 in the UK, that it was accounted for by the time required for the removal of the goods from the vehicle, but that was not the focus of the argument before us and I shall return to the point only at the end of this decision.[66]The allegedly impossible journeys were movements 17, 19, 22, 24, 29, 37, 38, 43, 44 and 57, all of goods said to be destined for the Aldi 5 warehouse. One of the journeys was undertaken by each of Mr Wild (24), Mr Parnham (37) and Mr Francis (44) and the remaining seven by Mr Blunsden. At [57] and [58] UT 1 said:
“[57] The evidence of Mr Cranny, Mr Waters and Mr Wild was that the 10 journey between Coquelles and Vaux-sur-Sûre took between 4 and 4½ hours. Mr Blunsden initially said that the journey to Vaux-sur-Sûre would take 4 or 4½ hours but then said that he could do it in 3½ hours in the best conditions and exceeding the speed limits as he sometimes did. Mr Parnham seemed less sure of the length of time required and accepted a time of 15 between 3½ and 4½ hours. At [453], the FTT calculated the time required for the round trip from Coquelles to Vaux-sur-Sûre and back to Coquelles as follows: ‘… if a driver took a 9 hour break before boarding a train at Folkestone, he could legally drive 4½ hours from Coquelles to Vaux 20 without a break, spend 15 minutes waiting and half an hour being unloaded and drive straight back to Coquelles giving a return time Coquelles to Coquelles of under 10 hours, however he would then need another long break.’ [58] Thus the FTT appears to have accepted that the minimum time for a 25 round trip was 9 hours 45 minutes. That calculation, however, ignored the fact that on many journeys, including all the ones alleged by HMRC to be impossible, the drivers were required to pick up a return load on the way back to Coquelles. That would add a minimum of 15 minutes waiting and 30 minutes being loaded to the journey time. Thus, adopting the findings of the 30 FTT, with the addition of 45 minutes for the return load, the minimum time required for the round trip would be 10½ hours.”
[67]UT 1 then examined some of the evidence about the journeys before coming to F-tT 1’s conclusions at [467] (quoted at para 56 above): “[67] Given the apparent inconsistency between the drivers’ evidence and 35 the evidence of journey timings, the task of the FTT was to weigh the competing evidence and decide which it preferred. In effect, it seems to us that this is what the FTT did in this paragraph. It is clear from reading the whole of [467] that the F-tT (which had the benefit of the experience of Mr Coles in the European logistics industry) analysed the timings of the 40 allegedly impossible movements in the light of the evidence of some of the drivers that they travelled when the roads were quiet and at speeds above the legal limits and concluded that the claimed journeys were not impossible. (And it is implicit that this analysis was extended to movement 24 driven by Mr Wild.) 45 [68] But this paragraph does not explain why the FTT concluded that the timings of the journeys shown on the schedule, apart from movement 29, were possible. [69] The FTT’s conclusion is that the timings offer ‘no support to Customs’ case’; whether or not that is a relevant conclusion, the issue the timings raise 20 is that of a challenge to the veracity of the drivers’ evidence. That challenge cannot simply be put aside by relying on the drivers’ evidence; that would be circular. The two must be resolved.” [original emphasis][68]The observation by F-tT 1 at [467] “that the journey timings offer no support to Customs’ case that the goods could not have 5 reached Aldi” and a number of other statements of a similar character were the foundation of HMRC’s case before UT 1 that F-tT 1 had wrongly placed the burden of proof on HMRC. As I have said, that ground of appeal was rejected, essentially on the basis that other comments made by F-tT 1 showed that they were throughout conscious that 10 the burden lay on SDM and that they had approached the assessment of the evidence correctly.[69]However, what UT 1 said at [69] highlights the essential issue before F-tT 1: were they to treat the timings revealed by the documentary evidence as reliable and determinative, and reject the drivers’ evidence if it could not be reconciled 15 with the timings; or, instead, did the drivers’ evidence provide a sufficient explanation of their ability to complete the journeys in less than what F-tT 1 had found was the normal minimum time of 9 hours 45 minutes (or 10 hours 30 minutes as adjusted by UT 1)? There was the additional complication of Mr Barlow’s argument that the times shown on the documents might not be accurate 20 or they may have been linked to the wrong movement. UT 1 went on to set out in more detail why they considered that F-tT 1 had failed to resolve the conflict properly, or at least to explain adequately why they had reached their conclusion that the drivers’ evidence was to be preferred: “[70] The simple recitation in that paragraph [ie [467]] of the factors of 25 traffic density and speed, does not seem to us to explain the significant gap between the minimum time (10½ hours) apparently required to complete the round trip when unloading and loading are taken into consideration and the actual times recorded in relation to some of the movements. Nor is it clear how these factors permitted the conclusion in Mr Wild’s case. The times 30 shown for the return trip from Coquelles to Vaux-sur-Sûre in movements 17, 19, 24, 29 and 37 are all materially less than the 10½ hours apparently required. The FTT does not expressly indicate whether or not it accepted that the drivers were travelling at a time of day when the roads were quiet and were exceeding the speed limits (and there was some apparently 35 contradictory evidence on this score); but even if it was implicitly accepted (without resolution of those conflicts), we cannot understand how the FTT concluded that those journeys could have been made in the times shown for those movements. [71] Mr Barlow says that there was evidence that dates on documents were 40 sometimes inaccurate, that Channel tunnel records were occasionally incorrect and that one driver might take another booking. If the FTT accepted this evidence and if it considered it relevant it might have formed part of the FTT’s explanation. But the fact is that there was no explanation of the resolution of very material conflicting evidence. 45 [72] Thus it seems to us that either the FTT’s conclusion could not have been reached on the evidence, or that the FTT has not adequately explained why it felt able to ignore the disparity between the evidence of how long the round trip would take and the evidence of the actual, much shorter, times taken in at least some of the movements identified by HMRC, or how it 50 reconciled any disparity with the drivers’ evidence … 21 [74] It seems to us that the absence of any explanation for accepting the evidence of the drivers that the journeys in movements 17, 19, 24, 29 and 37 took place as described in the face of other evidence that, on the FTT’s own calculation of the time required, showed that the journey times were impossible was a failure to give reasons on a material matter 5 - a matter vital to the conclusion reached by the FTT that all the journeys (apart from those to Germany and Latvia) resulted in delivery to Aldi. The failure to give reasons for accepting the evidence of the drivers was, in our view, an error of law.” 10 70. UT 1 then embarked on an examination of further evidence available to F-tT 1, some of which I shall also examine later, before allowing the appeal on the basis set out at [93]: “In our view, the FTT failed to give any or any adequate reasons for its conclusion that each of the ten allegedly impossible journeys resulted in 15 delivery of the consignments of spirits to Aldi. We allow the appeal on this ground and direct that the FTT’s decision be set aside and the case be remitted to a differently constituted First-tier Tribunal to determine the issue afresh in relation to the allegedly impossible journeys and to consider what effect its conclusion has on the evidence in relation to the other deliveries.” 20 UT 1’s later direction[71]As I have mentioned, UT 1 invited the parties to agree directions about the manner in which the new panel of the First-tier Tribunal was to deal with the remitted appeal. Paragraph 2 of the direction they made, after hearing the competing arguments required the new panel, F-tT 2, (so far as material now) to: 25 “a. determine whether all or any of the journeys described in the schedule of the ten allegedly impossible journeys produced at the hearing (other than movement 29) could not have taken place as described in the evidence of the drivers as recorded in the [F-tT 1] Decision by reference to the evidence that was before the First-tier Tribunal at the hearing, including the witness 30 statements, oral testimony as set out in the transcript and documents; b. in determining the issue at (a), the First-tier Tribunal shall have regard to the propositions of law and findings of fact (other than in relation to the allegedly impossible journeys) in the Decision; c. if it is found that all or any of the journeys could not have taken place 35 as described, consider what effect such finding has on the conclusion in the Decision that i. the goods carried on those journeys were delivered to Aldi in Belgium, and ii. the goods carried on other journeys, not alleged to be 40 impossible, were delivered to Aldi in Belgium; and take such steps as they consider just to determine the appeal either with or without hearing further evidence; and d. if it is found that all of the journeys could have taken place as described, to determine the appeal on the basis of the other findings 45 contained in the Decision.” 22 F-tT 2’s decision[72]The remitted appeal came before Judge Berner in June 2014. In his decision he related the history of the matter, as I have done, and then began an examination of the task which UT 1 had set for him. At [16] he described how F-tT 1 had concluded that they were required to decide between the competing 5 scenarios, and then added his own perception of the difficulties: “[17] The ‘competing scenarios’ were, first, what the original FTT described as HMRC’s scenario, that the goods never reached Aldi, which involved active participation by the drivers wherever the diversions occurred and 10 knowledge by the ringmasters of the diversion as to the individual movements of the drivers, and secondly, the scenario put forward by Mr Barlow on behalf of SDM, that the drivers were not involved and that the diversions or irregularities took place after arrival at Aldi. This was the way in which the original FTT described the position, but as the UT noted, at 15 [38], this was no more than a way of addressing the question whether SDM’s proposition (that the goods were taken to Aldi) was more likely than not – the ‘not’ being what was described as HMRC’s scenario. [18] As the UT, at [39], found it was right to do, the original FTT, at [442], considered the logical consequences of the diversions taking place before the 20 goods reached Aldi. It found that in those circumstances the drivers would have to have been involved. The ringmasters could only involve the drivers if they knew who the driver would be for each of the particular movements. On the evidence there appeared to be no way in which the ringmasters could have known the identity of the drivers of each movement if the goods were 25 never delivered at Aldi, and no explanation had been proffered by HMRC. [19] However, as the original FTT identified, at [445] and [446], there were on the other hand substantial difficulties with SDM’s scenario that the drivers were not involved in the irregularities and did deliver the goods to Aldi. For that to have been the case, the original FTT observed, there would 30 need to have been at least one (and probably more than one) dishonest employee at Aldi. The insiders could not have known when the goods would arrive. The insiders would have needed to ensure that the AADs and CMRs were not processed by anyone not involved in the conspiracy and that the consignments were not entered into Aldi’s stock records.” 35 73. At [141] Judge Berner returned to the choice between the two competing scenarios (albeit in the immediate context of his observations regarding movement 29, about which I shall have more to say later): “Mr Barlow submitted that the original FTT had found that either all or none of the deliveries had been made to Aldi. In other words, there was a distinct 40 choice between SDM’s proposition that the diversions had taken place through a conspiracy at Aldi, not involving the drivers, and that of HMRC, where the deliveries were not made to Aldi at all. I do not regard the choice as a binary one, and consequently I do not adopt the same approach as the original FTT. It is in my view quite possible for it to be found, on the 45 evidence, and having regard to the burden of proof, that certain deliveries are proved to have been made to Aldi, and certain have not been proved to have been so made. There is no need to speculate what happened to any deliveries in the latter category.” 23[74]It was for that reason that he examined each of the allegedly impossible journeys individually. By way of preliminary, he remarked at [32] that he was required to determine whether or not the allegedly impossible journeys, apart from 29, were in fact possible “by reference only to the evidence before the original FTT. It is only when dealing, in accordance with Direction 5 2c, with the broader effect of journeys found to be impossible that I would have to decide whether further evidence should be heard”.[75]One of the grounds of appeal on which HMRC were given permission is that, despite his identification in that paragraph of the limits on what he was able 10 to do, Judge Berner did not respect those limits, but instead relied on additional evidence not before F-tT 1 in determining whether the relevant deliveries could have been made. I shall deal with the arguments relating to that ground of appeal in due course, but it is appropriate that I record at this stage Judge Berner’s description of the difficulty of the task before him, as he perceived it, and the 15 compounding of that difficulty by the parties’ differing interpretations of what UT 1’s direction required. He identified the approach he intended to adopt in this way:
“[33] … this was not a re-hearing in the true sense. I had before me the relevant documentary evidence from the original hearing before the FTT, and a transcript of the hearing. I heard submissions from Mr Barlow and Ms 20 Simor on those materials. But, unlike the original FTT, I did not hear the witnesses. As directed by the UT, I have regard to the findings made by the original FTT, except in relation to the allegedly impossible journeys. [34] In these circumstances it was perhaps inevitable that disputes would arise as to the nature of the submissions made by the parties. This manifested 25 itself most particularly in the submission by Ms Simor that I should refuse to entertain any argument on the part of SDM that the timings on documents produced by SDM as part of the evidence of the drivers were not in fact correct. Ms Simor argued in this respect that no evidence had been adduced to support such a submission, and that no such evidence would be admissible 30 at a second hearing when it could have been provided at the first hearing. For his part, Mr Barlow argued that the relevant documents were produced by SDM to corroborate merely the fact that, as the drivers testified, they were in on the relevant occasions in Belgium, and no reliance had been placed on them by SDM to prove precise timings. 35 [35] As a general matter it is clear that the UT expected, and indeed directed, that this tribunal should re-consider the question of the allegedly impossible journeys by reference only to the evidence that had been before the original FTT. However, it was at the same time envisaged that this tribunal would have a hearing so that the parties could put forward their rival submissions 40 on that issue. The extent to which new evidence might nevertheless be admitted is a matter for this tribunal; I shall in that connection refer to certain evidence concerning the purchase of vignettes at Folkestone, and useful information derived from various searches of Google Maps. But submissions that go to the weight to be attached to particular evidence that 45 was before the original FTT are to be given due regard. It is inevitable, and accordingly must have been envisaged by the UT when it made its directions, that the parties’ submissions at this renewed hearing would not follow precisely the lines adopted at the original FTT hearing. I view SDM’s submissions concerning the reliability of document timings as falling into 24 that category, and it is on that basis that I do have regard to those submissions.”
[76]Consistently with that indication Judge Berner caused enquiries to be made by the parties about the information to be derived from the vignettes purchased at Folkestone, namely whether the time of purchase recorded on them 5 was that in the UK or in continental Europe. Surprisingly, as he remarked himself, the information was inconclusive but he resolved it by finding as a fact that the time shown on all the vignettes, whether they were purchased at Folkestone or on the continent, was continental, that is one hour ahead of UK time. That conclusion is 10 not now challenged but in any event, for the reasons Judge Berner gave at [40] (that check-in times were a more practical starting point), the evidence to be derived from the vignettes turned out to be largely immaterial, although it did have some corroborative value as I shall explain.[77]Judge Berner then dealt at [45] to [47], as he said he would, with Mr 15 Barlow’s argument that the timings shown on the various documents could not be relied upon as they might have been inaccurate. He recorded, but did not accept or reject, Mr Barlow’s argument that F-tT 1 had, at least impliedly, accepted that the timings might not be accurate by their conclusion that, notwithstanding documents which showed journey 29 to have been impossible within the time 20 available if they were assumed to be accurate in every particular, the driver’s evidence was to be preferred with the consequence that they rejected the timings. I interpose that I do not think the argument is quite right; F-tT 1 did not reject the timings, but accepted that there was an explanation (a reduction in the number of vehicles required to transport the consignment, and consequent confusion between 25 vehicles) which overcame the apparent impossibility.[78]Judge Berner mentioned at [46] a decision of the Privy Council to which Miss Simor referred him, Grace Shipping v C F Sharp & Co (Malaya) Pte Ltd [1987] 1 Lloyds Rep 207, in which, at pp 215-6, it identified the importance of documentary evidence in any case when, whether by passage of time or otherwise, 30 the recollections of witnesses might have become unreliable. He accepted, nevertheless, that one should not slavishly accept the accuracy of contemporaneous documents regardless of other evidence. The relevance of the timings was particularly acute in respect of movement 29, and in his discussion of that movement Judge Berner said, at [140]: 35 “Mr Barlow argued that accounting documents could be wrong. An assertion that the documents prove that the movement could not have taken place depends, he submitted, on the absolute accuracy of the times stated on the Eurotunnel account. Whilst it may be the case, and I accept, that documents may not be entirely reliable, the difficulty for Mr Barlow, as Ms Simor 40 pointed out, is that the documents he seeks to impugn as unreliable are the very documents put forward in support of the case that Mr Blunsden delivered the goods to Aldi. It would not in my view be right to ignore the content of such documents because of some supposed possibility of inaccuracy, for which no directly applicable evidence was available. The fact 45 that inaccuracies might have been discovered in similar documents from time to time does not lead to the conclusion that the accounts relevant to Movement 29 must have been inaccurate.” 25[79]Although he did not say so, it is apparent from his analysis of the evidence relating to the remaining allegedly impossible journeys that he treated the timings on the documents as reliable, meaning the best available evidence of the time at which an event occurred, thus implicitly rejecting Mr Barlow’s argument. I deduce that he did so primarily because the argument was 5 not supported by any evidence.[80]As he indicated at [35] was his intention, Judge Berner drew extensively on information to be derived from Google Maps. He did not explain why he had decided to admit the evidence, but he also did not record any objection from the 10 parties, who did not argue before us that they had raised any such objection. It seems that some Google Maps evidence was before F-tT 1, but they did not mention it. Judge Berner recognised, at [44], that there was no evidence before him of the reliability of the information to be gleaned from Google Maps, but by the time the appeal came before us both parties were content to accept that the 15 information was sufficiently accurate, though with some reservations on Mr Barlow’s part which I shall explore later. At [47] Judge Berner observed that he, like F-tT 1, might prefer the evidence of the drivers over the documentary evidence; but, although he did not put it in quite this way, he evidently considered that it was first necessary to establish what he referred to as a benchmark, or base, 20 for the time required to undertake a typical journey from Coquelles to Vaux-sur- Sûre and return. He began by rejecting a submission made by Miss Simor to the effect that the time required had already been conclusively determined:
“[48] On the question of timings generally, I should at this stage address a submission made by Ms Simor that the UT had, at [58], found that the 25 minimum time for a round trip from the Eurotunnel terminal at Coquelles, France, to the Aldi depot at Vaux-sur-Sûre and back to Coquelles, including the picking up of a return load on the way back to Coquelles, was 10½ hours, and that I must proceed on that basis, as it was binding on this tribunal. 30 [49] I do not accept that submission. The starting point for the figure of 10½ hours was the original FTT’s own finding, at [453], that a round trip, including only unloading at Vaux-sur-Sûre, and without a break during the trip, could legally be achieved in 9 hours 45 minutes. The decision of the original FTT has been set aside, and accordingly there is no appropriate 35 starting point for the UT’s own calculation. In any event, it is perfectly clear from the directions made by the UT on its referral back that the issue of the impossible journeys is to be considered by me only by reference to the evidence that was before the FTT, and the findings of the FTT (and not any of the UT) on matters other than the impossible journeys. I do not therefore 40 consider that my task is constrained as submitted by Ms Simor.”
