Skip to main content

Distributional Consequences of Monetary Policy in Emerging Economies: Dollarization, Domestic Inflation, and Income Divergence

Abstract

This paper explores the distributional effects of monetary policy in the context of a small open economy. Emerging markets are structurally different from developed economies. They are generally associated with greater financial frictions, underdeveloped financial markets, as well as both a high average level of dollarized assets and unequal access to them, among others. Thus, distributional effects of monetary policy in emerging economies require additional specifications. In particular, I show that wealthy households (represented by the top 10% of the income distribution), who are more able to save in foreign currencies, gain in purchasing power of their incomes by hedging against domestic inflation. At the same time, the poor households (represented by the bottom fifty percent of the income distribution) retain a larger share of liquid assets denominated in domestic currency, thus experiencing a greater burden of local currency inflation. I also show that contractionary monetary policy is associated with periods of higher income inequality in emerging markets that is likely to exacerbate the damaging impact of inflation on the bottom groups of the income distribution.

Introduction

It is well documented that income and wealth inequality have been dramatically increasing in the last several decades, even as poverty levels in major parts of the developing world have substantially fallen (Milanovic 2016; Saez et al. 2018). The causes of this rise of inequality are diverse, though not completely understood. However, it is clear that they partly result from deep underlying structures of inequality embedded in the global architecture of international finance. An important distinction between soft and hard currencies is a central characteristic of these structures of inequality. “Hard currencies,” in contrast to “soft currencies,” can easily be used to purchase goods, services, and assets from practically anyone in the world. Soft currencies can be used primarily only in the country of issue and must be typically exchanged for hard currencies for buying foreign-owned goods and services (Epstein 2018). This configuration leads to a significant institutional disadvantage and specific distributional schemes of developing and emerging economies in the global economic order.

Academic literature has long been indicating the issues of potential negative effects from an increase in financial dollarization and the conduct of domestic monetary policy. As such, domestic financial intermediation in many emerging economies allows for simultaneous circulation of two types of currencies: domestic and foreign. In fact, households in the developing world hold substantial amounts of foreign currency (or “financial dollarization”). That makes emerging markets exposed to external macroeconomic shocks and often prone to subsequent volatile waves of inflationary pressures. According to Levy-Yeyati (2006), despite the fact that monetary authorities in developing countries have succeeded in bringing down domestic inflation over time, the degree of deposit dollarization of banking system remains relatively high.

Hence, most medium and large developing economies historically have developed a combination of monetary policy mix consisting of inflation targeting (with flexible exchange rates) and countercyclical exchange rate intervention due to various experiences with dominant currencies in their respective regions (Levy-Yeyati 2019). The author also notes that whereas floating exchange rates help buffer the adverse external shocks, the same channel would be contractionary in heavily dollarized economies, which would be better off with more rigid exchange rate arrangements. Additionally, Acosta-Ormaechea and Coble (2011) found that traditional interest rate channel is effective in less dollarized economies, while in relatively highly dollarized economies, the exchange rate is more relevant in the transmission of monetary policy. As a result, savings accounts denominated in foreign currency can serve as a hedge against domestic income fluctuations. Moreover, the foreign currency deposits gain in value when inflation gradually erodes the purchasing power of local currency incomes. Thus, the concept of dollarization is an integral part in the long evolution of monetary policy in developing and emerging markets and should be included in the analysis of its distributional effects.

In this paper, I conduct empirical tests to analyze the distributional effects of monetary policy regimes in a small open economy framework. We presume that the key diverging mechanism is the presence of heterogeneous households with a differential level of financial dollarization, which are especially relevant for the study of monetary policy in emerging market economies. We follow the logic of the Prasad and Zhang (2015) model with an exogenous fraction of financially excluded (or rule-of-thumb) consumers coexisting with unconstrained households that are able to save in foreign currency and thereby smooth income fluctuations. Poor households divide their labor income between consumption and holding of domestic liquid assets. In emerging economies, this heterogeneous aspect in a household’s portfolio composition is a driving mechanism of differential response to inflation and interest rate changes.

I also show that the periods with relatively high inflation are positively related to the ratio of top incomes in a developing country. The paper presents evidence that increasing average level of inflation is negatively related to the share of the lowest incomes of population. I specifically concentrate on the evolution of the top 10% and bottom 50% of income shares as robust representatives of the concept of heterogeneous agents. At the same time, interest rate responses of monetary policy (represented by the hike in domestic short-term interest rates) exacerbate this unequal impact of inflation, further amplifying economic divergence between the top and bottom income shares in these countries.

I find that contractionary monetary policy is associated with periods of higher income inequality in emerging markets. For the empirical exercises, I combine the World Inequality Database, World Bank Data, and World Income Inequality Database on BRICS countries (Brazil, Russia, India, China, South Africa) to establish a benchmark model. Using panel data analysis with country- and time-level fixed effects, results show that inflation is positively related to the top 10% of income shares, and negatively related to the bottom 50% of income of income shares in BRICS countries. Also, I conclude that financial frictions, in addition to heterogeneous agents’ framework, play a key role in the relationship of foreign exchange participation and subsequent income redistribution. This is likely to exacerbate the damaging impact of inflation on the bottom 50% of income distribution, although the long-run disinflation goal might positively affect the low-income groups through reductions in unemployment and asset price changes.

Previously, the analysis of how monetary policy affects the distribution of income and wealth has traditionally focused on advanced economies with relevant theoretical assumptions of complete markets. However, emerging economies are structurally different from developed economies. They are generally associated with greater financial frictions, underdeveloped and incomplete financial markets, a high level of asset dollarization, and overall weak monetary transmission mechanisms. In addition, recent studies indicate that the distribution of foreign currency holdings is highly unequal. This raises new concerns about distributional effects of monetary policy in emerging countries.

