The effect of unconventional monetary policy on income inequality in the Eurozone

(1)

1

The effect of unconventional monetary policy

on income inequality in the Eurozone

L.T. Ruth

Theme: Monetary policy for open economies and

exchange rate fluctuations

Date: 8-8-2017

Abstract :

This thesis uses a dataset of 12 Eurozone countries in the period of 1999 to 2013 to test if the unconventional monetary policy of the ECB has had an effect on income inequality in these countries, analyzing both pre- and post-tax income. The size of the balance sheet of the ECB is used as the measurement for unconventional monetary policy. The regression results show that there is no significant relationship between the unconventional monetary policy of the ECB and income inequality. This paper contributes to the scarce research on the effects of unconventional monetary policy on income inequality.

E-mail: lucruth@gmail.com Student number: 10850740 Supervisor: Kostas Mavromatis Second Reader: Naomi Leefmans

(2)

2 Statement of Originality:

This document is written by Luc Ruth who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

1. Introduction

As a response to the 2008 financial crisis, the ECB reduced key interest rates substantially. In addition, non-standard measures had to be taken to support financing conditions and credit flow to the euro area economy over and beyond what could be achieved through reductions in the key interest rates alone. These measures stabilized the financial system and the economy. In 2010, the sovereign debt crisis prompted the ECB again to take unconventional measures to stabilize markets (European Central bank, 2010). During this time income inequality has risen in most countries in the Eurozone. One cause may be the unconventional monetary policy used in the wake of the crisis. This thesis examines the differing reasons for this increase across Europe, bolstering the research on this subject. The effects of the unconventional monetary policy on income inequality have been barely researched, due to the recentness of the events and data availability. In addition, policy insights are gained. Unconventional monetary policy does not seem to influence income inequality. Policies regarding social expenditure, tax progressivity, unemployment and education are more viable measures to influence income inequality.

To see if unconventional monetary policy has an effect on income inequality, this thesis looks at data from 12 Eurozone countries. The econometric model used is a multivariate regression with country fixed effects, controlling for conventional monetary policy, GDP growth, unemployment, social expenditure, educational attainment of the population, trade openness, government intervention, tax progressivity and unionization. Except for tax progressivity, unemployment and educational attainment, all variables are lagged by one year to account for the time it takes for the variables to have an effect. The Standardized World Income Inequality Database is used, which gives the possibility to measure the effects on both market and net income inequality (Solt, 2016). For robustness, the regression have also been done with lags for tax progressivity and unemployment, as well as with Gini coefficient from the OECD. To measure the unconventional monetary policy of the ECB the balance sheet size of the ECB is used, as due to the policy this has ballooned. For conventional monetary policy the marginal lending facility rate is used. GDP growth, unemployment and trade openness are taken from the World Bank. Unionization and government intervention data come from the ICTWSS. Tax progressivity information, educational attainment and social expenditure data comes from the OECD. Based on the results of the regressions, unconventional monetary policy does not seem to influence both pre- and post-tax income inequality. Policy makers should not use unconventional monetary policy to reduce inequality, nor does it seem to pose a risk of

(5)

5

increasing inequality. For combatting inequality, policies which reduce unemployment, increases in social expenditure and tax progressivity and investments in education seem to be more effective, as these all seem to have a significant effect. However, one has to take into account that the counterfactual outcome of having no unconventional monetary policy may increase inequality.

1.1 Justification

Income inequality is linked to a plethora of problems, for example lower overall happiness (Diener, Kesebir and Shigehiro, 2011), higher homicide rates, incarceration rates and crime (Hsieh and Pugh, 1993; Kaplan et al., 1996; Kennedy et al. 1996; Pickett et all, 2005; Pickett and Wilkinson, 2007), higher rates of obesity (Pickett et al., 2005a), mental health problems (Pickett and Wilkinson, 2007), lower self-rated health (Subramanian and Kawachi, 2006), lower social capital, lower social trust and lower status of women (Kawachi et al. 1997) and lower social mobility (Pickett and Wilkinson, 2007). Finally, higher income inequality has a negative effect on the upbringing of children, leading to lower school scores and higher dropout rates (Kaplan et al., 1996; Pickett and Wilkinson, 2007) and higher teenage births (Gold et al., 2001; Pickett et al., 2005b). While the effects of the unconventional monetary policy are largely unknown with respect to income inequality, it may have significant negative societal effects described above. In addition to the social importance of knowing the determinant of income inequality, not a lot of research has been done yet on the effects of unconventional monetary policy on income inequality. This thesis will therefore bolster the scarce research on this subjects.