[81]HMRC renewed the same argument in their application for permission to appeal to this tribunal from F-tT 2’s decision. It was rejected upon the slightly different basis that UT 1 had not made a finding at [58] (quoted at para 66 above) that 10½ hours was the minimum time required for the round trip. They had 45 instead merely examined F-tT 1’s reasoning on the point on the way to their conclusion that the reasoning should be reconsidered by F-tT 2. Had UT 1 made a finding that 10½ hours was the minimum they would no doubt have said so in their direction, but instead they left the point at large. I agree therefore with Judge 26 Berner that he was not constrained as HMRC contended, and that he was required to consider the matter afresh.[82]Judge Berner then described the question he had to consider, and the manner in which he was to approach it, in this way:
“[50] … the question before me is not one of likelihood 5 or of probability. It is whether the journeys described by the drivers could have been made. If such journeys were indeed possible (or, to put it another way, not impossible), then, however unlikely it might be that the journey took place as so described, the finding would be that the journey could have been made. 10 Thus, to the extent that it was possible that drivers would not comply with the law, by driving faster than the relevant speed limit, or by failing to take the necessary breaks, those factors should not inhibit a finding that a journey could have been made. Issues of compliance with the law are nonetheless relevant to the cases of individual drivers where those drivers gave evidence 15 as to their own behaviour. [51] The test of possibility must be considered by first establishing a benchmark time for the journey and comparing that with the evidence of the actual timings. In the absence of any evidence directly addressing the issue of impossibility of particular journeys, that benchmark time falls to be 20 ascertained from a consideration of all the relevant evidence. Having then applied the benchmark to the particular journey, if it is found that the journey could not have been completed by the time of check-in at Coquelles, a judgment has to be made as to whether that time could possibly have been made up, or whether there could be some other explanation for the apparent 25 impossibility.”
[83]He then explained that in order to establish a benchmark time for a complete trip he proposed to use the interval between check-in at Folkestone on the outward journey and check-in at Coquelles on the return journey, as those times could be established from the documents produced by Eurotunnel if they were assumed to 30 be reliable. He observed that the time taken to travel from the warehouse in the UK at which the goods had been picked up to Folkestone might be relevant in some cases, because of the requirement that drivers of heavy goods vehicles must take prescribed breaks, but it is evident that he regarded the check-in times as the most useful evidence and he did not, in fact, return to the journeys from the 35 despatching warehouse to Folkestone.[84]Judge Berner then proceeded to examine the available evidence about the time which might elapse between check-in at Folkestone and arrival at Coquelles on the outward leg. Two of the drivers said that, although on occasion the journey could take much longer, in ideal circumstances the crossing time could be as little 40 as an hour, and Mr Cranny (who also said that the shuttle ran, probably, every 30 to 40 minutes) said “you can be motorway to motorway within an hour”, which the judge took to mean that it would take an hour from check-in at Folkestone to departure from Coquelles. At [57] he said: “From this I conclude that the benchmark, or base, time that it would take 45 from check-in time at Folkestone to arrival at Coquelles is one hour. Whilst it is the case that it is likely that some or all of the journeys would have taken longer, the question is one of possibility, and not probability or likelihood.…” 27[85]He added that he included within that hour the time taken to purchase a vignette, in those cases in which the vignette had been bought at Folkestone. I need to mention at this point that we were shown a document produced by Eurotunnel itself, directed to users or prospective users of its freight service including the automatic check-in procedure adopted by the drivers 5 in this case. Judge Berner did not mention it but we were told that it was before him. The document describes the manner in which the procedure works, which I do not need to explore, and mentions that the service operates for 24 hours each day, and that at peak times there are six departures per hour, suggesting that Mr Cranny’s 10 evidence to F-tT 1 was not wholly accurate. It then states that “The transit time between the truck checking-in and its arrival on the motorway on the other side of the channel is 90 minutes”.[86]Judge Berner then turned to the available evidence about the time needed to drive from Coquelles to Vaux-sur-Sûre. F-tT 1 had heard from Mr Cranny, who 15 did not drive any of the loads but nevertheless spoke (from what source of information is unclear) about the necessary time, from Mr Parnham and Mr Blunsden who had driven some of the allegedly impossible journeys, and from Mr Waters, who had been engaged by Connie to take some of the remaining consignments to the Aldi warehouse. He described their evidence at [59]: 20 “Mr Cranny’s evidence was that the journey would possibly take about four or four and a half hours. Mr Waters said that, depending on the traffic, it would normally take about four and a half hours from Calais (Coquelles) to Vaux-sur-Sûre. Mr Blunsden’s evidence for the same journey was that it would take between four hours and four and a quarter hours, although he 25 also told the original FTT that it could probably be done in three and a half hours. Modern vehicles could cruise comfortably at 70 to 80 mph, and speed limiters on the vehicles could be turned off. Mr Parnham’s evidence was equivocal; his starting point, when it was suggested by Ms Simor in crossexamination that other evidence had suggested a journey time of four and a 30 half hours, was that the journey times would vary, but that it could take ‘four, four and a half hours, maybe five. Maybe three and a half.’ Challenged whether a journey time of three and a half hours would have been possible, and that such a journey would certainly be over the speed limit (90 kilometres per hour, or about 56 mph), Mr Parnham agreed that four and a 35 half hours sounded reasonable, but said ‘don’t hold me to it’. All this evidence was on the basis of what might be described as a typical such journey; none of it related to any of the actual journeys under consideration, nor did any of it specifically describe the shortest possible time.”[87]At [61] Judge Berner provided a table showing three permutations of 40 possible round trips. The table adopted variations of route in both the outward and return journeys which had been suggested by Miss Simor, together with travelling times for those routes derived from Google Maps. The aggregate distances shown, depending on the route chosen, varied between 735 and 757 km, and travelling times ranging from 6 hours 59 minutes to 7 hours 15 minutes; those times 45 included no allowance for the time required for unloading and re-loading the vehicle or for rest breaks. The average speed necessary if the distance was to be covered in the stated time is shown in the table to be between 101.4 kph (63 mph) and 107.3 kph (66.7 mph). The table also included a single journey from Vauxsur- Sûre to Coquelles via two staging posts, in order to cover one of the allegedly 28 impossible journeys; here the distance shown is 409 km, and the average speed on the fastest leg is 108.7 kph (67.5 mph) and on the slowest 95.8 kph (49.5 mph [sic: a closer approximation is 59.9 mph]). It is clear from what else he said that Judge Berner was conscious of the fact that the speed limit for a heavy goods vehicle on French and Belgian motorways is and at the relevant 5 time was 90 kph (56 mph), significantly lower than that for cars.[88]At [62]

Judge Berner made this observation:

“In seeking to ascertain a benchmark journey time for this purpose, it would be wrong in my view to place determinative weight on evidence of drivers 10 who were answering a question as to typical journey times. They were not asked, and their answers did not address, the real question of the quickest time at which the journey could be undertaken. Nor, when the question is one of possibility, and not probability or likelihood, would it be right to make any assumption that the journeys would be undertaken without any 15 transgression of the law, such as the speed limit. It must be remembered that the reason the question of the impossible journeys was referred back to this tribunal is that it is only if a journey is impossible that it can cast doubt on the acceptance by the original FTT of the reliability of the drivers’ evidence that the various loads had in fact been delivered to Aldi.” 20 89. He then went on to remark at [63] to [65] that some of the drivers who had given evidence accepted that they would sometimes exceed the speed limit, albeit others denied doing so, and that some did not always respect the requirements of breaks. He made it clear, indeed, that he would proceed from the assumption that the speed limit applicable to heavy goods vehicles on motorways in France and 25 Belgium was consistently broken by the drivers: “[66] In those circumstances, although I accept that the timings ascertained via Google Maps are directed at car journeys and not those by HGV vehicles, there was no evidence that would lead me to conclude that the average speeds revealed by the Google Maps information could not be 30 achieved by the drivers on the journeys at issue in this case. I note that Mr Blunsden at one point in his evidence said that drivers were ‘not always as slow as [these] Google maps make out’, but that was in the context of seeking to show that Movement 29, which was clearly an impossible journey, had taken place. On that basis, there is nothing to persuade me that 35 the base time I should apply for this purpose is in principle materially less than that provided by Google Maps. Furthermore I reach my conclusion notwithstanding that there was evidence that the vehicles would carry over 40 tonnes when loaded; there was no evidence what effect that might have, apart from the fact that such a load would make the vehicle more stable in a 40 cross-wind. To conclude that benchmark times different from those identified by Google Maps should be applied would, it seems to me, require reliance on my own inexpert impression of the speeds at which HGVs, as opposed to cars, could travel on the roads in France and Belgium, which would be inappropriate. 45 [67] Having regard therefore to all the relevant evidence in this connection, I conclude that it is reasonable, as a starting point, to apply as the base times for the journeys set out in the table the shortest of those which are described in the table. Accordingly, I do not accept that the possibility of the relevant journeys must be assessed, as HMRC have done, by reference to a journey 50 time between Coquelles and Vaux-sur-Sûre of four and a half hours. The 29 timing that I have concluded should be applied in that respect is 3 hrs 26 mins. I accept that, if the matter was being considered on the balance of probability, it is unlikely that it would be concluded that all the allegedly impossible journeys would have been achieved on that basis, but that is not the test I have been 5 directed to apply.”
[90]It was necessary also to factor in the time required to unload and load the vehicle. As I have said, in the case of each of the allegedly impossible journeys the driver had collected another load on his return journey from the Aldi warehouse to Coquelles. Mr Blunsden’s evidence before F-tT 1 was that 10 unloading might, if all went well, take less than 20 minutes, but he also said that it could take as much as an hour. Other drivers, too, gave estimates of the time but none suggested that less than 20 minutes was possible and most agreed that the exercise including waiting time might take up to an hour. Judge Berner adopted 20 minutes as the minimum time for unloading, but added a further five minutes 15 for the time occupied by completing the necessary paperwork. He assumed that the same time was needed for re-loading the vehicles.[91]Having identified what he perceived to be the correct approach to each element Judge Berner went on to examine eight of the allegedly impossible journeys in detail—he omitted movement 44 because, as he commented at [71], it 20 had been incorrectly identified as an impossible journey, and at this initial stage (that is, while following para 2a of UT 1’s direction) he was not required to reconsider movement 29. HMRC agree that he was right about movement 44. He allowed for the time taken to travel to and from the places at which fuel was shown to have been purchased, and in one case (movement 22) he allowed 15 25 minutes for the time taken to refuel. He mentioned that he had supplemented the Google Maps information with which he was provided with further Google Map research of his own. By this means he aimed to identify the base times for each stage of the relevant journey. He then said:
“[69] I have then sought to apply those base timings to the journeys as 30 described in the evidence before the original FTT. The aim has been to ascertain whether, using those base timings, the journey as described could have been undertaken and the driver could have returned to Coquelles by the time of his recorded check-in at the Eurotunnel check-in there. Where the application of the base timings has the result that the driver could have 35 arrived at Coquelles before the relevant time, I have concluded that such a journey was not impossible. [70] In those cases where the base timings lead to the result that the arrival at Coquelles on the return journey would, on that basis, have been after the check-in time, I have reviewed the case to check whether that leads to a 40 conclusion that the journey must be regarded as impossible. I set out the methodology I have adopted in that respect later when considering such journeys.”
[92]Of the eight movements he analysed on that basis, Judge Berner found that six could have been accomplished in the time available as determined from the 45 documents—that is, the driver could have travelled from Folkestone to Vaux-sur- Sûre to unload, and then back to Coquelles, picking up a load on the way, in the interval between his check-in on the outward leg and his check-in on the return without exceeding the base times Judge Berner had determined. In two cases, movements 17 and 24, he found that the journey could not have been achieved in 30 the time available if the base times he had determined represented the minimum possible times. However, he did not leave the matter there, but went on to consider whether, despite the apparent impossibility, the journeys might nevertheless have been accomplished.[93]Movement 57 was one of those which Judge Berner 5 found to be possible within the base times he had determined, and it serves to illustrate his initial approach. Mr Blunsden checked in at Folkestone at 19.40 (UK time) on 12 November 2006, buying a vignette a few minutes later. It was deduced that he could have arrived at Coquelles at 21.40 (continental time), although in this case 10 the precise time is unimportant because it was also assumed he drove overnight to Vaux-sur-Sûre, arriving there when the warehouse was closed. The evidence showed that it opened at about 7 am, and on any view arrival before then was possible. Mr Blunsden could therefore have taken a break while waiting for the warehouse to open. As I have explained, Judge Berner assumed a total time for 15 unloading the vehicle of 25 minutes. Thus if he discharged his load immediately the warehouse opened Mr Blunsden could have left by 07.25, in order to travel to Geer to pick up his return load; that journey, according to Judge Berner, would take 1 hour 11 minutes. After loading (25 minutes) he would be able to leave at 09.01 and drive for 2 hours 28 minutes to Coquelles, arriving in ample time for 20 his check-in at 15.30.[94]At [124] to [128] Judge Berner dealt with the apparent impossibility of movement 24, in respect of which the calculations he made showed that the driver, Mr Wild, would have arrived at Coquelles 13 minutes after his recorded check-in time. He resolved that difficulty by the expedient of assuming that Mr 25 Wild reduced his journey time by 18 minutes, to allow a small margin. Judge Berner recognised that completion of the journey by then implied an average speed of 110 kph, or 68.4 mph. He nevertheless considered that such a speed was within the cruising capacity of the vehicle and that the movement was not impossible. 3095. Movement 17 began for present purposes at 09.24 (UK time) on 29 September 2006 when Mr Blunsden checked in at Folkestone. He bought a vignette there at 10.32 (continental time), and therefore a few minutes after checkin. Judge Berner assumed a one-hour journey, and that Mr Blunsden arrived at Coquelles at 11.24, continental time. That conclusion was, he said, consistent with 35 Mr Blunsden’s purchase of fuel at Watou, at 12.35; the base journey time from Coquelles to Watou, as Judge Berner had calculated it, was 49 minutes. The journey time from Watou to Vaux-sur-Sûre, derived from Google Maps, was 2 hours 46 minutes (294 km at an average of 106.2 kph or 66 mph). That led to an assumed arrival time of 15.21. 4096. Assuming, again, an unloading time of 25 minutes Mr Blunsden could have left Aldi’s warehouse at about 15.45. He picked up a load on his return to Coquelles, although there was some dispute, or at least uncertainty, about the place where he did so. The available paperwork suggested that the load was picked up in Illkirch (which is near Strasbourg, and a long way from both Vaux45 sur-Sûre and Coquelles) but that it might have been brought by another driver to Bruges or Gullegem in order that Mr Blunsden could collect the trailer there, an explanation Judge Berner accepted for the purpose of his analysis. The difference 31 in journey time, if Illkirch is discarded as a possibility, would amount to only a few minutes and Judge Berner assumed 4 hours 13 minutes including the loading time for the return journey, giving Mr Blunsden an arrival time at Coquelles of about 19.55, too late for his recorded check-in time of 18.46.[97]The shortage of time in this case, 1 hour 9 minutes, could 5 not be so easily resolved as that in respect of movement 24. At [130] Judge Berner mentioned that his initial calculation, which resulted in a late arrival, required an average driving speed of 101.72 kph, or 63.2 mph, over the total journey of 746.4 km and then said: 10 “[131] To achieve the check-in time at Coquelles, the time taken for the journey would have to be some 1 hr 9 mins shorter. Taking the time at 6 hrs 10 mins would give an average speed of 121 kph (75.18 mph). [132] This requires a finer judgment than was needed for Movement 24. If, without more, I was asked to decide if such a journey was likely, I would 15 conclude that it was not. But in such a case the evidence of the driver could persuade me that the journey was indeed undertaken. However, the question is not one of likelihood, balancing the evidence, but whether a journey is impossible, so that the evidence of the driver in that respect is regarded as unreliable. Were the average speed to have exceeded 80 mph, which was the 20 limit to the comfortable cruising speed referred to in the evidence before the FTT, I would have concluded that the journey was impossible. But where the average speed falls within the 70 – 80 mph range, I cannot conclude that such a journey, although unlikely, is impossible. [133] I therefore find that, although it is a marginal case, Movement 17 was 25 not an impossible journey.”[98]Judge Berner then turned his attention to movement 29. He accepted, as had F-tT 1, that the documentary evidence showed that the journey supposedly made could not have been accomplished in the time available, assuming the documents were reliable. Mr Blunsden was again the driver. According to the documentary 30 evidence he checked in at Folkestone at 03.58 on 14 October 2006, and again at Coquelles at 12.00 the same day. If he arrived at Coquelles on the outward journey at 06.00 (allowing for the hour’s difference) he was left with no more than six hours to drive a minimum of 735 km, in the process discharging one load and picking up another. Assuming 25 minutes for each of loading and unloading, 35 there remained a maximum of 5 hours 10 minutes driving time, implying an average speed of about 142 kph or 89 mph. Although he did not expressly say so, Judge Berner evidently took the view (as had F-tT 1) that such an average speed was quite implausible. He recorded that Mr Blunsden had offered no explanation of the difficulty to F-tT 1, but had merely asserted that he had made the delivery 40 to Aldi.[99]By contrast with what he had said at [132] in relation to movement 17, the judge did not accept that the driver’s evidence was sufficient to displace the documents. He recorded Mr Barlow’s argument that the documents might have been wrong, but pointed out, as I have said, that it was SDM which had 45 introduced them, and that it had produced no evidence which cast any doubt on their accuracy. He also rejected the possible explanation accepted by F-tT 1 (that there might have been confusion about the lorry which carried the load) because, again, there was no evidence to support it and Mr Blunsden had in any event 32 insisted that he delivered the load to Aldi’s warehouse. For those reasons he concluded that movement 29 was to be distinguished from the others as the journey was indeed impossible in the time available.[100]It is apparent from what he said next that Judge Berner recognised that he needed to reconcile his acceptance of Mr Blunsden’s evidence 5 in respect of most of the movements for which he was responsible and his rejection of it in respect of movement 29. He addressed the point as follows:
“[144] The original FTT had the benefit of seeing the drivers, including Mr Blunsden, give evidence. I have had only the transcripts of the evidence they 10 gave. It is evident from the passage from the transcript of the evidence given by Mr Blunsden that his evidence did not comprise a detailed recollection of the particular journey in question. Of course it is possible that Mr Blunsden recalled the journey perfectly, and was not telling the truth when he asserted that the load had been delivered to Aldi. But that was not the conclusion 15 reached by the original FTT. Having decided themselves that Movement 29 was impossible, the original FTT nevertheless accepted the truth of Mr Blunsden’s evidence about all his other journeys (which, in common with my own findings in those respects, they found not to be impossible), concluding, at [463], that Mr Blunsden simply did not recall the journey 20 related to Movement 29. [145] Having considered the transcript of Mr Blunsden’s evidence, I find that there is no reason to conclude differently from the original FTT. It is not possible, at this distance from the evidence, to conclude that Mr Blunsden recalled Movement 29, that he knew that the goods had not been delivered to 25 Aldi, and that he deliberately misled the tribunal in that respect. Mr Blunsden did not say that he could not recall the journey, but that was the finding of the original FTT, having heard his evidence. His evidence that he had delivered the load must be regarded as simply an assertion on his part that all the loads carried by him that had been destined for Aldi had arrived 30 there. Although that cannot be accepted in relation to Movement 29 on the evidence in relation to that movement, that does not lead to the conclusion that Mr Blunsden’s evidence in relation to the other journeys must be regarded as untrue, or that it must be concluded that SDM have failed to discharge the burden of proving, in relation to those other journeys, that the 35 goods did arrive at Aldi.”

