Based on evidence from the industrial world, and particularly Europe, tax hikes have a significant negative effect on economic activity. The column shows that this empirical finding does not hold for a broader sample. In the developing world, higher taxes may be an effective way to raise revenues without reducing GDP. This is especially true in countries with low provision of public goods or commodity-dependent countries.
Alesina et al. (2015) claimed that “[f]iscal adjustments based upon spending cuts are much less costly, in terms of output losses, than tax-based ones ...”. Research during the past ten years has found large negative tax multipliers. Several individual and multi-country analyses that focus exclusively on industrial economies, mostly European, have built on Romer and Romer’s (2010) narrative approach and found tax multipliers ranging from about −1.7 to −5.1
This seemingly robust evidence has led to strong policy prescriptions in industrial European countries, as reflected in the above quote. We asked whether larger sample, including developing countries, shows large negative tax multipliers.
Non-linear effects of tax changes on output
We followed the narrative approach for a new dataset on value-added taxes for 51 countries (21 industrial and 30 developing) between 1970 and 2014 and, for the entire sample, we found a tax multiplier of −1.7 (Gunter et al. 2018). In industrial European economies, the tax multiplier we found is −3.6 (roughly in the mid-point of the range found in previous research) and statistically significant. But for the rest of the sample, the multiplier is −1.2 and not statistically significant.
Why would the tax multiplier vary so much? Theoretical distortionary and disincentive-based arguments (e.g. Jaimovich and Rebelo 2017) imply the effect of tax changes on output would be highly non-linear, and our results support this. As Figure 1 shows, the tax multiplier is essentially zero if the initial tax rate levels are very low, and becomes more negative as the initial tax rate and the size of the tax change increase.2
Figure 1 Non-linear cumulative tax multipliers after two years
This strongly suggests that the growing consensus pointing to large negative tax multipliers in industrial countries, particularly in Europe (e.g. Alesina et al. 2015), is driven mainly by high initial tax rates. Large negative tax multipliers are not a robust empirical regularity, especially in the developing world.
In a policy-oriented companion paper, we show that this novel finding has important policy implications related to revenue mobilisation, debt sustainability analysis, and the Laffer curve (Gunter et al. 2019).3
Tax multipliers for individual countries
The initial level of taxes varies greatly across countries, so it is natural that the potential output effect of changing tax rates varies too. Figure 2 shows that, given a country's current VAT rate, the tax multiplier could be statistically zero (light blue), moderate to high (yellow, orange, and red).
Figure 2 Tax multipliers after two years
Based on Figure 2, for half the world (88 of 175 countries) the tax multiplier is statistically zero (light blue). For example, tax changes would have virtually no effect on GDP in countries with low tax rates such as Angola, Costa Rica, Ecuador, Guatemala, Nigeria, and Paraguay. In contrast, the same tax increase (or decrease) would cause output to fall (or increase) in countries with relatively high VAT rates, including emerging markets like Argentina and Uruguay and, especially, many industrial European countries.
Revenue mobilisation in countries with low levels of provision of public goods and social and infrastructure gaps
The appropriate role and the size of government has been studied from many perspectives. For example, Wagner’s Law proposes a positive relationship between the size of government spending relative to GDP and real GDP per capita. As national income rises public spending tends to increase, both at extensive (new activities) and intensive (existing activities performed on a larger scale) margins. Figure 3 strongly supports this well-documented empirical regularity.
Figure 3 Relationship between GDP per capita and size of government spending relative to GDP
Countries above (or below) the fitted line are countries with a size of government spending larger (or smaller) than that of a typical country with the same level of income per capita.
If we define excess spending as the ratio G/GDP minus the predicted ratio G/GDP from the fitted line in Figure 3, it follows – from individual country examples in Figure 3 and more systematic evidence in Figure 4 – that countries with positive excess spending (like Honduras, Jamaica, Argentina, and Greece) tend to have higher VAT rates than those with negative excess spending (like Guatemala, Costa Rica, New Zealand, and Australia). This evidence suggests that countries such as Guatemala, with low provision of public goods for its degree of development, may be able to provide a typical level of public goods using revenues from an increase in the VAT rate, with little effect on economic activity.4
Figure 4 Relationship between ‘excess”’ spending and the VAT rate
Have you read?
Revenue mobilisation in commodity-dependent countries
Commodity prices plunged in 2014, and the price fall is increasingly perceived as permanent. So, most commodity-rich countries (especially those whose fiscal revenue structure depends excessively on commodity revenues) face a growing fiscal challenge. Figure 5, unsurprisingly, shows a strong association between an economy’s dependency on commodities (proxied by commodity GDP as a percentage of total GDP) and commodity revenue dependency (proxied by commodity revenues as a percentage of total revenues).
Figure 5 Relationship between commodity GDP (as percentage of GDP) and commodity revenues (as percentage of total revenues)
If we define excess commodity revenue dependency as commodity revenue minus the predicted commodity revenue from the fitted line of Figure 5, it follows – from country examples in Figure 5 and more systematic evidence from Figure 6 – that countries with positive excess commodity revenue dependency (like Malaysia, Yemen, and Nigeria) tend to have lower VAT rates (or no VAT at all, as in the cases of Bahrain, Brunei, and Iraq) than those with negative excess commodity revenue dependency (like Chile, Papua New Guinea, Trinidad and Tobago, and Republic of Congo).
Figure 6 Relationship between ‘excess’ commodity revenue dependency and the VAT rate
This evidence suggests that countries such as Nigeria could quickly mobilise revenues from non-commodity related activities by increasing their VAT rates with relatively little effect on economic activity.