If tariffs are reduced without complementary fiscal policies, trade liberalization may falter or fail
Many developing countries are trying to liberalize their trade regimes. But, despite broad agreement on the advantages of a more open economy, some of the reforms have been reversed or are encountering significant problems. These problems have been particularly evident in cases where the liberalization has been attempted mainly through a reduction in tariffs, without anticipating all the consequences of tariff cuts or coordinating them with other macroeconomic goals. If the budget worsens unexpectedly following a reduction in trade taxes, the government may be unable to wait for the longer-term gains from free trade to replace the lost revenue, and may feel forced to reimpose revenue-yielding restrictions. It may thereby unintentionally send a message to the private sector that a liberal trade regime does not “work.”
Such reversals, which are economically and politically costly, could be avoided by planning a fiscal policy consistent with the liberalization drive. As a first step in that direction, this article discusses the likely and complex impact of tariff cuts on the budget and offers some general policy guidelines to help the budget capture the long-term gains from freer trade. Besides the direct effects of the tariff reduction on the budget, the changes in the economy following the move toward free trade will also have effects on revenue and expenditure. Economic changes in response to liberalization—including complementary policies in other sectors, of which the most important tends to be exchange rate action—tend to evolve at different speeds. Hence, the short-run budget may differ significantly from the budget that eventually emerges when the liberalization has worked itself through the economy.
Changes in the economy
In the long run, the liberalized economy, with low or no barriers to the entry of imports, will look very different from the high-tariff economy: national output and income should be higher, as should the share of traded goods and the share of the country’s most abundant factor in national income. (In developing countries, typically, labor has gained at the expense of capital. Shifts toward outward-looking policies have tended to increase employment, since most countries that have reduced their tariffs significantly in recent years were labor-abundant and were not at full employment at the time they liberalized.)
Well-accepted arguments in favor of liberalization suggest that, eventually, the increase in national income will justify any adjustment costs, as the sectors whose income is expanding will be able to offset the effects of contraction in other sectors while still producing a net gain for the economy. The gain in national income is likely to take some time, however, and its speed will depend largely on two things: the flexibility with which workers and capital respond to the changed profit incentives in the export and import sectors; and the policies supporting the liberalization—notably the country’s exchange rate policy.
Devaluation is usually a particularly crucial accompaniment to a large-scale trade liberalization, in order to prevent the balance of payments from deteriorating and to offset employment losses in the formerly protected import-substitution industries by expanding the export sector. Without devaluation—with a fixed short-run nominal exchange rate—the reduction in tariffs would shift domestic demand from import-competing industries toward imports. The shift in demand from the previously protected industries would lead to production cuts and unemployment there, and to an overall contraction in activity, since labor would probably not be immediately absorbed elsewhere.
The rise in imports would erode the trade balance, making a real devaluation eventually inevitable. In an economy where the exchange rate is allowed to depreciate immediately—and, hence, where import growth is slower and the efficiency gains from liberalization emerge faster and with less transitional unemployment—the budget will diverge from the budget in an economy where the exchange rate is held fixed for as long as possible. Both of these cases are empirically important and are considered in what follows.
Effects on the budget
Direct impact. Trade tax revenue will, of course, bear the most direct impact of liberalization. However, the extent of the initial impact cannot be predicted in any simple way. Contrary to common perception, the reduction in tariff rates will not necessarily result in an immediate revenue loss (any more than an increase in the tariffs will necessarily raise revenue), but will depend on the way in which different parts of the economy respond to the rate change.
This article is based on a study, with Mario Blejer, to be published in a forthcoming volume, edited for the IMF by Vito Tanzi.
First, revenue will be affected by the price elasticity of the demand for imports. If tariffs are reduced but not fully eliminated, and the exchange rate remains fixed, the extra revenue from an increasing volume of imports is likely to offset some of the revenue loss from the cut in tariff rates. The impact of the direct price elasticity on revenue could be magnified if lower tariffs reduce incentives to evade trade taxes. However, while an increased volume of reported imports might generate a net revenue gain in the medium-term, the short-run price elasticity is likely to be very low, leading to a temporary fall in revenue which is worse than the expected long-run revenue loss.
Second, two income-related effects on imports can be expected. For a given level of activity, the reduction in tariffs will make people better off by raising their after-tax disposable income and consumption—including, presumably, their consumption of imports. Moreover, the more flexible the supply of import-substituting industries in the short run, the more these will reduce their production (rather than taking a price cut) and the more the market for imports will increase in response to a given tariff reduction. Any rise in import consumption, and any increase in import volume, will bolster trade tax revenues, though these gains may be small. On the other hand, if the contraction in domestic production is very large, aggregate demand is more likely to fall and, with it, consumption of tax-yielding imports. The net income effect on trade tax revenue therefore depends on countries’ particular circumstances.