Judge Berner set out his answer to the questions posed by UT 1’s direction at [147]:

“I have reached the following conclusions on the issues directed by the UT to be considered: 40 (1) Direction 2a. I have determined that none of the allegedly impossible journeys (other than Movement 29) could not have taken place as described in the evidence of the drivers as recorded in the original FTT’s decision; (2) Direction 2c. I conclude that, in relation to Movement 29, which 45 was an impossible journey, the effect of that finding is that the conclusion in the original FTT’s decision that SDM had discharged the burden of proof that Mr Blunsden had delivered that consignment to Aldi cannot be supported. 33 That conclusion does not, on the other hand, have any effect on the conclusions of the original FTT that, in respect of any other journey, the relevant goods were delivered to Aldi, whether those journeys were alleged to be impossible (which I have found could have taken place), or were journeys not alleged 5 to be impossible. (3) Direction 2d. I determine the appeal by dismissing the appeal in relation to Movements 29 and 65, and otherwise allowing the appeal.”
[102]The reference to movement 65 reflects what appears to be a slight confusion in UT 1’s decision. There had been no appeal against F-tT 1’s decision in respect 10 of that movement but UT 1 set aside the decision of F-tT 1 as a whole. The confusion is unimportant as SDM accepts that it is liable for the duty on the goods comprised in that movement. SDM’s proposed appeal[103]I put the heading of this section of my decision in that way because of a 15 disagreement between the parties about the interpretation of the relevant rule, namely rule 24(3)(e) of the Tribunal Procedure (Upper Tribunal) Rules 2008, which requires a respondent to an appeal to serve a response which states, among other things, “the grounds on which the respondent relies, including, in the case of an 20 appeal against the decision of another tribunal, any grounds on which the respondent was unsuccessful in the proceedings which are the subject of the appeal, but intends to rely in the appeal.”[104]Miss Simor argued that this provision did not allow a respondent to crossappeal, and that SDM had therefore to seek permission to appeal, which it had not 25 done, before we could entertain its arguments. Mr Barlow argued that, on the contrary, sub-para (e) made it clear that a respondent could appeal as of right, and that permission was not required.[105]It may be that the rule could be more clearly worded, but I am satisfied that Mr Barlow is right and that a largely successful appellant, as SDM was, before the 30 First-tier Tribunal can appeal, without first securing permission, against any finding in the First-tier Tribunal’s decision adverse to it once it has been served with notice of appeal to the Upper Tribunal by the party largely unsuccessful before the First-tier Tribunal.[106]SDM’s appeal is against F-tT 2’s finding in respect of movement 29, and 35 the focus of its challenge is Judge Berner’s statement, at [30], that UT 1 had “conclusively determined” that movement 29 could not have taken place as SDM claimed. Its argument is that UT 1 decided no more than that the journey could not have been accomplished in the time available if the documents were taken at face value (which is what F-tT 1 had also decided), and that in consequence Judge 40 Berner had incorrectly restricted the scope of the exercise he was required to conduct in respect of this movement. It seems to me that these are grounds on  The identity of the relevant sub-para was the subject of some debate in the parties’ skeleton arguments. The confusion seems to stem from an error in the version of the rules published on the tribunal’s website: the lettering of the sub-paragraphs jumps from (a) to (c). Thus what should appear on the website as (e) in fact appears as (f). 34 which SDM “was unsuccessful in the proceedings which are the subject of the appeal”, and that they fall within the scope of the rule. The rule does not impose a requirement of permission, and I see no basis on which such a requirement might be inferred. 5 HMRC’s submissions[107]The first of the two grounds on which permission to appeal was granted to HMRC is that Judge Berner did not, as UT 1’s direction required, confine himself to the evidence before F-tT 1. In essence, the argument as it was put to us is that he made findings based not on the evidence actually given, but on suppositions, 10 some based on what the witnesses might have said had they been asked different questions, which led him to unreal and unsustainable conclusions. The second is that he adopted a test of scientific or theoretical possibility, rather than realistic possibility, in determining whether or not (as was the ultimate question) the movements were properly discharged, even if discharge was to be established by 15 showing no more than that the goods arrived at the relevant warehouse. There is a good deal of overlap between those grounds and they were not treated as wholly discrete questions, by either side, in argument before us.[108]Miss Simor submitted that Judge Berner made a number of errors encompassed by those grounds of appeal. The first, to which I have already 20 alluded, was that he allowed SDM to contend that the timings on the documents it had itself submitted were not accurate, even though SDM had not made such a submission before F-tT 1 or UT 1. He was, she said, wrong to allow Mr Barlow to advance the argument that the documents might not be reliable. They were produced by SDM, as part of the evidence supporting its case that the goods were 25 delivered to their continental destinations. No submissions had been made to F-tT 1 or UT 1 that they were in any way inaccurate, nor could such submissions have been made as there was no evidence before F-tT 1 to show that the documents might have contained any errors. There was also no evidence (as opposed to assertion) of inaccuracy before Judge Berner. 30 109. Thus Judge Berner’s statement at [47] that “if the timings show a journey to have been impossible, the conflict between that and the evidence of the drivers might be resolved by preferring the evidence of the drivers to the stated timings” revealed an approach to the conflict which was not open to him. It was the very error in F-tT 1’s conclusions which UT 1 had identified at [69], and they were 35 right to say that the conflict between the timings and the drivers’ evidence could not be resolved by simply accepting the latter; at the very least the reason for doing so must be properly explained. But in any event the times stated on the documents had to be respected, because they represented the evidence before F-tT 1—thus the requirement imposed on Judge Berner was to determine whether the 40 journeys could have been completed within those times, and not to set the times aside if they were inconsistent with the drivers’ evidence.[110]The second error lay in the judge’s understanding, set out at [50], that he was required to decide whether the relevant journeys could theoretically have been made, “however unlikely it might be”. He should instead have asked himself 45 whether, on the balance of probabilities, the journeys could have taken place within the relevant times, having regard to the evidence before F-tT 1 and common sense. Instead, he had speculated about the evidence the witnesses might 35 have given, particularly about the speeds they could achieve on French and Belgian motorways, if they had been asked what was the shortest time in which they could drive from Coquelles to Vaux-sur-Sûre. He had failed to pay heed to the evidence actually given, that is what the drivers had said when asked what was the time necessary for 5 a typical journey.[111]That failing led him into the third error, which lay in his determination of the base times for the crossing from Folkestone to Coquelles, for the journey from Coquelles to Vaux-sur-Sûre and for unloading at Aldi’s warehouse. The mistake in his approach was revealed by what he said at [62] (quoted at para 88 above). In 10 essence, in the absence of direct evidence from the drivers of the shortest time in which a journey could be accomplished, Judge Berner had selected the shortest time any of the witnesses had mentioned as a possibility and had uncritically accepted that as the base time. In reality the drivers had given no more than brief and rather vague evidence on the topic. Thus, for example, Mr Blunsden had said 15 of the time which might elapse between check-in at Folkestone and arrival at Coquelles:
“On a good day, if everything went absolutely perfect, you could probably do it within an hour or an hour and a half.”
[112]The higher estimate was consistent with what was shown in the Eurotunnel 20 document (see para 85 above) and if Judge Berner was to determine the appeal by reference to base times, he should have adopted, for this part of the journey, the 90 minutes referred to in that document. The base time of 60 minutes he had accepted was not consistent with the evidence actually given to F-tT 1.[113]Similarly, when Mr Parnham was asked about the time needed to drive from 25 Coquelles to Vaux-sur-Sûre, the following exchange took place: “Mr Parnham: Four [hours], four and a half, maybe five. Maybe three and a half. Miss Simor: Is it possible in three and a half? I mean, that seems — Mr Parnham: I can’t remember to be honest.” 30 114. There was no warrant in UT 1’s direction for adopting the approach of assuming the minimum times mentioned in evidence given in that imprecise fashion, and it followed that the resulting finding, that the benchmark time, even before adjustment for movements 17 and 24, of a few minutes more or less than 3½ hours for the journey from Coquelles to Vaux-sur-Sûre or return could not be 35 regarded as consistent with the evidence given to F-tT 1. UT 1’s direction required the judge to adopt the evidence actually given to F-tT 1, which related to typical journey times; instead he had speculated about minimum possible times, and in doing so had accepted what were little more than guesses.[115]In similar vein, even though he recorded at [63] that the drivers had given 40 evidence that they did not exceed the speed limits (albeit Mr Blunsden had conceded that on occasion he might do so), Judge Berner’s conclusions necessarily implied not merely that the drivers might occasionally exceed the speed limit, but that they did so consistently and for long periods.[116]I have mentioned already that none of the tachograph discs relating to the 45 journeys was produced in evidence. Their absence was the topic of some debate 36 before F-tT 1 and was in particular dealt with in the course of the crossexamination of Mr Parnham. Miss Simor took us to a passage in the transcript in which Mr Parnham was asked to explain the absence of his tachograph discs, and it was put to him that he had failed to produce them because they would reveal not only that he participated in the diversions but also that he was 5 exceeding the speed limit. Mr Parnham agreed that he had attempted to complete the journeys as quickly as possible, because he had inadequate insurance cover for the load that he was carrying and was concerned about the risk of theft, but that he had done so by shortening or ignoring his break times, and not by exceeding the speed limit. 10 That evidence formed no basis for a finding that he had consistently exceeded the speed limit, and by a considerable margin.[117]The error of Judge Berner’s approach was, Miss Simor said, plain from his conclusions that movement 24 was possible if it was driven at an average speed for the whole journey of 110 kph, or 20 kph above the motorway speed limit, and 15 that movement 17 was possible at an average speed of 121 kph, or 31 kph over the limit. It was a fanciful proposition that any driver could have achieved such speeds. In fact, not all of the journey could be undertaken on motorways—there were non-motorway sections at the beginning and end of each journey—and that fact made the average speeds he determined even more implausible. 20 118. Although these were the extreme cases, the remaining allegedly impossible journeys, too, could not realistically have been accomplished in the times available. UT 1 had remitted the appeal to F-tT 2 in order that it could address the problem which UT 1 had identified at [69], that is that the drivers’ evidence could not be reconciled with the documents. That could not be done by simply 25 assuming, despite the lack of evidence before F-tT 1 supporting such a conclusion, that the drivers must have driven at sufficient speed to accomplish their journeys within the times shown by the documents.[119]At [66] Judge Berner remarked that “although I accept that the timings ascertained via Google Maps are directed at car journeys and not those by HGV 30 vehicles, there was no evidence that would lead me to conclude that the average speeds revealed by the Google Maps information could not be achieved by the drivers on the journeys at issue in this case.” Even SDM had made no submission that the times applicable to cars could be applied to heavy goods vehicles; on the contrary, it had accepted before F-tT 1 that the speed limit for such vehicles was 35 lower (90 kph rather than 120 kph in Belgium and 130 kph in France for cars) and that in consequence the maximum speed at which the lorries could travel was also lower; accordingly the Google Maps times, which were demonstrably those which could be achieved by cars, could not be treated as a reasonable guide to the times which might be required for a heavy goods vehicle to travel from one point to 40 another. It was in any event a matter of ordinary common sense that it would not be feasible to drive a heavy goods vehicle over a long distance at a constant speed of 20 or 30 kph above the permitted limit, as Judge Berner had found in two cases. Moreover, an average speed of 110 or 120 kph implied a higher speed over some parts of the journey to compensate for any need to slow down at other times. 45 Those findings could not stand, and they demonstrated that Judge Berner’s whole approach was flawed. 37[120]His exclusion of rest or break times from his calculations was likewise inconsistent with what the drivers had said. That was apparent from what

the judge himself recorded when describing their evidence at [65]:

“As regards the need for drivers to take breaks required by law, both Mr Cranny and Mr Waters confirmed the legal requirement 5 for minimum breaks. So too did Mr Blunsden. Mr Blunsden’s evidence was that if he had driven a full nine hours to his destination he would be required to take a break there of nine or eleven hours. He would never not bother to take this break and simply head straight back to Coquelles; he would park his vehicle 10 and go to sleep. Although from this evidence I can conclude that Mr Blunsden would generally take a break when he reached his destination, it does not persuade me that I should factor provision for breaks into the base time for journeys generally. It seems to me perfectly possible in practice that a driver could drive for a full nine hours without a break, and possibly longer 15 having regard to the evidence of Mr Parnham. Even Mr Blunsden, in reexamination, admitted that he would not always observe break times.”
