Keen: A VAT is a tax levied on all sales of commodities—goods and services—at every stage of production. The crucial element is that it credits taxes paid on inputs against taxes charged on outputs. By contrast, a simple turnover tax is levied on each transaction. This tax cascades through the chain of production and provides enormous incentives to, among other things, avoid the tax by vertical integration. Crediting allows the tax process to have a much more neutral impact on economic activity.
Ebrill: A combination of rising revenue needs—which implied a need for higher tax rates in general—and the degree to which earlier taxes, such as the turnover taxes, led to misallocations of resources. Also, most VATs operate as a tax on consumption, which is appealing from an economic perspective. But the rise in the VAT has truly been extraordinary (see map, page 156).
Ebrill: The department has been providing technical assistance on tax matters for many years, and the VAT is a major part of this. We wanted to be certain we were providing the right advice, and we wanted to know if there were areas where we should change our advice or at least make it more consistent.
Keen: Also, the VAT is hugely underresearched. To address our technical assistance concerns, we had to look at the VAT in a broader context and ask how it had fared and whether it had improved efficiency.
Keen: A key question was whether any elements of our advice have not worked. The study surveyed VAT experts on individual countries, and this allowed us to compare IMF advice and actual practice. Where there had been systematic differences between the two, we asked ourselves: Is our advice incorrect or are we not selling it effectively? The study yielded useful information in a number of areas.
Ebrill: We certainly found the devil was in the details. The exercise was much richer in insights on details of the tax than we had anticipated.
Ebrill: When the VAT first came into being in Europe, it was presumed to be a tax for sophisticated countries with sophisticated taxpayers. Now, the VAT is being implemented in a wide range of economies. We found that if the crediting mechanism is emphasized and if individual country circumstances are taken into account, the VAT can be applied just about anywhere. The only remaining questions have to do with the smallest open island economies.
The preconditions were one of the most important lessons of this exercise—and one we perhaps did not anticipate. Some tax administration reforms should be in place before a country embarks on a VAT, and these reforms are more difficult and more time-consuming to implement than we expected. Our results strongly suggest that self-assessment and a tradition of paying taxes are necessary preconditions for the effective implementation of a VAT.
Keen: We also came to appreciate that in some countries a VAT comes about naturally. The VAT seems a dramatic step, but in many cases a country with a broad-based tax notices problems with cascading and introduces crediting arrangements. In effect, the VAT serves as a tidying-up exercise. Of course, a VAT is a deepening of the tax process, but in a structural sense, conditions often dictate when a VAT emerges.
Ebrill: On some issues, the study provided definitive results; on others, the findings clarified our understanding of an issue but did not necessarily provide a clear-cut answer. In the debate over whether multiple rates or a single rate is preferable for a VAT, we found a genuine trade-off between tax administration considerations, which favor single rates, and tax policy considerations, which suggest taking account of differences in demand for various commodities. Most experts are still inclined toward a single rate, but certainly there are aspects of the issue that cannot be resolved definitively.
Keen: In contrast, we did make genuine progress on where to set the threshold for a VAT. The study found that we quite consistently recommended thresholds higher than those actually adopted. We asked ourselves what considerations drove those decisions and what constituted an optimal threshold for a VAT. With the benefit of simple modeling, we figured out the marginal costs and benefits of moving the threshold around.
A higher threshold saves on costs for both the tax authorities and the taxpayer. But a higher threshold also exacts a cost in terms of revenue and can induce distortions between companies above and below the threshold. Theory tends to back up the IMF’s preference for a high threshold, but why, then, haven’t we persuaded countries about this? We had perhaps tended to ignore the importance of imposing a simple tax on people below the VAT threshold, because the resulting revenue was not much. But such a tax may help politically to sustain a high threshold and thus be advantageous. This helped us bring policy and administrative perspectives together and think through a coherent strategy.
Ebrill: In the treatment of agriculture, we found our advice was not necessarily consistent across countries, and certainly the implementation of that advice was not. We realized we could come up with a much more coherent argument as to why agriculture should be treated like any other sector. Agriculture is a very difficult area to tax, but some of the devices used to exempt agriculture are causing even more problems.
Keen: The study also provided useful information on exemptions, which waive the tax charged on outputs but do not credit for amounts paid on inputs. In terms of tax neutrality, the VAT can be a thing of great beauty—at least to a certain strange kind of person. Even multiple rates preserve the elegance of a VAT by retaining the critical element of crediting. But exemptions truly muck it up. Once the cascading starts, the tax system loses any kind of transparency.
