10 Poverty, Fairness, and the VAT
- Liam Ebrill, Michael Keen, and Victoria Perry
- Published Date:
- November 2001
It is a common fear when a VAT is introduced or extended that there will be an adverse impact on poverty, or on the distribution of real income more generally. Many of the central issues raised by such a concern have been discussed at some length in Chapters 7 and 8 above, which considered the specific design issues of an appropriate rate structure and exemptions for the VAT. But the conceptual and practical considerations at stake are somewhat wider, and have in any event proved sufficiently widespread and powerful to merit explicit consideration. That is the task in this chapter.
The VAT in the Wider Tax System
It is potentially misleading to focus on the distributional impact of the VAT in isolation. What affects poverty and fairness is not the impact of any particular tax, but the impact of the tax system as a whole. This echoes one of the key points to emerge from the discussion of optimal VAT rate structures in Chapter 7: the importance for the appropriate design of the VAT of the range of instruments available to the government. More generally, it is not the tax side alone that matters for the alleviation of poverty and pursuit of fairness but public spending policies too: a regressive tax might conceivably be the best way to finance pro-poor expenditures, with the net effect being to relieve poverty.93
These broad points have played an important role in FAD advice. A common response to concern about the equity impact of the tax is that the proper role of the VAT is to raise revenue in an efficient manner, with distributional concerns best addressed by other tax or spending instruments. While the theory reviewed in Chapter 7 provides some support for this, it also emphasizes that it is only in quite extreme circumstances that distributional concerns become completely irrelevant to the design of the VAT. It has been seen, in particular, that the range of other instruments that the government is able to deploy is critical. The danger is that by simply referring in broad terms to the potential superiority of other instruments one may too easily overlook the restrictions that are faced in the use of such instruments, and so avoid confronting the real distributional issues. Indeed the departures from the broad-based, single-rate structure documented above suggest that the presumption of there being better-targeted instruments is not always found entirely compelling by the authorities.
Moreover, the VAT has become such a significant tax that it will to a large degree shape the impact of the tax-spending system as a whole. While care must thus be exercised not to forget the crucial role played by the rest of the tax-spending system, understanding the distributional impact of the VAT, viewed in isolation, must be part of informing and designing an overall strategy toward the relief of poverty.
Is Taxing Consumption Regressive?
A single positive rate of VAT applied to the broadest possible base—as advocated by FAD—is essentially a proportional tax on consumption. The burden of such a tax need not be distributed in that way, of course; but it is supposed in this section, as is usually done (if only implicitly), that it is.
Such a tax—and the arguments here apply to any flat tax on consumption, not just the VAT—will appear to be regressive when effective tax payments are related to current income: since the proportion of income that is spent tends to decrease with the level of income, so the proportion of income that is taken in paying a flat-rate VAT will decrease with income. In that sense, a broad-based, single-rate VAT—indeed any flat tax on consumption—is inherently regressive.
But this may not be the best way to think of progressivity. Welfare, presumably, ultimately depends on consumption, not on income. Thus higher income households are not necessarily better-off households. Rather, it is the manner in which tax liability varies with consumption that matters for assessing the distributional implications of a tax, not how it varies with income. The point is most easily seen by considering an idealized world in which all individuals can borrow or lend as much as they wish against their future income. The level of consumption that each individual chooses to sustain over time will then reflect their own assessment of their lifetime income, and so will be a better indicator of their lifetime welfare than is their current income. Some people who are poor in terms of their lifetime income, for example, may be observed in some particular period to have relatively high current income (perhaps because they are then youthful and productive in unskilled tasks). It would clearly be wrong to regard such a household as rich, and to interpret the low ratio of consumption tax payments to income for this household as indicative of regressivity. While the point that the distribution of the tax burden may be better viewed relative to consumption rather than income is thus especially clear in a context of perfect capital markets, the same point is valid more generally: the key observation is that it is consumption over the lifetime, not income over any relatively short interval of the lifetime, that matters most for welfare comparisons.94
This point is now quite widely appreciated for developed countries,95 but less so for developing. While it implies that a broad-based tax on consumption is not regressive, it does not imply that it is progressive. Rather it leaves the view of such a tax as essentially proportional and so neutral in its distributional effects.
Pursued further, an intergenerational perspective would suggest that the most important equity consequences of moving toward a VAT might not be those relating to distribution within generations but rather those concerning distribution across the generations.96 To the extent that current consumption is financed by past savings, a consumption tax is akin to a levy on the stock of savings, and so may bear most heavily on those who are relatively old at the time of the shift: those who are young will of course pay tax on their own future consumption, but the savings they have yet to make to finance that consumption will be untaxed.