[121]Similarly, at [76], the judge accepted 25 minutes—20 minutes for loading and five minutes for paperwork— representing Mr Blunsden’s evidence that this time was achievable if everything “really went well”. But the drivers, including 20 Mr Blunsden, had stated that it might well be necessary to wait while other vehicles were dealt with and that the exercise could take significantly longer, and it was unrealistic to conclude that 25 minutes could be routinely achieved, and was an appropriate time to adopt in respect of each movement. Miss Simor relied on this point, however, only as support for her primary argument that Judge 25 Berner’s approach to the assessment of the times required to drive the various journeys was flawed; she did not mount a discrete attack on the brevity of the loading and unloading times Judge Berner had adopted, or the limited extent to which he had allowed for rest breaks in his calculations.[122]In short, Miss Simor said, Judge Berner had failed to recognise that the test 30 was not whether the journeys might, theoretically, have been accomplished in the time available, but whether (as UT 1’s direction put it) they could have taken place “as described in the evidence of the drivers”. None of them claimed in their witness statements or in their oral evidence that they had driven at speeds as high as those determined by the judge, and none of them had claimed that they had 35 completely disregarded the requirement to take breaks at prescribed intervals. He was required to be satisfied, on the balance of probabilities, that the journeys did in fact take place, not that, if one made various assumptions for which there was no evidence before F-tT 1, they might have been possible. The conclusions the judge reached on journey times as a result of his wrong approach are quite 40 unrealistic. Miss Simor argued that it was necessary to take a common sense view, reflecting normal experience of traffic conditions, and not the artificial view he had in fact taken.[123]Judge Berner’s approach to the consequences which followed from his conclusion that movement 29 could not have been delivered as SDM claimed was 45 illustrative of his error. F-tT 1 had found that it was an impossible journey if the documentary evidence were to be regarded as determinative, but had gone on to accept Mr Blunsden’s assertion that the goods had nevertheless been delivered to the Aldi warehouse. Although Judge Berner had agreed with F-tT 1 that the documents demonstrated the impossibility of the journey and, unlike F-tT 1, he 38 had rejected the assertion of delivery (and with it the explanation of confusion in the identity of the vehicle which F-tT 1 had introduced) he did not go on to consider whether, in the face of Mr Blunsden’s claim that he had delivered this load when plainly he could not have done, the remainder of his evidence was reliable. Instead, he had adopted the same uncritical approach 5 as F-tT 1, a course which undermined the purpose behind UT 1’s remitting the matter for reconsideration. SDM’s submissions[124]In the respondent’s notice served in this appeal on behalf of SDM (drafted 10 by Mr Barlow) it is said that “it would be easy to forget that what is actually still in issue is whether [F-tT 1’s] finding that the movements did take place has been shown by [HMRC] to be an error of law”. In his skeleton argument Mr Barlow expanded on that point by arguing that UT 1 had been wrong to reach the decision they did: it was, he said, open to them to set aside the decision of F-tT 1 only if 15 they concluded that the findings of fact were perverse in the sense explained in Edwards v Bairstow [1956] AC 14, Secretary of State v SH (Sudan) [2007] UKHL 49, [2008] 1 AC 678 and Georgiou v Customs and Excise Commissioners [1996] STC 463. They had not done that; instead they had allowed HMRC’s appeal because they found that F-tT 1 had provided insufficient reasons for their findings 20 of fact, and they had done so even though insufficiency of reasons was not an argument advanced by HMRC before UT 1.[125]UT 1’s approach in this respect was wrong, Mr Barlow said, because the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009, by rule 35(3)(b), require that the tribunal provides “full written findings of fact and 25 reasons for the decision”. That requirement cannot be read in a way which places an obligation on the tribunal to provide reasons for its findings of fact, and it follows that a challenge may not be made to the decision merely because the findings of fact are unexplained. In any event Mr Barlow did not accept that F-tT 1’s reasons for accepting the drivers’ evidence was inadequately explained. 30 126. The authorities draw the same distinction. In Meek v Birmingham City Council [1987] IRLR 250 Sir Thomas Bingham MR said, at para 8 that:
“It has on a number of occasions been made plain that the decision of an Industrial Tribunal is not required to be an elaborate formalistic product of refined legal draftsmanship, but it must contain an outline of the story which 35 has given rise to the complaint and a summary of the Tribunal’s basic factual conclusions and a statement of the reasons which have led them to reach the conclusion which they do on those basic facts. The parties are entitled to be told why they have won or lost. There should be sufficient account of the facts and the reasoning to enable the EAT or on further appeal this court to 40 see whether any question of law arises ….”
[127]Similarly, in Farnocchia v Customs and Excise Commissioners [1994] STC 881 at p 884 Lord Hope, delivering the judgment of the Inner House of the Court of Session, observed that a decision is required to demonstrate: “… that the tribunal reached its decision in conformity with the statutory 45 procedure and the principles of natural justice, and the basis of the decision must be made clear … a value added tax tribunal must state what facts it found to be admitted or proved on all the matters it was required to decide 39 and what the reasons were for the decision which it reached on all points which were in issue in the appeal”.[128]Mr Barlow submitted that those passages made it clear that a distinction is to be drawn between the findings of fact, which must be stated clearly and as fully as the case requires, and the decision itself; the latter must 5 be reasoned but the former need not be. UT 1’s reliance on R (Iran) v Home Secretary [2005] EWCA Civ 982 for the contrary view was misplaced because even in that case, which in any event related to the differently worded rules of the Immigration and Asylum Tribunal, the court had not said that it was necessary to give reasons for findings 10 of fact. In addition, Mr Barlow argued, a failure to give reasons was not in itself a basis on which an appeal might be allowed. As the Court of Appeal made clear in English v Emery Reimbold & Strick Ltd [2002] EWCA Civ 605, [2002] 1 WLR 2409, the requirement was no more than that the fact-finding tribunal should explain its decision in sufficient detail as to enable the parties to know why they 15 have won or lost.[129]In Mountview Court Properties v Devlin (1970) 21 P & CR 689 at 695 Lord Parker CJ said:
“It seems to me that there was here a lack of adequate reasons for the decision … failure to give reasons pursuant to the duty imposed by section 20 12 of the Tribunals and Inquiries Act 1958 is not per se a ground on which the court could properly allow an appeal under section 9, the right of appeal being conferred on a person who is dissatisfied in point of law with a decision. That language, and, indeed, any analogous language found in the statutes giving a right of appeal on a point of law, to my mind connotes that 25 a successful appellant must demonstrate that the decision with which he is dissatisfied is itself vitiated by reason of the fact that it has been reached by an erroneous process of reasoning.”
[130]Mr Barlow emphasised the concluding words: it was incumbent on an appellant, if he was to succeed, to show that the decision under appeal was 30 reached by an erroneous process of reasoning. It is not enough that the appellate tribunal might have come to a different conclusion, and it is not enough that the reasoning is not set out—it must be shown to be erroneous. In addition, UT 1 had misunderstood s 12(2) of the Tribunals Courts and Enforcement Act 2007. They thought that the subsection required them to do one of two things, that is remake 35 the decision or remit the appeal. But the subsection expressly does not oblige the Upper Tribunal to set aside a decision of the First-tier tribunal even where an error of law has been found and when, as here, the error (if it was an error at all) amounted to nothing more than a failure to record reasons, the course of leaving F-tT 1’s conclusions undisturbed would have been appropriate. 40 131. I interpose that Miss Simor’s response to that argument was that Mr Barlow was quite wrong to say that we, sitting at the same level, have any jurisdiction to consider the correctness of UT 1’s decision. UT 1 rejected SDM’s subsequent application for a direction that their own decision be set aside, and SDM did not appeal against either the decision or the refusal to set it aside. Thus we could not 45 reconsider what UT 1 had said but must treat this appeal for what it is, an appeal against F-tT 2’s decision and nothing else. I shall return to this disagreement in due course. 40[132]Mr Barlow did not, of course, rely only on a critique of UT 1’s decision, but dealt fully with HMRC’s appeal against Judge Berner’s decision. He went to considerable trouble to persuade us that HMRC’s attack was on Judge Berner’s findings of fact, since both parties had agreed that the only issue before him was whether the goods in question in any or all of the movements 5 reached their stated destinations, and that was plainly a question of fact. The well-known authorities on the topic to which I have already referred show clearly that the essential question is not whether an appellate tribunal, hearing the same evidence, could or would come to the same conclusion but whether the finding was one which the 10 first-instance tribunal was entitled to make.[133]It was important to remember, said Mr Barlow, that only F-tT 1 had heard the witnesses’ live evidence. They had clearly accepted that evidence even when it conflicted with the documentary evidence. It was necessary to bear in mind that FtT 1 also heard HMRC’s evidence but nevertheless reached the conclusion that, 15 on the balance of probabilities, the goods destined for Belgium and Latvia had arrived in those countries. Against that background any appellate tribunal should be very careful before coming to conclusions contrary to those of the tribunal which had heard the live evidence. It was important too to remember that SDM had not failed in respect of the German consignment because F-tT 1 disbelieved 20 its evidence but simply because it could not produce any documents demonstrating regular arrival of the goods.[134]Mr Barlow added that, if one had to choose between the two competing scenarios, that advanced by HMRC was inherently the less probable. It was, he said, implausible that the criminals behind the fraud could discover which of the 25 drivers would take the loads, when they had been selected for the purpose only shortly before the goods were collected and when the drivers’ names had not been made known to the despatching warehouse. No theory, let alone evidence, was advanced which might explain how the criminals could have succeeded in identifying the drivers. By contrast, the proposition that one or more dishonest 30 employees at the Aldi warehouse could have diverted the goods immediately after their arrival, in collusion with the criminals, fitted with the evidence of the conspiracy which HMRC had obtained from the Belgian authorities and was the more probable position.[135]The kernel of his argument about movement 29 was that UT 1’s direction 35 required F-tT 2 to take account of all of the evidence before F-tT 1, and to pay heed, importantly, to the fact that F-tT 1 had believed the drivers’ evidence to the effect that they had delivered the goods. The differential treatment of movement 29 was necessary only because of the conclusion that, if the documents were right and were taken alone, it was impossible; it did not carry with it the implication 40 that the remaining evidence about that movement was to be disregarded. Instead, paras 2c and 2d of the direction, quite deliberately, did not exclude movement 29 from further consideration once F-tT 2 had reached its conclusions about the other nine (now eight) allegedly impossible journeys. HMRC’s response 45 136. HMRC resist SDM’s cross-appeal primarily because, Miss Simor said, it was not until the remitted appeal came before F-tT 2 that SDM claimed that the documents were inaccurate; it had put them forward to F-tT 1 and UT 1 as 41 evidence that the journeys did in fact take place, and had not suggested before either of those tribunals that the documents were not wholly reliable. Judge Berner was right to say that UT 1 had conclusively determined that movement 29 was impossible, a determination which SDM had not sought to appeal; it was therefore not open to it to argue before F-tT 2 or before us that it 5 was nevertheless possible. Contrary to Mr Barlow’s argument, UT 1’s direction distinguished between movement 29 and the remaining allegedly impossible journeys precisely because there was a distinction to be drawn.[137]At [65] UT 1 criticised F-tT 1 for their failure to explain why they “felt able 10 to ignore movement 29, which was clearly impossible on the documentary evidence of timings”. It was apparent from what UT 1 said at [77], that they were “unable on the material before us fairly to reach any conclusion as to whether or not any of the nine allegedly impossible journeys, apart from movement 29, were in fact possible”, that they recognised that there was a material difference between 15 movement 29 and the remaining allegedly impossible journeys. Against that background the omission of any reference to that movement in paras 2c and 2d of the direction could not be taken as an indication that F-tT 2 was nevertheless expected to reconsider it. Discussion 20 138. I deal first with Mr Barlow’s argument that UT 1 were wrong to criticise FtT 1 for their failure to provide adequate reasons for their conclusion that the various journeys, including movement 29, were possible. I agree with Miss Simor, broadly for the reasons she gave, that it is not open to us to reconsider UT 1’s decision. The appeal has reached us as one against F-tT 2’s decision, and that is 25 the only matter before us.[139]That short point is enough to dispose of the argument. I think it appropriate, however, to add two further observations. First, I do not agree with Mr Barlow that UT 1 set aside F-tT 1’s decision merely because they failed to explain why they preferred the evidence of the drivers to the documents where the documents 30 showed, or seemed to show, that the relevant journeys could not have been accomplished in the time available. The critical paragraph of UT 1’s decision in this respect is [69] (quoted at para 67 above) in which they said that the challenge to the drivers’ evidence which the documents posed could not be resolved by simply preferring the drivers’ evidence. I agree. This was not merely a lack of 35 explanation; it represents a failure to address and resolve a crucial issue in the case.[140]Second, I am unable to accept that the authorities to which Mr Barlow referred support the proposition that there is no requirement upon a tribunal to give reasons for its findings of fact. In many cases, such reasons might be 40 unnecessary either because they are obvious or because they can be readily gleaned from the findings themselves. But in others, the determination of which version of events is true, or which witness is to be preferred to another, may be critical to the decision. In such circumstances it would, in my judgment, be a dereliction of duty for the judge simply to set out his conclusions without any 45 reasoning. 42[141]Much the same was said by Patten LJ, with whom Hallett and Christopher Clarke LJJ agreed, in Weymont and another v Place [2015] EWCA Civ 289. After describing the recognised restrictions on the interference by an appellate tribunal with the findings of fact of an inferior tribunal, he made these observations: “[4] But the relative immunity of the trial judge’s 5 findings of fact to interference on appeal depends upon the trial process having been conducted in a way which confirms that the trial judge has properly considered and understood the evidence; has taken into account the criticisms of the evidence advanced by the parties’ legal representatives; and has reached a 10 balanced and objective conclusion about points on which differing or inconsistent evidence has been given in making the factual findings which form the basis of his decision. [5] An important aspect of this process is the production of a properly reasoned judgment which explains to the parties and to any wider readership 15 why the judge has reached the decision he has made. This includes making a reference to the issues in the case; the legal principles or test which have to be applied; and to why, in cases of conflicting factual evidence, the judge came to accept the evidence of particular witnesses in preference to that of others. 20 [6] The judge is not, of course, required to deal with every point raised in argument, however peripheral, or with every part of the evidence. The process of adjudication involves the identification and determination of relevant issues. But within those bounds the parties are entitled to have explained to them how the judge has determined their substantive rights and, 25 for that purpose, the judge is required to produce a fully reasoned judgment which does so: see English v Emery Reimbold & Strick Ltd [2002] EWCA Civ 605. The production of such a judgment not only satisfies the court’s duty to the parties but also imposes upon the judge the discipline of considering the detail of the evidence and the legal argument.” 30 142. Those observations are, in my view, quite inconsistent with Mr Barlow’s argument, which is to be rejected on this ground as well.[143]I have set out Mr Barlow’s proposition that the findings of fact made by both F-tT 2 and, so far as they remain relevant, F-tT 1 are, if not sacrosanct, susceptible of challenge only if they can be shown to be irrational fairly briefly 35 because it does not seem to me that the proposition is controversial. Moreover, as he argued himself, UT 1 did not reject the majority of F-tT 1’s findings of fact— on the contrary, they required F-tT 2 to respect them, at least at the first stage of its reconsideration—but remitted the appeal because of F-tT 1’s failure to explain how they had reconciled the documentary evidence showing that some of the 40 journeys were impossible, at least if the recorded times were correct, with the drivers’ evidence that they had nevertheless accomplished them. I should add for completeness that I agree with Mr Barlow that there is nothing relevant to this case in what was said by Lord Carnwath in Pendragon plc v Revenue and Customs Commissioners [2015] STC 1825 at [44] ff on the scope of the Upper 45 Tribunal’s jurisdiction in respect of the First-tier Tribunal’s findings of fact.[144]Although, as I have already said, UT 1’s direction might have been more clearly worded, their decision itself is clear. At [101] they said: 43 “We direct that the Decision be set aside and the case be remitted to a differently constituted First-tier Tribunal to determine whether the allegedly impossible journeys took place and consider what effect its conclusion has on the evidence in relation to the other deliveries.”[145]I read that as a direction that F-tT 1’s decision be set aside 5 in its entirety, and that it be re-made (by F-tT 2) in a prescribed manner. The prescription included the requirement that F-tT 2 should respect F-tT 1’s findings of fact at the first stage of the enquiry save in relation to the allegedly impossible journeys. If F-tT 2 were to conclude that some or all of those journeys were not possible it was 10 then to move on to para 2c of the direction which required it(i) to consider the impact of its conclusion that some or all of the allegedly impossible journeys were in fact not possible on F-tT 1’s determination that the goods comprised in those movements were delivered; and(ii) to consider the impact of its first stage conclusion on F-tT 1’s determination that the goods comprised in all the other 15 movements had been delivered. In other words, at the second stage of the enquiry F-tT 2 was required to revisit and, if appropriate, re-make, F-tT 1’s decisions about the remaining journeys, including movement 29 but (since there was no appeal in respect of it) not including movement 65.[146]I accept therefore that UT 1’s decision does represent an encroachment, 20 even if a limited encroachment, on the relative immunity, as Patten LJ described it, of F-tT 1’s findings of fact. But I do not think the encroachment takes Mr Barlow anywhere. First, SDM did not appeal against either the decision or the direction and, for the reasons I have already given, it is not open to us to reconsider them. Second, save for movement 29, Judge Berner did not revisit F-tT 25 1’s findings; rather, having reached the conclusion that, with the exception of movement 29, the allegedly impossible journeys were in fact possible, he did not need to reconsider their findings about the remaining journeys. Third, if it is accepted that Mr Barlow’s own argument to the effect that all of the journeys to Aldi’s warehouse took place or none did is right, it is inevitable that a conclusion 30 that one or more of the allegedly impossible journeys could not have taken place would alter the proper perception of the remaining journeys. I do not see how UT 1 could sensibly direct a reconsideration of the allegedly impossible journeys while preserving F-tT 1’s conclusions about the remaining journeys regardless of the outcome of the reconsideration when, as the extracts from F-tT 1’s decision 35 which I have set out show, they were themselves influenced by their conclusions relating to the allegedly impossible journeys in their appraisal of the journeys for which there was little evidence, and in particular no evidence from the driver.