The debate of academic and policy economists on VAT and tariff system

The literature on agency problems of representative democracies is extensive on the possible outcome of government behaviour trying to increase public expenditure when the political cost to do it is low. This is evident in the studies of Henderson, J. M. (1968); Niskanen, W. (1971); Oates, W. (1972); Barro, R. (1973); Nordhaus, W. (1975); Romer, T. and Rosenthal, H. (1980); King, D. (1984); Abrams, B. and Dougan, W. (1986); Schwab, R. M. and Zampelli, E. M. (1987); Dougan, (1988); Rogoff, K. and Sibert, A.(1988); Persson, T. and Svensson, L. (1989); Rosenthal, H. (1990); Alesina, A. and Tabellini, G. (1990); Tabellini, G. and Alesina, A. (1990); Alesina, A. and Drazen, A. (1991); Case, A. C., Hines, J. R. and Rosen, H. S. (1993). R. Bird has made his critique on the findings, citing new literature written by Dusek, J. in 2003 and also Slemrod, J. in 2005 that conclude that more expending leads to higher taxes and not the reverse.
Barnes, J. (2001) while commenting the "ideological underpinnings" of fundamental tax reform proposals, makes an interesting review of "leftwing" and "rightwing" arguments in favour of or against to consumption-based taxation versus income taxation. Representatives from the rightwing position seem to rely too much on arguments such as "intrusiveness" and high rates of the Income Tax "that threaten individual freedom" (as Barnes quoted from Hall and Rabushka, 1985, on this later assertion). Surprisingly, rightwing supporters seem to ignore how Socialist and Unitarian is a consumption-based tax like the VAT, where all population pay taxes without knowing whom to, where, when and why they are compelled to do so, due to the unknown tax incidence. On the other wind, leftwing supporters don't realise how difficult Income Tax could be to enlarge revenue collection for distributive policy objectives, particularly in a worldwide competitive scenario.
The IMF experts (2000) have estimated that 45 per cent of VAT revenue forgone by equity considerations benefited the richest 30 per cent population and only 15 per cent benefited the poorest 30 per cent. In Argentina, FIEL (1998) using Feldstein (1972) "distributive characteristic of goods", has estimated that 58 per cent of VAT expenditure due to some reduced tax rates (on meals and drugs) has benefited the upper 40 per cent richest families. In Italy, Liberati, P. (2001) has estimated the distributional and welfare effect of two legislative changes in VAT in 1995 and 1997, that fulfilling the EU directive, reduced to three the number of tax rates. Using two approaches - the Feldstein distributive characteristic of goods and the marginal dominance method (Mayshar and Yitzhaki (1995, 1996) demonstrated that a simpler "two tax rate" structure would have improved welfare even more. That is, a more proportional or uniform VAT is better for enhancing income redistribution objectives.
In recent years, many developing countries have moved away from non-tariff barriers and a highly variegated structure of tariffs to a few tariff rates that do not discriminate heavily across sectors. The extreme example is that of Chile which did away with all its quantitative restrictions and, with very few exceptions, instituted a single, uniform tariff in the late 1970s. During late 1980s, Mexico had replaced virtually all trade restrictions by three tariff rates. Yet another example is Bolivia, which has adhered more or less to a single tariff rate since the early 1980s.  
Is the replacement of non-tariff barriers by a single, uniform tariff a good idea? There are two schools of thought on this issue -- one led by policy economists and the other by academic economists. Policy economists, frustrated by the complexities of trade policy regimes in most developing countries, find the replacement of all trade restrictions by a single uniform tariff as the most effective instrument of minimising trade policy distortions.  By contrast, academic economists, working in the tradition of optimal tariff and tax literature, rarely think of a uniform tariff as a serious policy option. There are exceptions to this general tendency in both camps but they are sufficiently few to merely reinforce the rule.
