The Impact of International Agreements on Domestic Policy: An Analysis of Tariff Policy in African Countries

Article excerpt


This paper estimates a fixed effects tariff model to study the impact of the tariff reform provisions of international agreements on domestic tariffs, using a sample of eight Sub-Saharan African countries. The structure of the model explaining domestic tariff changed from the preagreement period to the postagreement period. However, the results indicate that for the most part, efforts by governments to adhere to tariff agreements failed in all but a few countries. Even for the countries in which the agreements appeared to be successful, the significance of the results is relatively weak. (JEL O0, O1)


International agreements can be used as precommitment mechanisms to enhance the credibility of domestic reform efforts [Boko, 1996]. However, for agreements to be effective, governments must adhere more closely to announced policies, and the private sector must view these announcements as more credible, thus modifying its behavior. Many African nations had precommitted to tariff reduction as part of the structural adjustment programs (SAPs) undertaken since the mid-1980s. However, did these precommitment agreements have a measurable impact on the domestic tariff policies of countries?

The credibility of a government-announced policy (although not explored here) is at the heart of the issue, and it has been dealt with at length in theoretical models. For instance, Kydland and Prescott [1977] discuss the importance of credibility as it applies to inflation-reduction policies. Their main finding was that in the absence of a binding, precommitment mechanism, monetary authorities have an incentive to renege on their announced policy. Thus, once the private sector has chosen a strategy consistent with an announced inflation policy, the monetary authorities can make a surprise increase in monetary policy to (attempt to) exploit the Phillips-curve relationship. Since rational private agents can see through the authorities' smokescreen, the announcement of a zero inflation policy is time inconsistent and, therefore, not credible. Further, Lapan [1988] discusses the time inconsistency of the optimal tariff in a model in which production decisions are made prior to consumption or trade decisions. He shows that the inability of the policy-active government to commit to a tariff before production decisions are made leads to an outcome that is Pareto inferior to the precommitment solution. The author also shows that if precommitment is not feasible, second-best policy instruments (production subsidies) are beneficial. However, what is really needed is an irrevocable precommitment mechanism.

Unfortunately, no direct empirical test is possible to address the question posed earlier. In reality, there is a vector of tariffs, not a single tariff, and international agreements will, at most, impose targets on average tariff rates (not on each component thereof). In addition, most countries use nontariff barriers and other invisible barriers that will implicitly be part of an agreement to liberalize trade but cannot be directly observed. Also, realized average tariff rates and trade flows will depend upon the realization of unpredictable variables (world prices, gross national product, domestic random variables, and the like) that cannot be part of the agreement. Finally, contrary to what a theoretical model might yield, satisfaction of the agreement is not predicated upon the (average) tariff rate exactly achieving some number. There is undoubtedly a range of possible outcomes that would allow the international agencies to conclude that the country is complying with its liberalization program.

For these reasons, and others, it is impossible to test directly whether African governments have completely adhered to the agreements into which they entered and whether these agreements have enhanced the credibility of government-announced policy (thereby affecting private behavior). …