[147]It is, I think, convenient to deal with the grounds on which HMRC were granted permission to appeal in reverse order. The basis of the second ground was 40 that Judge Berner asked himself whether any of the allegedly impossible journeys could have been accomplished by reference to a theoretical, rather than practical or reasonable, test. I agree with Miss Simor that this is what he did, as my description of his approach to movements 17 and 24 shows, and I also agree with her that it was not the task set by UT 1. It seems to me that the use in para 2a of 45 UT 1’s direction of the potentially ambiguous phrase “could not have taken place” may have led to some confusion about the nature of the task before F-tT 2. That confusion is, I think, reflected in what Judge Berner said in the final sentence of 44 [62] (quoted at para 88 above), to the effect that it was only if the journeys were impossible that doubt would be cast on the drivers’ evidence of delivery.[148]The true question in the appeal is not whether the journeys were possible, albeit that is a relevant factor, but whether SDM has shown that the goods were delivered. Mr Barlow argued during the course of the 5 hearing that HMRC had advanced several theories about the manner in which the goods might have been diverted, and where the diversions had taken place, but had advanced no evidence to support their theories or their argument that, of the competing scenarios, one— diversion before delivery—was to be preferred. In my judgment there is nothing 10 in that argument. As I have explained above, art 20.3 of the Directive requires the person said to be liable for the duty in the case of an apparently irregular departure to show that the movement was in fact correctly discharged; there is no burden on the taxing authority.[149]I also see no basis on which it might be argued that the standard to which 15 regular discharge is to be established is anything other than the conventional standard in civil proceedings in the United Kingdom, that is the balance of probabilities. Despite the potential for confusion I have identified I detect nothing in the body of UT 1’s decision, or in their detailed direction, which could realistically be interpreted as a requirement that F-tT 2 should determine whether 20 the allegedly impossible journeys were possible on anything other than that standard. It is plain that Judge Berner thought he had to apply a different standard because of his remark that in some cases, had he been reaching his conclusions on the balance of probabilities, he would determine that the journey was impossible whereas on the test he did apply he found it to be possible. 25 150. I return at this point to what Patten LJ said in Weymont v Place at [4], that the relative immunity from attack of an inferior tribunal’s findings of fact assumes that the findings were reached on the basis of a proper understanding of the task before that tribunal. With the greatest of respect to an experienced judge, I find it an irresistible conclusion that Judge Berner has misunderstood what was required 30 of him, that the findings he made are tainted by his application of the incorrect test, and that this ground of appeal is made out.[151]The other of the grounds on which HMRC were given permission to appeal relates to Judge Berner’s claimed failure to follow UT 1’s direction that he was to “determine whether all or any of the journeys … could not have taken place as 35 described in the evidence of the drivers” (emphasis added). Of course, the drivers said that they had delivered the goods to Aldi, and they also agreed that they might have exceeded the speed limit and that they might not have been entirely scrupulous about rest periods, and that was evidence Judge Berner could properly take into account. But I agree with Miss Simor that they did not say that they 40 drove consistently in excess of the speed limit, regardless of traffic conditions and the presence of police and speed cameras. Indeed, as Judge Berner himself said at [59] (quoted at para 86 above) they generally estimated the time required for a one-way journey at about 4½ hours, even if 3½ hours was mentioned as a possibility, and similarly a one-hour crossing of the Channel was mentioned as the 45 best possible time, rather than one which could be routinely achieved. Judge Berner has, however, addressed the question whether, if they were to ignore the speed limit throughout the journey, and on the assumption no delay was caused by 45 congestion or road works, they might have succeeded in travelling from Folkestone to Vaux-sur-Sûre and back to Coquelles sufficiently quickly to enable them to accomplish the entire journey within the times indicated by the available documents. In my view that approach, of focusing on the shortest time suggested for any particular part of the journey, was also an error in that 5 Judge Berner failed, as UT 1’s direction required, to consider all of the evidence of the drivers, but instead took account of parts of it while disregarding other parts. Similar criticism could be made of his failure to factor in any time for rest breaks.[152]In my judgment the two failings I have mentioned override Mr Barlow’s 10 argument based on the Edwards v Bairstow line of authority, for the reasons given by Patten LJ in Weymont v Place. Judge Berner’s findings of fact cannot be immune from attack when he has asked himself the wrong question, has applied an incorrect standard of proof and has examined the evidence selectively. In short, I am persuaded that he misunderstood the task required of him by UT 1 and for 15 those reasons too I would allow HMRC’s appeal.[153]Despite that conclusion I need to deal briefly with Miss Simor’s argument that Judge Berner wrongly allowed Mr Barlow to advance the submission that the timings might be inaccurate. It does not seem to me that Miss Simor is right to say that the reliability of the timings shown on the documents was not before UT 1, as 20 they mentioned the point at [71] (set out at para 69 above). Thus the argument does not appear to be factually correct, but even if it is I think it must be dismissed for other reasons.[154]Judge Berner explained why he had adopted the course he did at [34] and [35] (quoted at para 75 above), and in my view his explanation, coupled with the 25 ordinary bounds of judicial discretion and the application of the overriding objective of rule 2 of the First-tier Tribunal rules, is sufficient to justify his doing so. In addition, UT 1’s direction required him to pay regard to the evidence before F-tT 1 (which included the relevant documents); it did not confine him to the arguments advanced before F-tT 1. In any event his hearing Mr Barlow’s 30 argument made no difference to the outcome because, as I have already recorded, he rejected the submission for a lack of evidence to support it, and I detect no reason to think that he nevertheless treated the recorded times as anything other than a reliable reference point for the start or end of a journey, or for the timing of any other event, such as the purchase of fuel, for which the documents were the 35 evidence. He made the driving speeds and the intervals needed to cross the Channel or load and unload the lorries fit the recorded times, rather than the reverse.[155]Indeed, it seems to me that this was the trap into which both F-tT 1 and F-tT 2 fell. The essential error in F-tT 1’s decision, as UT 1 identified it, lay in their 40 failure to reconcile the evidence of the drivers that they had delivered the goods with the impossibility of their having done so in some cases if the documents they, or SDM, produced were correct. It is a factor of considerable importance that F-tT 1, who alone heard them give evidence, believed the drivers’ claims that the goods were delivered, but it is not, in my view, a factor which outweighs all 45 others. However plausible a witness may seem, his claim that he has done something when manifestly he could not have done must at the least lead to a close examination of his claim. F-tT 1 did not undertake that exercise—or if they 46 did, they did not explain how they reached their conclusion. Judge Berner did not fall into quite the same error but adopted what, in my view, is the impermissible approach of assuming that two conflicting parts of the evidence—the claims of delivery and the times shown on the documents—were accurate, and then selecting and in some cases distorting the remaining evidence 5 in order that the conflict could be resolved.[156]I have also come to the conclusion that SDM’s appeal in respect of movement 29 should be dismissed. It is true that Judge Berner said at [30] that “it has been conclusively determined that it could not” have been delivered, but it is 10 clear from what else he said, even if he did not make the point in the same paragraph, that it had been so determined only by reference to the documents. He made the observation at [30] as an explanation of why he did not intend to reconsider it in the same way as the other allegedly impossible journeys—indeed, UT 1’s direction did not permit him to do so. But at the second stage of the 15 enquiry he returned to movement 29 in order to consider whether he could accept, in the light of his other conclusions, that there was some evidence relating to that movement which would show that the documents were not determinative, because it was sufficient to demonstrate how the delivery had in fact been effected.[157]F-tT 1 accepted, as I have said, that there was a possible explanation though 20 it was based on what, in my view, was rather scant evidence. UT 1 set aside the entirety of F-tT 1’s decision and consequently remitted the re-making of the entire decision to F-tT 2: thus I disagree with Miss Simor’s argument that Judge Berner was not permitted to re-examine movement 29.[158]However, the conclusion he reached was one of fact, that SDM had not 25 discharged the burden of establishing that the goods were delivered. Mr Barlow did not argue that there was material evidence which Judge Berner overlooked or misunderstood or that for some other reason the conclusion was irrational in the Edwards v Bairstow sense. Rather, his argument was that, having concluded that the other deliveries to the Aldi warehouse had been made, Judge Berner ought to 30 have concluded that this consignment too was delivered; in other words, he should have adopted rather than, as he did, rejected the binary approach that all the consignments or none were delivered.[159]As I shall explain below I disagree with Judge Berner on the need for a binary approach and, had I concluded that Mr Blunsden’s other journeys were 35 possible but movement 29 was not, I too would be required to reconcile those divergent conclusions. However, as I have concluded that some of Mr Blunsden’s other journeys were either impossible or, on the balance of probabilities, did not take place I do not therefore need to resolve divergent conclusions. It is for those reasons that in my judgment SDM’s appeal in respect of this movement should 40 fail. Remaking the decision[160]As I have indicated the parties are agreed that if we should allow HMRC’s appeal we should re-make the decision. Since we are divided about the outcome that task falls to me. Before I embark on it I think it appropriate to make some 45 observations about the nature of the task and the constraints which have affected my approach to it. 47[161]I have throughout been conscious of the point that I have already made that I have not heard the oral evidence, and in particular the evidence of the drivers. That F-tT 1 believed their evidence in preference to the other evidence, even if they did not explain their reasons, is a factor to which I must necessarily attach considerable weight, and I must correspondingly be cautious 5 before coming to conclusions from which it follows that the drivers who claimed that they did deliver the goods gave untruthful evidence to F-tT 1. I am very conscious of the point made by Judge Cannan in his dissenting decision that it is not permissible to make findings of dishonesty, as a finding that the goods were diverted before 10 delivery necessarily implies, without first giving the person said to be dishonest an opportunity to deal with the allegation. It was, however, put to those of the drivers who gave evidence, in cross-examination, that some of their journeys were impossible and that the goods were not delivered to the Aldi warehouse. I have concluded that the combination of assessments based upon complicity in the 15 diversions and the cross-examination put the drivers sufficiently on notice of what was being said against them, and that they had an adequate opportunity of dealing with it. In addition, the issue in this appeal is not whether the drivers were party to a conspiracy, but whether SDM has discharged the burden of showing that the goods were delivered. For those reasons I have concluded that if I am satisfied 20 that the evidence of delivery at the Aldi warehouse cannot be true, I am bound to reject it even though I have not heard it myself. It should also be remembered that SDM invited us to remake the decision without hearing the drivers again.[162]I should add that I recognise the force of Mr Barlow’s argument, reflected in F-tT 1’s observation to the same effect, that the drivers were giving evidence of 25 events which had taken place four years earlier, and that most had not been asked about the deliveries until two years had gone by. If the journeys were uneventful it would be remarkable if they could remember very much about them, and I have borne that point in mind when examining what they said. On the other hand, I cannot overlook the fact that if HMRC are right the drivers did not deliver the 30 loads and, since they too have been assessed for the relevant duty, they have a motive of their own for giving untruthful evidence.[163]The first question I need to consider is whether I should approach the task of remaking the decision by following UT 1’s direction to F-tT 2, or instead I should consider the evidence at large, without any restrictions such as those imposed by 35 the direction. I have reached the conclusion that I should broadly follow the direction, by first examining the allegedly impossible journeys and then considering the impact of my conclusions about those journeys on the view to be taken of the remaining journeys, but that I should have regard to all of the evidence now available. I shall, therefore, and at each stage, take account not only 40 of the recorded times but also the additional evidence, not least that of the drivers, which was before F-tT 1, as well as information introduced later, particularly that derived from Google Maps.[164]I accept the principle of Mr Barlow’s argument that even if the times recorded in the documents should be found to be accurate there may be a credible 45 explanation of how the goods nevertheless arrived at their stated destination. I do not, however, accept his argument that the recorded times should be treated with caution, if not discounted altogether. I agree with UT 1 and F-tT 2 that as the documents were produced by SDM as evidence that the journeys took place on 48 particular days which could be matched to the collection of the goods from the UK warehouses it is for SDM to show, by something more than assertion about what might be the case, that the documents are accurate as to the day and inaccurate as to the time, or at least that the possibility that they are is so great that the reliability of the documents must be doubted. There is 5 no evidence of that kind; on the contrary, as I have already observed, the times at which the drivers purchased their vignettes and the times of their check-in closely match, the former being identifiably a few minutes after the latter, suggesting that both records are reasonably accurate. In addition, as Judge Berner said, the time shown on at least 10 one fuel purchase receipt can be matched to the journey.[165]Moreover, if Mr Barlow’s proposition is to benefit SDM, he must be able to demonstrate that the time available for the journey was, or at least could have been, greater than the recorded times indicate. For that to be so in the case of a journey whose feasibility is assessed by reference to the interval between check-in 15 at Folkestone and check-in at Coquelles, the check-in time at Folkestone shown on the Eurotunnel documents must be later than the true time, the time of check-in at Coquelles must be earlier than the true time, or both. I do not see how, in the absence of evidence supporting the argument, I could reasonably so conclude.[166]For similar reasons I reject the argument that the journey to which the 20 documents relate might have been incorrectly identified. SDM produced the documents for the precise purpose of showing that the journeys to which they appeared to relate took place. I do not see how it can now be plausibly argued, without some evidence to support it, that there might have been confusion between journeys. It is true that Mr Barlow was able to point to a Eurotunnel 25 invoice which SDM had queried because it appeared to show the same vehicle travelling through the tunnel twice in quick succession, and I accept that errors were possible. But that is not the same as saying that it has been shown that the documents are entirely discredited, and SDM has not even attempted to link the documents it produced to journeys other than those to which they appear, and 30 were initially claimed, to relate.[167]Although the underlying question in the appeal is not whether the relevant journeys were possible but whether SDM has discharged the burden of showing that the goods were delivered to the Aldi warehouse (I leave the other two consignments aside for the present), it will be apparent from what has gone before 35 that the focus of much of the debate, before F-tT 1, UT 1 and F-tT 2 and now before us, has been on the possibility of the journeys, upon the basis that if it can be demonstrated that the journey was possible the goods comprised in that movement are to be treated, without more, as having arrived. I have concluded that, even if that is not strictly the correct approach, it would be unfair, at this late 40 stage, to impose any greater burden on SDM than to show that the journeys were possible. Although Miss Simor did not formally concede that mere possibility was enough, she did not, as I understood her submissions, demur from the proposition that the appeal should be determined in that fashion.[168]By “possible” I mean realistically possible, and not possible only if one 45 assumes that the drivers persistently disregarded the law and had the good fortune never to encounter any delay: for the reasons I have given the burden on SDM is to show that they were possible on the balance of probability, taking all known 49 factors into account. Thus while I accept that, as Judge Berner said of himself, I have no expert knowledge of the speed at which a fully-laden (or, indeed, unladen) lorry might be driven on a motorway, I do not think it necessary or appropriate to leave aside my experience of ordinary traffic conditions or my knowledge derived from that same experience that, for 5 example, fully laden lorries are unable to accelerate as quickly as cars.[169]I think therefore that the most appropriate approach is to begin with the journey times, in order to determine, so far as the available evidence of times of check-in or fuel purchase makes it possible, which of the movements could have 10 taken place and which, if the timings alone were the only factor, could not. For the reasons I have given I shall take the recorded times, where available, as accurate. I shall then turn to the remaining evidence in respect of those, apparently impossible, journeys in order to see whether there is any evidence, or explanation, which makes it possible to conclude that they could have taken place. That, as I 15 understand it, is essentially what para 2a of UT 1’s direction required.[170]Although some criticism was made of Judge Berner’s use of the information he derived from Google Maps, substantial parts of the arguments advanced before us, by both sides, adopted the same information, albeit with different caveats. Miss Simor laid particular emphasis on the fact that (despite what Judge Berner 20 had said) the journey times shown by the Google Maps website were manifestly applicable to cars and not to heavy goods vehicles, and she argued that it was not permissible to adopt them uncritically as if they were. Mr Barlow’s caveat was that it should be borne in mind that, whichever route was adopted, almost the entirety of the journey from Coquelles to Vaux-sur-Sûre or return could be 25 accomplished by motorway, that the same was true of most of the detours to collect return loads, and that both the Coquelles terminal and the Aldi warehouse were in close proximity to motorway access points. With those caveats in mind I have concluded that it is appropriate to use Google Maps as a reference point.[171]The shortest direct route between the tunnel terminal at Coquelles and 30 Aldi’s warehouse at Vaux-sur-Sûre (which, as Mr Barlow said, is not in Vaux-sur- Sûre itself but on an industrial estate near to a motorway junction) is shown by Google Maps to be 349 km, but it involves the use of a route nationale for about 70 km and the estimated time required to cover that route is put by Google Maps at 4 hours 12 minutes, at an average speed of 83 kph. A longer route using 35 motorways for almost the whole journey extends to 363 km, but the estimated travelling time is 3 hours 30 minutes, at an average of 103.7 kph. I have reached the conclusion, particularly bearing in mind Mr Barlow’s argument that motorways were used as much as possible, that the drivers are more likely to have used the slightly longer route than the shorter route using the route nationale. I 40 agree with Miss Simor, even without evidence, that 3 hours 30 minutes cannot be the estimated time for a lorry limited to a maximum motorway speed of 90 kph, but is the time needed to undertake the journey in a car.