Initial TEC

A starting point of the analysis is to look at how tariff lines and imports are distributed by band. We focus here on EAC. It has a simple structure with 3 bands:
Initial Distribution by band

Initial Distribution by band

Initial Distribution by band

Initial Distribution by band

Initial Distribution by band for each sector

Initial Distribution by band for each sector

Analysis of scenarios

We consider 0 EAC sub-regions () and EAC as a block. This will lead to 0+1 optimization i.e. when the optimization is done at the regional level (maximization of regional welfare), or when “one” country will maximize its own welfare. These simulations are tagged as Scope=Delegation since the choice of the region, or of country c, is applied to the whole region (full implementation of the CET) .

We also consider the case where the choice of country c is only applied to country c. This case is tagged Scope=Unilateral. It basically defines the optimal tariff - assuming a band structure - that individual countries will select for themselves. We can use this to define a “distance” metrics (e.g. like a sum of weighted squared difference between the regional CET and the unilateral ones). But importantly, there is a significant difference, for the same country, between the Delegation tariffs and the Unilateral tariffs: one country c decides the regional tariffs, in the Delegation case, it can include the gains from the regional adoption, and therefore an additional terms of trade effects should take place. Therefore, for a country r, the tariffs associated with the Unilateral should be lower than the Delegation ones. Note: Assumption to be checked in the results

We have simulated 10 frameworks. A framework is defined by an option for public closure account and for market factor integration. By default we will focus on PUBC=2, VAT adjustment and MKIT=0, no market integration.

This leads to a total of 110 simulations, of which 0 have some errors in the last run.

** Warning: some results for Tanzania, Uganda and EAC_Others (the remainin countries) appear to be identical: same optimal values for tariff bands are obtained.. Need investigation **

Optimal tariffs value for each band

Based on this approach, the model provides us the optimal tariffs for each band. The following figure shows these results. We display at the same time the delegation [triangle], and the unilateral tariffs [dot]. We confirm the ranking between delegation and optimal tariffs for most cases.

Still, we should focus the analysis on the delegation case. The unilateral should be mainly seen as "benchmark, for more advanced analysis.

For the regional optimization, we only show “delegation” since the unilateral concept is useless here.

** On some graphs, tables, gdx, we have to be careful that some missing “values” are actually 0 due to the GAMS management of 0. So a “missing” dot on a graph should be interpreted as a 0 value.**

Optimal tariffs by band and by decider

Optimal tariffs by band and by decider

So, if we focus on the optimal regional tariffs, we see that from the initial 3 bands structure: 0%,10%,25%, we move to 6.27, 16.41, 16. So, we see a strong movement towards a convergence of the value of each band. Reducing tariff heterogeneity seems to be an important feature of the welfare maximization.

We also have a reversion between bands (see for some countries the hierarchy between band2 and band3). ** One explanation may be directly linked with the relative weights. To reduce an average, it mays make sense to reduce more the tariffs that have the largest weight, even if initially it is not the highest tariff in the aggregate**

We will review the impact of the tariff bands on the average protection in a next section, but we also notice that larger countries, like Kenya have a preference for higher tariffs, while smaller economies, e.g. Rwanda, tends to move tariffs to 0. This is also consistent with economic theory where optimal tariffs should decline with the size of the country.

** Analysis with alternative public closure and market integration should be done, but results are already included in the .GDX. **

Impacts on average tariffs

The following figure shows the impact of the tariff bands on average protection (trade weighted, base year trade values, excluding intra-regional trade flows). The regional average is decreasing from 13.89% to 12.63%. ** on this graph we see the issue indicated for TZA, UGA and OTH -> same dots.** Only for
Average tariffs pre and post optimization

Average tariffs pre and post optimization

Welfare impacts

Among the various variables we can track, and included in the saved GDX, we can look at the Equivalent Variation, expressed as a percentage of initial household consumption.

The results appear to be consistent with what we expect: * highest welfare for the region, EAC, is achieved when we maximize the welfare for teh block (decider=EAC). IT also leads to a welfare increase showing that the existing tariffs is not optimal; * Interestingly, the regional optimal (red dot) is also linked to higher welfare for all members; * Delegation at the country level, shows that in each case the “deciding” country can do better than the regional optimization, or any other partners. Kenya and Rwanda had the most extreme consequences on each other (normal since Kenya wants high tariffs while Rwanda wants low tariffs).
Welfare impacts

Welfare impacts