Ebrill: The study focuses on areas of difficulty—notably, auditing, self-assessment, and refunds. Auditing requires a rational and effective use of scarce resources. The study found that the audit function has often been underem-phasized in the implementation phase. Initially, taxpayers assume everyone is watching them, so this does not cause a problem. The problems come later, if taxpayers realize they are not being watched.
The study also looked closely at self-assessment. With self-assessment, taxpayers figure out their own tax bill rather than have the authorities do this for each citizen and assess accordingly. We found that introducing a VAT in countries that aren’t comfortable with self-assessment causes problems. The authorities then sometimes compound this problem by imposing more, not fewer, requirements on taxpayers. Clearly, countries uncomfortable with self-assessment should consider developing that capacity before pushing ahead with a VAT.
Keen: You build on what you already have, and education and taxpayer training help. You also keep the tax system simple. In countries with final withholding arrangements, initially most individual taxpayers would not have to file returns, and the number having to make arrangements for their tax liabilities would be minimized. It would be an incremental process. As people learn how to work with the system, it can become more sophisticated.
Ebrill: Refunds are clearly also a significant problem with the VAT in many countries. Exports should be zero rated, and the exporter should be entitled to whatever VAT is already embodied in the good. In practice, however, this is not working very well. Countries with fiscal problems are reluctant to part with the cash they have collected, and where there are concerns about fraud, the authorities may not be confident about their ability to identify genuine claims. Whatever the reason, many countries are not paying refunds. But the moment you do not pay refunds, you undermine the very principle of a VAT. Quantitatively, this can be a very large problem—in many developing countries, half of gross VAT receipts could be refundable.
Keen: Some countries restrict refunds to exporters or to capital goods. This reduces the apparent size of the problem but still jeopardizes the crediting mechanism. We have tried—though not entirely successfully—to develop a simple indicator for what the proper level of refunds in a country should be.
Ebrill: Going into the study, we asked ourselves whether a VAT was appropriate for small island economies. It is not clear, at least in simple small economies that import a lot of their consumption, that there is a huge gain from moving toward a VAT. A tariff may achieve the same result. But we found it was not so much a matter of the economy’s size as it was a question of how much of the economy’s consumption comes from imports, how much domestic production there is, and how complicated that domestic production is.
There are many small island economies where the VAT appears to work very well. Barbados, for example, has an increasingly sophisticated economy and a growing service sector, and it found that a VAT allowed it to broaden its tax base. The evidence suggests a VAT is more useful in more circumstances than some have argued.
Ebrill: Our study pulls together a lot of useful insights in a cogent fashion and allows us to ensure that our technical assistance program is as coherent and fruitful as possible. We have sent this study to the IMF’s Executive Directors and are hopeful we will receive comments from their national authorities. At some future stage, we are also considering regional conferences to meet with officials and identify precisely where the points of tension with the VAT are.
Keen: Two issues warrant more attention in the future: how VATs function in federal systems of government and the growing impact of E-commerce on VATs.
The VAT has functioned very well as countries have become more closely integrated in economic terms, because it so neatly removes indirect taxation from exports. It’s one reason why it has been so successful in the European Union. The irony is that the VAT has proved problematic when integration proceeds beyond a certain point. That’s because it becomes much more difficult to enforce zero rating on exports in the absence of border controls—it is simply harder to check that goods have actually been exported. One challenge now is how to run a VAT within increasingly integrated economic units that see themselves as a single country. This is why within Europe, for example, progress on the VAT has virtually come to a halt.
This issue comes up in other countries as well. It’s no coincidence that two of the biggest countries without a VAT are the United States and India, which are both federal countries. The issue has come up in our technical assistance work in the CIS [Commonwealth of Independent States], Bosnia, and India (which recently announced its plans to adopt state-level VATs). We have been doing conceptual work on this and see it as an important area for the future.
On the impact of E-commerce, the question is whether the Internet changes anything fundamental or simply intensifies existing problems. To a large extent, we see it intensifying existing problems—particularly the treatment of services. Under a VAT, it has always been difficult to treat services, because it is hard to intercept services as they cross borders. E-commerce, however, may make it harder to control some additional transactions. Once we digitize goods, for example, they become hard to intercept at the border.
The study on the value-added tax is with national authorities for comments. The study, with comments incorporated, will provide the basis of a forthcoming book by the authors.