[172]The detailed driving instructions offered by Google Maps for the preferred route indicate that (starting at the tunnel exit) one must first drive 1.8 km over 45 other roads before reaching the motorway. At the other end, the distance between the motorway exit and the Aldi warehouse is 1 km. If one were to assume that of the total journey, 2.8 km could be driven at 30 kph (allowing for bends, 50 roundabouts and similar obstacles and the need to give way to other traffic) and the remainder at the motorway speed limit for a heavy goods vehicle of 90 kph, the total time required for the journey amounts to 4 hours and 3 minutes. That time takes no account of congestion, road works or other impediments and makes no allowance for acceleration or braking. The difference between 5 the motorway speed limit and the estimated average speed for a car offered by Google Maps may not be capable of a direct, arithmetical, transfer to the estimate of the speed which might be achieved in a lorry but does suggest that a lorry too would not be able to maintain the maximum permitted speed and that the 4 hours and 3 minutes 10 at which I have arrived is an optimistic time.[173]I am not willing to do as Judge Berner did, and assume that the drivers exceeded the speed limit to a material degree. The transcript of his oral evidence shows that it was only reluctantly that Mr Blunsden accepted that on occasion he did speed, but he did not claim to have done so for prolonged periods or by a 15 substantial margin. As I have mentioned, Mr Parnham said he tried to complete the journeys quickly because he had inadequate insurance cover in case of theft, but then said that he would do so by shortening his breaks rather than by speeding. Other drivers gave evidence that they did not speed, because of the risk of being stopped by the police, a factor which would have lengthened the journey time and 20 in addition would have exposed them to fines, albeit there was some acceptance that speeding fines were an occupational hazard. It is also relevant that none of the drivers claimed to be under pressure to return to the UK especially quickly, and that they did not need to take any particular train but were able, as I have explained, to take the first available train after their arrival at the terminal. 25 174. F-tT 1 reached the conclusion, not from an analysis of information obtained Google Maps but from the evidence of the drivers themselves, that 4½ hours was required for a typical journey from Coquelles to Vaux-sur-Sûre. In reaching that conclusion they took into account factors such as congestion, and the drivers’ evidence about their propensity or otherwise to exceed the speed limit. Miss 30 Simor pointed out that even that time implied an average speed for the entire journey of 80 kph which, she suggested, was unlikely to be achieved in practice. She may well be right: any driver knows that, save in the very early hours of the morning, it is rarely possible to travel consistently at the maximum permitted speed and that the average speed for the journey will be below that maximum. It is 35 no doubt for that reason that the Google Maps estimates of the time required for a car assume that the average speed will be significantly lower than the applicable speed limit, though I accept that some drivers are likely to drive more quickly than others and that, as traffic conditions change, so too will the time required to undertake the journey. However, F-tT 1’s conclusion on this point is, in my view, 40 not unreasonable and, moreover, it is consistent with the evidence as they recorded it. I have said that 4 hours 3 minutes for the journey in a lorry is optimistic even if, in optimum conditions, it is possible; but for the reasons I have given I agree with F-tT 1’s estimate of 4½ hours as the time required for a typical journey. I have, nevertheless, not taken 4½ hours as an irreducible minimum in 45 the analysis which follows.[175]The return legs of the allegedly impossible journeys were not direct, as in each case the drivers picked up a load en route. I shall consider the times necessary for those journeys as I deal with them individually, but it is plain that in 51 no case could the travelling time be, realistically, less than 4½ hours plus the time needed to make a detour to collect the return load. Again, and in order to give SDM the benefit of any doubt when there is room for it, I have not treated that time as an irreducible minimum.[176]Judge Berner assumed that the time required to cross 5 the Channel was one hour from check-in. At [57] he explained that he derived that time from Mr Blunsden’s claim that the quickest time in which he had achieved the crossing was one hour, even though he and other drivers said that the time could on occasion be much longer, and Mr Parnham said that between two and three hours 10 was usually required. It requires no special knowledge to determine that the time would vary depending on the volume of traffic and, even if there were no queue, the time between check-in and departure of the next train. Thus even if one or two of the journeys might have taken as little as an hour, others are likely to have taken longer. I have no mechanism similar to Google Maps which might assist 15 and instead shall adopt the Eurotunnel brochure to which I have referred as the best evidence and take the 90 minutes indicated in it as the likely time. I have treated it as an irreducible minimum because, if I were instead to derive a time from the totality of the drivers’ evidence, it would be rather greater. F-tT 1 did not make a finding of the required time for the crossing. 20 177. As I have mentioned, Judge Berner’s estimate of the time required to load and unload a vehicle was derived from Mr Blunsden’s evidence that “if it went really well” it might be achieved in 20 minutes. He added 5 minutes for completion of the paperwork to arrive at his total allowance of 25 minutes. Mr Blunsden also said, however, that the unloading might take an hour, and Mr 25 Waters estimated between 45 and 60 minutes. Understandably none of the drivers could give a time for any individual load, and there was nothing beyond those rather vague estimates before F-tT 1. In my view Judge Berner’s estimate is not supported by the evidence—rather, as in the case of the time taken to cross the Channel, it is simply the shortest time any of the drivers mentioned as a 30 possibility—and a more realistic estimate, and one probably erring in SDM’s favour, is the 45 minutes suggested by Mr Waters. That was also the time accepted by F-tT 1 and UT 1. I cannot, of course, reject the possibility that the operation might have been achieved more quickly in any individual case, but I am satisfied that the evidence supports a finding that 45 minutes is required and, save 35 for a particular case to which I come later, does not support any shorter time.[178]Break times cannot be dealt with in quite the same way, as the need for, and required length of, a break depend upon an analysis of each individual journey. I shall return to this point, as necessary, later, but in general I have, like Judge Berner, disregarded break times because of the difficulty of building them into the 40 calculations.[179]Leaving break times to one side, Judge Berner’s approach was to assume not only short times for each stage of the journey but that in every case the driver had achieved all of those short times. Although it is possible that a driver might have such good fortune, real life experience indicates that it is unlikely and that, in 45 order to establish a typical rather than optimum time, it is necessary to build in some allowance for delays caused by a queue at the tunnel, traffic congestion and road works, waiting time at the warehouse while other vehicles are dealt with, and 52 so on. Even without such an allowance, and with no addition for break times or for the return detour, the aggregate of the realistic periods I have described for a journey from check-in at Folkestone to check-in at Coquelles is 12 hours. That would reduce to 11 hours (rounded down slightly) if instead I were to assume an optimistic rather than realistic assessment of the time required 5 to travel from Coquelles to Vaux-sur-Sûre or to return. I should add that the minimum time, before adjustment for breaks and the return journey detour, at which I have arrived is directly comparable with the 10½ hours (after adjustment by UT 1) at which F-tT 1 arrived, because that estimate takes no account of the 90 minutes 10 needed to cross the Channel. I emphasise that I do not take 12 (or 11) hours as determinative, but as the starting point for an examination of all the evidence relating to each of the allegedly impossible journeys. The allegedly impossible journeys[180]Movement 17 was the first of the allegedly impossible journeys examined 15 by Judge Berner, and is one which I have already described at para 95 above. It will be recalled that by Judge Berner’s calculation Mr Blunsden, who drove this load, would have arrived at Coquelles 1 hour and 9 minutes later than his recorded check-in time of 18.46. If instead the realistic times at which I have arrived are adopted, even without including an allowance for the detour to collect the return 20 load, Mr Blunsden could not have arrived at Coquelles before 22.24 (continental time), 12 hours after his check-in at Folkestone at 09.24 (UK time), and more than 3 hours 30 minutes later than his recorded check-in at Coquelles. In my view it is a fanciful proposition that this journey was feasible in the time available if the documents are correct; and in the absence of any evidence that the documents 25 were materially incorrect I must conclude that this journey did not result in the delivery of the consignment at Aldi’s warehouse. No amount of selection of the evidence, or assumption about plausible driving speeds or loading times, makes it possible to conclude, on the balance of probabilities, that this journey could have been undertaken as claimed. 30181. The second movement to be considered was 19, also driven by Mr Blunsden. He checked in at Folkestone at 20.20 (UK time) on 4 October 2006 and was recorded to have purchased a vignette at (it was assumed by Judge Berner in SDM’s favour) the French-Belgian border between Lille and Tournai at 08.24, continental time, the following day. The distance from Coquelles to the border at 35 that point is about 130 km, a distance which could be covered in about 1 hour 40 minutes at an average speed of 80 kph. It follows, if the place of purchase has been correctly identified, that either the Channel crossing took materially more than 90 minutes or that Mr Blunsden took a break, and there may have been a combination of the two. The distance from the border to the warehouse is about 40 230 km, implying a travelling time of roughly 2 hours 40 minutes assuming Mr Blunsden set off immediately after buying the vignette and was able to travel at the maximum permitted speed until he left the motorway. He would then arrive at the warehouse just after 11 am. If his average speed was 80 kph he would arrive at about 11.15. 45182. If one assumes the earlier arrival and 45 minutes for the unloading, Mr Blunsden could have set off on the return journey at about 11.45. He then went (if SDM’s case is correct) to Ooigem (about 245 km or 2 hours 45 minutes at an 53 average of 90 kph) to pick up another load (45 minutes) and from Ooigem to Coquelles (137 km or about 1 hour 35 minutes) in order to check in, arriving at about 16.50. That, in my view, is the earliest possible time of arrival. If Mr Blunsden was able to average only 80 kph, encountered a delay or took a break the arrival time would, of course, be later. The recorded check-5 in time was 15.37. Again, the only reasonable conclusion is that the goods could not have been delivered to the warehouse, and I do not agree with Judge Berner that this journey was possible.[183]There was a rather different issue in respect of movement 22, the 10 documentary evidence for which included a fuel purchase. Mr Blunsden checked in at Folkestone at 22.02 UK time on 6 October, buying a vignette a few minutes later. He bought fuel at Watou, just across the border from France into Belgium, at 10.38 the following day. There was evidence that he had made a telephone call from France at 10.04 and Miss Simor’s argument was that Mr Blunsden had 15 remained in France until at least 10.04 and had then driven into Belgium, and that the fuel was purchased on the outward journey. If that is correct Mr Blunsden could not conceivably have driven to Vaux-sur-Sûre and back to Coquelles by his recorded check-in time of 15.59.[184]Mr Blunsden, however, said that he had driven straight to the warehouse 20 from Coquelles and that he had purchased the fuel on the return journey. If that is correct, he could easily have reached the warehouse before its opening time of 7 am and have left, after discharging his load, at 07.45. He then needed to travel to Cuincy, France (250 km or 2 hours 50 minutes at 90 kph) to pick up the return load (45 minutes) and then to Watou (83 km or 55 minutes) to buy the fuel, 25 arriving at about 12.15. The fuel receipt, however, was timed at 10.38, approximately when Mr Blunsden would have arrived at Cuincy. It would certainly be possible to drive from Watou to Coquelles (71 km or 50 minutes) in time for the recorded check-in at Coquelles of 15.59, even if he did not arrive in Watou until 12.15, but it is quite clear, if the fuel receipt is correct, that this 30 journey cannot have taken place as claimed.[185]As it was Mr Blunsden who produced the receipt as evidence supporting his claim that he had delivered the load the totality of the available evidence suggests that the more likely explanation is that Mr Blunsden did not visit Vaux-sur-Sûre, and that his claim that this load was delivered is false. However, I recognise that it 35 is possible that the time shown on the fuel receipt is wrong, and I have therefore reconsidered this journey on the assumption that Mr Blunsden did buy fuel at Watou on his return journey, but at an unknown time.[186]If his evidence that he first travelled overnight to Vaux-sur-Sûre is correct, Mr Blunsden could have departed at 07.45 to drive to Cuincy, 253 km distant, 40 arriving at 10.55 if he maintained an average speed of 80 kph. After loading he could leave at 11.40 and drive to Watou, a distance of 88 km, in just over an hour, arriving at 12.45 or 12.50. According to Google Maps, driving from Cuincy to Coquelles via Watou adds about 16 km to the direct route, but I leave this oddity to one side. Mr Blunsden would then have time to buy the fuel and drive to 45 Coquelles before his recorded check-in at 15.59. It is therefore necessary to determine whether the apparent possibility of the journey, if the time on the fuel 54 receipt is left out of account, is undermined by Mr Blunsden’s phone call from France at 10.04.[187]According to Google Maps, the return journey from Vaux-sur-Sûre to Cuincy, assuming motorways were used, would take him across the border from Belgium into France at a point about 198 km from Vaux-5 sur-Sûre and 55 km from Cuincy. If he left the warehouse at 07.45, he would reach the border by about 10.14 assuming an average travelling speed of 80 kph, but at 09.57 if he could average 90 kph. Thus a call from France at 10.04 would be possible, but only if Mr Blunsden had been able to maintain the higher speed. 10 188. It follows that the journey was possible, but only if the time shown on the fuel receipt is wrong and Mr Blunsden was able to drive at what, in my view, is an unrealistically high average speed and any need for a rest break is ignored. Although Mr Blunsden could have driven overnight to Vaux-sur-Sûre, he would not have arrived there until about 05.00, giving him no more than two hours to 15 sleep. In my judgment the uncertainties about the feasibility of this journey are such that SDM cannot be said to have discharged the burden of showing, on the balance of probabilities, that the goods comprised in this movement were delivered.[189]I have also touched on the next of the allegedly impossible journeys 20 examined by Judge Berner, movement 24. He concluded that Mr Wild, who was the driver of that consignment, could have arrived at Coquelles in time for his check-in if he had travelled at 110 kph. The more realistic (as I have concluded) approach shows that the journey was impossible within the recorded times. Mr Wild checked in at Folkestone at 07.15, so could have arrived at Coquelles at 25 09.45 continental time. Adding 4 hours for the journey to Vaux-sur-Sûre (that is, assuming in Mr Wild’s favour an average speed of 90 kph), 45 minutes for unloading, 2 hours 45 minutes to Ooigem, where he too picked up a load, 45 minutes for loading and 1 hour 35 minutes for the journey to Coquelles leads to an arrival time, disregarding delays or breaks, of 19.35. The recorded check-in time 30 was 17.10. I do not see how it can realistically be argued that this journey was feasible. Even if the time taken to cross the Channel is reduced to 1 hour it was impossible. Moreover, Mr Wild would have had to drive for more than 8 hours with only two short breaks while his vehicle was unloaded and re-loaded.[190]The driver of movement 37 was Mr Parnham. He checked in at Folkestone 35 at 07.25 on 23 October 2006, so could have arrived at Coquelles (on my approach) at 09.55 continental time. The journey to Vaux-sur-Sûre (4 hours), unloading (45 minutes), trip to Geer for the return load (130 km or 1 hour 30 minutes travel and 45 minutes loading) and from Geer to Coquelles (273 km or just over 3 hours) would result, again disregarding delays and breaks and 40 assuming average motorway speeds of 90 kph, in his arriving at Coquelles at about 20.00. The recorded check-in time was 18.27 and, for the reasons I have given, I am not persuaded that the goods comprised in this movement can have been delivered to Aldi’s warehouse.[191]I must, however, deal with a further aspect of Mr Parnham’s evidence. He 45 produced a drawing of the warehouse which, Mr Barlow emphasised, was consistent with architects’ plans of it and was therefore an indication that Mr Parnham had indeed been there. F-tT 1 dealt with it very briefly, and did not 55 record when and in what circumstances Mr Parnham had prepared it, a topic on which his witness statement is also silent. It does not seem that F-tT 1 placed any weight on it. Judge Berner disregarded it because it was irrelevant to the question whether or not the journeys were possible.[192]I have concluded that I, too, should attach no weight to 5 the drawing. I do not know whether it was prepared from Mr Parnham’s recollection of what he observed when making a regular delivery to the warehouse, at a later visit after he had received an assessment for the duty on the loads which he drove, or in some other circumstance. In other words, I have no material from which I can determine 10 its evidential value. It is certainly insufficient to displace my finding that Mr Parnham cannot have delivered movement 37 as he claimed. It may be indicative of a visit by Mr Parnham to the warehouse but I agree with Judge Berner that it has no bearing on the possibility of this journey.[193]Mr Blunsden was the driver of movement 38. He checked in at Folkestone 15 at 20.30 on 23 October, and would have arrived in Coquelles at about 23.00 continental time. Judge Berner dealt with this movement on the assumption, urged on him by HMRC, that Mr Blunsden did not drive straight to Vaux-sur-Sûre, but took a break before doing so. That assumption was based on the fact that he is shown by his mobile phone records to have been in France when he made calls at 20 09.48 and 09.50 on 24 October. Judge Berner assumed (also at HMRC’s suggestion) a journey which began near to Dunkirk and proceeded to Vaux-sur- Sûre via Bruges and Brussels, which according to Google Maps would be the quickest route if Mr Blunsden was near Dunkirk when he made his calls. If one allows a shorter journey time to Vaux-sur-Sûre of 3 hours 30 minutes 25 (approximately 320 km at an average of 90 kph), 45 minutes for unloading, 1 hour 30 minutes to Geer, 45 minutes to pick up the return load and 3 hours from Geer to Coquelles Mr Blunsden would have arrived there at about 19.20. His recorded check-in time was 18.34 and, if the suggested route is right this journey too would be impossible. 30 194. However, although the telephone records identify the country from which a call was made they do not identify the place, and I have considered the possibility that Mr Blunsden took a different route, leaving France later in his journey. If he crossed, as was assumed for movement 19, between Lille and Tournai, he could have made his calls immediately before leaving France, and could have reached 35 Vaux-sur-Sûre by about 12.30, if he was able to drive at an average speed of 90 kph—at 80 kph he would arrive at about 12.45. After unloading he would be able to leave at about 13.15 (assuming faster driving) and travel the 130 km to Geer, to collect the return load, in about 90 minutes and leave there, after loading, at about 15.30. The distance from Geer to Coquelles is about 270 km requiring a minimum 40 travelling time of 3 hours—thus he could, just, have arrived in time to check in at 18.34. If his average driving speed was 80 kph he would plainly have arrived too late.[195]I cannot therefore say that this journey was as clearly impossible as some of the other movements, but its possibility is nevertheless dependent on the making 45 of various assumptions in SDM’s favour. As I have explained, those assumptions are not derived from the evidence before F-tT 1 and they are difficult to sustain as a matter of common sense. I am, therefore, not satisfied that SDM has shown, on 56 the balance of the probabilities, that the goods comprised in this movement were delivered to the warehouse.