WPS8720 Policy Research Working Paper 8720 Policy Implications of Non-linear Effects of Tax Changes on Output Samara Gunter Daniel Riera-Crichton Carlos Vegh Guillermo Vuletin Latin America and the Caribbean Region Office of the Chief Economist January 2019 Policy Research Working Paper 8720 Abstract An earlier paper titled “Non-linear effects of tax changes on change in the tax rate increase. This companion paper first output: The role of the initial level of taxation,” estimated shows that these findings have important policy implica- tax multipliers using (i) a novel dataset on value-added tions, given that the initial level of taxes varies greatly across taxes for 51 countries (21 industrial and 30 developing) countries and thus so will the potential output effect of for the period 1970–2014, and (ii) the so-called narra- changing tax rates. The paper then turns to some specific tive approach developed by Romer and Romer (2010) to policy applications. It focuses on the relevance of the argu- properly identify exogenous tax changes. The main finding ments for revenue mobilization in countries with low levels is that, in line with existing theoretical distortionary and of provision of public goods and social and infrastructure disincentive-based arguments, the effect of tax changes on gaps, as well as in commodity-dependent countries. The output is highly non-linear. The tax multiplier is essentially paper then considers some practical implications for the zero under relatively low/moderate initial tax rate levels standard debt sustainability analysis. Lastly, it evaluates the and more negative as the initial tax rate and the size of the implications of the findings for the Laffer curve. This paper is a product of the Office of the Chief Economist, Latin America and the Caribbean Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/ research. The authors may be contacted at gvuletin@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Policy Implications of Non-linear E¤ects of Tax Changes on Output Samara Gunter Daniel Riera-Crichton Colby College World Bank Carlos Vegh Guillermo Vuletin World Bank World Bank JEL Classi…cation: E32, E62, H20. Keywords: tax multiplier, tax policy, tax rate, value-added tax, non-linear, narrative. We w o u ld like to th a n k se m in a r p a rtic ip a nts a t th e Inte rn a tio n a l M o n e ta ry Fu n d , Inte r-A m e ric a n D e ve lo p m e nt B a n k , Wo rld B a n k , Fe d e ra l R e se rve B o a rd , E u ro p e a n S ta b ility M e ch a n ism , C e ntra l B a n k o f A rg e ntin a , C e ntra l B a n k o f C h ile , C e ntra l B a n k o f S p a in , U N E c o n o m ic C o m m issio n fo r L a tin A m e ric a a n d th e C a rib b e a n , G e o rg e Wa sh in g to n U n ive rsity, J o h n s H o p k in s U n ive rsity, W illia m s C o l- le g e , D av id so n C o lle g e , U n ive rsid a d N a c io n a l d e L a P la ta , S a o P a u lo S ch o o l o f E c o n o m ic s-G e tu lio Va rg a s Fo u n d a tio n , E sc o la S u p e rio r d’A d m in istra c ió i D ire c c ió d ’E m p re se s (E S A D E ), In stitu te o f E d u c a tio n a n d R e se a rch (IN S P E R ), Inte rn a tio n a l M a c ro Wo rk sh o p -R ID G E , L a tin A m e ric a n a n d C a rib b e a n E c o n o m ic A sso c ia tio n (L A C E A ), A n nu a l S y m p o siu m o f th e S p a n ish E c o n o m ic A sso c ia tio n , a n d X L IV M e e tin g o f th e N e tw o rk o f C e ntra l B a n k s a n d F in a n c e M in istrie s-ID B fo r m a ny h e lp fu l c o m m e nts a n d su g g e stio n s. We w o u ld a lso like to th a n k A lb e rto A le sin a , S ilv ia A lb riz io , L e o p o ld o A ve lla n , Fra n k B o h n , Fe rn a n d o B ro n e r, E d u a rd o C ava llo , J av ie r G a rc ia -C ic c o , A ito r E rc e , D av id e Fu rc e ri, V ito r G a sp a r, A le ja n d ro Iz q u ie rd o , H e rm a n K a m il, G ra c ie la K a m in sk y, A a rt K ra ay, G e ra rd o L ic a n d ro , A le ssa n d ro N o ta rp ie tro , P e te r M o ntie l, E d u a rd o M o ro n , Ila n N oy, P a b lo O tto n e llo , P e te r P e d ro n i, J av ie r P e re z , R o b e rto R a m o s, D av id R o b in so n , D ie g o S a rav ia , O le n a S tave le y -O ’C a rro ll, J ay S h a m b a u g h , H a m ilto n Tave ra s, Te re sa Te r-M in a ssia n , a n d M a rtin U rib e fo r h e lp fu l d isc u s- sio n s, a n d J o sé A n d ré e C a m a re n a Fo n se c a , D ie g o Frie d h e im , P a b lo H e rn a n d o -K a m in sk y, a n d L u is M o ra n o fo r e x c e lle nt re se a rch a ssista n c e . 