Financial dollarization in developing countries has become one of the most commonly observed characteristics in the literature. The share of foreign currency denominated bank deposits to the total deposits is a standard method to detect and quantify the level of dollarization of the financial system in a given country. In the presence of any sort of capital controls, the amount of remittances might also provide an indirect account of dollarization, because it measures the amount of private dollars, flowing into an emerging economy. According to Levy-Yeyati (2006), by the end of 2000, the mean dollarization ratio of all developing countries was 35%. At the same time, Ize and Levy-Yeyati (2003) argue that financial dollarization in developing countries is already beyond the phenomenon of currency substitution, and became more evident in the interest-bearing financial assets that leads to the mechanism of asset substitution. Moreover, monetary asset distribution and financial portfolios that can hedge against inflation are not uniform across individuals, which likely create asymmetric responses to various monetary policy shocks.

Thus, in middle- and low-income emerging countries, high rates of exclusion from formal financial institutions and underdeveloped financial systems severely constrain the ability of households to self-insure against idiosyncratic shocks. In particular, more than half of the population in emerging markets does not have access to the basic financial instruments, such as bank account (Prasad 2015). Most households manage their financial transaction using cash, making them more vulnerable to inflation. Additionally, there are limited financial alternatives available to insure against income shocks for the low-income groups. These factors reduce the capacity of low-income households to increase the level of precautionary savings. Ize and Levy-Yeyati (2003) have shown that large fractions of the population in developing countries, and particularly in Latin America, still save and borrow in foreign currencies. In a recent study by Drenik et al. (2018), the authors show that the likelihood of having assets in foreign currency increases with households’ income. As a result, these features of a small open economy are most likely to aggravate the differential income response and exacerbate the distributional consequences of monetary policy.

When investigating the welfare implications of monetary policy within developing countries, it is crucial to take into account the access to dollarized assets. Accordingly, rich households, who are mostly present in the tradable sector, hold more foreign currency and thus can alleviate the impact of an unexpected increase in domestic inflation, whereas poor households tend to hold more domestic liquid assets and accordingly are more exposed to inflation risks. Reinhart et al. (2003) found that there is a positive relationship between the likelihood of an inflationary past and the degree of dollarization. Similraly, Honohan (2007) documents that, over the period 2000–2004, countries that exhibit foreign currency deposits share of less than 50% were more successful in controlling inflation below 35% per annum from 1990 to 2005. Thus, the distributional effects in emerging economies along domestic inflation and interest rates dimensions institutionally linked to financial dollarization are of primary importance in this study.

The rest of the paper is organized as follows. Section two presents a literature review. Section three explains the main working mechanism of income distribution through financial dollarization and domestic currency inflation. Section four introduces the data and basic descriptive statistics to provide some analytical intuition for the results that I report in our empirical analysis. Next, I discuss the panel data approach of the empirical analysis and present the key results. Section five discusses policy implications and finally section six concludes.

Literature Review

Monetary policy is transmitted to the household sector by exerting three main effects: an income effect, a wealth effect, and a substitution effect. The interaction of these effects with certain dimensions of heterogeneity among households results in certain channels of monetary policy, which, in turn, affect inequality along various lines. Monetary policy may affect income inequality through income composition, earnings heterogeneity, portfolio composition, and unexpected inflation channels, which are comprised in theoretical models within distributional channels of monetary policy (Colciago et al. 2019). Thus, heterogeneous consequences of monetary policy principally depend on how households are distributed along relevant economic and financial dimensions. This paper effectively investigates the foreign and domestic currency with the related income heterogeneity dimensions within the household sector.

Thus, this study mainly principally relies on two previously discovered findings in the literature. The first is the impact of the monetary policy on households in advanced economies, and the second is the heterogeneous agents’ model, introduced mainly in the context of developed economies. For instance, Romer and Romer (1998), in one of the earliest contributions to the study of distributional effects of monetary policy, find a strong positive relationship between average inflation and inequality. The authors suggest that monetary policy aimed at low inflation levels is the most likely to result in improved conditions for the poor in the long run. More recently, in the presence of economies of scale in credit purchases, individuals with higher levels of consumption face lower transaction costs. Hence, the welfare cost of inflation would be substantially higher for low-income groups relative to their high-income counterparts, implying that inflation serves as a nonlinear tax on consumption Erosa and Ventura (2002).

The analysis of the distributive effects of monetary policy has been studied extensively regarding the distribution of households along various economic factors. For instance, Doepke and Schneider (2006) quantitatively assess the effects of surprise inflation in the postwar US economy on wealth redistribution by examining nominal asset positions on debtors and creditors. They report that even modest increases in inflation lead to a sizable increase in the redistribution of income and wealth. As an extension to this work, Doepke et al. (2015) consider the impact of housing price adjustments on aggregate demand. They argue that first-time house purchasing prices are minimally impacted by inflation, but in-demand housing prices respond significantly. Consequently, the result is that changes in housing prices, which largely hurt middle-age households and benefit elderly, rich groups, partially offset the direct redistributive effects of overall higher inflation. Meh and Terajima (2011) conducted an analogous exercise using Canadian household data, resulting in similar findings to Deopke and Schneider (2006), and claim that under inflation-targeting monetary regime, the maturity structure of nominal portfolios factors into the present-day value of gains and losses. Monnin (2014) documents a U-shaped link between long-run inflation and income inequality in ten OECD countries. As a result, low inflation rates are associated with higher-income inequality. As inflation rises, inequality declines, reaches a minimum with an inflation rate of about 13%, and then starts increasing again.

Regarding the additional monetary policy channels and their distributional income and wealth effects, Coibion et al. (2012), using detailed micro-level data on income and consumption, find that contractionary monetary shocks systematically raise inequality in labor earnings, total income, consumption, and total expenditures. Furthermore, monetary policy shocks account for a non-trivial component of the historical cyclical variation in income and consumption inequality in the USA. Moreover, Epstein and Montecino (2015) analyze distributional impacts of unconventional monetary policy. The authors conclude that quantitative easing likely contributed to increased inequality, despite having positive effects on employment and mortgage refinancing.

