Does Political Instability Lead to Higher Inflation? A Panel Data Analysis
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Contributor Notes

Author(s) E-Mail Address: aaisen@imf.org; fjveiga@eeg.uminho.pt

Economists generally accept the proposition that high inflation rates generate inefficiencies that reduce society's welfare and economic growth. However, determining the causes of the worldwide diversity of inflationary experiences is an important challenge not yet satisfactorily confronted by the profession. Based on a dataset covering around 100 countries for the period 1960-99 and using modern panel data econometric techniques to control for endogeneity, this paper shows that a higher degree of political instability is associated with higher inflation. The paper also draws relevant policy implications for the optimal design of inflation-stabilization programs and of the institutions favorable to price stability.

Abstract

Economists generally accept the proposition that high inflation rates generate inefficiencies that reduce society's welfare and economic growth. However, determining the causes of the worldwide diversity of inflationary experiences is an important challenge not yet satisfactorily confronted by the profession. Based on a dataset covering around 100 countries for the period 1960-99 and using modern panel data econometric techniques to control for endogeneity, this paper shows that a higher degree of political instability is associated with higher inflation. The paper also draws relevant policy implications for the optimal design of inflation-stabilization programs and of the institutions favorable to price stability.

I. Introduction

The main purpose of this paper is to empirically determine the main causes of the worldwide diversity of inflationary experiences, a challenge not yet satisfactorily confronted by the economics profession for two fundamental reasons. First, empirical models explaining inflation in the literature generally fail to account for inflation inertia and for the endogeneity of important economic and political variables affecting inflation. We use generalized method of moments system (GMM) estimation applied to dynamic panel data to address some of the econometric limitations of the ordinary-least-squares (OLS) models previously used in the literature. Second, several political variables used as explanatory variables in earlier studies were relatively poorer measures of political instability than those available in new datasets, such as the Database of Political Institutions (DPI) and the Cross National Time Series (CNTS) Data Archive. The use of these and other data sources, combined with modern econometric techniques, might provide more accurate estimations of the relationships between inflation levels and political instability.

Relying upon the theoretical literature and using a dataset covering around 100 countries for the period 1960–99, we investigate the main economic and political determinants of inflation. After controlling for the countries’ economic structures and for the behavior of economic variables that may influence inflation, we find that political instability leads to higher inflation. Moreover, the impact of political instability on inflation is much stronger for high-inflation than for moderate- and low-inflation countries, and also higher for developing countries than for industrial countries. Additionally, we find that institutions such as economic freedom and democracy are also important determinants of inflation. In particular, higher degrees of economic freedom and democracy are associated with lower inflation.

The paper is structured as follows. A survey of the empirical and theoretical literature on the relationship among inflation, political instability, and institutions is presented in Section II. The dataset and the empirical models are described in Section III. Section IV presents the empirical results, and Section V concludes the paper.

II. Political Instability, Institutions, and Inflation

Most economists acknowledge that differences in monetary and fiscal policies among countries are the main reasons behind the inflation variability they sustain. However, this explanation leads to a much deeper and fundamental question, which is why countries differ on the way they conduct fiscal and monetary policies. One of the many attempts that have been made to answer this question is based on the idea that structural features of a specific economy determine its government’s ability to collect taxes. Chelliah, Baas, and Kelly (1975), for example, provide evidence that countries with larger per capita nonexport income, that are more open to trade, and have larger mining sectors, but smaller agricultural sectors, have, on average, a higher “taxable capacity” or ease of collection. This view implies, among other things, that the countries’ ability to tax is technologically constrained by their stage of development and by the structure of their economies (e.g., size of the agricultural sector in GDP), and as tax collecting costs are high and tax evasion pervasive, countries might use the inflation tax more frequently. One interpretation is that governments in poor countries might find it optimal to rely more heavily on seigniorage instead of output taxes to finance their expenditures. In this connection, the Theory of Optimal Taxation (see Phelps, 1973; Végh, 1989; and Aizenman, 1992), according to which governments optimally equate the marginal cost of the inflation tax with that of output taxes, is consistent with the structural view of the determinants of inflation. Edwards and Tabellini (1991) and Cukierman, Edwards, and Tabellini (1992) fail to find evidence that this theory applies to developing countries. The empirical failure of the Theory of Optimal Taxation motivated the use of theoretical and empirical models focusing on the role played by political and institutional variables.

