Macroprudential policies and income inequality in former transition economies

See, among others, Beck et al., (2007), Roine et al., (2009), Delis et al., (2014) and Brei et al., (2018).

For instance, several studies indicate that an accommodative monetary policy leads to higher income and wealth inequality (e.g., Coibion et al. 2017). However, accommodative monetary policy could also lead to lower inequality through economic growth, employment gains and wage increases (Bivens 2015).

Financial stability is defined in terms of system resistance to extrinsic disturbances (Allen and Wood 2006; Padoa-Schioppa 2003) and/or the resilience of the financial system to disturbances originating from the system (Schinasi 2004).

Systemic risk is defined as the risk of interruption of the provision of financial services caused by the impairment of all or part of the value of the assets of credit institutions. This interruption may have a significant negative impact on the real economy as it deprives the funds needed to finance investment. See also Perotti and Suarez (2009), Borio and Drehmann (2009), and Hanson et al. (2011) for a more detailed discussion on the aim of macroprudential policies.

The theoretical literature on macroprudential policies and income inequality is also scant (see, e.g., Rubio and Carrasco-Gallego 2014; Clerc et al. 2015; Rabitsch and Punzi 2017; and Medicino et al. 2018). See also Colciago et al. (2019) for a review of the theoretical and empirical literature.

There seems to be a strong link between financial liberalization and income inequality. For example, Brei et al. (2018), using data for a panel of 97 economies over the period 1989–2012, find that this relationship is not monotonic. Up to a point, more finance reduces income inequality. Beyond that point, inequality rises if finance is expanded via market-based financing, while it does not when finance grows via bank lending.

For instance, Agenor et al. (2018) found that growth may be promoted by prudential policies whose goal is to counterbalance financial risks to the economy. At the same time, financial openness tends to reduce the growth benefits of these policies, because of either greater opportunity to borrow abroad or increased scope for cross-border leakages in regulation.

We use the four aforementioned inequality measures, provided by the comprehensive dataset of Lahoti et al. (2016), so as to account for different parts of income distribution. We would like to thank a referee for highlighting this point. See McGregor et al. (2019) for a detailed description of these widely used measures of income inequality.

The main determinants of income inequality in Central and Eastern European (CEE) economies can be categorized in the following categories (see Milanovic 1998, 1999; Rosser et al. 2000; Kaasa 2003; Mitra and Yemtsov 2006): (a) economic growth and technological change; (b) macroeconomic factors like inflation, unemployment, the size of government’s expenditure; (c) demographic factors (including human capital); (d) political and transition-related factors, including the size of the public sector and liberalization reforms; and (e) historical and cultural factors. Wan (2002) investigates the possible effect of growth on inequality in transition economies and finds that the rising inequality cannot be explained by standard growth models. In addition, Rosser et al. (2000), using empirical data for 16 transition economies between 1987–1989 and 1993–1994, find that income inequality is positively correlated with the share of output produced in the informal sector of former transition economies.

See Appendix (Table 7) for the list of the other European countries. We present results similar to those in Table 1 for the Atkinson index, the Palma Ratio and Theil’s T index in an online supplement.

For a more detailed description of the macroprudential measures, see Cerutti et al. (2017). In Table 8 in Appendix, we present a detailed description of the set of macroprudential measures used in our analysis.

The nexus between financial stability and inequality has been studied before (e.g., Rajan, 2010; Kumhof et al., 2015; Hauner, 2016), providing evidence that high levels of inequality hamper financial stability.

We use Theil’s T rather than Theil’s L index, with the former being more sensitive to changes in the upper tail of the distribution. We do so since in our study we already have measures which are more sensitive to changes in income at the bottom of the distribution e.g., the Atkinson measure, which has a sensitivity parameter \(\alpha = 2\) (the higher \(\alpha\), the more sensitive the Atkinson index becomes to the bottom of the income distribution).

These indices include measures of inequality discussed in Lahoti et al. (2016). The Gini coefficient is the most widely used measure of inequality, based on the Lorenz curve, a cumulative frequency curve that compares the distribution of income with the uniform distribution representing perfect equality. Theil’s T index belongs to the Generalized Entropy measures of inequality. Atkinson’s index belongs to a class of measures that are based on the aversion to inequality—the inequality aversion parameter has been set equal to 2 in our data. Finally, the Palma ratio is the ratio share of gross national income (GNI) belonging to the richest 10% of the population to the share belonging to the poorest 40%. For a more comprehensive discussion and description of the above measures of inequality, see McGregor et al. (2019) and  Haughton and Khandker (2009) “Handbook on Poverty and Inequality” , Chapter 6.

In our estimations, we drop outliers pertaining to extreme inflation values, as high inflation may distort severely measured inequality: we exclude observations for which the inflation rate is lower than -10% (deflation) and larger than 45%. This decision is based on observations for the specific group of countries we analyze. We did not observe any extreme values in the other variables included in our analysis and therefore kept all other observations in the sample.

This dummy takes value 1 when a country experiences a banking crisis. The empirical literature has not reached a consensus regarding the effect of financial crises on inequality. For instance, de Haan and Sturm (2017) find that the outbreak of banking crises increases inequality within the country. On the other hand, Roine et al. (2009) show that banking crises tend to lower the income shares of the rich and, thus, decrease income inequality.

Countercyclical Capital Buffers (CTC) and Capital Surcharges on Systemically Important Financial Institutions (SIFI) are excluded from our study since they are characterized by small variability, and, therefore, are dropped from all specifications.

We would like to thank an anonymous referee for raising this issue.

In an online supplement, we compare evidence from these estimations to those obtained from system GMM—albeit the later are somewhat problematic.

These results are not explicitly presented here for the sake of brevity. See Tables T.1 to T.72 in Online Appendix.

All the results of cross-sectional dependence tests are available upon request.

The results regarding the other covariates have already been discussed in the previous subsection and are reported in an online supplement for space conservation reasons.

In our data, LTV_CAP is present only in 8.6% of our sample, DTI in 10.5% and overall BOR are present in 13.9% of our sample—see Table 10 in Appendix.

An estimate of the standard error of (2) is given by \(\sqrt {\widehat{{{\text{Var}}\left( {\hat{\beta }_{1} } \right)}} + Z^{2} \widehat{{{\text{Var}}\left( {\hat{\beta }_{3} } \right)}} + 2Z\widehat{{{\text{Cov}}\left( {\hat{\beta }_{1} ,\hat{\beta }_{3} } \right)}}}\) which depends on the values that \(Z\) attains.

A graphical display of these effects for non-Gini inequality measures is provided in an online supplement for space conservation reasons.

See again the online supplement for details.