The second strand of the literature is to link the financial dollarization to the heterogeneity across households. This is of key importance to understand the distributional consequences of monetary policy in emerging markets economies. One of the methods to introduce heterogeneity is to model idiosyncratic labor income shocks, as in Krusell and Smith (2014). Gornemann et al. (2014) further this approach and build a new Keynesian model that features asset market incompleteness, heterogeneity in preferences and skills, a frictional labor market, and sticky prices. A key finding is that, due to monetary shocks, unemployed households tend to lose about four times as much as employed households. Overall, the authors predict that contractionary monetary shocks lead to an increase in all forms of inequality including income, wealth, consumption, and earnings.

When markets are not complete and agents differ in their ability to smooth consumption, their welfare depends on the nature of idiosyncratic shocks. Gali et al. (2004) specify an elegant model to allow for a fraction of consumers who do not borrow or save (non-Ricardian type), but instead follow a simple rule-of-thumb: Each period they consume their current labor income. As a result, the authors argue that when central banks respond to inflation shocks by the changes in the domestic interest rate, the size of the response coefficient should be an increasing function of the weight of rule-of-thumb consumers in the economy. A study by Campbell and Mankiw (1989) examined the fiscal policy consequences of the presence of non-Ricardian behavior among a substantial fraction of households in the USA. The authors estimate that nearly 50% of income accrues to consumers who are unable to smooth their consumption. Overall, in the presence of credit-constrained households, targeting only CPI inflation level is no longer welfare maximizing.

Additional type of heterogeneity, such as precautionary savings motive, implies that inflation increases capital accumulation and serves as a rationale for the observed hump-shaped relationship between inflation and capital accumulation. Algan and Ragot (2010) investigate the influence of inflation on aggregate capital accumulation in a heterogeneous agent, closed-economy model with debt limits. They argue that borrowing-constrained households are not able to rebalance their portfolio when inflation varies, and thus adjust their money holdings differently compared to unconstrained households. Bilbiie (2008) modified the rule-of-thumb consumers framework and studies how the presence of non-asset holders alters the slope of the IS curve and the determinacy of interest rate rules. The author argues that the central bank policy should be pursued with an eye on the aggregate demand side of the economy. Based on these findings, the extent to which agents participate in asset markets would become an important factor of the monetary policy stance. Recent studies by Areosa and Areosa (2016) examine optimal monetary policy by combining unskilled agents without access to the financial system with sticky prices. The authors conclude that a contractionary interest rate shock increases inequality, while inflation and the output gap decline (broadly consistent with Coibion et al. (2012). In addition, Auclert (2017) argues that rich and poor households have different marginal propensities to consume, and pinpoint three channels that amplify the distributional effects in the transmission mechanism of monetary policy to consumption.

The studies outlined above mainly characterize theoretical workings of large closed economies and greatly rely on data from the USA and other developed economies for empirical investigations. However, relative to the experience of developing countries, Prasad and Zhang (2015) stress the importance of identifying the differences such as incomplete and underdeveloped financial markets, low level of financial and credit access, and comparatively weak monetary transmission mechanisms. In a similar manner, Sunel (2018), using disaggregated data of high- and low-income households from Turkey, argues that the declining real interest rates following the recent disinflation period (1996–2014) resulted in a greater loss for low-income households. Since the term deposits can be reduced only to zero due to borrowing constraints, the additional rise in capital has to be equivalent to the fall in consumption. These findings echo the earlier literature by Easterly and Fischer (2001) and Mulligan and Sala-i-Martin (2000) that the poor hold a larger fraction of their financial assets in cash and thus are more vulnerable to domestic inflation.

In this paper, I contribute to the literature by conceptualizing comparative economic dynamics in the developing world in the presence of dollarization feature and hence specific distributional characteristics of monetary policy arising from those structural peculiarities. By considering a small open economy setting with a high level of dollarization in the households’ balance sheet, I argue that factoring in foreign currency denominated assets and liabilities in a domestic currency economy setting plays an important role in the distributional dynamics of monetary policy stance in the context of emerging markets. That would ensure that such important complexity needs to be included in the specification and study of monetary policy transmission in developing countries.

Empirical Analysis

Throughout the empirical analysis, the question I am asking is how do different income groups respond to the changes in domestic inflation and dollarization. In this section, I first outline the logic of income groups divergence and then introduce various data sources that I use for my empirical investigation. Then, I provide a set of descriptive statistics to establish an intuition for the final empirical results. Principally, I explain the panel data two-way fixed effects and panel VAR approaches to discuss the distributional effects of monetary policy, originated particularly from the evolution of inflation, dollarization level, and domestic interest rate changes.

An intuitive mechanism of income redistribution through financial dollarization can be schematically presented using the economic circular flow diagram in Fig. 1. In a simple closed-economy model, households demand various goods and services and supply their labor to firms. Firms produce goods and services and pay wages to workers in return. Now, if we introduce the leakage to the system through the savings mechanism and then inject those funds back to the circular flow in the form of dollarized assets (hard currency), it is clear that some groups of people will not only be hedged against inflation, but will also gain from domestic increases in prices and interest rates via enhanced capital accumulation. As a result, the aggregate demand should not remain unaffected, whereas the redistribution of incomes allows for higher purchasing power of dollarized consumers, while the purchasing power of local currency holders loses its value over time.

Fig. 1
figure 1

Source: Author’s calculations

Mechanism of income distribution through financial dollarization.

