Bases for tax policy and reform

Debate about taxes is hardly unusual in any country, particularly in the run-up to an election. After all, taxes impinge on the economic decision-making of households, so they inevitably loom large on a household’s list of concerns, and therefore become an election factor. Added to this is the inevitable political tinge of the debate, stemming from the ideologies of the policymakers regarding the objectives of taxation, and therefore the nature of instruments employed.

This last point is repeatedly demonstrated to an observer of American politics, with the marked differences between the two extremes of the Tea Party on the one hand and the staunch liberals in the Democratic Party on the other. Indeed, the ubiquity of the term “the one percent” during the past couple of years reflects concerns about increasing inequality, and by extension differing ideas about how best to deal with this situation, with tax proposals unsurprisingly at the forefront. This is only complicated further by considerations of the impact of tax reform on the stuttering economic recovery.

India too has seen certain debates regarding tax reform play out over the past few years. Attempts to implement the Goods and Services Tax (GST), leading to a discussion of the merits and demerits of the same, have dominated the discourse. Recently however, a radical proposal by an unheralded research institution has provoked a storm of opinions.

The institution, based in Pune, suggested a replacement of all extant taxes with a single banking transaction tax. Its affiliation with the BJP, and the espousal of this proposal by some BJP leaders, rid the institution of its obscurity and made the suggestion receive considerable print and airtime. There have already been several commentaries on the proposal in the national newspapers, most of which pour cold water (to put it mildly) on it. In this article, I hope to discuss why this is the case within a broader context of tax policy in India, using advances in the theory of taxation as a guide.

Theoretical underpinnings of tax policy

The choice of a policy and its subsequent evaluation are almost always based on a normative ideal, a benchmark that helps separate various proposals on terms that usually reflect efficiency considerations. As alluded to earlier, taxes are different from other matters of policy because of their redistributive implications. For example, most income taxes are progressive in nature, meaning that the average tax rate increases with income.

This is a consequence of redistributive concerns entering a policymaker’s objective, with the proceeds from the upper end of the income distribution helping to fund a safety net for low-income citizens.

Taxes are the major sources of government revenue in most economies, the proceeds of which are channelled to various types of programmes, ranging from income security to those aimed at boosting productivity, like infrastructural projects.

Taxes, however, distort prices and therefore result in inefficiency and deadweight losses. There is also considerable debate in the theoretical and empirical literature about the efficacy, and hence desirability of the government trying to go beyond its minimalist protectionist role and attempting to stabilise the economy through expansionary schemes.

Taxes play a crucial role in the celebrated second welfare theorem. According to the theorem, efficiency (described by the concept of Pareto optimality, i.e. no allocation can be unambiguously improved upon) can always be arrived at using a combination of prices and lump-sum transfers. However, such transfers, while obviously non-distorting, require staggering amounts of information regarding individual preferences to be known to the policymaker, which seems unreasonable.

So, policymakers are left with other instruments to help meet their objectives. The most common form of taxation, and one of the oldest, is the income tax, whose popularity can be explained by the relatively benign informational requirements needed to employ such a tax. However, taxing income distorts an individual’s labour supply decision on the intensive margin, i.e. concerning how many hours to work.

On the one hand, taxing income reduces the effective wage and hence makes working fewer hours (enjoying more ‘leisure’) more attractive. On the other, if the demand for leisure is normal (i.e. moves with income), a reduction in income due to the tax should reduce this leisure demand as well. The tax schedule (its variation with income) would then incorporate this labour supply effect while also focusing on the nature of the tax, for example, its progressivity.

The major theoretical breakthroughs came in the 1970s with the optimal taxation literature. Policymakers would maximise their objective of social welfare in order to hit their revenue target using the tax schedule as their instrument. Taxes could be direct (on income, and either linear or nonlinear) or indirect (on transactions, commodities).

In the case of linear indirect taxes, it turns out that the choice of tax on a particular good depends on its share in the consumption profile of those agents that the policymaker is most concerned about. If, as is common, it is uplifting the poor that dominates the welfare debate, then policymakers discourage less (i.e. do not tax as much) those goods which the poor consume more of.

Further, in a production economy (with several inputs and outputs), the famous production efficiency result of Peter Diamond and James Mirrlees shows that an optimal tax maintains efficiency in production, i.e. ensures that no reallocation of inputs across sectors could increase production. The idea here is that if the economy were not at the frontier, the policymaker could increase welfare by changing the tax structure of a good consumed by everyone. Further, the policymaker would choose to tax consumer prices and firm profits, so as not to distort production patterns and efficiency.

This result has several implications, most notably for the taxation of intermediate inputs, many of which are sector specific, produced by some firms and used by others. The intuition here is that producers in those sectors would then substitute away from the most efficient input choices, leading to inefficiency. A similar argument could be used for eliminating tariffs, which introduce a wedge between consumer and producer prices.

In developing countries, however, one should be careful while recommending policy based on such results, owing to the household level production that remains a feature there, which introduces a wedge between consumer and producer prices.