(6)

6

2. Literature review

2.1 Channels of influence from conventional monetary policy

In the literature, five channels are mentioned through which conventional monetary policy influences income inequality, summarized by Amaral (2017). The first channel is the inflation

tax channel. Increases in expected inflation disproportionately diminish the purchasing power

of households that rely primarily on cash to conduct their transactions. Empirics show that lower-income households use more cash to purchase goods and services. Erosa and Ventura (2002) find that expected inflation acts as a regressive consumption tax, thus increasing inequality.

The second transmission channel is the savings redistribution channel. When there is an increase in unexpected inflation, it lowers the real value of nominal assets and liabilities. This makes borrowers better off at the expense of lenders, since the real values of nominal debts decreases. To see what effect this has on income inequality, we have to look at the way these assets and their respective maturities are distributed across households. Long-term nominally denominated debt holders would gain, while short-term nominally denominated asset holders would lose. Doepke and Schneider (2006) do just that. They have mapped the assets holdings from the Survey of Consumer Finance into age and wealth categories and study the effect of a sustained unexpected increase in inflation. Their results how that middle-aged, middle-class households would experience the largest net wealth increase. The reason for this is that this group tend to hold long-term nominally denominated debt in the form of fixed-rate mortgages. Conversely, older, richer households lose the most, since they are the net savers with deposits and short-term denominated debt. Expansionary monetary policy would decrease inequality through this channel.

The third channel is the interest rate exposure channel. This relates to the redistribution due to changes in real interest rates. A fall in real interest rates would increase financial asset prices, to the extent that the interest rate used to discount future dividends decreases. However, to see what effect this has on inequality, one has to look again at households’ assets, liabilities and their durations. Net savers which hold primarily short-term duration assets and net borrowers with long maturity debt benefit from expansionary monetary policy, to the extent that it decreases real interest rates. Vice-versa, long-term net asset holders and short-term net debt holders lose in this case (Auclert, 2016).

(7)

7

The fourth channel is the earnings heterogeneity channel, which states that changes in monetary policy have the potential to affect labor earnings heterogeneously, depending on the position of the household in the income distribution. Heathcote, Perri and Violante (2009) conclude that changes in hourly wages primarily affect top earnings, while low earnings are mainly affected by changes in hours worked and the unemployment rate. How monetary policy affect inequality is dependent on how strongly it affects each of these variables. For example, if an expansionary policy reduces unemployment more than it increases hourly wages, inequality decreases. Evidence for this comes from Carpenter and Rodgers (2004), who find that an increase in the federal funds rate disproportionately reduce employment rates of lower-skilled workers and racial minorities, groups that are overrepresented in the lower income brackets. According to this channel, an expansionary monetary policy reduces inequality.

The final channel is the income composition channel. Households get income from different sources, which may respond differently to different sources. Lower income household rely more on transfer income, like unemployment benefits and social security, while median income households will rely more on labor income. Upper income households get relatively more business and capital income. An expansionary monetary policy may result in increased wages and lower unemployment, increasing inequality on the lower end of the distribution, as transfer income does not tend to change much. However, lower interest rates would also mean lower interest income for the higher income group, reducing inequality. Gornemann, Kuester and Nakajima (2012) find that a contractionary monetary policy increases inequality. This is primarily due to the effect on unemployment, as a contractionary shock tends to extend the unemployment spell.

(8)

8

2.2 Effects of unconventional monetary policy on income inequality

For unconventional monetary policy, the effects are less well researched. One way in which unconventional monetary policy could influence income inequality is through the effects of the asset purchases done by unconventional monetary policy. The asset purchases decrease the interest rates and increase the prices of the securities bought and of securities that are substitutable enough, through a portfolio effect. Gagnon et al. (2011) have argued that the LSAP purchase program of the Fed decreased long-term interest rates on a variety of securities, including securities which are not purchased. In addition, Rosa (2012) argued that this program had significant effects on the prices of US assets.