Data

For the empirical exercise, I utilize data from the World Inequality Database (WID) and the World Income Inequality Database (WIID). The World Inequality Database aims to overcome the limitation of country-level households surveys, which do not properly account for the very top of the income and wealth distribution. Therefore, the WID incorporates different data, such as national accounts, survey data, fiscal data, and wealth rankings, and merges them into a comparable order (World Inequality Database: Methodology 2018). The key parameter is the concept of national income (NI), rather than traditional measure of gross domestic product (GDP). GDP per capita reflects only aggregate values and is not designed to completely reveal the distribution of income across the various groups across a population. Moreover, the “national income” unit considers the depreciation of capital stock and net foreign income, both of which principally overvalue the basic gross domestic income statistics. The main innovation in these series is a comprehensive representation of the income distribution from the bottom to the very top (top 0.1%) grouping, using concepts consistent with macroeconomic national accounts.

Additionally, the World Income Inequality Database (WIID) collects data on economic inequality from 182 developed, developing, and transitioning economies beginning in the 1960s. The WIID systemizes its database mainly relying on the following resources: OECD Income Distribution Database as a benchmark model, Eurostat as the EU reference source on income distribution and social inclusion, LIS Cross-National Data Center (previously known as Luxembourg Income study), and the World Bank data (World Income Inequality Database: Version 3.4 2017). The data points are combined from the sources mentioned above and also include a variety of standardized income and population concepts, sample sizes, and statistical methods. The latest wave of the WIID (WIID 3.4) specifies the conceptual base for each observation. In line with “national income” concept of the WID, the WIID observations includes income items such as imputed rents for owner-occupied dwellers, imputed income from home production, and general in-kind income. Access to these variables ensures that observations are consistent with other publicly available historical records, a condition that is highly relevant to the context of developing countries assuming a high proportion of informal jobs. For the benchmark model of BRICS countries, I combine income distribution databases from the WID and WIID with a focus on the top 10% (fourth percentile) of income shares and bottom 50% (first two quintiles) of income distribution. I successfully merged these datasets due to methodological similarities in the pre-tax income measurements, which fills the gaps in country-year points of the WID dataset.

Variables on real interest rates, real exchange rates, inflation rates, and output per capita growth rates are computed and combined from the World Bank and BIS databases. For a benchmark model of BRICS countries (Brazil, Russia, India, China, South Africa), and for the consistency purposes, the data cover the years 1970 to 2018 with average annual frequency and are presented in Table 1.

Table 1 Data summary.

Empirical Analysis: Descriptive Statistics and Correlational Analysis

Figure 2 shows the evolution of the top 10% of national income shares in BRICS countries. As we can see from the graph, all five countries experienced a dramatic increase of the of top ten income brackets in the past decade. Also, it is crucial to note that the top 10% of income shares captured 40% or more of the total national income in all countries of interest. Generally, this is regarded as a very unusual path for developing countries and reflected a similar trend existing in the USA. In contrast, Fig. 3 shows that the bottom 50% of income shares declined in a similar manner over the same period and now account for less than 20% of total national income. Without factoring in wealth inequality and capital gains, the two trends in Figs. 2 and 3 display the evolution of income inequality in emerging markets, which continues to rise steeply. The severity of this effect can be compared to current Anglo-Saxon economies, including USA, Britain, Canada, and Australia (Piketty 2014). Accordingly, these trajectories in the development of top incomes in large, emerging economies in three different continents require a comprehensive investigation and analysis.

Fig. 2
figure 2

Source: Wealth Inequality Data, World Income Inequality Database

Top 10% of income shares in BRICS countries: 1970–2018. Note: This figure shows the plot of the dynamics of the top 10% of national income shares in Brazil, Russia, India, China, and South Africa during the period between 1970 and 2018.

Fig. 3
figure 3

Source: Wealth Inequality Data, World Income Inequality Database

Bottom 50% of income shares in BRICS countries: 1970–2018. Note: This figure shows the plot of the dynamics of the bottom 50% of national income shares in Brazil, Russia, India, China, and South Africa during the period between 1970 and 2018.

At the same time, there has been a structural decline in inflation in both advanced and developing countries in the last two decades. According to the country group classification of the International Monetary Fund, the mean annual consumer price index inflation rate in the world has fallen from 16% in 1975–1994 to 6% in 1995–2014 (Sunel 2018). While the magnitude of the decline of disinflation was only 5% in developed countries (from 7 to 2% on average between 1960–1990 and 2014), it has been 40% for emerging and developing economies (from 49% on average in 1979–1995 to 9% in 1996–2014). In addition, acknowledging the possible impact of financial dollarization and the differences in financial holdings between rich and poor households, it is useful to plot the relationship between the top and the bottom income subgroups and the average domestic inflation.

Figures 4 and 5 illustrate the scatter plots between the top 10% and bottom 50% of income shares against average inflation rates. In Fig. 5, top 10% of incomes are positively associated with average inflation rates, whereas in Fig. 6, the bottom 50% of income distribution is negatively correlated with average inflation levels. These descriptive findings enrich the discussion on income, saving, and consumption behaviors of households in emerging economies. Additionally, correlation effects obtained from Figs. 5 and 6 support earlier predictions of disproportionate effect of domestic inflation on the poor and the rich types of agents.

Fig. 4
figure 4

Source: Wealth Inequality Data, World Income Inequality Database, BIS

Top 10% of income shares and inflation in BRICS countries. Note: This figure shows the plot of the top 10% of national income shares and annual inflation rates in Brazil, Russia, India, China, and South Africa during the period between 1970 and 2018.

Fig. 5
figure 5

Source: Wealth Inequality Data, World Income Inequality Database, BIS

Bottom 50% of income shares and inflation in BRICS countries. Note: This figure shows the plot of the bottom 50% of national income shares and annual inflation rates in Brazil, Russia, India, China, and South Africa during the period between 1970 and 2018.

Fig. 6
figure 6

Source: Wealth Inequality Data, World Income Inequality Database

Heterogeneity across the income shares in BRICS countries. Note: This figure presents the distribution of the income shares in Brazil, Russia, India, China, and South Africa during the period between 1970 and 2018.