Cukierman, Edwards, and Tabellini (1992) develop a theoretical model whereby political instability and polarization determine the equilibrium efficiency of the tax system and the resulting combination of tax revenues and seigniorage governments use. They provide evidence to support the model showing that higher degrees of political instability and polarization lead to higher seigniorage revenues. Their measure of political instability is derived from a probit model that attempts to explain the likelihood that an incumbent government would remain in power. In the empirical analysis of Section IV, we employ alternative and more direct measures of political instability affecting seigniorage and inflation; we use variables that count the exact number of government crises or cabinet changes taking place in a particular year. Moreover, whereas they use a dummy variable for democratic regimes, we use the Polity Scale (ranged between -10 and +10) to measure different levels of democracy in different countries.2

Why should a greater number of cabinet changes or government crises lead to higher inflation? Frequent cabinet changes and government crises shorten the horizon of the members of government, since they are not certain that they will keep their posts during an entire term. The higher the probability of being replaced, the greater will the importance attributed to short-term objectives be. Therefore, it is difficult to maintain low inflation.

Paldam (1987) studies the relationship between inflation and political instability in eight Latin American countries. He argues theoretically—and examining the data (but without a formal econometric analysis)—that this relationship works both ways. The main connections from inflation to political instability would be related to the costs of inflation and to the responsibility hypothesis, according to which, people hold governments responsible for economic outcomes. The causality from politics to inflation is primarily related to the demand for public expenditures (which weak governments seldom resist) that are then financed by the inflation tax. Later on, when inflation has risen to high levels, it is much harder for a weak and unstable government to resist the political pressures asking for accommodating policies.

Some authors have also stressed the importance of institutions on economic performance. Acemoglu and others (2002) show that institutions are a very important element explaining volatility, crises, and growth, presenting evidence for a large cross-section of countries. They argue that poor macroeconomic performance is explained by weak institutions, such as the lack of a mechanism to ensure adequate contract enforcement and property rights, which, in turn, give rise to bad macroeconomic policies. We think that countries with weaker institutions not only have lower and more volatile growth but may also present higher inflation. In line with Cukierman, Edwards, and Tabellini (1992), we conjecture that economies with weaker institutions might be unable to build efficient tax systems leading them to use seigniorage more frequently as a source of revenue. In the next sections, in addition to the effects of political instability on inflation and seigniorage, we also estimate the effects of institutions such as economic freedom and democracy on those variables.

III. Data and Empirical Model

The dataset is composed of annual data on political, institutional, and economic variables for 178 countries,3 for the years 1960–99. The sources of political and institutional data are the Cross National Time Series Data Archive (CNTS); the Database of Political Institutions (DPI 3.0);4 the Polity IV dataset;5Gwartney and Lawson (2002);6 and the Freedom House ratings.7 Economic data was collected from the World Bank’s World Development Indicators (WDI) and Global Development Network Growth Database (GDN),8 the International Monetary Fund’s International Financial Statistics (IFS), the Penn World Tables (PWT 6.1), and the Organization for Economic Cooperation and Development (OECD) Statistical Compendium.

