The unilateral nature of a Generalized System of Preferences (GSP) would allow the UK to apply a preferential regime swiftly, but simply rolling over the scheme of the EU might affect developing countries’ access to the UK market.
Maintaining the lists of beneficiaries, eligible products and tariff margins is the first step towards ensuring that developing countries’ market access to the UK is unaffected by Brexit.
The maintenance of the current rules for preference removal (i.e. graduation) can be problematic. The uneven distribution of trade between the UK and the EU implies that the current import-share thresholds, which determine graduations, can provoke the loss of preferences, either in the UK or the EU, without any change in competitiveness (without any increase in imports). This would be the mechanical outcome of the separation of the UK from the EU27 block, which could be avoided by revising the graduation thresholds upwards.
Mechanical graduations in the UK could involve a substantial value of trade: in 2016, the sectors likely to be subject to mechanical graduations accounted for € 1.27 billion of UK imports, corresponding to approximately € 31.6 million in tariff preferences.
India is the country which would be most affected by the application of the current GSP structure in the UK. India and the Philippines could also be exposed to the risk of mechanical graduations in the block of EU-27 countries.
Exiting the European Union will give the UK back control of its trade policy. This implies that, other than determining what kind of trade relationship it will have with the EU, the UK will have to (a) decide whether to offer any unilateral trade preferences to countries which have relied on them so far and (b) negotiate new Free Trade Agreements (FTAs) with various trade partners around the world.
Currently, when exporting to the EU, developing countries do so under two types of regimes: FTAs, mainly the Economic Partnership Agreements (EPAs), and unilateral preferences, offered under the Generalized System of Preferences (GSP). The UK Government Department of International Trade recently published an update on its approach to transitioning existing trade arrangements and has so far signed continuity agreements with a number of countries. Importantly, the Government has also confirmed its intention to replicate the existing EU GSP scheme, in order to maintain the existing market access provided by these unilateral preference schemes.
This paper focuses on issues which relate to the application of a GSP scheme in the UK. Upholding the current level of access to the UK for developing countries should be a priority in any post-Brexit scenario, to avoid Brexit harming the living standards of people in poor countries (Mendez-Parra, 2017). The unilateral nature of GSP would allow the UK to apply such a preferential regime swiftly, without having to engage in the lengthy negotiations which characterize the process of transitioning existing FTAs in which the UK participates as a member of the EU.
This paper shows that, were the UK to roll over the EU’s GSP, there would still be changes in developing countries’ access to the UK market. I investigate the effect of Brexit on both sides of the coin, i.e. how market access through GSP would change in both the UK and EU-27 bloc, and, thereby, identify the elements of the GSP scheme that would have to be adjusted to replicate the status quo more closely.
Setting up a GSP tailored to the UK’s trade partners will start from rolling over the GSP currently applied by the EU. At least initially, applying the criteria which have so far determined the grant of unilateral trade preferences to developing countries when serving the UK appears a sensible move, in terms of both its viability in the short run and of continuity of regime.
a. The EU’s GSP
The GSP of the EU was established in 1971 and was founded on the idea of granting non-reciprocal and non-discriminatory preferential market access to developing countries, with the aim of increasing their export earnings, promoting their industrialization and accelerating their rates of economic growth (UNCTAD 1968). Since the 1970s, the GSP of the EU has evolved considerably, and has progressively made trade preferences more predictable, stable and limited to those countries most in need. As of today, the GSP features three sub-schemes, with increasing levels of market access in the EU: standard GSP, GSP+ and the Everything but Arms (EBA) initiative.
The standard GSP grants either zero or below Most Favoured Nation (MFN) import duties on about 66% of the tariff lines applied by the EU, to a list of fourteen beneficiary countries falling into the categories of low and lower-middle income countries, as defined by the World Bank. The GSP+ extends this preferential treatment, allowing duty-free imports of all the products covered by the standard GSP. Membership to the GSP+ is reserved to economically vulnerable countries and is conditional on the ratification of a list of 27 international conventions on sustainable development and good governance. Currently there are eight GSP+ members. Finally, the EBA initiative grants the most preferential treatment, as it allows duty-free, quote-free imports of all products shipped by 48 Least Developed Countries (LDCs), except arms and ammunitions.
The current GSP membership is the outcome of an evolution over time, determined by a series of reforms and the process of graduation from the scheme.