However, as numerous studies indicate, it is important to acknowledge that the effects of hyperinflation are different and damaging for both groups of incomes and the economy as a whole. In practice, emerging economies often exhibit rapid increases in inflation. In this paper, I abstract from the distributional analysis of hyperinflation and thus, exclude the data points with periods of apparent characterization of hyperinflation (inflation running higher than 200% per year).

Figure 6 (the top 10% of income shares and the bottom 50% of income shares) represents the country-level heterogeneity of income distributions in a comparable context. The boxes indicate the distribution of a specific income group in a given country. Thus, the length of the box displays the interquartile range of the income category, while a line subdividing the box represents the median. Hence, Brazilian data show very rigid movements in both categories, but also extremely high divergence between the two income subgroups of interest, which is substantially higher than the average levels in the sample. Additionally, both China and India have the most consistent observation points and also display similarity in the distribution of top 10% of income shares. However, the bottom group of income distribution in China experience greater divergence and lower average level than in India. It is important to note that following the dissolution of the Soviet Union, Russia experienced a significant rise in both income and wealth inequality. That eventually translated into large swings in the data points and overall explains a rather wide distribution of incomes in Russia. Next, South Africa contains the shortest income series among BRICS countries, but presents the highest level of inequality. In fact, top 10% of income shares in South Africa capture at least 50% of national income, whereas the bottom 50% earn at most 15% of total national income.

Although inflation seems to benefit the top income brackets disproportionally, it is important to keep in mind the effect of the domestic interest rate. Short-term interest rates are designed to respond primarily to inflation, which might eventually lead to the exacerbation of initial shocks. Standard policy reactions of monetary policy in both advanced and developing economies tend to increase the short-term interest rates to counteract increasing inflation. That could additionally lead to an increase in interest-bearing capital income of wealthy households and thus raise the share of incomes in the top 10%, while concurrently decreasing the bottom incomes. Thus, generating shocks to domestic interest rate and tracking the income response of different groups are primary focuses in the next section.

As I have mentioned earlier, we do not have data on the distribution of dollarized assets for our sample countries. However, using the BIS dataset, we can partially show the magnitude of the overall dollarization by depicting that the ratio of total liabilities within the countries, public and private, denominated in dollars. From Fig. 7, we can state that in all five cases, the average ratio of total liabilities denominated in dollars is 65%. That alone indicates a high level of financial dollarization in a country, which ultimately accrues to the top incomes.

Fig. 7
figure 7

Source: BIS

Share of total liabilities denominated in dollars in BRICS countries. Note: This figure shows the plot of the dynamics of the total liabilities denominated in US dollars in Brazil, Russia, India, China, and South Africa during the period between 1970 and 2018

Empirical Analysis: a Panel Data Approach

To empirically investigate the relationship between the distribution of income and monetary policy variables further, I perform panel regression analysis with country and year fixed effects (FE) within our sample from BRICS countries. Taking into account lengthy period of observed time series (1970–2018) and specific historical evolution of monetary policy variables in our set of countries, measuring the magnitude of monetary policy shocks is a complex issue. In particular, projections of inflation and interest rate expectations that are required for such kind of estimations are not available in the earlier periods. Rather, it is imperative to capture the long-run relationship between monetary policy indicators and income inequality that is likely to emerge from our long panel. An important assumption to the panel fixed effects model is that time-invariant characteristics are unique to the unit of observation and should not be correlated with other country characteristics. Each entity is different, and thus, both the entity error term and the constant should not be correlated with the others.

In our model, we specify the following equation:

$$ Y_{et} = \beta_{0} + \beta_{1} X_{1, \, it} + \beta_{2} X_{2, \, it} + \beta_{3} X_{3, \, it} + \beta_{4} X_{4, \, it} + \gamma_{1} E_{1} + \cdots + \gamma_{5} E_{5} + \sigma_{1} T_{1} + \cdots + \sigma_{n} T_{n} + \, u_{t} , $$

where Y(t) is the depended variable and represented by the top 10% of income share (or bottom 50%), X1 is the GDP per capita growth, X2 is the inflation rate, X3 is the real exchange rate, X4 is the interest rate, X5 is the dollarization rate, E(n) is the country dummy, T(n) is the time dummy variable, β(k) is the coefficient of the IVs, γ(n) is the coefficient for country dummies, σ(t) is the coefficient for binary time repressors, and ut is the error term.

Column 1 in Table 2 is a baseline OLS regression of Yt (top 10% of income share) on GDP per capita and inflation rate. This is just a benchmark specification that serves as a reference point to compare the changes that fixed effects model can induce when factoring in other factors. Therefore, without including any control mechanisms, the results from Column 1 show that the inflation variable is positive and significant. In the second specification, I add time-fixed effects and observe that inflation changes sign associated with top bracket of incomes. Including the full set of explanatory variables with both time- and country-level fixed effects, Column 3 controls for all time-invariant differences between the countries, so the estimated coefficients of the fixed effects model should not be biased because of the omitted time-invariant characteristics variables.

Table 2 Top 10% of income shares and monetary policy indicators

The results indicate that the principal two-way fixed effects approach establishes that an increase of average increase in inflation by 1% leads to a disproportional increase of the top 10% of income shares by 0.08 percentage points (Column 3), which is also statistically significant at the 1% significance level. The resulting positive signs and the magnitude of the coefficient suggest an important economic significance of the domestic currency inflation parameter and the top incomes in the developing world. In addition, we include a variable on dollarization, which is positive and significant at the 5% level. Next, I perform the same exercise on the bottom 50% of incomes shares, which can also serve as a robustness check for our earlier exercise.