The objective of our empirical exercise is to investigate the main political, institutional, and economic determinants of inflation across countries and time. This is done estimating dynamic panel data models for annual inflation levels (taken from the IFS). Since the level of inflation exhibits very high variability, its logarithm was used as our dependent variable. We hypothesize that it depends on the following explanatory variables:

  • lagged logarithm of inflation (IFS);

  • a set of variables representing political instability and institutions:

    • government crises (CNTS), a proxy for political instability, counts the number of rapidly developing situations in a year that threaten to bring the downfall of the present regime, excluding situations of revolt aimed at such overthrow;

    • cabinet changes (CNTS) is the other proxy for political instability—it counts the number of times in a year in which a new premier is named and/or 50 percent of the cabinet posts are occupied by new ministers;

    • index of economic freedom (Gwartney and Lawson, 2002): higher indexes are associated with smaller governments, stronger legal structure and security of property rights, access to sound money, greater freedom to exchange with foreigners, and more flexible regulations of credit, labor, and business; and

    • polity scale (Polity IV): from strongly autocratic (-10) to strongly democratic (10);

  • a set of economic structural variables that reflect characteristics of the countries that may affect their capacity to control inflation:

    • agriculture (in percent of GDP): share of the value added of agriculture in GDP (WDI); and

    • trade (in percent of GDP): openness to trade (WDI); and

  • variables accounting for economic performance and external shocks:

    • growth of real GDP per capita (PWT 6.1);

    • real overvaluation (GDN): real effective overvaluation of the national currency;

    • growth of oil prices (OECD): percentage annual change in oil prices; and

    • U.S. treasury bill rate (IFS): proxy for international interest rates.

Although we consider that high inflation results in most cases from high budget deficits that are monetized, we decided not to include money growth and deficits in our baseline model because, as stated in the previous section, we are searching for deeper determinants of inflation.

The empirical model for inflation levels can be summarized as follows:

Infit=αInfi,t1+Xi,tβ1+Wi,tβ2+vi+εiti=1,,Nt=1,,Ti(1)

where Inf stands for the inflation level of country i at time t, α is a parameter to be estimated, β1 and β2 are vectors of parameters to be estimated, X is a vector of strictly exogenous covariates, W is a vector of endogenous covariates, ν are country specific effects, and ε is the error term.

Substantial complications arise in the estimation of this model using OLS. In both the fixed and random effects settings, the difficulty is that the lagged dependent variable is correlated with the error term, even if we assume that the disturbances are not themselves autocorrelated. Arellano and Bond (1991) develop a GMM estimator that solves the problems referred to above. First differencing (1) removes νi and produces an equation estimable by instrumental variables:

D.InfIt=αD.Infi,t1+D.Xi,tβ1+D.Wi,tβ2+D.εit(2)

where D is the first difference operator and the variables and parameters are defined as in (1). The Arellano-Bond dynamic panel-data estimator uses the following instruments: levels of the dependent variable lagged two and more periods; levels of the endogenous variables lagged two and more periods; and the first differences of the strictly exogenous covariates, which are used as their own instruments.

More moment conditions are available if we assume that the explanatory variables are uncorrelated with the individual effects (see Arellano and Bover, 1995). In this case, lagged differences of these variables and of the dependent variable may also be valid instruments for the levels equation. The estimation then combines the set of moment conditions available for the first differenced equations with the additional moment conditions implied for the levels equation. Blundell and Bond (1998) show that this system GMM estimator is preferable to that of Arellano and Bond (1991) when the dependent variable and/or the independent variables are persistent.

IV. Empirical Results

The first empirical exercise consisted in finding the political, institutional, and economic determinants of inflation levels across countries and time. After dealing with annual inflation, we searched for the main determinants of seigniorage, so that our results could be compared more easily with those of Cukierman, Edwards, and Tabellini (1992).

The estimation results of the model described in the previous section using the method system-GMM for linear dynamic panel data models are shown in Table 1. The dependent variable is the first difference (D1) of Log (Inflation) and the explanatory variables are in first differences as well. Each estimated coefficient indicates the percentage change in the inflation rate that results from one unit change in the respective explanatory variable.9

Table 1.