The effects described above would be diluted by devaluation, which might enlarge short-run tariff revenue losses by containing import volumes at close to pre-liberalization levels or below (though there would be an offsetting valuation effect that would tend to raise the domestic currency value of the tax revenue somewhat). In some countries, tariff revenue losses have been partly offset by export taxes because traditional exports have risen after devaluation. Typically, though, imports have been responsible for a greater share of trade tax revenues. Whatever its net effect on short-run revenues, exchange rate adjustment is crucial to the longer-run success of a liberalization program.
The third important influence on trade tax revenue is the elasticity of substitution among imports. Even if total demand for imports rises following a liberalization, tariff revenues may still fall. Changes in imports’ relative price structure could induce consumers to switch to cheaper imports, particularly if the goods whose tariffs were being reduced had very close substitutes still subject to higher tariffs.
Fourth, the change in trade tax revenue will also be affected by the market structure of import trade. The import sector tends to be competitive, but in some liberalizing countries with only a few importers, these importers did not pass on the whole of the tariff reduction to domestic consumers. Instead, they collected windfall profits through their control of the market, by holding prices and import volumes constant and cashing in on the tariff rate reductions. The base remained little changed though the rate was lower, and so tariff revenue tended to fall.
Finally, in some countries, announcement of the impending liberalization gave traders an incentive to delay imports until tariffs were reduced, with the result that imports (and tariff revenue) surged dramatically in the wake of the liberalization. Unfortunately, this surge still represented a net revenue loss when considered over the whole fiscal period.
The economy’s response. The possible initial contraction of economic activity, its subsequent expansion, and the redistributions of income and consumption brought about by the effects of the liberalization on relative prices, affect important elements of the budget besides trade taxes.
The first of these is income tax revenue. In the long run, if national income rises, direct tax revenue should increase proportionally. However, the likely short-run contraction in activity and rise in unemployment could cause income tax proceeds to fall temporarily—with some offsetting revenue gain if importers become better off. A subsidiary point here is that if importers are less able than domestic producers to evade income taxes (since, particularly in developing countries, fuller records are kept of the sources of importers’ income) then a redistribution of income toward importers could reduce tax evasion as well.
The exchange rate devaluation could modify these effects on the income tax, given its tendency to redistribute gains in income toward exporters and away from importers. If importers were rich and concentrated, and exporters were poor and dispersed and outside the income tax base (as are many peasant farmers), exchange rate adjustments could make the income tax base smaller than under liberalization without devaluation.
The relative price changes that follow liberalization will also probably affect domestic consumption tax revenue, which will be depressed anyway immediately after the liberalization if consumption is responsive to a drop in income. On the other hand, value added tax (VAT) and excises on imports should rise if imports rise. Many countries, however, have not imposed domestic consumption taxes on imports at as high a rate as on domestic goods, so that the shift from domestic to imported goods that accompanies the fall in tariffs leads to a drop in indirect tax revenues. Moreover, people may shift toward liberalized imports from imports whose tariffs remain high—typically luxury goods that also bear the highest excise tax or graduated VAT rates; and so a move away from these will lower excise/VAT revenue.
The larger the share of consumption taxes in total taxes, the higher total tax revenue after a tariff reduction will generally tend to be in countries where importers are competitive, since all previous tax revenue from tariffs will now be kept by consumers. But, where importers are oligopolistic and do not pass on the tariff reduction to consumers, revenue will be higher the more taxes are skewed toward income and profits taxes, since some (or all) of the previous revenue from tariffs will be incorporated instead in importers’ profits.
Besides its impact on revenue, liberalization is also likely to affect government expenditure, particularly outlays on goods and services. Since governments typically do not pay tariffs on their imports, a tariff cut will not directly reduce government’s bill for goods and services. However, there are indirect budgetary consequences related to exchange rate policy. If the nominal exchange rate remains fixed, cheaper inputs to private sector production should feed through into reduced costs of public sector domestic purchases of goods and services. Then, as soon as the exchange rate begins to depreciate, the cost of government imports will rise, in both the current and capital budgets. Finally, even if the liberalization will simplify the customs administration in the long run, it will entail short-run administrative costs—for training personnel and rewriting the tariff, revising tax returns and related forms and software, and undertaking the necessary information campaign.
Transitional costs. The transitional period during which the economy adjusts, perhaps painfully, to its new relative prices may itself impose special costs on the government budget. Some of these are built directly into the budget—such as open-ended commitments to unemployment benefits and income-maintenance schemes, whose fiscal burden rises as the economy contracts. Other costs may be incurred to speed up the transition and reduce the private sector’s adjustment costs—including government-sponsored redundancy schemes, retraining schemes, incentives for investment in the export sector, and even temporary subsidies to producers of importables who are now being forced to produce at the world price.