The effects of the LSAP program on income inequality influence in some channels the same way as conventional monetary policy, as both cause a fall in the interest rates. Keeping portfolios constant, households see decreases in debt servicing and interest income. The effect of these decreases is dependent on who is holding interest-sensitive assets and liabilities. Younger household tend to hold more interest-sensitive liabilities, while older households have more interest-sensitive assets. As richer households tend to be older and relatively more dependent on interest income, inequality should drop. However, households also adjust their portfolio over time and firms take advantage of the reduced debt burden and cheaper credit. Aggregate real activity improves as well and affects other types of income like wages, so the final effect is ambiguous (Amaral, 2017). As asset prices are affected, wealth inequality should be impacted a well. How it is impacted depend on the relative price changes in each asset class and the distribution of different types of assets and liabilities across different interest groups.

Domanski, Scatigna and Zabai (2016) analyze surveys of household finances for five European countries as well as the United States and find that the main forces changing wealth inequality since the start of the Great Recession have been changes in equity valuations and changes in the prices of houses. The largest part of all assets in the bottom 80 percent of wealth distribution in the United States is in the form of real estates, while financial assets are relatively more important for the very top of the distribution. This would mean that that increases in real estate prices will be inequality decreasing, while increases in other asset prices will be inequality increasing.

Conversely, poorer households tend to be more leveraged, as they borrow more to finance assets, again primarily in the form of housing. When asset prices change, these households’ net wealth changes more than the wealth of richer, less leveraged households. In addition, net wealth is also affected by the distribution of liabilities and how the changes in

(9)

9

interest rates and prices impact the values of these liabilities. Domanski, Scatigna and Zabai find that these changes in asset prices and interest rates have increased wealth inequality in the United States. However, O’Farrell, Rawdanowicz and Inaba (2016) find insignificant effects. None of these studies test the counterfactual outcome. Bivens (2015) does this and argues that compared to the no stimulus alternative, the LSAP program has reduced inequality significantly, primarily through its effects on output stabilization.

2.3 Empirical results of monetary effects

Coibion et al. (2012) use data from the Consumer Expenditure Survey from between 1980 and 2008 for the USA. The CEX data used is of high quality and allows them to differentiate between different types of inequality. They also distinguish different types of monetary shocks and how these relate to inequality. For sudden contractionary monetary policy, income inequality increases. To explain this, they found that the labor income in the higher percentiles increased, while labor income in the lower percentile groups decreased. In addition, they check for more structural changes monetary policy changes by trying to estimate the Fed’s target inflation rate. Their conclusion remains the same however: contractionary monetary policy leads to increased inequality. However, their data only goes up to 2008, thus leaving out the unconventional monetary policy which took place afterwards. In addition, they did not have wealth data on hand, and gain from wealth had thus be left out.

Saiki and Frost (2014) do look at unconventional monetary policy in Japan. They use a vector auto regression model to estimate these effects. Their conclusion is that through the income composition channel, expansionary monetary policy increases income inequality. They also say that the effect may be even bigger in Europe and the US, where households have an even larger portion of their savings in equities and bonds than in Japan. Davtyan (2016) comes to the same conclusion in her research about the effect of monetary shocks on income inequality in the USA. She found that contractionary monetary policy reduces income inequality. To get to this she uses a VECM approach with the Gini index and the 90-10 percentile approach as measures of income inequality. These results are supported by Albanesi (2007), who finds that there is a correlation between high inflation and higher inequality. This is based on a model where in equilibrium low income household hold more cash as a fraction of total purchases. Expansionary monetary policy would thus lead to more income inequality. Finally, Domanski, Scatigna and Zabai (2016) use household survey data of six advanced economies to look at the recent changes in wealth inequality by looking at the effect on households assets and liabilities.

(10)

10

With this data, they run a simulation to determine what kind of effect the recent monetary policy had on asset prices and interest rates. They conclude that wealth inequality has risen. Interest rates had a negligible impact, but the rise in equity prices was the main driver of this rise. However, a recovery in housing prices has partially offset this effect. Caution has to be taken with these results. The changes in assets and liabilities are not taken into account. In addition, human capital in the form of future labor income and pension rights is not taken into account.