In Table 3, the bottom 50% of income shares, instead of the top 10%, serves as the dependent variable. According to the hypothesis presented earlier in the paper, the bottom 50% of incomes should respond negatively to inflation, since the poor hold more liquid cash and have a larger exposure to the general increases in prices. Again, Column 1 in Table 3 displays the results of baseline OLS regressions with different control variables. It is important to mention that the coefficients demonstrate negative association between the average inflation levels and the bottom of income distribution.

Table 3 Bottom 50% of income shares and monetary policy indicators

Importantly, Column 3 extends the model by adding country-level fixed effects and time-fixed effects, respectively. Column 3 specification with full set of control variables and two-way fixed effects model displays that an increase of average inflation by 1% leads to a disproportional decrease on the bottom 50% of income shares by 0.03 percentage points, a finding that statistically significant at 1% significance level. Moreover, dollarization parameter is also negatively associated with the lower incomes. These results lend support to the theoretical predictions that poor households save in domestic currency or do not save at all, and in times of macroeconomic shocks are not able to smooth their income fluctuations. The differential response by the top and the bottom of income distribution suggests that positive trend inflation causes these two groups to diverge in developing countries over time.

As a robustness check, I run earlier two-way fixed effects model regressions, but with first and second lags included. The strategy includes estimating the coefficients on lags of average inflation and dollarization and discussing whether the lagged effect of monetary policy variables produces any substantial deviations from our baseline results. This method will indicate the consistency patterns of previously obtained results, which might be otherwise overturned by some unobservable heterogeneous country- of time-related characteristics.

Table 4 contains the main results obtained from the both groups of incomes with a special interest on the first and second lags of inflation. Specification 1 reports country- and time-fixed effects regression outcomes for the top 10% of income shares. Specification 2 reports fixed effects model results for the bottom 50% of income shares. Accordingly, Column 1 produces negligible decline in the first and second lags of inflation. All the coefficients from three specifications remain in the range of positive 0.08 percentage points and are statistically significant, which makes our earlier baseline findings robust. Dollarization rate coefficients also remain positive and significant when adding lagged specification. Column 2 considering lower-income brackets reports even stronger consistency of the lagged estimates of inflation. Both first and second lags of inflation do not show any sizable deviations from the baseline estimates. All measurements continue to demonstrate negative sing and the magnitude of 0.03–0.04 percentage points. Also, dollarization rate coefficients remain negative and significant. Overall, Table 4 findings show that our country- and time-fixed effects model regression outcomes are robust to additional specifications, and underline the fact of divergent responses from the top and bottom incomes shares to both inflation and dollarization levels.

Table 4 Panel regression with inflation and dollarization lags

Empirical Analysis: a Panel VAR Approach

To analyze the long-run relationship between domestic interest rate and its differential effects on the distribution of incomes in developing economies, and as an additional consistency and robustness exercise, I propose to use panel VAR approach. In typical VAR models, all variables are treated as endogenous and interdependent, if not specifically designed as an exogenous parameter. This setup allows us to investigate the impulse response functions (IRFs) of different shocks and the ways it affects other systemic macroeconomic variables. Panel VARs seem particularly suited to address the aim of the paper as they are able to capture both static and dynamic interdependencies, and account for cross-sectional dynamic heterogeneities among other functions (Fabio and Ciccarelli 2013). In this paper, I simulate a shock to domestic inflation, domestic interest rate, and dollarization level to explore the responses of the top 10% and bottom 50% of incomes shares in our sample of developing countries.

Now, I describe the main structure of the panel VAR as follows:

$$ Y_{it} = \, A_{0i} \left( t \right) \, + \, A_{i} \left( l \right)Y_{t - 1} + F_{i} \left( l \right)W_{t - 1} + u_{it} , $$

where Yi,t is the vector of the variables described in the basic identification scheme as Yi,t. Then, A0i(t) contains all the deterministic components of the data (constant terms, seasonal dummies, and deterministic polynomials in time). Ai(l) and Fi(l) are the polynomials in the lag operators, which are heterogeneous across units. Wt−1 represents the vector of exogenous variables. uit are the identically and independently distributed errors. In computing impulse response functions, we identify shocks using a Cholesky decomposition. The data run annually from 1970 to 2018 with all variables measured in levels. However, our estimation avoids the periods of hyperinflation for each country separately, which usually occurs when it exceeds 200% a year.

Then, I run a panel VAR model for a typical emerging market economy in the following order: {GDP per capita growth, inflation, interest rate, exchange rate, dollarization, top 10%/bottom 50% of incomes} with the first lag included. A variable that is higher in the ordering causes contemporaneous changes in subsequent variables. Variables that are lower in the ordering can affect previous variables only with lags. Thus, the interest rate variable appears before income variables, which emphasizes its exogenous nature relative to households’ income—a key assumption of this exercise. This way, we are interested to see the impact of monetary variables on income dynamics and position income brackets in the end of the order. In addition, we perform Fisher-type panel unit-root tests to confirm stationarity of our panel.

Next, I generate a 1-standard deviation shocks to domestic inflation, dollarization, and interest rates and trace out the effect on the top 10% of incomes shares and bottom 50% of income shares correspondingly. Figure 8 demonstrates that an increase in domestic interest rates by a central bank steadily increases the share of top decile of incomes. This finding can be qualitatively interpreted as follows: Rich households will enjoy the higher interest rate return on their financial assets, including the possibility of converting back their foreign currency holdings into domestic currency. These results also can explain a slow increase in top incomes as annual returns take time to accumulate the capital gains.