Mitigating the Distributional Impact
The discussion so far has assumed a single-rate VAT levied on essentially all domestic consumption. In fact, measures can be taken to mitigate the distributional impact of the VAT by taxing different commodities at different rates and/or by introducing exemptions. These measures were discussed at length in Chapters 7 and 8. The essential conclusion, recall, is that the redistribution which can be achieved by these means is limited and can come at significant administrative and other cost (especially for exemptions). While the VAT can be significantly compromised by turning it to pursue distributional objectives, however, there may be circumstances—likely to be ones in which government has few other tools available to it—in which these considerations should not be ignored.
The use of other tax instruments to offset VAT effects can be quite explicit. Canada, for instance, introduced an explicit tax credit at the time of introducing the GST; and both Canada and Australia have introduced tax credits for first-time house buyers to help offset the increase in house prices expected upon the introduction of VAT on sales of new houses. It may also be appropriate to take measures to address the intergenerational consequences of the VAT, perhaps an increase in pensions—brought about automatically if these are indexed to consumer prices—or (less well targeted) a credit for elderly taxpayers.
Another common concern at the time of introducing a VAT is that traders will use the opportunity to bring about an unwarranted price increase. How well founded a fear is this is not clear: if traders can exploit monopoly power when the VAT is introduced, why not exploit it even without the VAT? Perhaps the introduction of a VAT provides a focal point for tacit collusion. In any event, this is commonly a real concern among the public. A number of countries, including South Africa and the United Kingdom, have sought to reassure consumers on this score by establishing temporary watchdog bodies at the time of introducing their VATs, giving them a mandate to receive, investigate, and perhaps publicize complaints of unfair pricing.
The distribution of VAT payments is likely to be quite sensitive to patterns of consumer preferences and to the rate structure and exemptions built into the VAT itself. While the empirical evidence does show significant diversity of experience, it also points to some general observations.97
As one would expect, VATs that are characterized by a broad base and simple rate structure appear regressive when viewed relative to some measure of current income and broadly proportional when viewed relative to total consumption. OECD (1988), for instance, finds this to be true of three of the four countries for which it presents comparable figures: Denmark, the Netherlands, and Sweden.
It is also clear, however, that rate differentiation and exemptions can make the VAT a progressive tax, at least, when viewed relative to consumption. For the United Kingdom, for instance—whose VAT is marked by extensive zero-rating of food and other items important in the budgets of the less well off—OECD (1988) finds that the proportion of consumption taken in VAT rises from 6.3 percent of total consumption for the decile of the population with the lowest consumption to 9.5 percent for the decile with the highest consumption.98Kay and Davis (1985) reach a similar conclusion, finding the distribution of VAT payments in the United Kingdom to be progressive in the sense that as income increases so a larger proportion of spending is on items that are fully taxed. Similarly, Ballard and Shoven (1987) find, for a computable general equilibrium model calibrated to the United States, that rate differentiation can substantially reduce the harm that low income groups suffer from moving away from an income tax and toward a VAT: in one of their simulations, for instance, moving from a single rate to a differentiated VAT benefits the poorest decile by nearly ½ percent of their lifetime income (and indeed turns a loss for them into a gain).
While studies of VAT incidence for developing countries are scant, the growing evidence is that there too, the VAT can be strongly progressive when viewed relative to consumption. Younger and others (1999), for instance, find that in Madagascar the distribution of VAT payments is more equally distributed than is the total consumption: that is, the share of all VAT that is paid by the poor is smaller than their share of total consumption.99Younger and Sahn (1998) reach the same qualitative conclusion for Côte d’Ivoire, Guinea, and Tanzania. Indeed, not only does the VAT emerge from these studies as progressive, it emerges as more progressive than several other taxes. Importantly, it proves more progressive, in particular, than the trade taxes that in many contexts it has replaced. This certainly cannot be taken for granted: for Bangladesh, for instance, Hossain (1995) finds that while zero-rating food would greatly mitigate the adverse impact of replacing the preexisting indirect tax regime by a VAT, it would not eliminate it. Nevertheless, recent work is challenging any notion that the VAT is an inherently regressive tax.
There is thus little doubt that by introducing rate differentiation and exemption the VAT can in many cases be made less regressive than a proportional tax on consumption, and that indeed it has in some cases been turned into arguably quite a significantly progressive tax. Whether it is appropriate to complicate the VAT in this way is of course another question (that has been discussed in earlier chapters); Ballard and Shoven (1987), for instance, also find that the distributional benefit of rate differentiation comes at a substantial efficiency cost. Moreover, the empirical evidence does tend to confirm the basic point that, despite its potential for progressivity, the VAT is generally less progressive than wage or salary taxes.
The Treatment of Small Traders
The issues addressed so far relate to the impact on the real incomes of those who purchase taxed commodities. A second and distinct set of concerns relates to the impact on those who sell them.