Government-financed adjustment schemes are a special case of the generalized pressure that may emerge on the government to introduce an expansionary fiscal policy, if transitional unemployment becomes widespread. The government may feel bound, say, to expand its demand, or set up employment-generating projects. Unfortunately, this kind of reaction would not only worsen the public finances but could well slow the speed of adjustment, as it would reduce incentives for workers and capital to shift to the export sector.
The policy challenge
In the long run, given the improvement in national income that follows trade liberalization, there is no reason why the budget balance should be worse than when the economy was protected. There is, however, a policy challenge: to create the conditions that ensure that the general rise in income is indeed reflected in the budget. Several possible shortcomings of the tax structure could prevent the economy’s revenue potential from being realized, and these may need to be addressed by adjustments in policy.
- Not all income may be taxed. In many countries, governments have been slow to recognize emerging tax bases, such as income in the new export industries stimulated by the liberalization, and slow to remove exemptions such as tax holidays to exporters designed to offset disincentives created by the previous tariff regime. For the budget to capture the income gains associated with liberalization, it is important for the tax system to be broadly based. A redefinition—and widening—of the income tax base could be a strategically important complement to trade liberalization.
- Income from different sources may be taxed differently. For instance, owners of capital might be taxed more heavily than wage earners. If so, the redistribution of income that follows trade liberalization will affect income tax yield, perhaps unfavorably. In general, the more neutral is the tax system across income sources, the smaller this effect will be.
- Even if income from different sources is taxed equally, progressivity or regressivity in the tax structure could mean that the budget gains a less-than-proportionate share of the increase in national income induced by liberalization. This would occur, for instance, if the tax structure was progressive and liberalization made the distribution of national income more egalitarian—as might well be the case in a labor-abundant economy where capital was concentrated in the hands of relatively few. Thus, again, in assessing the revenue impact of a liberalization, policymakers may need to be aware of intersectoral differences in income distribution, perhaps adjusting the progressivity of the tax structure to take these differences into account.
- National expenditure may be expected to rise with national income, so that revenue from consumption taxes should also be higher than before the liberalization. Moreover, any movement toward a more equal income distribution will tend to increase total consumption further (because the average propensity to consume will rise). However, the increases in consumption tax yields will be realized only if the consumption basket of the relatively poor is taxed at the same rate as that of the relatively rich. In many developing countries, a large share of indirect tax revenue comes from a small number of luxuries that are easily taxable, such as cars. In such a case, a redistribution associated with trade liberalization could have unforeseen depressing effects on revenue from indirect taxes unless policymakers take care to widen the base of their consumption taxes.
- Where, as is often the case, the increase in national income is accompanied by an increase in employment, policymakers need to ensure that payroll tax receipts and social security payments rise proportionally, as well as would revenues from other taxes linked to wages (such as health or education levies). On the expenditure side, total spending can be cut by the amount that unemployment benefits fall.
Two revenue-related costs of liberalization, which tend to reduce the budget balance, appear to be unavoidable, even under the most prudent of policy regimes. First, although the share of traded goods in national income is higher after the liberalization, in practice, revenue from trade taxes normally falls. This is because exports are hardly ever taxed as heavily as imports, since the disincentive effect of an export tax is more direct and easily perceived. Hence, trade liberalization usually implies a shift in production from traded goods in the tax base to traded goods “outside” the tax base.
Second, exports and imports are relatively easy to tax, but to collect revenue successfully from income and domestic consumption taxes requires a relatively sophisticated accounting system. Hence, some of the increase in national income may be expected to go to a new generation of tax collectors and auditors—reducing the gain to the government.
In sum, even if liberalization seems certain to yield long-run gains, the decision to liberalize may not be straightforward since, besides the political difficulties of removing protection, the economy may face a temporary drop in income and a deterioration in the budget balance (as well as in the trade balance) in the short term.
This does not imply that liberalization should be postponed until the economy is in balance. Indeed, it could be difficult to find a “balance” as long as the economy retains its trade distortions. Rather, the short-run costs of liberalization should be identified from the catalogue of effects listed here, so that they may be budgeted for, explicitly. If the likely size and duration of adjustment costs can be estimated—and, for example, a period can be agreed on during which the budget balance should return to “normal”—the task of liberalization will be much easier.
With regard to tax policy, the greater significance of the domestic tax base in a low-tariff economy will make it more crucial that the tax system be broad and neutral. Given that the aim of liberalization is to open the economy, it would be inadvisable to try to offset declines in import tariff revenue by taxing the growing export sector. Rather, policymakers might devote their energies to streamlining the operations of the domestic tax system, to control its overall administrative costs.
On the expenditure side, the government might examine its employment policy. If employment is expanding in the economy at large, a government could use the opportunity to shed its excess labor. Finally, the structure of contributions and expenditures on income-maintenance programs could be aligned to minimize the impact on the budget balance of the government’s support to the sectors that lose during a liberalization.