1 Introduction In Gunter et al. (2018), we estimate tax multipliers using (i) a novel dataset on value-added taxes for 51 countries (21 industrial and 30 developing) for the period 1970-2014 and (ii) the narrative approach developed by Romer and Romer (2010) to properly identify exogenous tax changes. The main empirical …nding is that there are strong non-linear e¤ects of taxation on output. In particular, and in line with existing theoretical distortionary and disincentive-based arguments, we show that the e¤ect of tax changes on output is essentially zero for relatively low initial tax rate levels or small tax changes and becomes increasingly negative under higher initial tax rates or larger tax changes. The reason is that the distortion imposed by taxation on economic activity is directly, and non-linearly, related to both the initial level of the tax rate and the size of the tax change. Figure 1 reproduces a key result from our paper. This …gure reports the estimated tax multiplier after two years, evaluated at alternative initial tax rate levels and di¤erent sizes of the tax change. We can see that the most negative multipliers occur for the highest values in both axes. In other words, while the fall in output associated with increasing revenues by $1 tends to be zero for low levels of the initial tax rate and small tax hikes, this cost becomes much larger as the initial tax rate and the size of the tax hike increase. Hence, the evidence shows that the output e¤ect of tax changes is, indeed, highly non-linear. Figure 1. Role of initial tax rate level and size of tax rate change Notes: Dark blue represents a statistically zero tax multiplier. Source: Gunter, Riera-Crichton, Vegh, and Vuletin (2018). 2 These …ndings have important policy implications given that the initial level of taxes varies greatly across countries and thus so will the potential output e¤ect of changing tax rates. Figure 2 shows that, given countries’current VAT rate, the tax multiplier could be statistically zero (dark blue color), or moderate to high (yellow, orange, and red colors). For example, a 1.5-percentage-point increase in the VAT rate would essentially not a¤ect GDP in countries with low tax rates such as Angola, Costa Rica, Guatemala, Ecuador, Nigeria, and Paraguay. In contrast, the same tax increase (decrease) would cause output to fall (increase) in economies with relatively high VAT rates, including some emerging markets like Argentina and Uruguay and, especially, many industrial European countries. For example, according to our estimates, the 1-percentage-point increase that took place in Greece in June 2016 (when the VAT rate changed from 23 to 24 percent) should, in and on itself, have reduced GDP by about 1.75 percent by June 2018. In contrast, the 2-percentage-point increase that took place in Ecuador in January 2000 (when the VAT changed from 10 to 12 percent) should not, in and on itself, have a¤ected GDP. Figure 2. Tax multipliers for countries around the world Notes: Tax multipliers are calculated based on a 1.5-percentage-point change in the VAT rate. Light blue indicates statistically zero tax multipliers. Source: Gunter, Riera-Crichton, Vegh, and Vuletin (2018). This companion paper discusses some relevant policy applications that result from the non-linear e¤ects of tax changes on output. We …rst focus on the relevance of our arguments for revenue mobilization in countries with low levels of provision of public goods and social and infrastructure gaps (Section 2) as well as in commodity-dependent countries (Section 3). Section 4 then explores some practical implications for debt sustainability analysis (DSA). Section 5 analyzes the non-linear implications in terms of the La¤er curve. Finally, Section 6 o¤ers some concluding remarks. 