[196]Mr Blunsden was also the driver of movement 43. He checked in at Folkestone at 18.43 on 1 November, but the precise time is unimportant as his evidence was that he drove overnight to the warehouse, arriving 5 in time to take a rest break before it opened at 7 am. If he left at 07.45 after discharging his load he would have been able to arrive in Geer to pick up the return load (again assuming 90 kph) at 09.15, and leave at 10.00. According to F-tT 1, Mr Blunsden himself put the travelling time to Geer at 1½ hours, but said that the loading there was 10 very efficient, taking no more than 15 minutes. If that is right he could have left by 09.30. It seems that he then travelled to Veurne (which is on a direct route from Geer to Coquelles) to purchase fuel. The distance from Geer to Veurne is (according to Google Maps) 204 km, all motorway apart from the distance from the Geer warehouse to the motorway, implying a journey time of 2 hours 16 15 minutes and an arrival at Veurne of about 12.15 if he left Geer at 10.00, but 11.45 if he left Geer at 09.30. According to F-tT 1, Mr Blunsden himself estimated the driving time at between 2 and 2½ hours, implying that his earliest possible arrival time at Veurne was 11.30. The fuel purchase is recorded by F-tT 1, to have taken place at 11.25, the time shown on the receipt Mr Blunsden produced. That time is 20 not wholly implausible, but does imply a very quick journey time. If the fuel was indeed purchased at 11.25 Mr Blunsden would have been able to drive the 71 km (50 minutes) to Coquelles in good time for his recorded check-in at 13.10.[197]If, instead, it is assumed that after leaving Vaux-sur-Sûre at 07.45 Mr Blunsden drove at an average of 80 kph to Geer (arriving at 09.25), was loaded in 25 the 15 minutes he claimed and then drove at an average of 80 kph to Veurne he would arrive there at about 12.00, after he is shown to have bought fuel. A stop for refuelling of 15 minutes followed by a journey of 71 km at 80 kph would also enable him to arrive at Coquelles just in time for his recorded check-in at 13.10.[198]Thus this journey, too, was possible, but only if one assumes in SDM’s 30 favour that Mr Blunsden was able to achieve all of a very quick set of times and in the case of more realistic driving speeds that the time shown on the fuel receipt, whose accuracy Mr Blunsden did not challenge as he gave his evidence, is wrong. In my judgment the probability is that the journey did not take place, though I cannot say it was clearly impossible. 35199. Movement 57 was identified by UT 1 as an impossible journey only because they had no information about the time available, rather than because the apparent time was too short. It was also driven by Mr Blunsden. It too began (if Mr Blunsden is right) with an overnight journey to the warehouse where he took the opportunity to rest before it opened at 7 am. After discharge of his load he could 40 have left at 07.45 to travel to Geer (1 hour 40 minutes at 80 kph), pick up a load (45 minutes) and drive to Coquelles (3 hours 23 minutes) arriving at about 13.30. That estimate builds in no allowance for delays or further breaks, but as the recorded time of check-in was 15.30 there was ample margin for such contingencies and this journey was plainly possible. 45200. Before F-tT 2 Miss Simor argued that the journey as described made no allowance for a break, and that without a break Mr Blunsden would have been driving for ten hours without rest. Judge Berner rejected that argument, recording 57 that Mr Blunsden could have slept for about six hours while waiting for the warehouse to open (which would undermine the argument that he needed to drive for 10 unbroken hours) and that there was time available for him to take a short break by interrupting his return journey to Coquelles. While I disagree with Judge Berner on matters of detail—I consider Mr Blunsden would 5 have had rather less than six hours waiting at the warehouse and the time available for a break on the return was somewhat limited—I agree with him that this factor does not lead to the conclusion that the journey was impossible.[201]After dealing with the other allegedly impossible journeys and concluding 10 that they were possible Judge Berner turned his attention to movement 29 which, as I have said, he concluded was not possible; he therefore decided that SDM had not discharged the burden of showing that the goods comprised in that movement had arrived.[202]Although, in view of my other conclusions, the point assumes relatively 15 little importance I should say that I have encountered some difficulty in reconciling Judge Berner’s conclusions in respect of movement 29 with his acceptance of Mr Blunsden’s claim that the other loads he drove were in fact delivered. As [144] and [145] of his decision (set out at para 100 above) show Judge Berner accepted, as had F-tT 1, that Mr Blunsden did not recall the load. As 20 I have said before, an absence of recollection would be entirely understandable had this been one of several routine loads correctly delivered to the warehouse, but Judge Berner decided that it was not. The documents relating to the movement which SDM supplied show that it was Mr Blunsden’s lorry which transported the goods, and that he invoiced SDM for the carriage; there is no scope for a mistake 25 in the identity of the driver. If, as Judge Berner found, Mr Blunsden did not deliver the load to its stated destination it obviously follows that he did something else with it.[203]I do not accept that he could simply have forgotten that fact: an experienced driver, as he apparently was, would be well aware of the significance of the 30 diversion of a load. Even if the load was in fact delivered there was an oddity in its manner of delivery, either of the kind suggested by F-tT 1 or another about which I see no need to speculate. There is no evidence that Mr Blunsden reported any problem; on the contrary, he maintained before F-tT 1 that he delivered all of the loads for which he was the driver to their intended destination uneventfully. If 35 Mr Blunsden’s claim that he delivered this load is rejected, doubt must be cast on his claim to have delivered the other loads for which he was the driver. I do not agree with Judge Berner that the doubt can be resolved by simply accepting what Mr Blunsden said about the other loads and brushing aside—indeed, ignoring— the absence of any explanation of this movement. As I have mentioned, Judge 40 Berner made a positive finding that this load was not delivered but the position would be much the same if he had merely said that SDM had not discharged the burden of showing that the goods were delivered. In my judgment an explanation has to be found if the non-delivery of this load is not to put in doubt Mr Blunsden’s claim that all the other loads for which he was responsible were 45 delivered. 58 Conclusions[204]I come at this point to Mr Barlow’s argument that either all of the loads were delivered to the Aldi warehouse or none was. If that proposition is correct it follows that, if I am right in my other conclusions, HMRC’s appeal should be allowed (save perhaps in respect of movement 12, destined 5 for Latvia) since I have determined that, on the balance of probabilities, six of the movements (17, 19, 22, 24, 29 and 37) could not, or are not shown to, have taken place as claimed, and that there are serious doubts about two others (38 and 43) which might have been possible only if certain assumptions for which there was no clear evidence 10 are made; indeed, I have already said that I think it unlikely that either of those movements took place as claimed. Of the nine allegedly impossible journeys— like Judge Berner I leave movement 44 out of account—only one (57) was plainly possible. However, rather than simply accept Mr Barlow’s proposition that this is an “all or nothing” case I have decided that I should examine it. 15205. At first sight it seems possible that some of the drivers could have delivered their loads to the warehouse, while others did not. However, if one accepts that one or other of the two competing scenarios must be, at least in general terms, correct, I agree that logically a binary approach is necessary, and in this too I differ from Judge Berner. The alternative, that some loads were delivered to the 20 warehouse by drivers not involved in the fraud and then diverted while others were not delivered there at all, seems to me to import not only the implausibilities of each of the scenarios as they were urged upon us by the parties but also some implausibilities of its own. Neither party suggested that the alternative was a realistic possibility. 25206. I accept, without repeating the reasons, that both of the competing scenarios give rise to questions for which there is no obvious answer. Neither is supported by clear evidence and instead the choice between them, if choice is to be made, must be based on impression. For that reason I do not treat the choice as determinative, but as a means of checking the realism of my other conclusions. 30207. Had I to make a choice I would encounter greater difficulty with Mr Barlow’s theory that the goods were delivered and then diverted, because of the drivers’ own evidence. They all said in their witness statements that the goods were unloaded from their vehicles in a loading bay to which they had been directed, and in plain view, and that they had entered the office in the building in 35 order to have their paperwork dealt with; none said that the goods were unloaded in an area which was not overlooked or that the papers were taken from them outside and later returned, a course which would have enabled a dishonest Aldi employee to falsify the stamps and signatures while not making the drivers suspicious. Some also spoke of a security guard at the entrance to the warehouse 40 compound though others said there was no such guard; if there was a guard it seems likely that part of his job was to check incoming and outgoing traffic. Unless there was a fraud involving the entire Aldi staff, a proposition for which there is no evidence at all, it is difficult if not impossible to understand how, in those circumstances, the arrival and subsequent departure of 63 consignments of 45 unexpected goods could have been concealed from the honest staff at the warehouse, and it is equally difficult to understand how the involvement of even 59 some of the Aldi staff was not discovered in the course of the Belgian investigation.[208]I do not, therefore, find anything in a comparison of the competing scenarios which undermines the conclusion that, if some of the loads were not delivered, the probability is that none of them was. I accept 5 that SDM would not be able to produce genuinely receipted copy 3 (or copy 2) AADs or CMRs whichever scenario is correct, but there are other factors which, taken together, seem to me to be significant. The first is that none of the drivers was able to produce a tachograph disc relating to the relevant movements. The drivers offered 10 various explanations for their absence—destruction to avoid detection of a failure to take the requisite breaks, or theft, for example. In any one case the reason might well be true, but I find it surprising, though it too is by no means a determinative factor, that not one disc could be found.[209]I have also examined the evidence relating to the other movements, not 15 involving apparently impossible journeys. The documentation is in many cases incomplete, though there are differences from one movement to another. In some cases, for example, there is no evidence of collection from the despatching warehouse, in others no evidence of a cross-Channel journey, in others still no evidence of a return load. Occasionally more than one category of document is 20 absent. Of course, the absence of documents does not demonstrate that the goods were not delivered but it does make it difficult if not impossible to conclude that that there is sufficient secondary or alternative evidence of delivery to demonstrate that, on the balance of probabilities, they were.[210]I should mention, if only briefly, Miss Simor’s argument that the documents 25 in several cases show a lengthy interval, measured sometimes in days, between collection of the goods from the despatching warehouse and the check-in of the vehicle at Folkestone, an interval which allowed for the illicit removal of the goods. As I have not heard the drivers’ explanations of them I do not find it possible to attach a great deal of weight to the delays, though I do find it 30 surprising that if their evidence is correct they were willing to leave unattended, sometimes for several days, loads of valuable goods which Mr Parnham in particular recognised were vulnerable to theft.[211]For the reasons I have given I have concluded that the better view is that all of the movements destined for the Aldi warehouse arrived there, or none did. I am 35 satisfied, on the balance of probabilities, that some of the 63 consignments supposedly sent to the Aldi warehouse at Vaux-sur-Sûre loads cannot have arrived there. In consequence SDM has failed, in my judgment, to discharge the burden of showing that any of the 63 movements was properly discharged or, for the sake of completeness, that the place at which the goods were diverted was in an identified 40 or identifiable Member State other than the United Kingdom. It follows that it is liable, as guarantor, for the duty on those consignments.[212]I come, finally, to movement 12 to Latvia. As I have said, it differed from the others in the identities of the vendor, the purchaser and the despatching and receiving warehouses, and the driver, Mr Bunce, was responsible for no other 45 load, although he, or Connie which he managed, did provide some of the drivers of those movements. There was, again, no receipted copy 3 AAD although there was an apparently genuinely receipted CMR; there was no further relevant 60 documentary evidence. Judge Berner, whose decision is the subject of this appeal, did not deal with the movement at all, but instead allowed SDM’s appeal in respect of it, one must assume on the basis that if he was right in finding that the remaining movements (apart from 29 and 65) were properly discharged, so too was movement 12. I need to consider, therefore, what 5 was said about this movement by F-tT 1.[213]At [466] they said:
“Although Mr Bunce’s evidence overall was not impressive, we found his evidence in respect of the actual movement to Latvia to be credible. He was 10 able to describe how he undertook the journey and to give details of the premises. It was a six-day return journey to a remote destination entirely outwith his usual work. It is not unreasonable that he should remember the journey, even after such a long period of time. The fee paid by SDM to Connie was £1,425 compared with £475 for Movement 11 to Aldi. His 15 evidence was that he did the longer trips himself. A signed CMR was produced stamped ‘UNISTOCK Latvija Riga’. It was never suggested that either the stamp or the signature might be forgeries or where or how they might have been applied other than at Unistock.”
[214]Perusal of the transcript of Mr Bunce’s oral evidence bears out F-tT 1’s 20 description of it as unimpressive. He was vague in much of what he said of the various movements in which he had had some involvement, and claimed to have forgotten a good deal, including the detail of the beginning of this journey. Indeed, what he said on the topic differed from time to time. However, as F-tT 1 said, he did give a clear description of most of the journey, which they found to be 25 credible. I do not think it is open to me, when I have not heard Mr Bunce, to come to a different conclusion when there is no reason to think that the journey was impossible and when there is some documentary evidence, even if it is incomplete, to support it. I accept that the Latvian authorities have said that the load did not arrive, but I also accept that there may have been some confusion 30 about the identity of the consignment. My conclusion in respect of this movement is that SDM have discharged the burden of showing that it was properly delivered.[215]For those reasons I would allow HMRC’s appeal in respect of movements 1 to 11, 13 to 28 and 30 to 64 in each case inclusive, and dismiss it in relation to movement 12. I would dismiss SDM’s appeal in respect of movement 29. As I 35 have explained, movement 65 is not before us. JUDGE CANNAN[216]I gratefully adopt the detailed description of the circumstances in which this appeal comes before us and of the submissions of the parties set out at [1] to [137] 40 of Judge Bishopp’s decision. I agree with Judge Bishopp for the reasons he gives at [138] to [146] that Mr Barlow’s submissions criticising the decision of UT 1 are to be rejected. I also agree with Judge Bishopp for the reasons he gives at [153] and [154] that Miss Simor’s criticism of F-tT 2 for allowing Mr Barlow to make a submission that the times on the documents might be inaccurate is also to be 45 rejected. 61[217]I have reached a different conclusion from Judge Bishopp in relation to the two grounds of appeal which HMRC have permission to pursue in this appeal. I too have found it convenient to deal with those grounds in reverse order.[218]The basis of the second ground was that in following paragraphs 2(a) and (b) of the directions given by UT 1 (“Stage 1”), Judge Berner 5 wrongly asked himself whether any of the allegedly impossible journeys could have been accomplished by reference to a theoretical, rather than a practical or reasonable test of possibility.[219]The starting point in relation to this ground of appeal is the decision of UT 10 1. At [101] it remitted the case to F-tT 2 to determine whether the allegedly impossible journeys took place, and to consider what effect its conclusion on that issue had in relation to the evidence of other deliveries which were not alleged to be impossible. UT 1 envisaged a two stage approach, albeit an approach which was subsequently refined: 15a. The first stage was clearly to determine whether on the balance of probabilities the allegedly impossible journeys actually took place with the goods being delivered to Aldi.b. The second stage was to consider the effect of those findings on the other journeys and deliveries. 20 220. This approach reflected HMRC’s case before F-tT 1, namely that it was unlikely the goods were delivered to Aldi because, amongst other reasons, the alleged journeys said to support delivery were in some cases impossible by reference to the times identified in the documentation.[221]The circumstances in which UT 1 came to make its directions and the terms 25 of the directions it gave have been set out by Judge Bishopp. At the hearing before UT 1 to consider those directions Mr Barlow had submitted that no witnesses should be called at the remitted hearing, but that F-tT 2 should examine the documentary evidence in relation to the allegedly impossible journeys and take into account F-tT 1’s findings of fact. Miss Simor had submitted that there should 30 be a complete re-hearing.[222]In the event, UT 1 directed a modified two stage approach, which at paragraph 2(c) (“Stage 2”) left it to F-tT 2 to decide whether and to what extent it needed to hear further evidence. The directions themselves also included a Note from UT 1 as follows: 35 “We recognise that these directions leave open the possibility that the Firsttier Tribunal may decide that it needs to rehear the evidence …”[223]We were told that there were no submissions to Judge Berner as to how he should approach Stage 1.[224]Civil courts and tribunals do not generally resolve issues by reference to 40 what is possible or impossible. They are concerned with the balance of probabilities and what is likely or unlikely. In some circumstances assessing what is likely or unlikely can involve consideration of what is possible or impossible.[225]The question of whether or not some of the journeys were impossible appears to have arisen for the first time in HMRC’s cross-examination of the 45 drivers. That led to the submission before F-tT 1 that at least some of the journeys 62 were impossible and therefore F-tT 1 should find that it was unlikely the goods for those journeys were delivered, and by inference the goods for the other journeys. On analysis HMRC’s case in this respect was simply that taking into account the timings on the documents, it was unlikely the journeys were made and the goods delivered to 5 the Aldi warehouse.[226]The ground of appeal which succeeded before UT 1 was what was described as “the narrower ground” of Ground 3. Ground 3 was that some of the journeys to Aldi were impossible which meant that F-tT 1 was not entitled to find that any of the movements involved deliveries of goods to Aldi. That ground of appeal 10 succeeded, but only to the extent that F-tT 1 had failed to give reasons, either for finding that the journeys were possible, or as to how it was able to reconcile movement 29 which F-tT 1 had itself found to be impossible.[227]Following the hearing before UT 1, consideration was given to the fact that the appeal could not be remitted to F-tT 1 as originally constituted. The focus was 15 by then on the allegedly impossible journeys, both as to whether or not they were possible, and if not what effect such a finding would have, together with the impossibility of movement 29, on the finding that on a balance of probabilities the deliveries took place.[228]UT 1 clearly wanted to avoid a re-hearing unless it was unavoidable. It was 20 for that reason that it set out a prescriptive approach to be followed by F-tT 2 whilst at the same time recognising that a re-hearing might turn out to be necessary.[229]With respect to UT 1, and indeed with the benefit of hindsight, I accept Miss Simor’s submission that such an approach was likely to lead to difficulties. 25 Having said that neither party sought to appeal the directions made by UT 1. Further, neither party nor indeed F-tT 2 sought any clarification or variation to those directions. It falls to be considered therefore whether F-tT 2 properly followed those directions.