Introduction: a debate re-opened?
The Financial Times reported on 26 July that there is:
a ‘live debate’ in government about whether Britain should quit the customs union
The FT points out that staying inside the customs union:
could restrict Dr Fox’s ability to strike new trade deals[,] or prevent them altogether.
Indeed, that seems to be rather understating the case since ‘Common customs duties on imports from outside the EU’ is one of the four founding principles of the EU Customs Union. And it seems hard to envisage entering into trade negotiations with tariffs on imports to the UK off the table.
[ADDED 18 SEP 2016: It should be noted as a matter of precision that, contrary to what is often said, Turkey is not part of the EU Customs Union. Rather, it formed a Customs Union with the EU in 1995. It may be that the debate in Government is not quite as reported, but is about whether we should enter into a Customs Union agreement with the r-EU. But the same consequences would follow for our ability to enter into free trade agreements with third countries.]
Surely the decision to create a new Department of International Trade must have been based on an assumption that we would not remain in the Customs Union? According to its web-site, the new Department is responsible for:
- developing, coordinating and delivering a new trade and investment policy to promote UK business across the globe
- developing and negotiating free trade agreements and market access deals with non-EU countries
- negotiating plurilateral trade deals (focused on specific sectors or products)
- providing operational support for exports and facilitating inward and outward investment
Clearly, it would not be able to fulfill all these responsibilities if we were to remain in the EU Customs Union. Moreover, Theresa May would hardly have appointed Liam Fox, who in his recent campaign to become leader of the Conservative Party had advocated the creation of a new department of trade, presumably precisely because of an anticipated post-Brexit freedom to enter into preferential trade agreements with non-EU countries, if she had not expected that the UK would be able to exercise that freedom.
Be that as it may, and whether or not this amounts to the re-opening of a matter that had previously been closed, it would appear that there is now a debate within government about the merits of remaining within the Customs Union.
‘Up to 24%’ versus ‘2 to 24%’
The same article in the FT gives an alarming estimate of the potential cost of the reimposition of border controls:
It may be worth emphasising that HM Treasury’s claim is not (say) that the transaction costs of trade could increase by 24%, but that the extra costs could amount to 24% of the value of the goods themselves. This is an astonishing claim, which is worthy of investigation as to the basis upon which it is founded.
Contrary to the impression given by the wording, the quoted words are from the Treasury report, and not from the OECD study. The Treasury report is its April 2016 ‘ long-term economic analysis of EU membership and its alternatives’. At paragraph 1.26, we do indeed find:
The OECD has estimated that crossing the border, documentation and other delays can increase the transaction costs of trade by up to 24% of the value of traded goods.
Reference no. 18 is to OECD Trade Policy Paper No. 150, published in 2013, and entitled ‘Trade Costs: What Have We Learned? A Synthesis Report’ (henceforth ‘OECD Trade Costs 2013’). In the Executive Summary is the following statement:
Crossing the border, documentation and customs compliance requirements, lengthy administrative procedures and other delays can increase transaction costs an estimated 2 to 24% of the value of traded goods.
It seems to me that quoting the 24% and not the 2% is being selective with data in an inappropriate way, since for most readers, it will skew the idea formed of likely costs towards the upper end of that range.
The basis of the ‘24%’
Where does the high end of the OECD’s range of estimates come from? The short answer is that it is a figure for possible costs in some developing countries, deriving from a World Bank survey of 80 countries, and is not relevant to the question that the Treasury was addressing, which is the possible cost of the reintroduction of custom controls between the UK and r-EU. Therefore it should not have been cited at all. Rather, the 2% figure at the low end of the range would have been more in line with estimates found in the literature for developed countries. To demonstrate this requires tracing the source of the OECD’s estimates back through several stages.
As has been said, the Treasury paper quoted from the Executive Summary of OECD Trade Costs 2013. The study concerns the entirety of trade costs from exporter to importer, which are helpfully broken down into costs before the border, at the border, and beyond the border, with transport costs running through from start to end of the trade chain:
The term TTC is apparently (see paras. 60-61) used in this report specifically for at-the border Trade Transation Costs, which are sub-divided into three, as shown in the central part of the above diagram:
The text reads (para. 65):
Among TTCs, a common distinction is usually made between direct TTCs, namely those related to border and customs procedures, and indirect and hidden TTCs, including delays at the border, opportunity costs, bribery, corruption and informal trade. Existing literature commonly estimates that direct TTCs amount to 2-15% of the value of traded goods, while indirect TTCs range between 1% and 24% of that value (OECD, 2009, chap.1).
As is clear from the diagram, indirect TTCs include ‘delays at the border’ and ‘opportunity costs’ (should this perhaps be the ‘opportunity costs of delays at the border’?), while the hidden TTCs are those resulting from bribery, corruption and informal trade. No estimate is given here for hidden TTCs, but there are estimates of a range of 2 to 15% of the value of traded goods for direct costs, and 1 to 24% for indirect costs. It should be noted that while tariffs are included in the diagram among the ‘direct costs’, they are not included among the direct TTCs.