Among academic economists, Corden (1958, 1985) and Harberger (1990), both acutely aware of the limitations of a uniform tariff as the "optimal" structure, strongly favour it. In his academic writings, Corden (1971, 1974) has systematically shown why tariff uniformity is non-optimal under most circumstances. Yet, in his policy writings and policy advice, he favours uniformity with strong conviction. In the policy world, economists trained in the tradition of modern optimal taxation theory a la Diamond and Mirrlees (1971), have been persistently opposed to tariff uniformity. [Dahl, Devarajan and van Wijnbergen (1986), Heady and Mitra (1985) and Mitra (1987)]  
It is argued that the differences between the two sides are the result of the failure of the latter to appreciate the complexities of policy making in developing countries and the advocacy by the former of a uniform tariff for wrong reasons. Academic economists rely primarily on theoretical models to show why under most circumstances the optimal structure of tariffs is non-uniform but fail to address how this structure is to be calculated and implemented in practice. Policy economists, finding theoretically derived optimal structures to be too complex to be of practical value, fall back on the uniform tariff as a practical solution to the problem of minimising distortions.
To keep the arguments in sharp focus, it is important to clarify the context of the debate at the outset. Uniform tariffs are usually advocated in the context of a small, open economy. For many developing countries, this is a reasonable assumption. Obvious exceptions to this are exporters of primary products who have a very large share in the world market.   Analytically, the assumption allows us to abstract from positive optimal tariffs resulting purely from market power in the world markets.
In the absence of exogenously specified non-economic objectives or political constraints resulting from, say, lobbying, the optimal trade policy for a small open economy is complete free trade. Therefore, any discussion of positive optimal tariffs must presuppose the existence of such objectives or constraints. Two objectives that have played an important role in the academic literature as well as policy discussions are protection and government revenue. In addition, the income distribution objective has received some attention. In principle, trade taxes are rarely the first-best policy instrument. For example, when revenue is the objective, consumption or value added taxes are a superior instrument. In the favourite model of trade theorists and policy makers, factor endowments capital, labour and land are assumed to be fixed. In this model, a uniform consumption or value added tax on all commodities leaves relative prices entirely undistorted from the world prices. Therefore, revenue can be raised at zero cost; essentially the uniform consumption or value added tax amounts to a lump sum tax. Unfortunately, the same cannot be accomplished via trade taxes. A uniform trade tax on all goods amounts to a tax on imports and subsidy on exports (negative imports). If trade is balanced, the subsidy paid on exports offsets the revenue raised on imports. To raise positive revenue, a uniform trade tax cannot be imposed on all trade (including all exports) and relative prices must be distorted.  The focus on trade taxes in this context presupposes that domestic taxes are not available due to an absence of collection machinery. For many African and South Asian countries, this is a realistic assumption. Similarly, when protection is the objective, the superior instrument is a production subsidy. However, due to fiscal considerations, it is not a feasible option in reality. Because tariffs generate revenue while providing protection, in practice, they remain a preferred policy instrument.
ACADEMIC ECONOMISTS' OBJECTIONS TO A UNIFORM VAT & TARIFF: As already noted, the standard context in which uniform tariffs are advocated is that of a small open economy. It is further assumed that there are constant returns to scale in all sectors and no distortions anywhere in the economy. These assumptions are a good starting point because in their absence, the case against a uniform tariff is trivial.   The global optimum in this setting is achieved by setting all tariffs uniformly at rate 0, i.e. free trade. To make the tariff issue substantive, we must introduce a noneconomic objective. We begin with the revenue objective.
THE REVENUE OBJECTIVE: Suppose the government wishes to raise a certain amount of revenue through trade taxes. A distortion must now be introduced and global optimum sacrificed. The issue is what is the least distortionary set of trade taxes to raise the specified revenue? Clearly, tariffs that move the economy the least from the free trade equilibrium while raising the required revenue will do the trick. Ignoring cross price effects, this amounts to taxing imports with low import demand elasticity more heavily and those with high elasticity less heavily. For each dollar raised in revenue, the movement away from the optimum is less for goods with low elasticity than those with high elasticity. Therefore, it makes sense to introduce a proportionately larger tariff distortion in the former than in the latter. This is indeed the essence of the well-known Ramsey (1927) result which states that when lump sum taxes are not available, revenue raising taxes should be levied in inverse proportion to the elasticity of demand. Because import demand elasticities are usually different across commodities, optimal revenue raising tariffs will be non-uniform. Cross-price effects only strengthen this point.