[230]In directing F-tT 2 to consider whether an allegedly impossible journey 30 “could not have taken place”, UT 1 must have been directing consideration of whether such a journey was reasonably possible. In that sense I agree with Miss Simor that the direction is not concerned with what she has described as “theoretical possibility”. UT 1 clearly chose those words in direction 2(a) carefully. Hence when it came to give directions for Stage 2 in paragraphs 2(c) 35 and (d) it was concerned with findings either that the journeys “could not have taken place” or “could have taken place”. In the event that F-tT 2 found that the journeys could not have taken place it was required in paragraph 2(c)(i) to consider the effect of that finding on the conclusion that the goods carried in the allegedly impossible journeys were delivered to Aldi. In paragraph 2(c)(ii) F-tT 2 40 was required to consider the effect on the conclusion that the goods on all the other journeys were delivered to Aldi.[231]Stage 2 of the process required F-tT 2 to consider what amounted to the real issue in the whole appeal, namely whether on a balance of probabilities the goods were delivered to the Aldi warehouse. UT 1 left to F-tT 2 the task of determining 45 what evidence it might need to hear at Stage 2 in order to resolve that issue, both in relation to the allegedly impossible journeys and in the other journeys. 63[232]UT 1 must have considered that F-tT 2 might be in a position at Stage 1 to reconcile the allegedly impossible journeys to the evidence heard by F-tT 1. If that was the case, then it anticipated that F-tT 2 would give reasons as to why it was the case, which would deal with the error of law it had identified. That process was paragraph 2(d) of the directions and formed 5 part of Stage 2.[233]UT 1 was not asking F-tT 2 to consider at Stage 1 whether it was more likely than not that the journeys did in fact take place. That was Stage 2 and could only be answered in the light of all the evidence available, including any new evidence that F-tT 2 considered necessary in order to determine the issue. 10 234. Judge Berner conducted the enquiry required by Stage 1. He was concerned with matters of possibility and not probability because that is how UT 1 had directed him to approach the task. Both parties sought to establish before him the earliest time by which a driver could return to Coquelles in time for the known check-in time. Various evidence was considered in relation to crossing times, 15 routes, rest breaks, speed and loading times.[235]Stage 1 required F-tT 2 to have regard to the evidence of the drivers in relation to the allegedly impossible journeys. It also required F-tT 2 to have regard to the findings of fact made by F-tT 1, although not those findings it had made in relation to those journeys (paragraph 2(b) of the direction). In other words, F-tT 2 20 was not to be constrained by the findings of F-tT 1 in relation to the allegedly impossible journeys. That is plainly because UT 1 had already found that F-tT 1 had failed to give adequate reasons as to why, apart from movement 29, those journeys were possible.[236]Judge Berner was in a difficult position because in many respects there was 25 a lack of clarity in the evidence of individual drivers. For example as to the time from checking in at Folkestone to exiting at Coquelles, as to whether they took rest breaks, and as to whether they exceeded the speed limit. I do not see how that lack of clarity could be resolved without making findings of fact, and indeed without hearing the evidence of the drivers themselves. However it was not 30 intended that F-tT 2 should hear further evidence at Stage 1.[237]Judge Berner was only concerned at Stage 1 with what was reasonably possible. I consider that it was therefore necessary for him to identify the quickest possible journey time for each movement, having regard in particular to the evidence of the driver for that movement. He clearly recognised that what was 35 possible in terms of the direction had to be reasonably possible. Hence at [132] he stated that he would not have been prepared to accept a cruising speed in excess of 80 mph. The evidence of Mr Blunsden was that 70-80mph was the maximum comfortable cruising speed of the vehicles. That evidence had not been challenged, nor indeed was it questioned by Mr Coles the member in F-tT 1 who 40 was recorded as having considerable experience of the European transport industry.[238]Similarly, Judge Berner’s application of benchmark times was expressly by reference to what he considered was reasonable. At [67] he stated: “ Having regard therefore to all the relevant evidence in this connection, I 45 conclude that it is reasonable, as a starting point, to apply as the base times for the journeys set out in the table the shortest of those which are described in the table.” 64[239]In my judgment therefore the first ground of appeal is not made out. Judge Berner was applying the test of reasonable possibility required by the directions of UT 1.[240]The second ground on which permission to appeal was granted was that Judge Berner did not confine himself to the evidence before 5 F-tT 1, as required by the direction of UT 1. Essentially Miss Simor’s submission was that at Stage 1 Judge Berner did not consider whether the journeys could not have taken place “as described in the evidence of the drivers”. Instead she submits that he took account of parts of the drivers’ evidence whilst selectively disregarding other 10 parts of their evidence. Under this ground of appeal Miss Simor submitted that the Judge was wrong:a. to find that the drivers had not been asked or given evidence on the quickest time a journey could be undertaken.b. to find that even though certain drivers’ evidence was that they did not 15 drive in excess of speed limits, they may in fact have done so.c. to permit SDM to argue that the documents may not be accurate.[241]Miss Simor also submitted that in considering the issue of possibility or impossibility, Judge Berner adopted a wholly unrealistic approach, in particular in his consideration of the average speed at which HGVs could travel when fully 20 laden. It was submitted that the average speeds of 68 mph and 75mph required to complete movements 24 and 17 respectively were wholly unrealistic and led to a perverse conclusion that those journeys were possible.[242]In her skeleton argument Miss Simor criticised Judge Berner’s approach to the evidence and as to whether rest breaks should be factored into the timings. In 25 her oral submissions she also criticised Judge Berner’s allocation of the time for loading and unloading. In the event however Miss Simor did not significantly rely on those submissions save in so far as it might be necessary for us to remake the decision. In light of the fact that Judge Berner was seeking to identify the quickest possible journey time I see no reason to criticise his decision in this regard. There 30 was evidence to justify his approach to rest breaks and the time for loading and unloading.[243]Miss Simor had provided Judge Berner with a print from the Eurotunnel website which stated that “the transit time between the truck checking-in and its arrival on the motorway on the other side of the Channel is 90 minutes”. That 35 evidence was also made available to us and Miss Simor submitted that one hour might be a matter of possibility, but 90 minutes was a matter of probability. For the reasons given above Judge Berner was concerned with matters of possibility, so he was right to take a crossing time of one hour for the purposes of Stage 1.[244]Miss Simor’s principal submission was the speed at which the journeys 40 would have to be driven meant that they were impossible in light of the timings on the documents.[245]In carrying out his enquiry at Stage 1, Judge Berner considered all the documentary evidence in relation to the allegedly impossible journeys, together with the transcripts of the drivers’ oral evidence. It is clear that Judge Berner did 45 take into account evidence which was not before F-tT 1. In some respects he was 65 encouraged to do so by the parties. Firstly he was provided by Miss Simor with the Eurotunnel evidence as to the crossing time from Folkestone to Coquelles. Secondly the Google Maps which were provided to Judge Berner do not appear to be the same Google Maps which were before F-tT 1. Further, they were not comprehensive and Judge Berner himself obtained information 5 from Google Maps to complete the picture. Thirdly, both parties obtained evidence as to the timings of vignettes. Neither party criticised the Judge for taking any of this evidence into account.[246]Judge Berner’s justification for admitting new evidence at this stage was set 10 out at [35]. I agree with Judge Bishopp that Judge Berner’s explanation, coupled with the ordinary bounds of judicial discretion and application of the overriding objective is sufficient to justify his doing so. I have no doubt that if UT 1 had been asked to vary its direction accordingly it would have done so.[247]The drivers’ evidence relevant to the quickest time in which the journeys 15 could be undertaken was essentially their estimates of times for various stages of the journeys and as to whether they observed speed limits. Unsurprisingly, the evidence given by the drivers as to journey times and speeds was all different. There was also the evidence from Google maps. All this evidence was considered by Judge Berner at [58] to [64]. 20 248. The first point to be made is that I agree with Judge Berner at [59] where he says that the drivers did not in their evidence specifically describe the shortest possible time for the journeys to Aldi. I do not accept that in those circumstances Judge Berner was bound to rely solely on evidence as to typical journey times. Whilst the legal burden was at all times on SDM to establish on the balance of 25 probabilities that the journeys took place, it was HMRC which adopted the position that some of the journeys could not have taken place because they were impossible by reference to the timings on the documents. The question of the impossible journeys was relevant to the evidential burden of establishing that the deliveries were made, referred to by Judge Berner at [42] of his decision. No 30 doubt that was why Miss Simor sought to cross examine the drivers with a view to establishing that some of the journeys were impossible. However the focus of the evidence before F-tT 1 was not clearly on the quickest time possible for various stages of journeys as it was before Judge Berner.[249]The findings of F-tT 1 as to whether the journeys were possible would 35 clearly depend on the drivers’ answers to questions intended to identify the shortest time in which the journeys could take place. It is only then that F-tT 1 or indeed F-tT 2 could say whether the journeys were impossible. The fact that the drivers’ evidence did not specifically deal with the shortest possible time for the journeys, by which I mean the shortest reasonably possible time, may be because 40 the issue was not so well defined before F-tT 1 as it has become at subsequent hearings. For example, it was not put to Mr Blunsden that an average speed of 75mph might be required to complete movement 17 in time for check-in at Coquelles. Since the hearing before F-tT 1 there has been a much more detailed and forensic analysis of routes, times and speeds of the journeys. 45 250. The detailed evidence from the drivers as to the time it would take to drive between Coquelles and Vaux-sur-Sûre was summarised by Judge Berner at [59]. I accept that as a fair summary of the drivers’ evidence given in cross examination. 66 It was on the basis of a typical journey, rather than the shortest possible time for a journey.[251]The evidence as to whether the drivers ever broke the speed limit is summarised by Judge Berner at [63] and [64]. Mr Blunsden drove 6 of the 8 allegedly impossible journeys we are concerned with 5 (movements 17, 19, 22, 38, 43 and 57). He was also identified as the driver for movement 29. Mr Blunsden was not asked in cross examination whether or not he broke the speed limit. He did say in re-examination that he did not always observe the speed limits.[252]A review of the transcripts shows that Mr Blunsden was asked in detail 10 about movements 17, 29 and 43 and it was put to him that those movements were impossible. The only questions he was asked about movement 57 were in relation to the trailer that was used. He was asked no questions about movement 38. He was not asked about timings on the route he claimed to have taken on movement 19 or in relation to that part of his route which involved a fuel stop at Watou on 15 movement 22.[253]Mr Wild drove one of the allegedly impossible journeys (movement 24). He did not give oral evidence and therefore could not be asked about it. His witness statement did not deal with timings because the issue only arose at the hearing.[254]Mr Parnham drove one of the allegedly impossible journeys (movement 37). 20 He was asked about speeding in cross examination, albeit in the context of the destruction of his tachographs and in the light of his witness statement where he stated that he would drive to Aldi as quickly as possible because his insurance cover did not cover transport of alcoholic drinks.[255]In cross examination Mr Parnham was vague about journey times for 25 various stages of the journey from Folkestone to Vaux-sur-Sûre returning via Geer. It appears that he was involved in only 5 movements to Aldi all of which took this route. His evidence included reference to journey times of 3½ hours between Coquelles and Vaux-sur-Sûre, an hour and a bit from Vaux-sur-Sûre to Geer and 3 hours from Geer to Coquelles. He gave those answers in relation to the 30 journey generally and not by reference to specific movements. He also said “don’t hold me right to these without checking on a map”.[256]It was not put to Mr Parnham that those times were impossible, although it was suggested in passing that a journey time of 3½ hours between Coquelles and Vaux-sur-Sûre would involve breaking the speed limit. Later in the cross 35 examination Mr Parnham was taken specifically to movement 37. The times put to him as impossible were not the shortest times he had referred to in his earlier evidence. What was put to him as impossible was 4½ to 5 hours from Coquelles to Vaux-sur-Sûre, 1½ hours from Vaux-sur-Sûre to Geer and 3½ to 4 hours from Geer to Coquelles. He maintained that he had made the delivery to Aldi. 40 257. I agree with Judge Berner’s conclusion at [64] that in its proper context Mr Parnham’s evidence does not remove the possibility that he might, on movement 37, have exceeded the speed limit.[258]Miss Simor submitted that Judge Berner had failed to take into account factors such as speed cameras, acceleration and deceleration, road works and other 45 likely road conditions. However there was no evidence at all before Judge Berner as to specific road conditions. He was required to consider what was reasonably 67 possible and was entitled to assume optimal driving conditions. It is not impossible that a small number of journeys could be made under optimal driving conditions.[259]The third criticism made of Judge Berner’s decision under this ground is that he permitted SDM to argue that the documents were not 5 accurate. As I have already said, I agree with Judge Bishopp for the reasons he gives that this criticism should be rejected.[260]Finally, Miss Simor submitted that it was perverse for Judge Berner to conclude that movements 17 and 24 were possible in circumstances where the 10 average speed required to complete each journey was 75 mph and 68 mph respectively. As I have already stated, Judge Berner was entitled to approach his task on the basis that it was possible the drivers exceeded the speed limit, which for HGVs on a motorway was 90 kph or 56 mph. There was no evidence as to whether on a small number of occasions they might have consistently exceeded 15 the speed limit because that question was never asked. There was however unchallenged evidence from Mr Blunsden that modern HGVs could cruise comfortably at 70-80 mph. It was accepted, save for very short link roads at each end, that the journey from Coquelles to Aldi’s warehouse at Vaux-sur-Sûre was motorway to motorway. In the absence of any evidence as to specific road 20 conditions I consider Judge Berner was entitled to treat it as possible that the vehicles could have cruised at such an average speed.[261]Judger Berner recognised at [67] and [132] that if the question had been whether it was likely that movement 17 could be achieved in the times indicated by the documents the answer by reference to the evidence before F-tT 1 would be 25 no. However he did note that evidence from the driver, which I take to mean evidence at a rehearing, could persuade him that the journey was undertaken. That would be a matter for Stage 2 of the directions. In the circumstances he found that movement 17 was possible and I am not satisfied that he was wrong to do so.[262]For these reasons I do not consider that the second ground of appeal is made 30 out. Judge Berner’s decision was consistent with the underlying evidence and his conclusion that the journeys were possible was not perverse.[263]Judge Berner rightly went on to consider Stage 2 in relation to movement 29 which F-tT 1 and UT 1 had found was impossible. Judge Berner reconciled the impossibility of that movement with F-tT 1’s finding that the goods had been 35 delivered in relation to all the other movements to Aldi, albeit for different reasons from F-tT 1. HMRC do not appeal against Judge Berner’s approach to movement 29 at Stage 2. In particular they do not appeal Judge Berner’s decision to reject what he described as the binary approach of both parties. As previously indicated, this was the proposition of both parties that either all the goods were 40 delivered to the Aldi warehouse or none of the goods were delivered.[264]If I had been satisfied that F-tT 2 had fallen into error and was required to remake the decision then I would want to hear further evidence from the drivers in order to make my own findings. As UT 1 stated at [77], one of the reasons it did not reach any conclusion on whether the journeys were impossible was that it had 45 not heard the witnesses. That point assumes added significance where the result of finding that the journeys were impossible must be that there was a conspiracy involving the individual drivers and the directors of SDM who F-tT 1 have 68 previously found to be honest witnesses. I do not consider that it would be right to decide the principal issue as to delivery in this appeal by reference to the burden of proof. In general terms, either the drivers were telling the truth, or they were lying.[265]Henderson J noted in Ingenious Games LLP 5 v Commissioners of HM Revenue & Customs [2015] UKUT 0105 (TCC) at [65]:
“ … as the FTT rightly recognised, it is not open to the tribunal to make a finding of dishonesty in relation to a witness unless (at least) the allegation has been put to him fairly and squarely in cross examination, together with 10 the evidence supporting the allegation, and the witness has been given a fair opportunity to respond to it.”
[266]If there is to be a forensic analysis of journey times, I consider that Mr Blunsden and Mr Parnham would be entitled to have that analysis fairly put to them in cross examination. The question for those drivers, in light of the specific 15 material and analysis this tribunal has been taken to, would be to explain in detail how each journey was possible. Those questions were put to Mr Blunsden in relation to movement 29, but they were not put in relation to all the other movements now said to be impossible. Certainly it does not appear to me that the detailed analysis now relied on by HMRC as to why movement 37 was impossible 20 was put to Mr Parnham.[267]I am conscious that Mr Barlow on behalf of SDM had been strenuously resisting any form of re-hearing. Indeed both parties before this tribunal invited us to remake the decision in the event that the appeal succeeded. For the reasons given above I do not think it appropriate to do so without hearing evidence from 25 the drivers.[268]Finally, I must deal with SDM’s cross-appeal in relation to movement 29. Mr Barlow argued that having concluded the other journeys were not impossible, Judge Berner ought to have concluded not only that those deliveries were made but also that the delivery for movement 29 was made. 30 269. In dealing with Stage 2, Judge Berner rejected the binary approach put forward by both parties. Whether he was right to do so, in my view, depends on the context of movement 29. For example F-tT 1 was not satisfied that movement 65 to Germany was delivered, but the context of that movement was different to the other movements under consideration. UT 1 did not anticipate that further 35 evidence would be adduced in relation to paragraph 2(d) of the directions were FtT 2 to find that the allegedly impossible journeys were in fact possible. It was to have regard solely to “the other findings contained in the Decision [of F-tT 1]”.[270]In this context it seems to me that UT 1 did not consider that the binary approach should necessarily apply in relation to movement 29. Otherwise there 40 would have been no purpose in remitting the appeal. I consider that it was open to Judge Berner to find that in relation to one journey, movement 29, SDM had not satisfied the burden of showing that the goods were delivered to Aldi. It had therefore not established that the diversion occurred outside the UK and its appeal in relation to that movement fell to be dismissed. 69[271]In my view therefore both the appeal and the cross appeal should be dismissed. However in relation to the appeal that is subject to Judge Bishopp’s casting vote. 5 DISPOSITION[272]By the exercise of Judge Bishopp’s casting vote, HMRC’s appeal is allowed in respect of movements 1 to 11, 13 to 28 and 30 to 64, in each case inclusive, but dismissed in respect of movement 12. SDM’s appeal in respect of movement 29 is dismissed. 10 273. Any application for costs must be made within the prescribed time limit but need not be accompanied by a schedule of the amounts claimed. Colin Bishopp 15 Upper Tribunal Judge Jonathan Cannan Upper Tribunal Judge 20 Release date: 19 November 2015 Signed on original

Cited in 4 later judgments