2.4 Other influences on income inequality

Governments are also said to be able to influence income inequality through policy. Rueda points to three ways governments can influence wage inequality: minimum wages, the generosity of the welfare state and through public sector employment (Rueda, 2008). Minimum wages would increase wages at the lower end of the wage distribution, decreasing the difference between high and low wages. However, it is also argued that a higher minimum wage would lead to higher unemployment, in turn increasing inequality. With regard to welfare state generosity, a more generous welfare state would lead to a (higher) reservation wage. Because of this reservation wage, a higher wage is necessary to get people to work, decreasing inequality. Another way to influence inequality more directly is through public sector employment. The government is the biggest employer in industrialized countries. By setting the wages for its employees, the government can influence inequality for its employees and thus a large part of the workforce. It is generally acknowledged that the size of the public sector is related to wage compression, also because the public sector is more sheltered from competitive pressure. With regards to welfare state generosity, a more generous welfare state would redistribute more to the poor and would mean more insurance against labor market risks. The effect of the welfare state on inequality is twofold. First, it reduces peoples dependence on labor income, as they are insured against these risks. Second, the welfare state increases peoples access to public services which allow low income workers to increase their income. Finally, governments influence inequality through taxes. More progressive taxes reduce inequality as higher incomes pay a relatively higher percentage of their income than low wages, compressing the wage distribution.

Doerrenberg and Peichl (2014) find that there is a significant correlation between social expenditure and inequality. Their results are backed up by Niehues (2010) who finds that higher social spending is associated with lower income inequality. In addition, Lee (2005) finds that larger public sectors are correlated with lower post tax income inequality in democracies.

(11)

11

Further positive effects are found by Roine et al. (2009), Cálderon and Chong (2009), while Garciá-Peñalosa and Checchi (2008) find a negative effect.

A more progressive tax system reduces inequality according to Doerrenberg and Peichl (2014), Cooper and Duncan and Sabrianova-Peter (2008). Additionally, higher top marginal tax rates reduce income inequality according to Roine et al (2009). Redistributive policy lowers inequality according to Smeeding and Brandolini (2007), while Garciá-Peñalosa and Checchi (2008) find that a lower tax wedge reduces income inequality.

Globalization is also found to have an effect on income inequality. This is due to increased competition in various sectors of the economy. Low skilled workers are competing with low skilled workers from other countries. Since low skilled workers have become more abundant due to globalization, wages would go down for this group in sectors which are exposed to globalization. For higher skilled workers, the inverse is true. Globalization increases demand for these workers, thus increasing their wages. Both increase the wage differences between the groups and thus income inequality. Atkinson (2003) concludes that this effect is indeed happening for post-tax income inequality (2003). Borjas (1994) finds the same effect, at least for unskilled workers in northern countries. Jaumotte, Lall and Papageorgiou (2008) examine the effects of both trade and financial globalization, where they find that trade globalization reduces inequality while financial globalization increase inequality. Finally, Wren (2008) find that especially low-skilled workers are affected by globalization.

Education is put forward as a mitigating factor in income inequality, as a more educated workforce would reduce the amount of low skilled, low paid workers in a population and increase the amount of high skilled workers. This would increase wages for the former group

and decrease it for the later, due to demand effects. Calderon and Chong find a negative relation

between education and income inequality (Calderón and Chong, 2009). Alderson and Nielson (2001) also find this effect. Iversen and Stephens (2008) find that education has an effect on the income distribution in a country, where the type of human capital regime determines the type of education. They distinguish three regimes of human capital formation: social democratic, liberal and Christian-democratic. The social democratic is characterized by a high level of spending on day care, preschool, primary, secondary, and higher education, active labor market policy and vocational education. It also includes moderate levels of employment protection. This results in high levels of industry- and occupation- specific skills and high levels of general skills, especially at the middle and the lower end of the distribution. The Christian-democratic regime is characterized by high levels of vocational education and employment

(12)