Graduation was introduced in 1995 (European Union, 1994), as a means of removing from the GSP those countries and sectors which are no longer considered in need of a preferential treatment. This exclusion mechanism has undergone various modifications over time and its current form works both at the country and the sector level:
This second, competitiveness based, graduation mechanism currently applies only to standard GSP members. The import share threshold was initially set at 15% (12.5% for textiles) and has been revised upwards over time because of country-graduations, or reforms of the scheme which reduced its membership base. These threshold revisions are meant to avoid that, as some countries leave the programme, other beneficiaries would mechanically graduate because of the reduction of the total out of which the shares are calculated.
c. GSP+ eligibility criteria
Standard GSP beneficiaries can apply for the more preferential GSP+ status, under a set of conditions. Other than having ratified the list of 27 international conventions, GSP+ applicants must be identified as vulnerable economies. Vulnerability is expressed in terms of a country’s size, i.e. the country’s share of total EU GSP imports being less than 6.5% (European Union, 2015b), and concentration of the export portfolio, i.e. the share of the seven largest sections in total EU GSP imports from that country being larger than 75% (European Union, 2012).
d. Rolling over the EU’s GSP to the UK
To allow continuity in market access to developing countries post-Brexit, the UK will roll over the scheme as currently applied by the EU. The working of the scheme rests on the interplay of a variety of factors which the UK would have to maintain in its own GSP:
i. List of beneficiaries across standard GSP, GSP+ and EBA; list of products eligible for preferential treatment; Preferential margins, with respect to the MFN tariffs;
ii. Rules for graduation from GSP;
iii. GSP+ eligibility criteria.
Maintaining the lists of beneficiaries and eligible products, as well as the preferential margins granted, is the first step towards ensuring that market access to the UK is unaffected by Brexit. In particular, it would ensure continuity of regime for EBA beneficiaries (Mendez-Parra, 2017).
The maintenance of the current graduation rules can, however, be problematic. The separation of the UK from the EU27 block may split the market served by developing countries, such that some of their import-shares may exceed the graduation thresholds, either in the UK or the EU27. Preference removal could therefore occur without any change in competitiveness, but could be the mechanical outcome of Brexit. These graduations, henceforth referred to as mechanical graduations, will be determined by the import-shares in the EU28 pre-Brexit and, crucially, the UK-EU27 trade distribution. The countries more at risk of graduating are those with larger import-shares pre-Brexit, i.e. close to the graduation threshold, and those whose trade is unevenly distributed between the UK and the EU27. Trade distribution is a key aspect of how rolling over the EU’s GSP will affect future UK GSP beneficiaries, as Brexit would only leave import-shares (and graduations based on this criterion) unchanged if only trade were evenly spread across the UK and the EU27.
The maintenance of the current GSP+ eligibility rules could be another issue. These criteria are applied by the EU at the moment of evaluating a GSP+ applicant (European Union, 2012). After having received GSP+ status, a country can leave the GSP+ scheme only through a country-graduation, or because of violations in the implementation of the list of international conventions. Changes in vulnerability which occur post GSP+ accession do not cause the loss of GSP+ status. When setting up its own GSP+ programme, the UK will therefore have to verify that all current GSP+ members comply with the current vulnerability thresholds in the UK. The uneven UK-EU27 trade distribution could determine a change in a country’s eligibility, similarly to mechanical graduations (for the size criterion, a country might account for more than 6.5% of UK’s GSP imports, while accounting for less than 6.5% of EU’s GSP imports). In addition, the current GSP+ members’ size and concentration might have changed since their accession to the EU’s programme, and if the UK were to adjust the entry criteria to accommodate for this, new GSP+ applications could not easily be rejected.
Two empirical exercises have been performed in this paper. First, I calculated which country-sector pairs would lose GSP preferences because of mechanical graduations either in the UK or the EU27, if the UK rolled over the EU’s GSP with its current graduation rules. Second, I computed the figures which determine the vulnerability status of a GSP+ beneficiary, i.e. concentration and size, for UK, the EU28 and the EU27. These figures allow me to assess: a) whether current GSP+ members would qualify for GSP+ status in the UK post-Brexit, were the same vulnerability criteria maintained; b) which other countries could apply for GSP+ in the UK, were the UK to adjust these criteria to accommodate the change in vulnerability of the current GSP+ members since their accession to the programme.
Annual data on EU imports at the CN-8 digit product level were extracted from COMEXT, for the period 2001-2017. These data contain information on the CN product code, the value imported, the reporting country and the country of origin of the shipment.