3 2 Policy implications I: On the relationship between the size of the government and economic development The role and the size of the government in an economy have been studied from di¤erent perspectives in the literature. Some of the main determinants of government spending as a proportion of GDP include trade openness (Rodrik, 1998), country size (Alesina and Wacziarg, 1998), degree of economic development (Wagner, 1883, 1893; Easterly and Rebelo, 1993), political organization (Persson and Tabellini, 1999; Milesi-Ferretti, Perotti, and Rostagno, 2002), and business cycle volatility (Fatas and Mihov, 2001). Within this broad set of theories, one that has received considerable attention focuses on the relation- ship between the size of the government and the degree of economic development. In fact, one of the best-established empirical regularities in public …nance is the existence of a positive relationship between the size of government spending (relative to GDP) and real GDP per capita. This so-called s Law, also known as the law of increasing state activities, is named after the German econo- Wagner’ s empirical analysis of Western Europe at the end of the 19th century. Wagner mist Adolph Wagner’ (1883, 1893) pointed out that, as national income rises, public spending tends to increase, both at the extensive margin (i.e., new activities are undertaken) and at the intensive margin (i.e., existing activities are performed on a larger scale). Speci…cally, Wagner argued that the public sector would take over and expand administrative, regulatory, and protective activities previously performed by the private sector because, as nations develop, they face increased complexity in terms of legal and administrative frameworks. Wagner also predicted the expansion of public expenditures related to culture and welfare based on the presumption that, as income rises, citizens will increase the demand for services, such as education, public health, old age pension, and other social protection programs. In fact, these types of goods and services are viewed as luxury goods (i.e., the income elasticity of demand exceeds unity) and have, in principle, more characteristics of public goods than private goods, in which case public provision seems to be the norm. Finally, Wagner was of the view that govern- ment intervention would be required to manage and …nance natural monopolies and ensure the smooth operation of market forces. While not uncontested, these arguments were later re…ned by, among others, Peacock and Wiseman (1967), Musgrave (1969), and Bird (1971). s law and a large, and more recent, body of empirical Figure 3 shows that, as predicted by Wagner’ evidence, there is indeed a very strong association between GDP per capita and the size of government spending (relative to GDP).1 Based on this framework, countries above (below) the …tted line are 1 Most of the recent empirical evidence regarding the existence of Wagner’ s law is based on country-speci…c analyses as opposed to cross-sectional analyses (like the one shown in Figure 3). While, strictly speaking, a cross-sectional analysis is technically less accurate (as it does not control for other country determinants that may be constant over time), this more 4 countries with a size of government spending larger (smaller) than that of a typical country with the same level of income per capita.2 For example, Honduras – a lower-middle-income country, based on the World Bank income classi…cation, with a GDP per capita of $4,785 – has a ratio of public s –an upper-middle-income country with a GDP per spending to GDP of 23 percent, while Jamaica’ capita of $8,528 –is 27 percent. However, Costa Rica –an upper-middle-income country with a GDP per capita of $14,471 (i.e., a per capita income 3 times as high and 70 percent larger than that of Honduras and Jamaica, respectively) –has a ratio of public spending to GDP of just 18 percent. Figure 3. Relationship between GDP per capita and size of government spending relative to GDP Notes: Data correspond to the year 2015. Total number of countries in the sample is 107. Sources: Data for government spending over GDP ratio from WEO-IMF. Data for GDPpc from Penn World Tables. Interestingly, de…ning “excess”spending as the ratio G/GDP minus the predicted ratio G/GDP from the …tted line in Figure 3, it follows, from individual country examples in Figure 3 and more system- atic evidence from 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). global cross-section analytical framework should allow us to draw some relevant policy insights in a more transparent way. 2 For comparison purposes, GDP per capita is PPP-based. Figures are in constant 2011 PPP dollars. 