12

protection. It also has medium levels of public spending on primary, secondary and tertiary education and active labor market policy. This results in high levels of firm- and industry specific skills and moderately high levels of general skills at the lower end. Finally, the liberal regime is characterized by low levels of spending on day care and preschool, and moderate spending on primary, secondary and tertiary education. In addition, it has low levels of employment protection, active labor market policy and vocational education. Because of the low levels of public spending, there is substantial private spending on higher education and in

some countries day care. This culminates in low levels of specific skills and in low general

skills at the bottom, but in the top general skills are on par with to those in the social democratic regime. Iversen and Stephens conclude that there is high wage inequality in countries with a liberal regime, high differentiation between secure insiders and insecure outsiders in the Christian-democratic regime and a relative absence of these divisions in the social democratic regime. Higher education attainment would then reduce income inequality.

Finally, unions are said to influence income inequality. According to Rueda (2008), unions may take over government roles when these do not step up to protect employees. The effect of unions on wage inequality depends on which workers are more organized: higher or lower paid workers. When highly paid workers are more organized, wage inequality might

be increased due to these unions. García-Peñalos and Checchi (2008) conclude that labor market

institutions are important for reducing inequality. This is because unions lead to higher unemployment for nonunion workers, which in turn leads to higher inequality. Calderón and Chong (2009) however find that trade union membership has an inequality reducing effect. Rueda (2000) found that for countries where corporatism is high, right wing governments employ more people in the public sector. He also found that when corporatism is high, governments do not significantly influence minimum wages. Lastly, Doerrenberg and Peichl (2014) find a significant negative relationship between union density and inequality.

(13)

13

3. Data and Methodology

3.1 Hypotheses

Taking the previous literature into account, two hypotheses will be stated. Following research from Frost and Saiki and Davtyan, expansionary unconventional monetary policy is expected to increase income inequality. Two types of income inequality are researched in this thesis, pre- and post-tax inequality, therefore the hypotheses will be:

H1: Expansionary unconventional monetary policy increases post-tax income inequality. H2: Expansionary unconventional monetary policy increases pre-tax income inequality.

3.2 Model specification

This research uses a panel regression which includes country fixed effects. The dataset consists of 12 Eurozone countries, being Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, and Spain. Data for all measures is from 1999 up until and including 2013. Not all current Eurozone members are included in the analysis, as some have a very low amount of observations (<5). As not all data is available for every country for every year for all variables, the final dataset is an unbalanced panel. The effect of unconventional monetary policy on income inequality is estimated using the regression model below.

Yi, t = β1 UMPi, t-1 + β2 CMPi, t-1 + β3 Growthi, t-1 + β4 Unemp i, t + β5 SocExpi, t-1

+ β6 TaxProgi, t + β7 Tradei, t + β8 Unionsi, t-1 + β9 Intervi, t-1 + β10 Lowi, t + β11 Highi, t + αi + 𝜀i, t

The left hand side variable captures the Gini coefficient of inequality for a certain country i in

year t. It may constitute market or net inequality. UMPi, t-1 contains the log of the balance sheet

of the ECB, lagged by one year, capturing the effect of its unconventional monetary policy and

is the main explanatory variable of interest. CMPi, t-1 is the policy rate of the ECB and measures

the conventional monetary policy of the ECB lagged by one year. Growth measures the state of

the economy using the GDP growth as a measurement lagged by a year. Unemp i, t captures the

unemployment rate in percentages, while SocExpi, t-1 capture the social expenditure of the

government lagged by year. TaxProgi, t measures the progressivity of the tax policy in a country.

Unemployment and tax progressivity are not lagged as they would have a direct effect in that

year. Tradei, t is a control variable capturing the openness of a country to trade and thus

globalization pressures. Unionsi, t-1 measures the union density in a country lagged by one year,

(14)

14

again lagged by a year. Lowi, t and Highi, t capture the effect of different educational levels of

the population. Finally, αi are the entity-specific intercepts due to the fixed effects regression

(15)