Results for Yearly Inflation 1/

article image
Source: System-GMM estimations for dynamic panel-data models (using Stata 8.2).Notes: (i) Two-step results using robust standard errors corrected for finite samples (using Windmeijer’s, 2000, correction); (ii) Sargan tests never reject the validity of the over-identifying restrictions; and (iii) significance level at which the null hypothesis is rejected: a=1 percent; b=5 percent, and c=10 percent.

Columns 1 to 4 report the results of estimations performed for the full sample, while Column 5 reports the results obtained when using a sample including only developing countries; t-statistics are in parenthesis.

D1 stands for first difference and LD1 for one-time lagged first difference.

All explanatory variables described in the previous section were included in the estimation reported in Column 1. Considering that the high correlation of the index of economic freedom with the polity scale, agriculture (in percent of GDP), and trade (in percent of GDP),10 may lead to problems of collinearity between independent variables, that index was not included in the model of Column 2.11 Since agriculture (in percent of GDP) was not statistically significant, it was not included in the estimation reported in Column 3. The number of cabinet changes that occur within a year was used in the model of Column 4 instead of government crises. Finally, Column 5 reports the results of the estimation of the model of Column 3 only for developing countries.

The results reported in Table 1 confirm the hypothesis that political instability leads to higher inflation, and show that the effects are sizeable: an additional government crisis increases the inflation rate by 16.1 percent (Column 3),12 and a cabinet change leads to an increase of 9.1 percent (Column 4). The effect of a government crisis is even higher when only developing countries are considered (increases inflation by 20.2 percent, Column 5). Economic freedom also has important effects on inflation: a move of one point up the scale (towards greater freedom) reduces the inflation rate by roughly 25 percent (Column 1). Democracy has a small impact on inflation, as an additional point in the polity scale reduces the inflation rate by only 0.5 percent to 0.7 percent (Columns 3 and 4), and this variable is not statistically significant when only developing countries are considered (Column 5). Concerning the economic variables, trade (in percent of GDP), real overvaluation, and the growth of oil prices have relatively small impacts on inflation rates. Nevertheless, they have the expected signs: greater openness to trade and real overvaluation of the currency decrease inflation, while higher oil prices increase it. The marginal effects of the growth of real GDP per capita and of the U.S. Treasury Bill rate are higher: when the treasury bill rate goes up by one percentage point, the inflation rate increases by roughly 3 percent; and when the growth rate of real GDP per capita is one point higher, inflation decreases by at least 2 percent.13

Columns 1 and 3 of Table 2 show the results of estimations in which government crises and cabinet changes are interacted with dummy variables accounting for inflation below and above 50 percent. The interactions with inflation below 50 percent are not statistically significant, while those with inflation ≥50 are highly statistically significant and have positive signs.14 It is also worth noting that the estimated coefficients for these interactions are much higher than those found in Table 1: when the inflation rate is high or very high, an additional government crisis increases it by 84.5 percent (see Column 1), and an additional cabinet change increases it by 97.3 percent (see Column 3). Thus, the effect of political instability on inflation is much stronger in situations of high or very high inflation than for the entire sample. The same applies to the growth rate of M2, which was used instead of inflation in the model of Column 5.

Table 2.

Results Using Interaction Variables 1/

article image
Source: System-GMM estimations for dynamic panel-data models (using Stata 8.2).Notes: (i) Two-step results using robust standard errors corrected for finite samples (using Windmeijer’s, 2000, correction); (ii) Sargan tests never reject the validity of the over-identifying restrictions; and (iii) significance level at which the null hypothesis is rejected: a=1 percent; b=5 percent, and c=10 percent.

Columns 1 to 4 report the results of estimations performed for the full sample, while Column 5 reports the results obtained when using a sample including only developing countries; t-statistics are in parenthesis.

D1 stands for first difference and LD1 for one-time lagged first difference.