Information on the list of products eligible for GSP, GSP+ and EBA was constructed with data from WITS-TRAINS, which provides raw data on the 12-digit tariff lines eligible for GSP treatment in the EU. The tariff data were then aggregated at the 8-digit level.
Detailed time-varying information on GSP, GSP+ and EBA membership status, as well as graduation episodes, was retrieved from EU regulations.
i. Mechanical graduations
Since the GSP reform of 2005, the EU decides on country-sector graduations every three years. The decision is based on the calculation of import shares, which is performed in the following way. For graduations applying to the 2017-2019 period, for instance:
Following this methodology, I computed the import-shares that will determine graduations from EU’s GSP in 2020-22 (the precise details are given in equations (1) and (2) in the Online Appendix). To identify mechanical graduations, I then re-calculated these import-shares separating import flows directed to the UK and the EU27.
The European Commission released the decision about 2020-22 graduation at the same time this paper was being produced (European Union, 2019). This allowed me to verify the correctness of the calculations performed here. The graduation decision of the EU includes the shipments reported by the UK in 2015-17, hence, as far as the 2020-22 period is concerned, EU-27 import-shares will not apply.
Post-Brexit, the UK will roll over the EU’s GSP and will have to decide whether to also apply the decision about the (future) 2020-22 graduations or to compute its own import-shares. The former case appears more likely, given the intention of the UK to avoid changes to the current level of market access. However, even if the UK replicated the 2020-22 graduations decided at the EU28 level, calculating what the UK import-shares would be and which sectors would mechanically graduate under the current GSP rules sheds light on the direction in which the future UK GSP will have to change, to maintain the promise of unchanged market access post-Brexit for developing countries.
ii. Vulnerability criteria for GSP+
The two vulnerability criteria of trade concentration and size are computed as follows.
i. Mechanical graduations
Table 1 shows countries and sectors which will graduate in the EU’s GSP for the 2020-22 period, together with mechanical graduations which would occur should the calculation be performed separately for the UK and the EU27. As mentioned above, the EU27 graduations will not apply in 2020-22 and, most likely, neither will the UK graduations. Table 1 is therefore informative about the scenario which might present itself in 2023 (for the next round of graduations), due to the uneven distribution of trade between the EU27 and the UK. In order to show the magnitude of the effects, Table 1 reports the value of imports in the sections subject to graduations, the value of GSP preferences (i.e. the value imported multiplied by the difference between MFN tariffs and GSP tariffs), and the import share of the section which determined (or would determine) the graduation.
The main finding is that both the UK and the EU would have to revise the graduation thresholds upwards to avoid mechanical graduations post Brexit. In Table 1, the light rows show mechanical graduations in the UK, the dark rows show those in the EU.
India is the country which would be most affected by the application of the current GSP structure in the UK. This would be the joint outcome of the strong overall competitiveness of India among GSP beneficiaries, and the relevance of the UK as a market for India’s exporters. For the 2020-22 period, India will graduate from the GSP of the EU in eight sectors, and these graduations would all be confirmed in the UK’s GSP. This is not surprising, as India is by far the largest exporter in these sectors amongst GSP countries, regardless of the destination of the shipments. Importantly, however, India’s trade is unevenly distributed between the UK and the EU, with UK’s import shares exceeding EU’s shares in most sectors. Table 1 shows that in six sectors where India receives GSP treatment when exporting to the EU, it would graduate from UK’s GSP. To put this finding in perspective, the graduations imposed under the EU’s GSP scheme would affect approximately 30% of the UK’s GSP imports from India. The additional mechanical graduations would bring this figure up to about 54%, implying that mechanical graduations alone would account for about 24% of UK’s GSP imports from India.
Nigeria is also potentially affected by graduations in the UK, as it would lose GSP preferences in the mineral products section, which accounts for 13% of the UK’s GSP imports from Nigeria.
Mechanical graduations in the UK could involve a substantial value of trade: in 2016 only, the sectors potentially subject to mechanical graduations accounted for € 1.27 billion of UK imports, corresponding to € 31.6 million in tariff preferences. Overall, these graduations would be worth approximately three times as much, since they apply to three years intervals.