PROTECTION OBJECTIVE: Next, consider the protection objective. If the objective takes the form of protection targets by sectors as in infant-industry protection, a uniform tariff cannot be the right instrument. The objective in this case is to provide differential protection across sectors which a uniform tariff aims to avoid.  The most favourable form of a protection objective for the uniform tariff is to require that the overall value added in the domestic import-competing sectors, measured at world prices, be above the free trade level. This objective can be best achieved by a uniform subsidy on value added to all sectors. Such a subsidy creates equal distortion at the margin in all sectors and creates no by-product distortions in the economy. The cost of achieving the objective is minimised.  As already noted, tariffs are a more convenient instrument of protection because they raise revenue while subsidies impose a fiscal burden on the economy. Proponents of uniform tariffs argue that absence of subsidies in a uniform tariff is the least distortionary instrument for achieving the protection objective. A uniform nominal tariff, applying equally to final goods and inputs, results in an equal ad valorem subsidy to value added in import-competing sectors. In a trade theorist's jargon, a uniform tariff leads to equal effective protection across all import-competing sectors.
Opponents of uniform tariffs are quick to point out at least four major problems with these plausible sounding conclusions. First, if some import competing sectors use one or more exportables or non-tradables as inputs, a uniform nominal tariff fails to equalise effective protection across sectors. No tax is paid on exportables and nontradables used as inputs. Therefore, a uniform nominal tariff protects the value added in sectors using exportables or nontradables as inputs more than in other sectors. The marginal cost of protection is higher in the former than the latter sectors. The cost of providing the same overall protection to value added can be reduced by applying a lower nominal tariff on sectors using exportables and nontradables as inputs and higher nominal tariff on other sectors.
Second, if one or more imported inputs are used in some exportables or nontradables, a uniform effective rate of protection no longer minimises the distortion cost of protecting value added in import competing sectors. In addition to creating the desired distortion, i.e. a uniform effective protection in import competing sectors, the uniform tariff now also distorts production in exportable and nontradable sectors using imported inputs. Lowering the tariffs on inputs used in exportables and nontradables and raising them on inputs used exclusively in import competing sectors can reduce the distortion cost. This will shift the distortion away from where it is not desired (exportables and nontradables) towards where it is desired (import competing goods).
Third, tariffs distort not merely production but also consumption. If we assume that no imported inputs are used in exportables or nontradables and no exportables and nontradables are used in the production of import-competing goods, a uniform nominal tariff will coincide with uniform effective protection and  moreover, minimise the distortion in production. Yet, it will not minimise the overall cost of the protection objective. Since the byproduct distortion in consumption is not desired, overall protecting the goods with inelastic consumption demand more than others can lower distortion costs. This change will increase the distortion cost in production, lower it in consumption, and up to a point lower it overall. Uniform tariffs whether nominal or effective are non optimal.
Finally, if tariff evasion is possible via smuggling, even if the conditions for a uniform nominal tariff to be optimal are satisfied, its adoption will fail to yield the optimum. Not all goods can be smuggled with equal ease: automobiles are far more difficult to hide in a suitcase than wristwatches. A uniform nominal tariff on the books will translate into a non-uniform nominal tariff in practice.
INCOME DISTRIBUTION OBJECTIVE: For completeness, one may also note the objection to a uniform tariff based on the income distribution objective. Sometimes, tariffs are used to curb the consumption of luxury goods. In this context, the case against a uniform tariff needs no elaboration. One point to note, however, is that a consumption tax on luxury goods may not be as infeasible as a general consumption or value added tax. Therefore, it may often be worth giving serious consideration to a consumption tax for income distribution purposes. [The second instalment will appear on Saturday, April 02, 2016]
This is a part of the Conference Paper titled "From Differential to Uniform Rate System: The Political Economy of Reforming Value Added Tax System in Bangladesh" presented at the Members Conference of the Institute of Chartered Accountants of Bangladesh. Mr Nojibur Rahman, Secretary, Internal Resources Division and Chairman of NBR, was the chief guest while Barrister Jahangir Hossain, member, VAT Policy, was the special guest. Jamaluddin Ahmed, PhD, FCA, is Chairman, Emerging Credit Rating Ltd.
[email protected]