In Columns 2 and 4, respectively, government crises and cabinet changes were interacted with dummy variables representing industrial and developing countries. Results are in line with the ones described above. The interactions with industrial countries are not statistically significant while those with developing countries are. Thus, the positive relationship between political instability and inflation found in Table 1 is true essentially for developing countries.15

Excessive money growth leading to high inflation is generally caused by attempts to extract large seigniorage revenues. According to Cukierman, Edwards, and Tabellini (1992) countries with a more unstable and polarized political system will have more inefficient tax structures and, thus, will rely more heavily on seigniorage. We test the hypothesis that greater political instability increases seigniorage in the estimations whose results are reported in Table 3. In order to make our results more easily comparable to those of Cukierman, Edwards, and Tabellini (1992), we also define seigniorage as the ratio of the change in reserve money (IFS, line 14) to total government revenues (IFS, line 81).16 The explanatory variables are those used in the estimations for inflation. Since the first lag of seigniorage is not statistically significant when included, we have a static panel data model that can be estimated by the within groups (fixed effects) estimator without incurring in problems of inconsistency.17 In order to account for the possible endogeneity of cabinet changes, government crises, growth of real GDP per capita, and real overvaluation, these variables were lagged one period.

Table 3.

Results for Seigniorage 1/

article image
Notes: (i) Panel regressions controlling for fixed effects (using TSP 4.5); (ii) seigniorage, the dependent variable, was defined as the ratio of the change in reserve money (IFS, line 14) to total government revenues (IFS, line 81); and (iii) significance level at which the null hypothesis is rejected: a=1percent, b=5 percent, and c=10 percent.

Models estimated with a constant; t-statistics are in parenthesis.

Results regarding political instability are very similar to those obtained for inflation: cabinet changes and government crises lead to higher seigniorage (Columns 1 and 2); these effects are statistically significant when inflation is high or very high, but not when it is below 50 percent (Columns 3 and 5); and this link between political instability is present in developing countries but not in industrialized ones (Columns 4 and 6). Democracy does not seem to affect seigniorage, as the polity scale is not statistically significant. Regarding economic variables, a larger agricultural sector, real overvaluation, lower oil prices and higher treasury bill rates are associated with higher seigniorage. Although real overvaluation and growth of oil prices do not have the expected signs, the estimated coefficients are rather small and only marginally significant for the latter. That is, their effects on seigniorage are negligible.18

V. Conclusions

Using the dynamic panel data system-GMM estimator and the within-groups (fixed-effects) estimator on a sample covering around 100 countries analyzed for the period 1960–99, this paper finds that a higher degree of political instability, measured using several political and institutional variables, generates higher inflation rates and seigniorage. Higher numbers of cabinet changes or government crises measure not only political instability but also economic policy variability, since every new cabinet that takes over power might have a new set of preferences regarding inflation and unemployment levels. In addition, since every new government is inserted in a very unstable political and institutional environment, it is also very likely to be removed in a short period. These perverse mechanisms greatly affect the way governments conduct monetary and fiscal policies, generating higher inflation and seigniorage. We have also shown that the mechanisms indicated previously are more pervasive and stronger in developing and, especially, in high-inflation (above 50 percent) countries than in developed and low-inflation countries.

The results above are in line with those obtained by Cukierman, Edwards, and Tabellini (1992) regarding the positive relationship among political instability, inflation, and seigniorage. Given the costs in terms of economic growth and welfare generated by high inflation levels and volatility, we believe that this is an important contribution, not only for positive economics, but also in a normative way. Policy makers in developing countries should be aware that it is essential to reform institutions and create viable mechanisms conducive to long-run price stability. Besides, inflation-stabilization efforts may be only temporarily effective if they do not include serious fiscal and political reforms. Our results imply that reforms aimed at reducing political instability and increasing economic freedom and democracy would surely help reduce inflation.

References

  • Acemoglu, Daron, Simon Johnson, James Robinson, and Yunyong Thaicharoen, 2002, “Institutional Causes, Macroeconomic Symptoms: Volatility, Crises and Growth.” NBER Working Paper No. 9124, (Cambridge, Massachusetts: National Bureau of Economic Research..