Fig. 8
figure 8

Inflation rate shocks to the top 10% and the bottom 50% of income shares

At the same time, in Fig. 9, we observe that the income shares of the bottom 50% of population suffer instantaneously, as they get cut off from accessible credit by higher interest rates. Specifically, the baseline VAR analysis demonstrates that a positive shock of one standard deviation to the interest rate increases the share of income held by the top 10% by 0.018 percentage points over a 5-year horizon. Moreover, the same shock to interest rates lowers the share of income held by the bottom 50% by about 0.002 percentage points over a 5-year horizon. Analogously, as depicted in Fig. 10, we simulate a shock to dollarization level, which results in the gain to the top incomes and negatively affects the bottom incomes. Hence, again we claim that the distributional consequences of monetary policy to the extent arise from the heterogeneity in domestic and foreign monetary asset holdings, which eventually translate into different magnitudes of domestic inflation tax. These findings relate and deepen earlier established results of divergent paths of incomes caused by changes in domestic inflation, which generally lead to the widening of income gap in emerging market economies.

Fig. 9
figure 9

Interest rate shock to the top 10% and the bottom 50% of income shares

Fig. 10
figure 10

Dollarization shock to the top 10% and the bottom 50% of income shares

Policy Discussion

Our cross-sectional evidences of contractionary monetary policy suggest greater distributional effects for a small open economy with a high level of financial dollarization. Also, this paper specifically indicates unanticipated escalations in domestic inflation, originated from the exchange rate depreciation episodes or other foreign macroeconomic shocks. Thus, in fact, the more central bank manages the nominal exchange rate, the smoother the relative prices behave—effectively diminishing differential response of both financially excluded and financially included households. However, this mechanism is far different from targeting the exchange rate specifically. Hence, it is important to distinguish between exchange rate targeting policies and frameworks managing the exchange rate fluctuations. As a result, one of the main policy recommendations that should be practically addressed in developing and emerging economies to decrease economic inequality in the long term is a policy of enhanced exchange rate management.

Overall, domestic inflation is thought to reflect a sovereign risk premium for emerging economies (Min 1999; Akitoby and Stratmann 2008). Taking into account a greater spread between inflation in advanced and developing countries, emerging economy borrowing rates in the international markets are indeed higher. The periods of inflation surges due to external shocks in developing countries already start from the relatively higher inflation rates. Therefore, interest rates hikes are not supposed to bring inflation down per se, but bring inflation back to the previous structurally higher levels. This mechanism can partially explain the general ineffectiveness of contractionary monetary policy (and of inflation-targeting framework, in general) in emerging economies to significantly decrease the level of inflation to the macroeconomic benefits of all citizens.

Hence, the demanding policy challenge today is to organize the regimes that expand national policy autonomy within the currency jurisdictions of emerging economies. For that purpose, it is essential to renegotiate the strictures on capital controls in existing bilateral and multination trade and investment agreements. In general, developing countries often supplement monetary policy with capital controls, incurring varying degrees of effectiveness. This approach can reduce capital flows and, correspondingly, reduce exposure to external macroeconomic shocks and domestic currency volatility. When faced with capital controls or other forms of capital acquiring constraint, financially included households tend to spend less of their income on foreign asset purchases. Finally, it is suggestive to develop frameworks for international cooperation and gradually shift toward mutual currency invoicing mechanisms in bilateral financial and trade arrangements between BRICS countries.

Conclusion

When studying the general trend of economic inequality and the organization of monetary policy in emerging economies, it is crucial to take into account the access to dollarized assets and the distribution of foreign currency holdings. Our findings show that controlling for country- and time-fixed effects, we observe divergent relationships of the top and bottom income shares with the domestic inflation and dollarization levels. Central banks in the developing world, and of BRICS economies particularly, should collectively start engaging in greater integration strategies to facilitate the transactions in their national currencies. This will help to strengthen the value of domestic currencies and simultaneously to de-dollarize the highest brackets of their respective income distributions, which will ultimately flow back and start contributing to the benefit of their respective domestic economies.

This paper investigates the distributional consequences of monetary policy in an emerging economy, explicitly originated from the changes in local inflation and short-term domestic interest rates. The outcome leads to a strong divergence of the top and bottom income shares. In this work, I found that modest increases in inflation primarily hurt the bottom of income distribution, since the poor households principally hold more domestic liquid assets. At the same time, rich households, which mainly save in foreign currency-denominated monetary assets, are effectively hedging against domestic inflation as well as protecting the purchasing power of their incomes. This operational mechanism portrays an upward redistribution of incomes, which facilitates greater economic inequality in the context of a developing economy with a high level of financial dollarization.

References

  • Acosta-Ormaechea, Santiago, and David Coble. 2011. Monetary Transmission in Dollarized and Non-Dollarized Economies: The Cases of Chile, New Zealand, Peru and Uruguay. IMF Working Papers WP/11/87.

  • Akitoby, Bernardin, and Thomas Stratmann. 2008. Fiscal policy and financial markets. Economic Journal 118: 1971–1985.

    Article  Google Scholar 

  • Algan, Yann, and Xavier Ragot. 2010. Monetary policy with heterogenous agents and borrowing constraints. Review of Economic Dynamics 13: 295–316.

    Article  Google Scholar 

  • Areosa, Waldyr, and Marta Areosa. 2016. The inequality channel of monetary transmission. Journal of Macroeconomics 48: 214–230.

    Article  Google Scholar 

  • Auclert, Adrien . 2017. Monetary Policy and the Redistribution Channel. NBER Working Paper No. 23451.

  • Bilbiie, Florin. 2008. Limited asset market participation, monetary policy and (inverted) aggregate demand logic. Journal of Economic Theory 140(1): 162–196.

    Article  Google Scholar 

  • Campbell, John, and Gregory Mankiw. 1989. Consumption, income, and interest rates: Reinterpreting the time series evidence. NBER Macroeconomic Annual 4: 185–216.

    Article  Google Scholar 

  • Coibion, Olivier, Yuriy Gorodnichenko, Lorenz Kueng, and John Silvia. 2012. Innocent Bystanders? Monetary Policy and Inequality in the US. NBER Working Paper No. 18170.