The figure of 24% appears only one more time in the report, at para. 130, again as the high end of the range of estimates for indirect TTCs. This leaves a puzzle as to the provenance of the 2 to 24% range of estimates given in the Executive Summary, which was for ‘Crossing the border, documentation and customs compliance requirements, lengthy administrative procedures and other delays’, which would seem to be inclusive of both direct and indirect costs. One would surely have expected therefore the range of estimates given there to have been higher, perhaps even increased by simple addition to 3 to 39%, since there does not seem to be any very obvious reason why a high direct cost should not coincide with a high indirect cost.
Leaving aside this apparent mismatch between the body of the report and the Executive Summary, it is clear that both the high estimates, of 15% for direct TTCs, and of 24% for indirect TTCs, are of concern, since if the cost of a reintroduction of a customs border between the UK and the EU even approached either one of these values, there would be major hindrances to trade. As indicated by para. 65, shown above, of OECD Trade Costs 2013, both estimates derive from an OECD study published in 2009. The earlier study, entitled ‘Overcoming Border Bottlenecks: The Costs and Benefits of Trade Facilitation’ (henceforth ‘OECD Bottlenecks 2009’), addressed (p. 3) ‘the negative impact of inefficient border procedures’, and sought to ‘shed light on the economic significance of overcoming border bottlenecks through trade facilitation’.
P. Walkenhurst and T. Yasui, the authors of Chapter 1 of OECD Bottlenecks 2009, make reference (p. 21) to previous estimates of direct TTCs:
They cite a survey of the literature contained in a study published in 2002, again by the OECD, as indicating that
direct TTCs involved in export and import procedures amount to 2-15% of the value of traded goods
and this is clearly the source of the high end estimate of 15% for direct TTCs that we are looking for. A footnote is added to the effect that some of these studies in the 2002 OECD survey may include elements of indirect TTCs also. Another survey by SWEPRO found the same range, but other studies gave estimates of only about 1% of the value of traded goods.
It is not clear to me whether ‘export and import procedures’ refers only to at-the-border procedures, or to the whole export and import process from exporter to importer. As will be seen below, this is a crucial distinction.
The 2002 OECD study, entitled ‘Business Benefits of Trade Facilitation’ (henceforth ‘OECD Facilitation 2002’), observed (para. 1) that:
As tariff levels have declined through successive GATT/WTO rounds, and global supply chains have come to dominate production patterns, growing attention has been directed to the remaining cost factors that are important for international competitiveness, in particular those incurred by trade formalities and procedures.
One of the aims of the paper (para. 3), which had its origins in a UK proposal, was to review existing literature on ‘the business costs of trade transactions that would be reduced by trade facilitation’. The scope of the paper, unlike OECD Bottlenecks 2009, is on the whole trade process from seller to buyer but, unlike OECD Trade Costs 2013, it is concerned (para. 5) with ‘procedures and associated information flows required to move goods internationally from seller to buyer and to pass payment in the other direction’, and not with the physical transport of goods. The transactions covered include (p. 6, ‘contact’ should presumably read ‘contract’ in line 1):
It will be noted that ‘border control and clearance’ is only one of six named types of transaction. Clearly, it is necessary to avoid conflating estimates of the cost of all six of these trade transaction costs, with estimates of at-the-border trade transaction costs, especially since these latter are sometimes, as we have seen in OECD Trade Costs 2013, rather confusingly termed TTCs, rather than, say, BTTCs for at-the-Border Trade Transaction Costs, the which term I will henceforth use for clarity. In contrast to the narrower usage, OECD Facilitation 2002 stated (para. 10) that ‘trade transaction costs’:
cover the preparation and presentation of all prescribed documentation or electronic equivalents and the provisions of any required explanations, authentifications and supporting supplementary information to service commercial and official procedures at all stages of the physical movement of the goods from consignor to consignee and the movement of related means of payment in the opposite direction.
It would appear that part of the explanation of high estimates for direct BTTCs is indeed that they have been confused with TTCs. OECD Facilitation 2002 explains:
The US National Committee on International Trade Documentation (US NCTID) figures for total trade transaction costs were later, via an interpretation of them by UNCTAD, presented by Staples and succeeding authors as estimates for the cost of customs formalities. As the authors point out the resulting estimates of 7-10% were ‘misleading and an exaggeration’. But where does the even higher figure of 15% come from?
A possible answer may be gleaned from para. 32 of the same paper:
As I interpret this, US NCITD made estimates, 7.5% in each case, of:
i) total transaction costs from exporter to importer of goods exported from the USA;
ii) total transaction costs from exporter to importer of goods imported to the USA.
There is no need to double this figures, as ‘subsequent researchers’ appear to have done. The figure of 7.5% is for the whole process from exporter to importer, not just for the ‘exporting end’ of it. So this appears to be a classic case of double-counting, so far as I can see.