The implications of the production efficiency result might seem surprising considering the near ubiquity of Value Added Taxation (VAT). The usage of VAT is often put down to concerns regarding the size of the base of the tax system (i.e. its share in revenue), which is particularly valid in a developing country with a large informal sector. However, VAT maintains a uniform tax rate across sectors, and works through a system of cross-enforcement, allowing the buyer to receive a credit that he can deduct at the time of payment of his own taxes. Unlike sales taxes, which are levied only on the finished article and hence are vulnerable to total loss if there is evasion, VAT allows partial recovery but the efficacy of the cross-enforcement process does depend on the line remaining unbroken. By incentivising parties to correctly report their transactions, proponents of the VAT claim efficiency gains. However, a recent line of thought suggests that such forms of taxation would drive firms and agents to the informal sector and away from the paper trail, an effect that needs to be incorporated particularly by developing country policymakers. Tariffs on the other hand affect the formal and informal sector alike.

The original breakthrough in the optimal taxation literature (due to Mirrlees) pertained to income tax. In this framework, incomes were assumed to be dependent on unobserved ability or productivity, with the more able individuals earning a higher income. Keeping in mind the redistributive environment, more able individuals would wish to underreport their incomes so as to reduce their tax burden. This would further constrain a policymaker, who would also keep the labour supply effect in mind. Under certain assumptions regarding the nature of individual preferences, the marginal tax rate would depend negatively on the elasticity of labour supply (as expected), on the distribution of productivities (due to the disincentive effect which would be stronger for more productive individuals) and also on the redistributive concerns of the policymaker.

The marginal tax rate for the top productivity bracket is zero, due to the disincentive effect and the fact that raising the tax rate for them doesn’t add revenue because there is no segment above them. Recent work by authors like Emmanuel Saez challenges this result, by pointing out that the disincentive effect would have to be large to keep the tax rate low, which isn’t borne out in the empirical literature. Further, the applicability of this result too is dubious. At the lower end of the income distribution, where participation in the labour market is a concern, means-tested benefits tend to be higher and are gradually phased out to counter disincentive effects on the slightly better off.

A major result due to Anthony Atkinson and Joseph Stiglitz showed the superfluity of indirect taxes in an environment where nonlinear direct taxes were feasible. Their result depended crucially on assumptions regarding preferences (in particular, the separability of consumption and leisure in the utility function), barring which the earlier channels of correlation between good consumption and leisure determine tax rates across goods. Consequently, using the direct tax alone can achieve the desired outcome. Note that the result of a uniform tax rate across goods is hardly borne out by the VAT in practice, suggesting that costs of evasion and administrative costs too should be considered.

Thus far, we have referred to income only as earned income from work, not derived from bequests or from savings. Capital income taxation is an important and controversial topic, however, and relevant for developing countries like India (where the corporation tax contributes significantly to revenue). The dominant theoretical suggestion (as usual, under certain conditions) here is that there be no tax on savings, for that would distort investment decisions by introducing a growing wedge over time between savings and returns. This neglects situations where the economy has abundant capital on account of over accumulation and hence can benefit from a tax (‘dynamic inefficiency’). The nature of the tax would then depend on the preferences of current vs. future generations. However, the same outcome could equally result from social security systems or through government debt. Further, for wider horizons, the zero capital taxation result is even more established. Credit constraints could overturn this result though, as taxing savings and providing transfers could serve as a means of insuring agents with uncertain incomes in a scenario where consumption is correlated with savings. Also, issues of classification of capital income (which sometimes seems a reward for labour) perhaps motivate a tax.

Finally, while the optimal taxation literature envisages a tabula rasa on which to frame policy, there is a perhaps more relevant literature on tax reform, which accepts the existence of certain policy measures and seeks to find the best way (if any) to improve on the status quo. This requires a consideration of all the effects of undertaking a marginal reform, i.e. essentially a cost-benefit analysis.

Also relevant for India is the research on taxes in a federal system. In general, the argument is that ‘local level’ taxation overlooks several externalities that would be incorporated by a benevolent planner (often elusive in most political scenes).

The optimal taxation literature has been quite influential in the framing of policies across the world, particularly in developed nations. Besides the already mentioned role of the informal sector in developing economies that is often neglected, issues regarding framework, choice of instruments and administration costs too should be incorporated before the model can be applied realistically.

A growing body of work has tried to single out the importance of building administrative and fiscal capacity (either by expanding the network of monitoring & collection or by figuring out ways of expanding the tax net) for a nation’s development. Improving fiscal capacity benefits citizens when they value public spending more, but it also makes evasion harder and might be expensive. This line of research also suggests that the increasing adoption of income taxes by countries over time is due to improvements in living standards, which increase the gains from taxing people and hence make adoption easier. There might also be a complementarity between improvements in legal and fiscal capacity. Over time, governments build up their capacity, and rely predominantly on income tax and VAT/sales tax for revenue. This research also covers political economic considerations that play a major role not only in the development of such capacity, but also sometimes in the nature of proposals themselves, while developing the key components of evasion costs and the benefits from public spending.