  • Colciago, A., A. Samarina, and J. Haan. 2019. Central bank policies and income and wealth inequality: A survey. Journal of Economic Surveys 33(4): 1199–1231.

    Article  Google Scholar 

  • Doepke, Matthias, and Martin Schneider. 2006. Inflation and the distribution of nominal wealth. Journal of Political Economy 114(6): 1069–1097.

    Article  Google Scholar 

  • Doepke, Matthias, Martin Schneider, and Veronika Selezneva. 2015. Distributional Effects of Monetary Policy. Working Paper #14.

  • Drenik, Andres, Gustavo Pereira, and Diego Perez. 2018. Wealth redistribution after exchange rate devaluations. American Economic Association Papers and Proceedings. 108: 552–556.

    Google Scholar 

  • Easterly, William, and Stanley Fischer. 2001. Inflation and the poor. Journal of Money, Credit, and Banking 32: 160–178.

    Article  Google Scholar 

  • Epstein, Gerald. 2018. The Political Economy of International Finance in an Age of Inequality: Soft Currencies, Hard Landings. Northampton, MA: Edward Elgar Publishing.

    Book  Google Scholar 

  • Epstein, Gerald, and Juan Montecino. 2015. Did Quantitative Easing Increase Income Inequality? Institute for New Economic Thinking Working Paper Series No. 28.

  • Erosa, Andres, and Gustavo Ventura. 2002. On inflation as a regressive consumption tax. Journal of Monetary Economics 49: 761–795.

    Article  Google Scholar 

  • Fabio, Canova, and Matteo Ciccarelli. 2013. Panel Vector Autoregressive Models: A Survey. Working Paper Series NO 1507, January.

  • Gali, Jordi, David Lopez-Salido, and Javier Valles. 2004. Rule-of-thumb consumers and the design of interest rate rules. Journal of Money, Credit, and Banking 36(4): 739–763.

    Article  Google Scholar 

  • Gornemann, Nils, Keith Kuester, and Makoto Nakajima. 2014. Doves for the Rich, Hawks for the Poor? Distributional Consequences of Monetary Policy. Working paper, Federal Reserve Bank of Philadelphia.

  • Honohan, Patrick. 2007. Dollarization and Exchange Rate Fluctuations. World Bank Policy Research Working Paper No. 4172.

  • Ize, Alain, and Eduardo Levy-Yeyati. 2003. Financial dollarization. Journal of International Economics 59: 323–347.

    Article  Google Scholar 

  • Krusell, Per, and Anthony Smith. 2014. The redistributive effects of financial deregulation. Journal of Monetary Economics 68: 867–896.

    Google Scholar 

  • Levy-Yeyati, E. 2006. Financial dollarization: Evaluating the consequences. Economic Policy 21: 61–118.

    Google Scholar 

  • Levy-Yeyati, Eduardo. 2019. Exchange rate policies and economic development. In Economic Development.

  • Meh, Cesaire, and Yaz Terajima. 2011. Inflation, nominal portfolios, and wealth redistribution in Canada. Canadian Journal of Economics 44(4): 1369–1402.

    Article  Google Scholar 

  • Milanovic, B. 2016. Global Inequality: A New Approach for the Age of Globalization. Cambridge, MA: Harvard University Press.

    Book  Google Scholar 

  • Min, Hong. 1999. Determinants of Emerging Market Bond Spread: Do Economic Fundamentals Matter? World Bank Policy Research Working Papers.

  • Monnin, Pierre. 2014. Inflation and Income Inequality in Developed Economies. CEP Working Paper 2014/1.

  • Mulligan, Casey, and Xavier Sala-i-martin. 2000. Extensive margins and the demand for money at low interest rates. Journal of Political Economy 108(5): 961–991.

    Article  Google Scholar 

  • Piketty, Thomas. 2014. Capital in the twenty-first century. Cambridge, MA: The Belknap Press of Harvard University Press.

    Book  Google Scholar 

  • Prasad, Eswar. 2015. Distributional Effects of Monetary Policy in Emerging Market Economies. IZA Discussion Paper No. 9272, August.

  • Prasad, Eswar, and Boyang Zhang. 2015. Distributional Effects of Monetary Policy in Emerging Market Economies. NBER Working Paper No. 21471.

  • Reinhart C.M., K.S. Rogoff, and M. Savastano. 2003. Addicted to Dollars. NBER Working Papers Series No. 10015.

  • Romer, Christina, and David Romer. 1998. Monetary policy and the well-being of the poor. NBER Working Paper No. 6793.

  • Saez, E.F.A., G. Zucman, L. Chancel, and T. Piketty. 2018. World Inequality Report 2018. Accessed February 7, 2020. wir2018.wid.world.

  • Sunel, Enel. 2018. Welfare consequences of gradual disinflation in emerging economies. Journal of Money, Credit, and Banking 50: 705–757.

    Article  Google Scholar 

  • 2017. World Income Inequality Database: Version 3.4. UNU-WIDER-UNDP project.

  • 2018. World Inequality Database: Methodology. October 1. Accessed December 7, 2017. https://wid.world/methodology/#library-browse-by-country.

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Zhandos Ybrayev.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This paper is a revised version of Chapter 1 of my PhD dissertation at University of Massachusetts-Amherst. I thank my advisors Gerald Epstein, Michael Ash, Adam Honig, and Shouvik Chakraborty for their constant guidance and support. I also thank the anonymous referee for the detailed revisions and helpful comments.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Ybrayev, Z. Distributional Consequences of Monetary Policy in Emerging Economies: Dollarization, Domestic Inflation, and Income Divergence. Comp Econ Stud 64, 186–210 (2022). https://doi.org/10.1057/s41294-021-00163-2

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1057/s41294-021-00163-2

Keywords

  • Monetary policy
  • Inequality
  • Income distribution
  • Emerging economies

JEL Classification

  • D31
  • E52
  • E63
  • F33
  • O11