Advancing the evaluation of tax reforms: A new IMF database

Era Dabla-Norris, Ruud de Mooij 07 August 2018



In 2016, Laura Kawano and Joel Slemrod published an important article that criticises prior studies estimating Laffer curves for corporate income tax (Kawano and Slemrod 2016). Those studies typically regressed corporate tax revenue on corporate tax rates in a quadratic relationship, and then inferred from that the revenue-maximising tax rate. Kawano and Slemrod, however, show that these studies ignore information about corporate tax bases – which systematically vary with rates – and thereby attenuate the relationship between corporate tax rates and corporate tax revenues. Correcting for this omission raises the revenue-maximising tax rate from roughly 30% to 50%.

Kawano and Slemrod’s study illustrates a key problem in tax research, namely, that quantitative information is often only available for tax rates, not tax bases. To address this weakness, studies have used alternative indicators of tax, such as realised tax payments, to measure the combined effect of tax rates and tax bases. For example, studies on taxation and economic growth have typically used of tax-to-GDP ratios as the independent variable to measure tax burdens. However, such regressions are likely to yield biased estimates since the tax-to-GDP ratios themselves depend on GDP growth.

To address those methodological problems, the IMF has developed a new Tax Policy Reform Database (TPRD) that was made publicly available for research in April 2018. It uses text-mining techniques to extract granular information on past tax policy reforms in 23 countries during four decades, using more than 900 OECD reports and 37,000 news clips from International Bureau of Fiscal Documentation (IBFD). The TPRD reports information on tax rate and base changes, reform announcement and implementation dates, whether the measures are implemented over multiple years or not, whether the reform was major or minor, and if they were implemented as part of a fiscal consolidation package. 

The TPRD can be used to infer systematic trends in taxation and comparative analysis across countries. For example, it reveals that the 23 countries implemented, on average, 130 tax reforms over the past few decades, mostly in personal and corporate income tax systems. France and Italy top the list with over 200 reforms, while Brazil and Czech Republic had the lowest number at fewer than 50 reforms (Figure 1). The majority of reforms in personal and corporate income taxes were base changes; in VAT, the majority were rate changes. The average time lag between announcement and implementation was 150 days, but the variation is large. Interestingly, unlike common belief, corporate tax rate reductions have recently come along with a narrowing of corporate tax bases. Amaglobeli et al (2018) provide more detailed information.

Figure 1 Frequency of tax reforms by country and tax type

Source: Amaglobeli et al. (2018)

The TPRD also lends itself to addressing some methodological problems of existing tax research and allows researchers to analyse new questions. A few examples. 

Let us start with taxation and growth. 

  • First, the TPRD contains information on whether tax reforms have been implemented in years of fiscal consolidation (during which tax reform is more likely to be unrelated to the business cycle). This circumvents endogeneity problems and thus is more likely to result in unbiased estimates. 
  • Second, the TPRD distinguishes not only various tax categories, but also tax rate versus tax base changes. The latter turns out to be important (see Dabla-Norris and Lima 2018). Indeed, when normalised for the same fiscal impact, rate increases are associated with significantly larger negative effects on long-term growth (both output and employment) compared to base broadening reforms. Moreover, short-term tax multipliers turn out to be significantly larger if they are driven by changes in tax rates compared to changes in tax bases.

Figure 2 Growth effect of a base broadening versus rate increase, equivalent to 1% of GDP

Source: Dabla-Norris and Lima (2018)

Another example is strategic tax competition. Several prior studies had estimated fiscal reaction functions by regressing the tax rate of a country on a weighted average of other countries’ rates – using instrumental variables to address endogeneity concerns. The TPRD allows to enrich such analysis in several ways, as shown by Hebous et al. (2018): 

  • by looking at how countries respond to announcements and implementations; 
  • by exploring the importance of controlling for tax base changes; and 
  • by considering whether countries also compete by using their tax bases. 

Many other tax questions can also be addressed using the TPRD, some of which IMF staff are currently exploring. For instance, we plan to look at the political economy of tax reforms: How do tax reforms affect elections? Do elections affect reform announcements, implementations and reversals? We are also analysing whether single tax reforms have different implications from tax reform packages; and we are looking at specific effects of tax reforms during economic crises. 

Authors’ note: More information on the Tax Policy Reform Database can be found at


Kawano, L and J Slemrod (2016), “How do corporate tax bases change when corporate tax rates change? With implications for the tax rate elasticity of corporate tax revenues”, International Tax and Public Finance 23(3): 401-433.

Amaglobeli, D, V Crispolti, E Dabla-Norris, P Karnane and F Misch (2018), “Tax Policy Measures in Advanced and Emerging Economies: A Novel Database”, IMF Working Paper 18/110.

Dabla-Norris, E, and F Lima (2018), “Macroeconomic Effects of Tax Changes: Evidence from Fiscal Consolidations”, IMF Working Paper, forthcoming.



Topics:  Frontiers of economic research Taxation

Tags:  tax reforms, IMF, tax rates, tax bases

Division Chief, Fiscal Affairs Department, IMF

Deputy Division Chief for the Tax Policy Division of IMF’s Fiscal Affairs Department


CEPR Policy Research