Discussion Paper

No. 2018-57 | July 19, 2018
(Submitted as Global Solutions Paper)


Governments use tax expenditures to boost investment, innovation and employment. However, these schemes are largely opaque, costly and often ineffective in reaching their stated goals. They also frequently trigger unwanted side effects. In order to improve the performance of these tools, the authors present three concrete policy proposals: First, governments should increase transparency on tax benefits. G20 members should take the lead on this with frequent and comprehensive tax expenditure reports. Second, G20 governments should improve the design of tax incentives with the aim of minimizing the generation of windfall profits and negative spillover effects within and across (in particular, on poorer) countries. Third, governments should phase out tax expenditures that are environmentally harmful, including tax incentives for fossil fuels and other schemes that promote an unsustainable use of natural resources.

JEL Classification:

H2, H87, N4


  • Downloads: 702


Cite As

Agustin Redonda, Santiago Diaz de Sarralde, Mark Hallerberg, Lise Johnson, Ariel Melamud, Ricardo Rozemberg, Jakob Schwab, and Christian von Haldenwang (2018). Tax expenditure and the treatment of tax incentives for investment. Economics Discussion Papers, No 2018-57, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2018-57

Comments and Questions

Michael Thoene - Referee report
December 20, 2018 - 11:06

(i) Is the contribution of the paper potentially significant?

The contribution of the paper is potentially significant.

The paper's subject is of great importance for modern fiscal governance. As the term "tax expenditures" makes very clear, tax subsidies are government expenditures that are handled via the ...[more]

... tax system. With regard to budget transparency and the comprehensiveness of democratically approved public budgets, tax expenditures always pose a greater problems than their counterparts on the expenditure side, direct subsidies. Nevertheless, there are numerous constellations in which tax expenditures are clearly superior to direct subsidies in instrumental and allocative terms. In this respect, across-the-board demands regarding tax subsidies - e.g. the recommendation to convert all of them into direct subsidies – may be well-intentioned, but lack a practical perspective.
All the higher is in the importance of modern instruments for creating an integrated an integrated subsidy transparency for fiscal governance. They have to reflect the potential substitutionality of direct and fiscal subsidies and addresses the high degree of dependence of the concept of tax expenditure on the underlying tax system. The last aspect, in particular, is crucial for international comparisons of tax expenditures. To put it simply, tax concessions are always the exception to the rule - the exception to the regular tax burden. To the extent that tax systems differ from country to country - and even under EU tax harmonization this is very well the case - the exceptions to these different tax rules are difficult to compare. This problem has permeated the practice-oriented literature since the fundamental work of Paul R. McDaniel and Stanley S. Surrey (1985).
With references to recent good practice in Latin America, the paper by Redonda et al. shows a very appreciable sensitivity for this problem and for the necessary steps still to be taken (Proposal 1). This also applies to the other dimensions dealt with in Proposal 2 and Proposal 3. The underlying problems are vividly described in each case and proposals are put forward which one always likes to agree with from an economic point of view. Promoting more transparency for and more cost-benefit-analysis of tax expenditures is always a good idea. The authors give instructive examples for both. The phasing out of environmentally harmful tax expenditures is also a very laudable postulate.

But here, paper falls short of its potential. It describes problems and proposes solutions that may sometimes be technically difficult. But these problems are not reflected in their political economy, i.e. at the source. The paper discusses the necessary prerequisites for solving problems, but remains rather silent on the sufficient prerequisites. This weakens the significant contributions of the paper.

The example of environmentally harmful tax expenditures is the most obvious case: To propose the gradual abolition of these subsidies in the G20 is, as mentioned, the right thing to do. But it is not new for the G20. In fact, there are major political obstacles to implementing a phase-out. They primarily stem from the importance of these tax concessions in distributional terms. In most countries there are also various other political and economic obstacles to abolition. Political hurdles and trade-offs prevent the G20 from making much progress on environmentally harmful subsidies, not a lack of knowledge about environmental damages or fiscal benefits of such a policy.
The criticized deficits in transparency and cost-benefit analyses of tax expenditures are also due to clear political obstacles. Where transparency is lacking, there may also be a lack of the best instruments. But more often than not there are also "good" political reasons for governments and parliaments to avoid more transparency.
In light of stylized facts, the struggle for more transparency becomes both, more difficult and more important. If we, as economists and “experts”, make proposals against this backdrop, the credibility and, most importantly, the acceptance of our proposals increase significantly, when they display a clear awareness of such real-world obstacles and political trade-offs. From the perspective of the policy-consulting economist, this is the "sufficient condition" for contributing to a real solution. Without it, policymakers quickly qualify expert proposals as overly simplified and detached - and for this reason alone, the proposals are not taken seriously.

For the proposals of the paper of Redonda et al that clearly would be an undeserved fate. Their three proposals, when implemented, are likely to make tax expenditures more transparent, less frequently used and therefore more efficient instruments.

(ii) Is the analysis correct?

The paper takes a programmatic, not an analytical approach. Main comments see above.

Agustin Redonda - Reply to anonymous referee report
January 09, 2019 - 20:39

We would like to thank this anonymous referee for her valuable comments. We highly appreciate that readers with knowledge in the field take the time to leave a referee that will certainly help us improve the paper. Two reactions triggered by the report follow:
1) We agree that TEs can ...[more]

... be preferred to direct spending. Indeed, under certain conditions, TEs could be more cost-effective than direct spending and may hence be the best option to pursue a specific public policy goal. In particular, as discussed by Toder (2000), TEs may be preferred to direct spending when eligibility conditions are directly linked to tax return data, when it is more important to maximize the number of beneficiaries than to minimize excess claims or when the policy objective is to incentivize a clear and broadly defined activity by reducing its net price. Moreover, there exist TEs that are effective and efficient in reaching their stated goals. The US Earned Income Tax Credit (EITC) is a case in point. The EITC is a refundable tax credit seeking to strengthen incentives to work and increase children’s economic opportunity by boosting the income for poor households. The EITC has effectively lifted more than 6 million people out of poverty since the observed positive long-term effects on high-school and college graduation, on employment and wages are largest for children from the poorest households (Bastian and Miclemore, 2018). Moreover, Bastian and Jones (2018) find that, besides being one of the largest provisions in the US – it distributes around $70 billion a year to almost 30 million lower-income families – the EITC “helps pay for itself”. As shown by the authors, the EITC significantly increases labor supply, which then leads beneficiaries to pay more in taxes and receive less in public assistance, than they otherwise would have. Some tax incentives to boost R&D and innovation provide further illustration (see, e.g. Bloom et al., 2002; Rao, 2016 and Thomsom, 2017).
To sum up, we agree that the paper could be clearer by acknowledging that TEs can be effective and - under certain circumstances - even preferred to direct spending programs. This said, we do not recommend "to convert all [TEs] into direct subsidies" and, in addition, this argument does not contradict our main recommendations. Indeed, it makes the case for TE evaluation a stronger one, since this will help governments worldwide to identify those TEs that are effective and hence worth supporting, and those that are not value for money and should hence be eliminated.
2) The point regarding the political obstacles for TE reform is a fair one and well taken. The political economy of TEs is a very relevant topic indeed since, very often the lack of o political will and/or the lobbying by pressure groups are among the mains reasons to introduce new TEs and, more importantly, to keep in place provisions even when they had been proven to be ineffective. Yet, we believe that this discussion is beyond the scope of the article and may be an interesting perspective for future research in the field.