Although the EU has already decided about graduations for the 2020-22 period, it is worth noting that, regardless of what the UK will decide about its own GSP, Congo, India and the Philippines could be further affected by Brexit, as they could mechanically graduate in the block of EU-27 countries. These EU-27 graduations will not apply in the 2020-22 period, but it is nonetheless informative to inspect these import-shares, because they reveal that which exports are oriented more towards the EU27 than the UK. This might cause future graduations from the EU’s GSP, in the (smaller) post-Brexit EU-27 market.
ii. GSP+ vulnerability criteria
Table 2 presents the findings of the second exercise and reports the trade concentration and size figures by which the vulnerability is assessed in the context of EU’s GSP. Panel A focuses on current members of the EU’s GSP+, whose vulnerability has to be verified at the moment of entry into the GSP+ scheme of the UK; Panel B reports the corresponding figures for standard GSP members, which might come into play were the UK to revise the entry criteria to ensure all current EU GSP+ members are eligible for the UK scheme.
Beginning with the current GSP+ members, all countries report trade concentration shares well above the 75% threshold imposed by the EU. These shares are even higher for the UK, implying that in this dimension Brexit makes no difference.
The size of GSP+ members, as their share in total EU GSP imports, requires more attention, however. The current EU shares of Pakistan is 6.49%, which is above the 2% entry threshold which needed to apply when Pakistan made its GSP+ application, implying that Pakistan grew since it entered the scheme. Pakistan’s share is even higher for the UK, suggesting, similarly to the case of mechanical graduations, that its trade is oriented more towards the UK than the EU. Taken together, these figures suggest that the size criterion will have to be revised upwards, if all current EU GSP+ members are to qualify for UK’s GSP+ scheme.
A second aspect to consider is that the revision of the size criterion from 6.5% to, say, 8.5% (to ensure that Pakistan is eligible for GSP+ in the UK) could trigger a new set of applications from standard GSP beneficiaries which comply with the revised entry requirements. Except for India and Vietnam, standard GSP beneficiaries would all be considered small under the revised size criterion in UK’s GSP+. The concentration criterion, however, would prevent the application of Indonesia.
It is important to note that the revision of the vulnerability criterion for GSP+ applications would, in practice, make no difference. The three largest standard GSP beneficiaries, India, Vietnam and Indonesia, would still be excluded from GSP+ under an 8.5% criterion and all remaining GSP members are small enough to qualify under the 2% criterion. Regardless of Brexit, therefore, the reason why 11 out of 14 standard GSP members have not yet applied for (or received) GSP+ treatment must lie in the other condition imposed by the EU, which is the ratification of a list of 27 international conventions. 
The EU regulations specify that GSP+ eligibility arises not only from the ratification of the conventions, but also from their effective implementation, which must be maintained, and which is regularly monitored by the European Commission. Inspection of the participants to the conventions shows that most GSP members have not ratified the entire lot of treaties, except for Congo and Nigeria. These latter countries, therefore, could apply for GSP+ in both the EU and the UK, but the fact that they have not received such status so far, hints to a defect in implementation of the principles contained in the conventions, which would make them unsuitable GSP+ applicants in the post-Brexit scenario too.
To ensure that the countries currently receiving preferences in the EU’s GSP will do so also in the UK’s GSP, the UK should revise the graduation thresholds upwards. Inspection of Table 1, however, reveals that there is no single threshold which would preserve the status quo, because increasing the thresholds to avoid mechanical graduations would de-graduate other sections. Despite this difficulty, I attempt to provide some guidance on how to minimize disruptions post-Brexit.
Currently, there are three graduation thresholds: 57% for most sections, 47.5% for textiles, and 17.5% for live plants, animal and vegetable fats and oils, and mineral product. Threshold revisions, therefore, could be done separately for these three groups.
Three mechanical graduations in the UK concern sections with a 17.5% threshold. To avoid these, the UK could increase the threshold to 77%, but this would de-graduate section 3 for Indonesia, which currently graduates in the EU’s GSP. The other four mechanical graduations in the UK concern sections with a 57% threshold: to avoid these, the UK could increase the graduation threshold to 70%, but this would de-graduate two additional sections (15-b and 17-b for India).