15 3.3 Data description

3.3.1 Dependent variable

The dependent variable encompasses the measure of income inequality. The measure is constructed as a Gini coefficient, ranging from 0 to 1, where 0 is no inequality and 1 is full inequality. The best dataset for both comparability and coverage is het Standardized World Income Inequality Database (SWIID) from Solt. This dataset incorporates data from multiple other databases. These include the United Nations University’s World Income Inequality Database, the OECD Income Distribution Database, the Socio-Economic Database for Latin America and the Caribbean of the CEDLAS and the World Bank, EUROstat, the World Bank’s PovcalNet, the UN Economic Commission for Latin America and the Caribbean, national statistical offices around the world, academic studies. The data provided by the Luxembourg Income study is employed as the standard. It incorporates comparable Gini indices of net and market income inequality for 176 countries from 1960 to 2014. It also includes information about the relative as well as the absolute redistribution. The dataset allows to check two different kinds of inequality: net and market inequality. Market inequality constitutes differences in income before taxes and social transfers, while net inequality measures income differences after these transfers. The Solt dataset achieves its extensive coverage by using imputation to replace missing data. For this thesis, Gini coefficients are gathered by extracting their values after 100 imputations from this dataset and regressing the explanatory variables on these values.

3.3.2 Independent variables

The measure for conventional monetary policy will be the policy rate of the ECB in percentages of the marginal lending facility, following Joyce, Miles, Scott and Vayanos (2012). Data is taken from the European Central Bank and is available for 1999 to 2016.

The measure for unconventional monetary policy will be the balance sheet of the ECB, as this ballooned during this period. It includes the total assets of the central bank as a percentage of GDP, again following Joyce, Miles, Scott and Vayanos (2012). Data is taken from the European Central Bank. Data is available for 1999 to 2014.

3.3.3 Control variables

GDP growth is included as due to higher economic activity financial income may

increase, as well as wages overall, both influencing income inequality. Data is taken from the World Bank. It constitutes the annual percentage growth rate of GDP at market prices based on constant local currency. GDP is the sum of gross value added by all resident producers in the

(16)

16

economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. The data is available for 1999 up until 2015 for all countries in the dataset.

The unemployment rate is another source of income inequality, as higher unemployment would lead to higher income inequality if one income group would be disproportionally laid off, lowering their income vis-à-vis other income groups. The data is mainly taken from Eurostat. Unemployment rate is the number of unemployed people as a percentage of the labor force, where the labor force consists of the unemployed plus those in paid or self-employment. Unemployed people are those who report that they are without work, that they are available for work and that they have taken active steps to find work in the last four weeks. The data is available for 1999 to 2016.

Social expenditure influences primarily net income inequality, as higher social

expenditure constitutes a higher amount of transfers and thus more redistribution from higher to lower income groups. data is gained from the OECD. It is operationalized as social expenditure per capita to account for differences in population size between countries. Policy areas covered by this measure are old age, survivors, incapacity-related benefits, health, family, active labor market programs, unemployment, housing and other social policy areas. The values are comparable between countries and thus allow for cross-country comparisons.

Tax progressivity capture the degree of progressivity in the tax burden for different

income groups. A more progressive tax system would lower net income inequality, as higher income would lose relatively more money in tax. A more progressive tax may also lead to a lower amount of people willing to work extra hours for higher wage or willing to work hard for higher paying jobs, both reducing market and net income inequality. Tax progressivity will be calculated following Arnold (2008). He constructs an index of tax progressivity based on the concept of residual progressivity and is calculated as follows;

𝐼𝑛𝑑𝑒𝑥 𝑜𝑓 𝑝𝑟𝑜𝑔𝑟𝑒𝑠𝑠𝑖𝑣𝑖𝑡𝑦 = 1 − 100 − 𝑀𝑎𝑟𝑔𝑖𝑛𝑎𝑙 𝑇𝑎𝑥 𝑅𝑎𝑡𝑒

100 − 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑇𝑎𝑥 𝑅𝑎𝑡𝑒

The marginal tax rate is gained from the OECD. It constitutes the marginal personal income tax and social security contribution rates on gross labor income. The average tax rate is taken from the same database and has the same definition, except that is has for average tax rates. The value is bound between 0 and 1, where a higher value means more progressivity. Data is available for 1999 to 2016.