5 Figure 4. Relationship between “excess” spending and VAT rate Notes: Data correspond to the year 2015. Total number of countries in the sample is 107. Sources: Data for government spending over GDP ratio from WEO-IMF. Data for GDPpc from Penn World Tables. This evidence suggests that countries such as Guatemala, with a low level of provision of public goods for its degree of development, may reach a more typical level of provision of public goods by collecting more revenues from increases in the VAT rate with little e¤ect on economic activity. Note that this analysis is quite conservative regarding the positive e¤ect of these …scal changes since it does not include the potential output e¤ect of higher government spending (associated with larger …scal revenues). In fact, in the 2016 Article IV Consultation for Guatemala, the IMF argued that [w]hile the …rst-best option is to quickly mobilize revenue to cover the originally planned budget spending in 2016, a temporary relaxation of the overall de…cit (to its historical average of 2 percent of GDP) could be justi…ed to prevent further cuts in social and capital spending in the event of revenue shortfalls. The acute problem of extreme poverty and malnutrition as well as the dire infrastructure and security needs could warrant maintaining a de…cit at 2 percent of GDP, or even modestly higher, over the medium term to allow the time needed for revenue mobilization. This would not jeopardize …scal sustainability (Annex III) and be consistent with current implementation capacity...[Moreover,] tax policy measures are also needed... Reform of tax and customs administration is a key priority, both to raise revenue and reduce corruption. 6 This view about the importance of revenue mobilization in countries with low levels of provision of public goods and social and infrastructure gaps is, indeed, in line with the evidence presented in Figures 3 and 4 In a similar vein, and mainly due to the need to reduce the …scal de…cit, Costa Rica’s …scal authorities sent to Congress in 2016 a draft law to increase the VAT rate by one percentage point in 2016 (i.e., from 13 to 14 percent) and another one in 2017 (from 14 to 15 percent). 3 Policy implications II: On revenue mobilization in commodity- dependent countries Since the fall in commodity prices in 2014, most commodity-rich countries (especially those whose …scal revenue structure depends “excessively” upon commodity revenues) have been facing a growing …scal challenge as the price fall has increasingly been perceived as more permanent than temporary in nature. Indeed, Figure 5 shows, for a sample of 55 countries for which commodity revenue data were available, that the share of commodity revenues (as percentage of total revenues) tracks very well that of commodity prices.3 Since commodity prices are expected to remain low in the years to come, so will commodity revenues. This, in turn, shows the serious challenge faced by commodity-rich countries and particularly by those with high commodity revenue dependency. s dependency on commodities (proxied by the Figure 6 shows the association between the economy’ commodity GDP as a percentage of total GDP) and the commodity revenue dependency (proxied by commodity revenues as a percentage of total revenues). Not surprisingly, the …tted line shows that the higher is commodity GDP, the larger is the share of commodity revenues in total revenues.4 For example, in a country like Kuwait, where oil production represents about 60 percent of total GDP, oil revenues constitute about 90 percent of total revenues. In contrast, in the Philippines, oil production represents less than 1 percent in both GDP and …scal revenues. One can view countries above the …tted line as countries whose …scal revenue structure depends “excessively”upon commodity revenues relative to a typical country with the same level of economic dependency. The opposite is true for countries below the …tted line. For example, while in Papua New Guinea commodity production represents about 32 percent of GDP and revenue dependency is about 34 percent, in Nigeria commodity 3 Commodity revenues include income from mining and/or hydrocarbons. The list of countries and commodity cov- erage comprises the following. For both mining and/or hydrocarbons commodities: Australia, Bolivia, Brazil, Canada, Colombia, DRC, Indonesia, Mauritania, Papua New Guinea, and Vietnam. For hydrocarbons commodities: Algeria, Angola, Azerbaijan, Bahrain, Brunei, Cameroon, Chad, Congo Republic, Ecuador, Equatorial Guinea, Iran, Iraq, Ivory Coast, Kazakhstan, Kuwait, Libya, Malaysia, Mexico, Myanmar, Namibia, Niger, Nigeria, Norway, Oman, Philippines, Qatar, Russia, Saudi Arabia, Sudan, Syria, Timor-Leste, Trinidad and Tobago, United Arab Emirates, United King- dom, Uzbekistan, Venezuela, and Yemen. For mining commodities: Botswana, Chile, Ghana, Guinea, Guyana, Kyrgyz Republic, Lesotho, Mongolia, Peru, Sierra Leone, Tanzania, and Zambia. 