The other four mechanical graduations in the UK concern sections with a 57% threshold: to avoid these, the UK could increase the graduation threshold to 70%, but this would de-graduate two sections (15-b and 17-b) which currently graduate in the EU’s GSP. These de-graduations might be problematic, not necessarily because of the lost tariff revenue, which the UK might be willing to incur, but because of the increased competition among GSP beneficiaries coming from India and Indonesia being able to export to the UK on GSP rather than MFN terms. For section 3, the countries most affected by the readmission of Indonesia would be the Solomon Islands, Sri Lanka, the Philippines and Kenya, with import shares of 11%, 8.3%, 5.5% and 2%,respectively; for section 15-b the countries most affected by the readmission of India would be Mozambique and Vietnam, with import shares of 20% and 11%, respectively; for section 17-b the countries with the largest import shares after India are Cambodia (10%), Indonesia (8%), the Philippines (5.5%) and Bangladesh (4.7%).
It is hard to foresee the reaction of India and Indonesia’s competitors to the decision of the UK to grant India GSP preferences in these three product sections (aimed at avoiding mechanical graduations in other sections), and it is definitely beyond the scope of this paper to speculate on the likely WTO legality of such a decision. What I can do, however, is compare the value of UK imports potentially affected by India and Indonesia’s competition due to a threshold revision (i.e. the value of UK GSP imports from competitors in section 3, 15-b and 17-b) with that affected by mechanical graduations. With 2016 data, the former amounts to approximately € 410 million, and the latter to € 1.27 billion. In terms of trade shares, imports of the affected competitors in sections 3, 15-b and 17-b account for about 3.1% of their overall GSP imports in the UK; the mechanical graduations that could be avoided by raising the import shares are instead worth 24% of India’s UK GSP imports. From a purely monetary perspective, therefore, it would be worth revising the thresholds upwards and avoiding mechanical graduations. There should certainly be, however, also other considerations factored in the adjustment the UK’s GSP post-Brexit, among which, not least, the non-discriminatory nature (and WTO compatibility) of such adjustments.
In 1995, when graduation was first introduced, the EU exploited a specialization index to graduate competitive sectors from GSP beneficiaries, but also applied a 25% import-share threshold to exclude very competitive sectors that did not meet the graduation conditions according to the specialization index (European Union, 1995). The 2005 reform left only the import-share as a graduation criterion, but lowered the threshold to 15%.
The reasons for a 15% versus a 25% threshold must be found in the balance between political economy and development considerations made by the EU at that time, which held the country-section pairs in between those shares to be competitive enough without GSP preferences. Table 3 shows that setting a 15% threshold in 2005 graduated many of the big trade partners of the EU, i.e. China, Brazil, Russia, South Africa, in a variety of industries (prominently wood and wood articles, section IX, and vehicles and other transport equipment, section XVII) thereby reducing competitive pressure on EU firms active in those sectors. An even lower import share, however, would have excluded from the GSP trade partners at a lower level of development, such as India, Pakistan and the Philippines. This trade-off between protecting EU industries from import competition and granting trade preferences to foster development and industrialization could have therefore led to the adoption of this specific 15% import-share limit.
Since 2005, however, this balance of political economy versus development goals has not been amended, and the graduation threshold has only been revised upwards, to ensure proportionality of preferential treatment in case some beneficiary left the programme (European Union, 2015a).
This analytical study shows how Brexit could cause harm to developing countries even if all the current GSP features were rolled over by the UK. The uneven trade distribution between the UK and the EU implies that the current import-share thresholds, determining graduations from the GSP, can provoke the loss of preferences in a number of sectors either in the UK, or the EU.
The balance of political economy versus development aid, established by the EU over a decade ago, could now be maintained by the UK, in the event that it separates itself from the EU. Rolling over the GSP of the EU goes quite a long way in ensuring that preferential market access to the UK is unchanged for many trade partners, especially those not subject to the threat of country-sector graduations (GSP+ and EBA countries). The vulnerability thresholds that determine eligibility to the GSP+ regime will need to be revised upwards in the UK’s scheme, to ensure that three members of the current scheme (Pakistan, the Philippines and Sri Lanka) are not excluded from the GSP+ programme of the UK. For the standard GSP beneficiaries, the UK would have to consider also the graduation mechanism. Changing the import-share threshold, even if only by a handful of percentage points, could make a big difference in terms of preferential market access and, were the UK to roll over the EU’s GSP, it would have to revise these thresholds upwards to attempt to maintain unchanged market access post-Brexit for all current GSP beneficiaries. Unfortunately, avoiding graduations by raising the thresholds would re-admit into the GSP some of the country-sector pairs that had previously been graduated, thereby distorting the competition among GSP members. This trade-off will force the UK to make a decision which, one way or another, implies that Brexit will take its toll also on some developing countries.
