What is fair SACU customs compensation to the BLNS?

Yash Ramkolowan

Discussions on the Southern African Customs Union (SACU) inevitably fall back to the revenue sharing formula used to distribute excise and customs revenues across the five SACU member states. The largest share of revenues distributed to Botswana, Lesotho, Namibia, Swaziland (BLNS) is from the customs component of the pool and the rationale for this distribution is particularly contested. The message from the BLNS, however, is almost always consistent: the share of revenue from the common customs SACU pool is justifiable and “fair” given the de facto control that South Africa exerts over the region’s tariff policy.[1]

The assumption of fairness in the way that customs revenues are distributed between SACU member states has been a key sticking point in SACU negotiations, and a major reason for the current deadlock. Using 2014 baseline data, one can estimate the likely import duties that each of the BLNS would collect from the existing CET. This can be compared to the customs revenues that would be distributed from the SACU customs pool.[2] 

As shown in
 Figure 1, SACU’s total customs revenue would amount to roughly R44 billion based on total extra-SACU imports and the applied imported weighted tariff for 2014. If SACU’s customs revenues were split between member states according to each country’s extra-SACU imports (and the tariff applied on these imports), South Africa would retain 96% of customs duties collected.

Figure 1 BLNS customs duties based on the SACU formula versus implied duties based on BLNS imports
Source: Own calculations based on data from ITC Trademap and WITS TRAINS database

However, based on the SACU customs sharing formula, the BLNS would receive more than 83% of import duties collected. Put differently, in 2014 the BLNS would have to raise their individual aggregate tariff on non-SACU imports by between 7 and 34 times in order to achieve the same level of duty collection.

Botswana, for example would need to raise its aggregate import-weighted tariff from just over 2% to more than 75%, ceteris paribus. Of course, we know (and have previously illustrated) that some South African exports to the region consist of re-exported goods, originating from outside of SACU. This would underestimate the import duties due to the BLNS. Nevertheless, even if we assume that all BLNS imports were from outside of SACU[3], the conclusion is clear: the BLNS currently receive significantly more customs duties from the SACU customs revenue pool than the revenue “due” to them based on their extra-SACU imports.

But is this fair? The economic arguments for additional compensation to the BLNS stem primarily from the trade diversion and price raising effects that an externally imposed trade policy can have. In the case of trade diversion, the core argument is that the BLNS import goods from high-cost South African producers rather than lower-cost global producers because of the common external tariff (CET) duties applied on imports outside of SACU.[4] Similarly, the BLNS are forced to import higher-priced goods overall. This is ultimately detrimental to consumers in these countries and has additional negative welfare impacts in terms of inefficient production and resource allocation.

Dealing first with the issue of trade diversion, economic theory is clear that free trade agreements (and customs unions) can have significant trade diversion effects on member states. This may be especially true in the case of SACU, which is dominated by a single large economy. The key question is therefore not whether SACU member states should be entitled to compensation for trade diversion effects, but the quantum of such compensation.

Leith (1992)[5] provides an eloquent way of identifying this quantum. Under this approach the issue of trade diversion is addressed by “taxing” the source of production within the customs union. Within SACU, member states experiencing intra-SACU deficits would therefore be compensated by member states with intra-SACU trade surpluses. This compensation would be equivalent to the implied customs revenue based on the CET. Each member state would be compensated based on the most-favoured-nation (MFN) duty that they would expect to receive as if these net imports originated outside of SACU.[6]

A more complex (and difficult to estimate) area of compensation is the possible price raising effect of the CET. In addition to the direct negative impact that higher prices have on consumer welfare, higher prices may skew the allocation of resources across the economy. Resources may be directed to sectors that are not necessarily efficient without import protection. This area of compensation assumes that the BLNS would forego customs duty revenue and allow tariffs to fall to zero if they were outside of the customs union. Identifying the precise economy-wide impact of the CET on the BLNS would usually require the use of a complex economic model. For our own simplified analysis, compensation of twice the tariff duty would seem to present more than “fair” compensation for any potential price raising effects.[7]

A summary of the three compensation components is provided in Figure 2 First, the BLNS would receive their own share of customs duties based on their extra-SACU imports. This is equivalent to the government revenue that each country would earn if it applied the CET on extra-SACU imports. Second, the BLNS are all net importers within SACU and would therefore be compensated (by South Africa, as the only net exporter within SACU) for trade diversion effects through the Leith (1992) formula. This makes up the largest compensation component for the BLNS. Finally, the BLNS would be compensated for the price raising effects that the externally imposed CET has on consumers and the wider economy.