4 Unlike Figure 3, Figure 6 suggests that a non-linear estimation would not increase the predictive power relative to the linear estimation. 7 production represents about 18 percent of GDP (i.e., almost half of that of Papua New Guinea) and revenue dependency is about 77 percent (i.e., more than twice as that observed in Papua New Guinea). Figure 5. Relationship between …scal commodity revenues and commodity prices Notes: Total number of countries in the sample is 55. Price of minerals includes copper, aluminum, iron, ore, tin, nickel, zinc, lead, and uranium price indices. Price of oil is simple average of three spot prices (Dated Brent, West Texas Intermediate, and the Dubai Fateh). Sources: Data for commodity revenues (as % of total revenues) based on authors’ computations using IMF and IDB information. Price of minerals and oil from IMF, PMETA, and POILAPSP. Figure 6. Relationship between commodity GDP (as % of GDP) and commodity revenues (as % of total revenues) Notes: Data correspond to the period 2010-2013. Total number of countries in the sample is 55. Sources: Data for commodity revenues (as % of total revenues) and commodity GDP (as % of GDP) is based on authors’ computations, using IMF and IDB information. 8 Interestingly, de…ning “excess”commodity revenue dependency as commodity revenue (as percentage of GDP) minus the predicted commodity revenue (as percentage of GDP) from the …tted line of Figure 6, it follows, from individual country examples in Figure 6 and more systematic evidence from Figure 7, that countries with “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 7. Relationship between “excess” commodity revenue dependency and VAT rate Notes: Data correspond to the period 2010-2013. Total number of countries in the sample is 55. Sources: Data for commodity revenues (as % of total revenues) and commodity GDP (as % of GDP) based on authors’computations, using IMF and IDB information. This evidence suggests that countries, such as Nigeria, with “excessive” dependency on commodity revenues could quickly mobilize revenues from non-commodity related activities by increasing their VAT rates with relatively little e¤ect on economic activity. This evidence is in line with the IMF s Article IV Consultation, which “urged a gradual increase in the VAT rate, further 2016 Nigeria’ improvements in revenue administration, and a broadening of the tax base.” It should be noted that s largest economies and has one of the lowest VAT rates in the world at 5 Nigeria is one of Africa’ s VAT rate is much lower than that of other members of the Economic percent. In particular, Nigeria’ Community of West African States (ECOWAS). The average VAT rate in the ECOWAS (excluding Nigeria) is 16.5 percent, with rates between 10 and 20 percent. In fact, several proposals have been submitted to the Nigerian Congress to deal with the new normal of low oil prices, with new VAT rate 9 proposals ranging from 7.5 to 10 percent. This view about the importance of revenue mobilization in countries whose …scal revenue structure depends “excessively” on commodities is in line with the evidence presented in Figures 6 and 7. Indeed, several commodity-dependent economies including Nigeria, Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and United Arab Emirates (with no VAT or GST taxes) are planning to increase/implement a VAT rate. 4 Policy implications III: Debt sustainability analysis (DSA) This section explores some practical implications of the non-linear e¤ect of tax changes on output for DSA. The most basic equation that characterizes the evolution of public debt dynamics at a horizon h in DSA is given by5 1+ dt+h = dt+h 1 (rt+h gt+h ); (1) 1 + t+h where d, r, and g denote public debt, revenue, and primary spending, respectively, expressed as a percentage of GDP, and is the real interest rate paid on public debt, typically assumed to be constant. Notice that ; the growth rate of output, is assumed to be exogenous. While is sometimes allowed to change over time (to take into account the debt implications of di¤erent growth paths), such a path is set exogenously and does not capture the endogenous response to, for example, a change in tax policy. Given our …ndings regarding the non-linear response of output to tax changes, equation (1) should, ideally, internalize