(17)

17

Openness to trade captures the effect of globalization and competitive pressure from foreign industries and their workers on domestic wages. More open economies would experience more competitive pressure on wages, where scarce, highly skilled people would have increased wages and abundant, low skilled workers see their wages reduced. Both of these effects would mean a higher amount of income inequality. Openness to trade is defined as the percentage of GDP due to trade. It is the calculated sum of exports and imports of goods and services measured as a share of the gross domestic product of a country. It is a good indicator of openness to trade, as a higher percentage in this case would correlate with a higher openness to trade. The data is gained from the World Bank, which has an extensive database covering a wide range of countries and is available for .

Since unions bargain wage deals for their workers, unions should lead to lower income inequality. As said, this is primarily dependent on which groups these union primarily represent: high income groups or low income groups. If they are mainly representing high income groups, income inequality may be increased. In addition, if they do represent lower income groups and bargain for higher wages for these groups, it may lead to less people being employed. This in turn would also lead to higher inequality. To measure the effect of unions, the union density

rate is used. Data is gained from the Institutional Characteristics of Trade Unions, Wage

Setting , State intervention and Social pacts database, or ICTWSS database for short. The union density rate is defined as the net union membership as a proportion of wage earners in employment. The scale ranges from 0 to 100. Data is attainable from 1999 up to 2013.

In addition to unions, governments may feel the need to ensure their citizens fair wages, intervening in the economy. Government intervention also comes from the ICTWSS database. It constitutes the government intervention in wage bargaining. The variable can have five values. 1 means that governments do not influence wage bargaining. 2 means that the government influences wage bargaining by providing an institutional framework of consultation and information exchange. 3 means that governments influence wage bargaining indirectly through price ceilings, indexation, tax measures and minimum wages. 4 means that the governments participate in wage bargaining directly like in social pacts. 5 means that governments impose private sector wages settlement, places ceiling on bargaining outcomes or suspends bargaining. A higher values on this scale thus means increased government involvement. Data is available for 1999 to 2013.

Finally, differences in educational attainment may lead to differences in wages, as higher educated people generally earn more than their lower educated counterparts. To capture

(18)

18

this effect, two variables are used: the percentage of persons having only attained primary education and the percentage of people having attained tertiary education. Data is taken from the OECD and consists of the gross enrolment ratio in tertiary education, regardless of age, expressed as the percentage of the population of the five-year age group following on from secondary school leaving. Gross enrolment is taken as we want to include all people that have had a tertiary education, opposed to net enrolment, which measures the percentage of people enrolled at the moment.

3.4 Descriptive statistics

Table 1 – Summary statistics

Variable Obs Mean Std. Dev. Min Max

Gini net ₁₈₄ _28.72291 _3.263823 _22.74555 _36.44239

Gini market ₁₈₄ _46.50638 _2.964266 _35.38126 _55.68162

ECB Balance

sheet 190 13.97966 .493944 13.23234 14.74133

ECB Policy Rate 191 3.126229 1.445113 .52342 5.27671

GDP growth 191 1.636817 3.094642 -9.132494 10.86297 Unemployment 202 8.64604 4.630056 1.9 27.5 Social expenditure 191 23.67094 3.8381 12.567 31.938 Trade openness 202 108.8579 75.71985 44.72518 438.1569 Union density 166 31.45178 17.55448 7.547652 76.27162 Government intervention 180 3.177778 .8662225 2 5 Tax progressivity 180 .1950017 .0750525 .0336332 .5653648 Low educated 174 35.86554 16.59647 13.09658 81.159 High educated 174 26.37212 8.121414 8.682372 45.93559

Table 1 shows the summary statistics of the variables used. The variable for pre-tax inequality is always higher than post-tax inequality, which can be explained by progressive taxes and transfers.

(19)