Figure 2 Fair compensation to the BLNS for trade diversion and price raising effects.

Source: Own calculations based on data from ITC Trademap and WITS TRAINS database

Comparing this total estimate of “fair” compensation to the customs revenues that the BLNS would receive from the SACU customs pool (Figure 3) is revealing. For the BLNS, compensation under the current SACU customs revenue sharing arrangements is substantially higher than what would be considered “fair”. For Botswana, SACU customs distribution may be close to four times more than what might be considered fair; for Swaziland, this may be more than six times the estimate of fair compensation. For Lesotho and Namibia the order of additional compensation (beyond what may be considered fair) is less dramatic, but still significant.

Figure 3 Comparing “fair” compensation to distribution of the SACU customs pool under current arrangements
Source: Own calculations based on data from ITC Trademap and WITS TRAINS database.

This basic analysis demonstrates that under the current customs revenue sharing formula the BLNS are more than compensated for ceding their tariff policy to South Africa. In fact, the magnitude of current SACU customs transfers probably goes far beyond what may be considered “fair” compensation. The analysis does not consider the possible benefits of the SACU customs union for the BLNS (including, for example, trade creation impacts) and the true extent of “fair” compensation may be lower than demonstrated above.

So, what does this all mean for SACU? Considering the current deadlock in SACU negotiations (particularly around the revenue sharing arrangement and the creation of a regional tariff board), it does need to be acknowledged that the BLNS have been more than fairly compensated for the externalisation of their tariff policy. Acceding to this point may finally ensure that the BLNS are able to play a greater role in determining SACU’s trade and industrial policy. The analysis also suggests that the transparency and clarity of revenue distribution for “compensation” and “development” purposes can be improved. This could finally move SACU beyond a short-term and narrow focus on fiscal compensation, and toward a regional agenda that is more substantively concerned with the common and long-term development needs of all SACU member states.

[1] The focus here is on the distribution of the SACU customs pool only. 15% of total SACU excise collections (the “development” component of the revenue sharing formula) is also distributed between SACU member states for development and economic convergence purposes. The formula for distribution aims to ensure that comparatively less developed SACU member states receive a proportionately higher share of the development component.

[2] 2014 trade and tariff data is used to provide aggregate revenue estimates. The overall estimate of customs revenue would differ from actual revenues due to duty rebates, customs leakage etc.

[3] The counterfactual to this is that under the current SACU formula the BLNS would actually receive less customs revenue if a higher proportion of their imports was found to originate from outside of SACU.

[4] The elimination of tariff barriers between SACU member states will also have positive trade creation impacts, which would reduce the level of compensation required to pay to the BLNS. However, in the absence of the SACU customs union, SACU member states would still trade under the SADC free trade agreement (FTA) where import duties between SADC member states should be close to zero. Moving from one zero-tariff trade regime to another may still have trade creation implications if, for example, there are stringent rules of origin in one regional agreement but not in the other. For illustrative purposes the trade creation implications of the SACU customs union are ignored.

[5] Leith, J.C. 1992. The Static Welfare Economics of a Small Developing Country’s Membership in a Customs Union: Botswana in the Southern African Customs Union. World Development, Vol. 20, No. 7, pp. 1021–8.


[6] The issue of BLNS imports originating from outside of SACU, rather than South Africa, has implications for “fair” compensation. As the share of extra-SACU imports in BLNS total imports rises, compensation to the BLNS for trade diversion impacts would fall. At the extreme end, assuming all BLNS imports originated outside of SACU, the BLNS would, in fact, have a net trade surplus with South Africa. This would imply that they would need to compensate South Africa for trade diversion impacts.

[7] Under partial equilibrium analysis, point estimates of the “dead weight loss” would be equal to government import duty revenue where the aggregate supply and demand curves are perfectly (unit) elastic. Compensation of double the collection of import duties would seem to be more than sufficient for the imposition of a tariff from this perspective. To reiterate, the level of such compensation is directly related to the extent to which the BLNS would eliminate tariff barriers in the absence of the CET.