the endogenous response of output. Naturally, the same would be true of any other …scal shock such as, for example, an increase or decrease in government spending. While the DSA could, in principle, also endogenize the response to any other type of shock, these endogeneity considerations seem more obvious when dealing with variables that enter directly equation (1). To analyze the e¤ects of tax rate changes on output in the context of the DSA, we will rely on a simple, back-of-the-envelope type of analysis that enriches equation (1) by including the e¤ect of tax changes not only on revenue collection (i.e., r), but also on output growth (i.e., ). To characterize the e¤ect of changes in tax rates in a simple way, we assume that Rt = tax ratet Yt ; (2) where R and Y are revenues and output, respectively. Moreover, we will assume that the growth rate 5 This basic approach does not take into account the currency of denomination of debt (e.g., Diaz-Alvarado, Izquierdo, and Panizza, 2004) or the uncertainty of the debt-to-GDP ratio by exploiting the variance and covariance among debt determinants (e.g., Arizala, Castro, Cavallo, and Powell, 2008). 10 at horizon h depends on the change in tax rate at h = 0. Speci…cally, @ t+h t+h = + ; (3) @ (tax ratet ) where is an exogenous source of change in growth and the second term on the right-hand side of (3) captures the e¤ect on economic growth at horizon h of the change in the tax rate h periods before. For example, when evaluating t+h at h = 0, one would be endogenizing the impact e¤ect of a tax change on economic growth. We will now resort to equations (1)-(3) and use a set of initial conditions for the variables involved to evaluate the e¤ect of alternative approaches to DSA. Speci…cally, we will assume an initial stock of public debt of 50 percent of GDP, an exogenous growth rate of 2 percent (i.e., = 0:02), a real interest rate of 3 percent, and a primary de…cit of 2 percent. When evaluating the e¤ects of tax changes on growth and debt-to-GDP ratio, we will focus on the e¤ects, after 2 years, of an increase in the VAT rate of 1.5 percentage points (as in Figure 2). Based on equation (1), Figure 8 shows, as a benchmark, that if there were no change in the VAT rate, the debt-to-GDP ratio would increase by 5 percentage points (i.e., from 50 to 55 percent) after 2 years. Naturally, in this …rst benchmark scenario, there would be no change in output. In the second scenario, we evaluate the e¤ects of a 1.5-percentage-point tax hike on the debt-to-GDP ratio, without considering its e¤ect on economic activity (the typical assumption in DSA). Given equation (2), this would imply an increase in revenues (expressed as a percentage of GDP) of 1.5 percent and, as a result, a reduction of 1.5 percent in the 2 percent primary de…cit (i.e., the new post-tax hike primary de…cit equals 0.5 percent). Such a reduction in the primary …scal de…cit lowers the increase in debt/GDP by 3 percentage points. In other words, the debt-to-GDP ratio increases from 50 to 52 percent (under the tax hike scenario) compared to 55 percent (under the no tax change scenario). In the third scenario, economic activity is allowed to be a¤ected by the tax hike in a linear way (i.e., s (2018) estimations, according to conventional linear estimation strategies). Based on Gunter et al ’ output falls by a total of 0.9 percent after 2 years. This fall in economic activity increases the debt- to-GDP ratio which, in turn, partially o¤sets the positive e¤ect on debt reduction of the fall in the primary …scal de…cit. Speci…cally, as indicated in Figure 8, the debt-to-GDP ratio increases by 2.4 percentage points in this case, compared to 2.0 in Scenario 2. But this begs the question: is this the proper way of endogenizing the role of tax changes on output in the DSA, given the non-linear e¤ect of tax changes on output? Certainly not. To illustrate this, scenario 4 in Figure 8 compares two cases depending on the initial VAT rate: (i) the initial VAT rate is set at a low level (12 percent) and (ii) the initial VAT rate is set at a high level (24 percent). The di¤erence between the two cases is 11 striking. When the initial VAT rate is low, the growth e¤ect is virtually zero (i.e., output falls after two years by just 0.1 percent). Hence, the debt-to-GDP ratio increases by 2 percentage points in this case, the same as in scenario 2. In contrast, when the initial VAT rate is high, output falls after two