In particular, it is generally agreed that the costs of complying with the VAT are likely to include a significant fixed component, and so may bear most heavily on smaller traders. There is again an incidence question: part of this burden may be passed on to consumers in the form of higher prices.
Clearly, a critical role is played here by the choice of threshold above which registration for the VAT is compulsory. This is discussed in the next chapter, but some points bearing on equity require emphasis here. Most obviously, the level at which the threshold is set can be raised, if felt necessary, to exclude from the requirement to charge the tax those traders for which compliance would be especially burdensome. Focussing on those traders who sell directly to consumers, the operation of the threshold conveys a double benefit on the smallest traders who remain below it: not only are they spared the compliance burden of their competitors above the threshold, but they are also spared the requirement to charge tax. The distributional impact of the VAT on traders as a group may thus be rather complex: while small traders just above the threshold may suffer through a disproportionately high compliance burden, those who are even smaller benefit by acquiring a distinct competitive edge over taxed traders (paying tax on their inputs, like their competitors, but escaping tax on their own value added). Thus it is far from obvious that all small traders are harmed by the VAT (compared to, say, a retail sales tax levied on all traders, or indeed to no indirect tax at all).
This picture is more complicated, however, in respect of traders who sell primarily to other traders rather than direct to final consumers. While such traders avoid the requirement to charge tax by remaining below the threshold, they must also pass on, or bear themselves, the tax that they have been charged on their own inputs. This latter disadvantage could be avoided by registering voluntarily for the VAT—but then the compliance costs are again incurred. Small traders wishing to sell mainly to other registered traders can indeed be placed in a difficult position under a VAT.
Note that even those below the threshold who sell to final consumers are still liable to pay tax on their inputs, which is then not recovered. This ability of the VAT to reach enterprises below the threshold by the taxation of their inputs enables the burden of the VAT to be spread more widely and, hence, potentially, more fairly than that of other taxes.100 Those who should register for the VAT but seek to evade tax by failing to do so, in particular, will be taxed under the VAT to the extent of their taxed inputs. This quiet taxation of hard-to-tax sectors through VAT on their inputs can indeed be deployed deliberately to impact on such sectors. As already noted, the agricultural sector, for instance, often proves hard to tax through the income tax, land, or wealth taxes: taxing the sector through its inputs may be acceptable as a third-best method of ensuring some equity in the distribution of the tax burden between this sector and others. This improved intersectoral equity might at first sight seem to imply some reduction in intrasectoral equity, as larger farmers will be able to recover their input taxes against output tax. But in doing so those larger farmers will ultimately pay tax not only in relation to their inputs but also on their own value added, so that smaller farmers will be favored; only in the case—a potentially important one—in which agricultural outputs are taxed at a sufficiently low rate that registered producers are due refunds (perhaps because they are exported) will the taxation of agricultural inputs disfavor small farmers.
It seems that the VAT is quite widely perceived—especially by those with relatively little experience of it—as an especially regressive and unfair tax. Quite where this belief comes from, however, is far from clear.
The “ideal” VAT is a tax on consumption; and a strong case can be made that since an individual’s consumption is one of the best observable indicators of their living standard, so consumption is potentially one of the most equitable of tax bases. True, a uniform tax on all forms of consumption takes the same proportionate amount from those with low consumption as it does from those with high, so that it is not in that sense a progressive tax. But some progressivity can be (and commonly is) introduced by exempting key products, and perhaps by some rate differentiation—though we have argued that the equity gains to be achieved in this way may be quite limited relative to the revenue foregone. These are in any event objections that apply not only to the VAT, but to other broad-based sales taxes, such as a retail sales tax. Indeed the treatment of small traders under the VAT is generally more favorable under a VAT than under alternative forms of consumption tax, since those selling to final consumers—small retailers in particular—are placed at a competitive advantage relative to larger.
To some extent those who perceive the VAT as regressive may be implicitly comparing it with a progressive personal income tax. But this not only overlooks inequities associated with income taxation that are not as widely understood as they should be (the inequity some see, for instance, in taxing more heavily those whose income is highly variable relative to those with more stable receipts). It also ignores the considerable difficulty that developing countries experience in levying an effective personal income tax, one of the most challenging taxes to administer. Indeed the adoption of a VAT is often intended as the first stage in a reform process that will ultimately lead to effective income taxation.
The more general point, however, is that few taxes are very well suited to the pursuit of equity objectives. Expenditure policies can often be far better targeted to these aims, and in that context the first duty of taxation is to raise revenue with as little distortion of economic activity as possible. It is important not to overstate this point: many developing countries also face severe limitations on the effectiveness of their spending policies, and these may properly temper tax policy advice. The key point, however, is that it is in this wider and more difficult setting that the potential of the VAT to alleviate poverty and enhance fairness must be assessed.