Difference in scope
Para. 32 above gives four main purposes for which documentation is required, giving rise to associated costs: financing, insuring, shipping and customs. Earlier, as shown above, ‘border control and clearance’ was listed as one of six types of transaction. Let us, for sake of argument, split the difference and assume that costs are distributed between five main forms of transaction, of which customs are one. Then the US NTCID figure of 7.5% for total direct TTCs would yield a value for customs costs of 7.5 ÷ 5 = 1.5% of the total value of the traded goods, one tenth of the 15% figure that entered into the literature.
Changes over time
OECD Facilitation 2002 points out (p. 11) that ‘a number of developments in business practices and in the related regulatory environment have considerably altered the transaction process and certainly affected the ensuing costs’. The first five of these:
give some impression of the degree of harmonisation and automation that had been attained even by 1994. It seems to me that the result of these and later technological and regulatory developments would tend to be in the direction of the reduction of direct TTCs, including BTTCs.
An EU study
For the sake of completeness, OECD Facilitation 2002 does contain one other reference to an estimate of 15% for TTCs:
The description of the Cost 306 project shows no evidence of original research on TTCs. Rather, it was a feasibility project ‘To research and demonstrate how EDI (Electronic Data Interchange) and other computer techniques (in a standardised environment) could be employed instead of the currently used paper document procedures to the benefit of the transport sector.’ It is likely that the figures cited in the report were derived from previous research. Most importantly, as for the US NCTID figure of 7.5%, these are total TTCs, not BTTCs.
Direct TTCs: Conclusion
In conclusion, there is no sound basis for the OECD Trade Costs 2013 (para. 65) high end estimate of 15% for direct TTCs ‘related to border and customs procedures’. Their probable provenance is the US NCTID study which, in the first place, is more than four decades old, secondly, made estimates for total TTCs, not for border BTTCs, and whose figure of 7.5%, thirdly, appears to have been subjected to a double-counting by subsequent researchers, at least so far as I can see.
It was for indirect TTCs, and indeed for BTTCs as I have been calling them, that OECD Trade Costs 2013 (para. 65) gave its range of estimates of between 1 and 24% of the value of traded goods. Its source again was OECD Bottlenecks 2009, where we find at page 21-2:
The authors take border waiting time data from the World Bank and combine it with Hummel’s estimate of value per day lost to yield a range of estimates for the indirect cost of time delays at the border. While the methodology may be open to criticism in detail, the basic idea may yet be sound. In a 2012 version of the paper, with the same title ‘Time as a Trade Barrier’, Hummels and Schaur suggest precisely such an application of their research:
Finally, our results are relevant to the increased emphasis on trade facilitation – identifying regulatory or other nontariff barriers to trade – in trade negotiations and among aid and development groups such as USAID and the World Bank. Many efforts to facilitate trade, such as streamlining customs procedures or improving port infrastructure, generate benefits measured in days saved. With our estimates of the value of each day saved one can then calculate the monetary benefits of these initiatives and how they compare to the cost incurred.
The main problem, with regard to the matter at hand, that is, the Treasury’s citation of figures for the potential cost of the reintroduction of customs controls between the UK and the EU, is that the high end estimate of 24%, which derives from a waiting time of 24 days, is not relevant to the EU at all, but to other parts of the world. Batra et al’s 2003 World Bank study, which was based on the 2000 World Business Environment Survey (WBES) of 80 nations, explains at page 40 that:
It is not clear, to me at least, where the reference to waiting times of 24 days comes from, but possibly it is from the WBES data set (which was included in a CD-ROM with Basra 2003), and possibly it is for the single country of the 80 that had the longest waiting times at ports and customs. Since the enlargement of the EU, it seems likely that Central and Eastern European member states will now have waiting times for non-EU imports closer to those of what were then the OECD nations. And these were 2 days on average, with a median generally of less than half a day.
Taking the figure of 2 days for OECD nations and using the same methodology as OECD Bottlenecks 2009 would give an estimated figure of 2% of the value of traded goods for indirect BTTCs, as contrasted to the 24% that seems to have eventually emerged into the OECD Trade Costs 2013 Executive Summary, and thence to the Treasury paper.
It is important to note also that this World Bank data is for the time ‘for goods to be processed through ports and customs’, and not through customs only. Because the UK has only one land border with r-EU, and because it is not in Schengen, trade with the EU is already generally subject to processing times at sea ports, airports, and their border controls. This time should in principle, it seems to me, be subtracted from the total to give the extra time due to customs controls themselves. If, for sake of argument, half of the time taken is for customs controls, then their indirect cost would be 1%.
It is not my intention here to provide an estimate for the potential cost of the reintroduction of customs controls between the UK and r-EU. Only a detailed examination of the procedures employed at borders for different kinds of cargo from countries with different trade relationships with the EU could yield a realistic estimate. My purpose has been only to demonstrate that the value of ‘up to 24%’ given by HM Treasury had no valid basis.