years by 2.8 percent. This drop in output largely o¤sets the positive e¤ect in debt reduction of the fall in the primary …scal de…cit. Speci…cally, the rise of 1.5 percentage points in the VAT rate (from 24 to 25.5 percent) reduces the debt-to-GDP ratio by less than 2 percentage points compared to the case of “no change in VAT” (scenario 1). Figure 8. Change in debt-to-GDP ratio and cumulative output change under di¤erent model assumptions (two years after the tax shock) Source: Authors’computations. These back-of-the-envelope calculations clearly suggest that endogenizing the e¤ect of tax changes on output may have important implications for debt sustainability, particularly when the e¤ect on output is non-linear. 5 Policy implications IV: La¤er curve The La¤er curve is a theory developed by supply-side economist Arthur La¤er. His main point is that the relationship between tax rates and tax revenues is non-linear and that the size of the tax base is a¤ected by the response of economic agents, depending on the level of the initial tax rate. In line with our non-linear arguments, and a very large tradition in public …nance that builds on the role of distortionary arguments, La¤er argued that when tax rates are “low enough”, economic activity is 12 not much distorted to begin with and that, in such a context, higher tax rates should be associated with larger tax revenues. Why? Because the size of the tax base does not decrease (increase) much in response to an increase (decrease) in tax rates. In contrast, when tax rates are “high enough” (which, in this supply-side type of arguments, is a proxy for high levels of distortion), further increases (decreases) in the tax rate could actually decrease (increase) revenues. This sort of counterintuitive association between tax rates and revenues re‡ects, based on supply-side arguments, the large response of economic agents to high initial levels of distortion. Relying on our non-linear …ndings, Figure 9 illustrates that, indeed, this seems to be the case for the VAT. Considering, as in Figures 2 and 8, an increase in the VAT of 1.5 percentage points, the solid red line shows the corresponding increase in revenues as a percentage of GDP for di¤erent initial values of the VAT rate. Figure 9. Relation between initial VAT rate and change in revenues as percentage of GDP in response to an increase in VAT rate of 1.5 percentage points Notes: The solid red line indicates the change in revenues (as percentage of GDP) in response to a 1.5-percentage-point increase in the VAT rate, taking into account the non-linear …ndings and assuming (without loss of generality) that revenues are given by the VAT rate times GDP. Black dots re‡ect the number of countries for each initial level of VAT rate (as of November 2017). Speci…cally, the size of the dots indicates the number of countries that, as of November 2017, have each level of VAT rate. Small-sized dots indicate VAT rates present in one to 6 countries, medium-sized dots indicate VAT rates present in 7 to 11 countries, and large-sized dots indicate VAT rates present in 18 to 24 countries. The total number of countries having a VAT rate is 172. The median (and average) VAT rate as of November 2017 is 16 percent, the minimum VAT rate (for countries having such a tax) is 5 percent, and the maximum VAT rate is 27 percent (for only one country). Source: Authors’computations. 13 For su¢ ciently low initial levels of the VAT rate, the increase in revenues (as a percentage of GDP) is around 1.5 percentage points (re‡ecting the fact that the tax base does not respond much). In contrast, the increase in revenues falls with higher initial levels of the VAT rate. In fact, for VAT rates higher than 32 percent, an increase in the VAT rate would actually decrease revenues as the response of the tax base more than o¤sets the direct e¤ect of the tax rate change. Figure 9 also shows the number of countries that, as of November 2017, have di¤erent VAT rates as well as their implied location in the estimated relationship. For a total of 172 countries with VAT taxation, the size of the dots re‡ects the number of countries having di¤erent current VAT rates. Small-sized dots indicate VAT rates present in 1 to 6 countries, medium-sized dots indicate VAT rates present in 7 to 11 countries, and large-sized dots indicate VAT rates present in 18 to 24 countries. The median and average VAT rate, 16 percent, is far from the VAT rate of 32 percent. While