    • Search Google Scholar
    • Export Citation
  • Aizenman, Joshua, 1992, “Competitive Externalities and the Optimal Seignoriage,” Journal of Money, Credit and Banking, Vol. 24, No. 1, pp. 6171.

    • Search Google Scholar
    • Export Citation
  • Arellano, Manuel, and Stephen Bond, 1991, “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations,” Review of Economic Studies, Vol. 58, pp. 27797.

    • Search Google Scholar
    • Export Citation
  • Arellano, Manuel, and Olympia Bover, 1995, “Another Look at the Instrumental Variable Estimation of Error-Component Models,” Journal of Econometrics, Vol. 68, pp 2951.

    • Search Google Scholar
    • Export Citation
  • Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer, and Patrick Walsh, 2001, “New Tools in Comparative Political Economy: The Database of Political Institutions,” World Bank Economic Review, Vol. 15, No. 1, pp. 16576

    • Search Google Scholar
    • Export Citation
  • Blundell, Richard, and Stephen Bond, 1998, “Initial Conditions and Moment Restrictions in Dynamic Panel Data Models,” Journal of Econometrics, Vol. 87, pp. 11543.

    • Search Google Scholar
    • Export Citation
  • Chelliah, Raja, Hessel Baas, and Margaret Kelly, 1975, “Tax Ratios and Tax Effort in Developing Countries 1969–71,” Staff Papers, International Monetary Fund, March, Vol. 22, pp. 187205 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Cukierman, Alex, Sebastian Edwards, and Guido Tabellini, 1992, “Seignioriage and Political Instability,” American Economic Review, Vol. 82, No. 3, pp. 53755.

    • Search Google Scholar
    • Export Citation
  • Edwards, Sebastian, and Guido Tabellini, 1991, “Explaining Fiscal Policy and Inflation in Developing Countries,” Journal of International Money and Finance, Vol. 10, pp. S16S48.

    • Search Google Scholar
    • Export Citation
  • Gwartney, James, and Robert Lawson, 2002, Economic Freedom of the World, 2002 Annual Report (Vancouver: Fraser Institute).

  • Paldam, Martin, 1987, “Inflation and Political Instability in Eight Latin American Countries 1946–83,” Public Choice Vol. 52, pp. 14368.

    • Search Google Scholar
    • Export Citation
  • Phelps, Edmund, 1973, “Inflation in the Theory of Public Finance,” Swedish Journal of Economics Vol. 75, pp. 6782.

  • Végh, Carlos, 1989, “Government Spending and Inflationary Finance: A Public Finance Approach,” Staff Papers, International Monetary Fund, Vol. 36, pp. 65777.

    • Search Google Scholar
    • Export Citation
  • Windmeijer, Frank, 2000, “A Finite Sample Correction for the Variance of Linear Two-Step GMM Estimators,” Working Paper Series N° W00/19 (London: Institute of Fiscal Studies).

    • Search Google Scholar
    • Export Citation
1

The authors wish to thank Carlos Végh, Henry Chappell, and three anonymous referees for very helpful comments. Francisco Veiga also wishes to express his gratitude for the financial support of the Portuguese Foundation for Science and Technology (FCT), under research grant POCTI/32491/ECO/2000 (partially funded by Fondo Europeo de Desarrollo Regional (FEDER)), and acknowledge the research assistance of Helena Fernandes. The corresponding author is Francisco José Veiga, Departamento de Economia, Escola de Economia e Gestão, Universidade do Minho, P-4710-057 Braga, Portugal. Tel.: +351-253604534; Fax: +351-253676375; E-mail: fjveiga@eeg.uminho.pt.