 The EU has also signed non-EPA FTAs with some developing countries, e.g. Peru, Central America (Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, Panama), Colombia, Vietnam.
 https://www.gov.uk/government/publications/existing-trade-agreements-if-the-uk-leaves-the-eu-without-a-deal/existing-trade-agreements-if-the-uk-leaves-the-eu-without-a-deal, page last updated on 02/04/2019.
 At the time of writing, the UK has signed trade agreements with the CARIFORUM bloc, Chile, the Eastern and Southern Africa bloc (Madagascar, Mauritius, Seychelles and Zimbabwe), Faroe Islands, Israel, Pacific states (Fiji and Papua New Guinea), the Palestinian Authority and Switzerland.
 Resolution 21(II) on “Preferential or Free Entry of Exports of Manufactures and Semi-Manufactures of Developing Countries to the Developed Countries”, UNCTAD meeting 1968.
 This is the number of beneficiaries as of the time of writing. The membership base changed considerably over time and is now at its lowest since the launch of the programme. Tables with members of the Standard GSP, GSP+ and EBA initiative are in the Online Appendix.
 The share of tariff lines eligible for the duty-free GSP+ treatment is virtually the same as that for standard GSP (66%), with the difference that about 50% of standard GSP tariffs, although lower than MFN, do not go to zero.
 See subsection 2.c for a definition of vulnerability in GSP context.
 The identification of LDCs follows the long-standing UN definition, which is based on the three main criteria of income, human assets and economic vulnerability.
 Crucially the reform of 2014, which halved the overall GSP membership from 177 to 88 countries.
 The last country to graduate for income reasons was Paraguay, in 2019 (European Union 2018). Ghana, Ivory Coast and Swaziland where excluded from GSP in 2019 because they started benefiting from preferences under alternative agreements with the EU (European Union, 2018).
 In the GSP terminology, sections are sectors in which products have been classified for the purposes of calculating GSP related statistics. Since the 2014 GSP reform, there are 32 product sections, which derive from the 21 Harmonized System (HS) sections classification.
 For EBA beneficiaries, the country-sector graduation mechanism never applied, while it was removed for GSP+ members in the 2014 reform (European Union 2012). The country-graduation based on the level of income per capita or the presence of alternative preferential agreements applies instead to all GSP members.
 The thresholds were increased to 17.5% (14.5% for textiles) in 2014 because of the exclusion from GSP of upper-middle income countries, and further to 57% (47.5% for textiles; 17.5% for live plants, animal or vegetable oils, and mineral products) in 2015 after the graduation of China, Ecuador, the Maldives and Thailand (European Union 2012, 2015a).
 Mendez-Parra addresses the issue of a UK GSP scheme, with an aim similar to mine, i.e. informing on how to design a UK GSP which ensures no country is worse off post-Brexit. His paper, however, is oriented on African countries, and takes a broader perspective, analysing which membership basis, product coverage and preference depth the UK’s GSP should have. I, on the other side, focus on the specific issues relating to rolling over the EU’s GSP on the UK.
 Without any increase in imports.
 So far this happened only once, for Sri Lanka, between 2010 and 2017.
 A further technical issue in the application of the EU’s GSP scheme in the UK concerns the rules of origin which determine the grant of GSP tariffs to developing countries’ shipments when directed to the UK. I briefly outline this issue in the Online Appendix.
 To the best of my knowledge, this methodology is not described in such detail in any European Union publication about the GSP. I have therefore derived this method myself, by attempting to reproduce the graduation decisions that the EU took in the past, i.e. in 2005, 2008, 2012 and 2015, i.e. to calculate the import-shares that determined graduations given the threshold that applied in each period. Observance of all the points in section 3b allows me to reproduce these graduation episodes.
 In one instance my calculations do not correspond to the graduation decisions taken by the EU. See Table 1 and section 3 in the Online Appendix for more details.
 The details are provided in the Online Appendix, in equations (3) to (6)
 The tariff margin calculation is performed at the 8-digit product level; and then aggregated up to the product-section level. On 13th March 2019 the UK Government announced a provisional MFN tariff schedule that would apply in case Brexit happened without a deal. The examination of this tariff schedule by Gasiorek and Magntorn Garrett (2019) UKTPO Briefing Paper 29, reveals that the UK will set a large number of tariffs to zero (95% of the tariff lines, compared to the 26% of the EU MFN schedule). If the provisional UK MFN schedule applied, therefore, developing countries in receipt of GSP preferences would be affected by “preferences erosion”, i.e. the reduction of the value of preferential market access and, implicitly, an increase in competition from countries which do not benefit from preference. However, since the likelihood of a no-deal Brexit appears very small, in this paper I decided to use the current EU MFN schedule for the calculation tariff margin.