Extension to India

India too has moved on to the agreed-upon combination of income taxes and VAT, triggered in particular by the turmoil of 1991. In the following two decades, policymakers brought some much needed clarity and rationality to the tax system, which had often seemed ad hoc and not very efficient, and reduced various rates that had hitherto prompted evasion. This was in addition to measures that raised fiscal capacity and the tax base. Income taxes were progressively divided into three brackets in the early 1990s, subsequently revised following the 1997 budget. Corporate taxes too were rationalised, with the top bracket corresponding to the top income tax bracket. Dividend taxation and the taxation of profits (both of which were suggested by theories above) were simplified. The presence of various potentially distorting inducements for businesses remains an issue (although government support to industry could be justified in an environment where foreign competitors enjoyed benefits from their respective governments).

There are still quite a few areas where reform is required, especially in a situation where certain measures are introduced for brief periods to gain revenue whilst potentially distorting the economy. A key development has been the adoption of VAT in 2005, which unites several excise and sales taxes under its rubric while also improving efficiency and perhaps increasing revenue. Asymmetric treatment of goods and services has prompted steps towards the adoption of a Goods and Services Tax (GST). Progress is stalled, however, due to concerns about revenue sharing between the Centre and States and the proposed dual nature of the system. (Non-agricultural) income and wealth taxes, corporation tax and customs taxes dominate the Centre’s revenue sources, while sales taxes and agricultural income & wealth dominate state revenue sources.

The theoretical background provided makes the banking transaction tax proposal all the more curious. Interestingly, the government did introduce a temporary transaction tax of 0.1% on ‘large’ transactions in 2007, which it revoked in 2009, with the short-term objective of raising revenue while also discouraging the use of black money by creating a paper trail. Potential distortions caused by the move do not seem to have been considered. In particular, the black money network could have spawned its own informal financial system as a consequence.

The recent proposal suggests a 2% tax on all transactions, and makes grand proclamations regarding how such a tax could replace all other forms of tax. Much ink has already been spilt on this topic, including on this website, which point out the various flaws. As one would suspect, the revenue estimates are giddily optimistic at best, with some empirical research suggesting that revenue might even decline following adoption of such a tax, reflecting the reduced role of cash transactions as an economy grows richer.

Further, while some Latin American countries like Argentina, Brazil, and Peru among others, did institute a transactions tax, it was temporary and was by no means intended to replace the existing tax systems. Concerns about the impact on the informal economy remain relevant. Further, taxing transactions could even have implications for the choice of savings instrument preferred by agents, which could even distort the investment process. Some commentators have also emphasised the potentially harmful impact of removal of inducements on small businesses.

The gradual implementation of the UID based cash transfer system could have been a motivation for the proposal, but the intended scheme ignores a founding tenet of taxation, namely progressivity, by treating all transactions identically (and therefore being a flat tax). Even if targeted transfers could achieve most of their objectives, a moralistic justification remains elusive (note the slight similarity here with the ‘Buffett tax’ debate in the US).

Finally, remembering the research on fiscal capacity described earlier, surely the role of the government should be to develop said capacity, with the taxable basis naturally rising with growth and the concomitant structural transformation of the economy? With the various issues highlighted above, a transactions tax could scarcely hope to supplant the consensus methods of taxation, which too were inferior to the non-distortionary transfers. The endorsement by some (but thankfully not all) in the BJP seems therefore to be opportunistic and misguided. Tax reform is indeed important, but the advice of the various pertinent commissions set up should rather be heeded as India moves up the development curve.


References for the curious reader

Most of the theoretical references are quite technical. Two good sources nevertheless are the textbook, The Economics of Taxation (2nd edition, 2010), MIT Press, by Bernard Salani ; and the still fantastic review article, ‘Pareto Efficient and Optimal Taxation and the New New Welfare Economics’ in the Handbook of Public Economics (Vol. 2, 1987) by Joseph Stiglitz. Two relatively simpler articles are ‘Optimal Taxation in Theory and Practice’ (2009) by N. Gregory Mankiw et al. and a partial rejoinder, ‘The Case for a Progressive Tax: From Basic Research to Policy Recommendations’ by Peter Diamond and Emmanuel Saez (2011), both in the Journal of Economic Perspectives.

For applications to developing countries, see ‘Taxation for developing countries’ by Ehtisham Ahmad and Nicholas Stern in the Handbook of Development Economics (Vol.2, 1989); and the very readable ‘Taxation and Development’ by Robin Burgess and Stern in the Journal of Economic Literature (1993). A good source for the fiscal capacity literature is the very recent survey, ‘Taxation and Development’ by Tim Besley and Torsten Persson in the Handbook of Public Economics (Vol.5, 2013).

A good overview on India’s experience with tax reform is ‘Tax System Reform in India’ by M. Govinda Rao and R. Kavita Rao in Roger Gordon (ed.) (2010), Taxation in developing countries: Six case studies and policy implications, Columbia University Press. On VAT, see ‘The value added tax: Its causes and consequences’ by Michael Keen and B. Lockwood in the Journal of Development Economics (Vol.92, 2010), which contains references on the impact of the informal sector on VAT, as does the Besley-Persson article.

Lalit Contractor has a MPhil in Economics from Oxford University and is curious about Economics and its interactions with politics and society.