2

Three additional shortcomings of the analysis in Cukierman, Edwards, and Tabellini (1992) are the presence of endogeneity in some explanatory variables, the absence of explanatory variables accounting for inflation inertia, and the use of a cross-sectional dataset using averages from 1971 to 1982 for only 79 countries. We use System-GMM estimation applied to dynamic panel data covering the period 1960–99 with annual data for around 100 countries. This methodology allows us to fully address the above-mentioned shortcomings and to use additional information provided by the changes of the different variables over time to account for the developments in inflation in each country.

3

Missing values for some variables reduce the number of countries to at most 97 in the estimations for inflation and to 66 in those for seigniorage.

4

On this database, see Beck and others (2001). Available on the Internet through Philip Keefer’s page in the World Bank’s site (http://www.worldbank.org/research/bios/pkeefer.htm).

6

Available on the Internet (http://www.freetheworld.com/release.html). This report presents data on the index of economic freedom and its components for the years 1970, 1975, 1980, 1985, 1990, 1995, and 2000. In order to avoid a great number of missing values in our sample, straight-line interpolation was used to generate annual data.

7

Available on the Internet (http://www.freedomhouse.org/ratings/).

9

Since cabinet changes, government crises, growth of real GDP per capita, and real overvaluation can be affected by inflation, they were treated as endogenous. As done for lagged inflation, their lagged values two and three periods were used as instruments in the first difference equations and their once lagged first differences were used as instruments in the levels equation.

10

These correlations are, respectively, 51 percent, -56 percent, and 40 percent. The complete correlation matrix is available from the authors upon request.

11

Several changes in results occur: government crises becomes more significant; the polity scale changes sign and becomes statistically significant; agriculture (in percent of GDP) changes sign; trade (in percent of GDP) becomes statistically significant; and the U.S. treasury bill rate, becomes more significant.

12

Thus, if the inflation rate is at its sample mean of 51.98 percent, a government crisis will push it to 60.35 percent, that is, the inflation rate will increase by 8.37 percentage points.

13

A series of robustness tests not shown here were also performed. These consisted in adding more variables to the model of Column 3 of Table 1 or in replacing some variables for reasonable alternatives. We found that the following changes lead to lower inflation: greater executive constraints, more political rights, and more civil liberties (when each of these variables replaces the polity scale); and higher real GDP growth (when used instead of growth of real GDP per capita). Proxies for ideological polarization, urbanization, currency inside banks, GDP growth of main trading partners, the exchange rate regime, and central bank independence, were not statistically significant. We also estimated models in which a dummy variable for each year/decade or region was included in order to control for timespecific or region specific effects. Then, we performed estimations for alternative samples: first, excluding extreme values of inflation (annual rates above 1000 percent); and, then, excluding Latin America. Results (available from the authors upon request) were very similar to those shown in Table 1.

14

Similar results are obtained when the Index of Political Cohesion (DPI, variable Ipcoh), or the fractionalization ratio (DPI, variable Frac) are used instead of government crises or cabinet changes.

15

We also investigated the main determinants of inflation volatility in a series of estimations whose results are not shown here, but are available from the authors upon request. We performed several within groups (fixed effects) estimations for a panel of the logarithm of standard deviations of inflation for three-year periods. The main findings were that inflation becomes more volatile at higher levels and that greater political instability, less economic freedom, greater ideological polarization, and greater fragmentation of the parties’ shares in parliament lead to higher inflation volatility.

16

Results are practically the same when we define seigniorage as the ratio of the change in reserve money (IFS, line 14) to nominal GDP (IFS, line 99b). Results are available from the authors upon request.

17

Hausmann tests indicate that the fixed effects specification is preferable to a random effects model and to a simple OLS model.

18

When included, the index of economic freedom is highly statistically significant and has a negative sign, as expected, indicating that greater economic freedom leads to lower seigniorage. Robustness tests such as those performed for inflation were conducted, and results remained essentially the same.

Does Political Instability Lead to Higher Inflation? A Panel Data Analysis
Author: Mr. Ari Aisen and Mr. Francisco José Veiga