the minimum VAT rate is 5 percent, the maximum VAT rate is 27 percent (for only one country). In other words, no country in the world has a VAT rate above 32 percent, indicating that no country could reduce the VAT rate and still increase the corresponding revenue collection. In sum, while the La¤er curve is theoretically plausible in our framework, in practice no country is on the “wrong”side of the La¤er curve. This evidence suggests that countries like, for example, Nigeria, with “excessive” dependency on commodity …scal revenues could quickly mobilize revenues from non-commodity related activities by increasing their VAT rates with relatively little e¤ect on economic activity. 6 Concluding remarks Based on our paper titled “Non-linear e¤ects of tax changes on output: The role of the initial level of taxation,” this companion paper discusses some relevant policy implications. We …rst focused on the relevance of our arguments for revenue mobilization in countries with low levels of provision of public goods and social and infrastructure gaps (measured as deviations from the predicted value s law). Since such countries have typically low initial tax rates, we concluded that based on Wagner’ they could increase the VAT rate, and thus tax collection, at little or no cost and thus raise productive government spending. We then focused on commodity-dependent countries and showed that countries with “excessive” de- pendency on commodity revenues tend to have a relative low VAT rate or none at all. Hence, as in the previous case, these countries could increase the VAT tax rate at little or no cost and thus reduce their dependency on commodity revenues. Thirdly, we explored some practical implications of our empirical …ndings for debt sustainability analysis. In particular, we showed that, when initial VAT 14 rates are high, a tax hike is much less e¤ective in preventing a rise in public debt if account is taken of the negative e¤ect of higher VAT rates on output. Finally, we showed the implications for the La¤er curve of the non-linear response of output to increases in VAT rates and concluded that no country in the world could actually reduce the VAT rate and still increase total revenues. References [1] Alesina, A., and R. Wacziarg (1998). “Openness, country size and government,”Journal of Public Economics, Vol. 69, Issue 3, pp. 305-321. [2] Arizala, F., C. Castro, E. Cavallo, and A. Powell (2008). “Debt sustainability fan charts: Com- bining multivariate regression analysis and external forecasts,” Draft Working Paper (IDB). [3] Bird, R. (1971). “Wagner’s‘law’of expanding state activity,” Public Finance, Vol. 26, No. 1, pp. 1-26. [4] Diaz-Alvarado, C., A. Izquierdo, and U. Panizza (2004). “Fiscal sustainability in emerging market countries with an application to Ecuador,”Research Department Working Paper No. 511 (IDB). [5] Easterly, W., and S. Rebelo (1993). “Fiscal policy and economic growth: An empirical investi- gation,” Journal of Monetary Economics, Vol. 32, Issue 3, pp. 417-458. [6] Fatas, A., and I. Mihov (2001). “Government size and automatic stabilizers: International and intranational evidence,” Journal of International Economics, Vol. 55, Issue 1, pp. 3-28. [7] Gunter, S., D. Riera-Crichton, C. Vegh, and G. Vuletin (2018). “Non-linear e¤ects of tax changes on output: The role of the initial level of taxation,” Policy Research Working Paper No. 8668 (World Bank). [8] Milesi-Ferretti, G., R. Perotti, and M. Rostagno (2002). “Electoral systems and public spending,” The Quarterly Journal of Economics, Vol. 117, Issue 2, pp. 609-657. [9] Musgrave, R. (1969). Fiscal systems (New Haven: Yale University Press). [10] Peacock, A., and J. Wiseman (1967). The growth of public expenditure in the United Kingdom (London: Allen and Unwin). [11] Persson, T., and G. Tabellini (1999). “The size and scope of government: Comparative politics with rational politicians,” European Economic Review, Vol. 43, Issues 4-6, pp. 699-735. 15 [12] Rodrik, D. (1998). “Why do more open economies have bigger governments?,”Journal of Political Economy, Vol. 106, No. 5, pp. 997-1032. [13] Romer, C., and D. Romer (2010). “The macroeconomic e¤ects of tax changes: Estimates based on a new measure of …scal shocks,” American Economic Review, Vol. 100, No. 3, pp. 763-801. [14] Wagner, A. (1883). Finanzwissenschaft (Leipzig: C.F. Winter). [15] Wagner, A. (1893). Grundlegung der politischen oekonomie (Leipzig: C.F. Winter). 16