 These shares are computed for 2016, out of imports of GSP eligible products. If these shares were computed out of total UK imports, under both GSP and MFN tariffs, EU’s graduations would affect 22% of India’s imports, and mechanical graduations a further 17%.
 In the interest of brevity, only the countries whose shares are problematic are shown. Tables with figures for all GSP+ and standard GSP members are provided in the Online Appendix
 The size criterion was revised upwards from 2% to 6.5% in 2015, whereas Pakistan joined the GSP+ in 2014.
 This monitoring activity would have to be undertaken by the UK too, to evaluate both new GSP+ applications and the effective implementation of the conventions by current GSP+ members. This would be an additional burden on the UK, arising from its decision to leave the EU
 Similar adjustments might have to be undertaken also by the EU in its own GSP scheme, but, given the focus of this paper on a possible UK GSP and the fact that these adjustments would be substantially easier in the EU scheme, this section analyses possible adjustments to the UK GSP only.
 The de-graduation of sections 15-b and 17-b would imply a loss of tariff revenue for the UK of approximately € 21.5 million per year.
 Also India, with an import share of 19.9%, would be affected by Indonesia’s competition in section 3. However, if Indonesia were not re-admitted into the GSP because of a threshold change, India would mechanically graduate and lose its GSP preferences in section 3. For this reason, I did not include India among the countries which would be negatively affected by the de-graduation of Indonesia in this section.
 Indonesia and Pakistan have, respectively, a 2.4% and 1.7% import share in section 15-b and would be less affected by India’s competition. Despite the large import share, Vietnam might also be relatively less affected than Mozambique, as it signed an FTA with the EU in 2018 (and should therefore leave the GSP within two years of the entry into force of the FTA).
 Other, smaller, competitors in section 17-b are Angola (2.5%), Vietnam (2.2%), Sri Lanka (1.4%) and Nigeria (1.3%).
 It should be noted that there is considerable heterogeneity behind this aggregate figure, with Mozambique’s share being close to 80% and that of all other countries being around or less than 10%.
 The 2016 share of UK GSP imports from Nigeria affected by the mechanical graduation in section 5 is barely noticeable. This is because UK imports from Nigeria in section 5 for 2016 seem to be exceptionally low, relative to other years.
 The change in the threshold must have also reflected the reduced graduation threat due to the removal of the specialization index.
European Union, 1994, “Council Regulation (EC) No 3281/94”, Official Journal of the European Union, L 348/1, 19.12.1994
European Union, 2005, “Council Regulation (EC) No 980/2005”, Official Journal of the European Union, L 169/1, 27.06.2005
European Union, 2012, “Council Regulation (EU) No 978/2012”, Official Journal of the European Union, L 303/1, 25.10.2012.
European Union, 2015a, “Council Regulation (EU) No 2015/1978”, Official Journal of the European Union, L 289/1, 05.11.2015.
European Union, 2015b, “Council Regulation (EU) No 2015/602”, Official Journal of the European Union, L 100/8, 17.4.2015.
European Union, 2016, “Council Regulation (EU) No 2016/330”, Official Journal of the European Union, L 289/1, 08.03.2016.
European Union, 2018, “Council Regulation (EU) No 2018/148”, Official Journal of the European Union, L 26/8, 31.01.2018.
European Union, 2019, “Council Regulation (EU) No 2019/249”, Official Journal of the European Union, L 42/6, 13.02.2019.
Gasiorek M. and P. Holmes, 2017. “Grandfathering: What Appears Bilateral is Trilateral”, UKTPO Briefing Paper No. 13.
Gasiorek M. and J. Magntorn Garrett, 2019. “Briefing Paper 29 – Deal or ‘No Deal’? The economic consequences of the UK’s ‘No Deal’ tariffs”, UKTPO Briefing Paper No. 29.
Mendez-Parra, M., 2017. “Designing a new UK preferences regime post-Brexit.” Working Paper 521. London: ODI
UNCTAD, 2015. GSP – Handbook on the Scheme of the European Union.
Click here for the Online Appendix.