19 3.5 Correlations

Table 2 – Correlations between explanatory variables

ECB Bala nc e sheet ECB P olicy Ra te GDP growth Une m plo

yment Social expendit

u re Tra d e openn ess Union densi ty Governm ent interven ti onTax pro gr essi

vity Low educa

te d Hi gh ed uca te d

ECB Balance sheet 1.0000

ECB Policy Rate -0.5682 1.0000

GDP growth -0.2834 0.4429 1.0000 Unemployment 0.1891 -0.3252 -0.3873 1.0000 Social expenditure 0.3486 -0.4308 -0.4465 0.2751 1.0000 Trade openness 0.1729 -0.1403 0.1735 -0.3280 -0.2742 1.0000 Union density -0.0347 -0.0040 0.0926 -0.1630 0.0254 0.2582 1.0000 Government intervention _{-0.0180 -0.0774 0.0445} _0.1446 _{-0.1652 0.3356} _0.3505 _1.0000 Tax progressivity 0.1526 -0.2193 -0.0587 -0.0975 0.0351 0.4582 0.2885 0.0292 1.0000 Low educated -0.2015 0.1896 0.0215 0.3001 -0.3394 -0.2065 -0.2198 0.2099 -0.3681 1.0000 High educated _0.3502 _{-0.3450 -0.0262 0.0327} _0.0796 _0.4784 _0.3473 _{0.3962 0.3414} _{-0.5272 1.0000}

(20)

20

Table 2 shows the correlations between the explanatory variables. The ECB policy rate is highly negatively correlated with the ECB balance sheet, which is expected as unconventional monetary policy is used when there is a very low policy rate. Social expenditure is negatively correlated with the policy rate as well as with GDP growth. Since social expenditure is measured as a percentage of GDP, a growing GDP would indeed be negatively correlated with social expenditure if nominal expenditure stays the same. Furthermore, percentage of the population that has achieved tertiary education is positively correlated with trade openness and highly negatively correlated with percentage of persons with only lower education. This is expected, since people that attain tertiary education are not included anymore in the primary education group.

(21)

21

4. Analysis

4.1 Links with post-income tax inequality

Table 3 – Regression done on post-tax income inequality

Variables Coef. Std. Err. P>t [95% Conf. Interval]

ECB Balance sheet _0.1925 _0.3770074 _0.611 _-0.5536 _0.9387

ECB Policy Rate _-0.1895 _0.1147759 _0.101 _-0.4166 _0.0376

GDP growth _-0.0753 _0.0474007 _0.114 _-0.1692 _0.0184 Unemployment _0.1673 _0.0497154 _0.001*** _0.0689 _0.2657 Social expenditure _-0.3761 _0.1295184 _0.004*** _-0.6325 _-0.1198 Union density _-0.0639 _0.0798147 _0.424 _-0.2219 _0.0939 Trade openness _-0.0089 _0.0112665 _0.427 _-0.0312 _0.0133 Government intervention -0.0833 0.1894252 0.661 -0.4582 0.2916 Tax progressivity _-4.3849 _1.698486 _0.011** _-7.7467 _-1.0231 Low educated _0.2069 _.0647319 _0.002*** _0.0787 _0.3350 High educated _0.1794 _.0924135 _0.054* _-0.0034 _0.3624 constant _26.1457 _8.626793 _0.003 _9.0709 _43.2206 N 147 sigma_u _2.7310 sigma_e _0.9925 rho _0.8833 R2_-within _0.3301 R2_-between _0.5680 R2_-overall _0.5326

(22)

22

Table 3 shows the results from the regression done on post-tax income inequality. The main variable of interest is the size of the ECB balance sheet, which is not significant here. Unconventional monetary policy would not seem to have a significant effect on post-tax income inequality. As the first hypothesis was ‘Expansionary unconventional monetary policy

increases post-tax income inequality’, the first hypothesis has to be rejected. This may be

because unconventional monetary policy does not have an effect on the different channels influencing income inequality in the theory, but it may also be that the differing channels are cancelling each other out, since some have opposite effects. Overall however, there is no significant effect, going against research from Saiki and Frost, and Davtyan, who found significant inequality increasing effects. Looking at conventional monetary policy, the policy rate of the ECB also does not seem to have a significant effect on income inequality, although it is just barely insignificant. Doerrenberg and Peichl found that different inequality variables can yield significantly different results. A slightly different dataset might therefor show a significant correlation. Again, the insignificance might be because of the different channels cancelling each other out. Unemployment is highly significant and inequality increasing. The obvious explanation for this is that people without work do not have a wage, thus increasing inequality. Social expenditure is also highly significant and inequality decreasing, bolstering similar results found by Doerrenberg and Peichl, Lee, Roine et al., and Calderón and Chong, but going against findings by García-Peñalosa and Checchi. This can be explained due to transfers of money benefitting the poor. Tax progressivity is significant at the 5% level and inequality decreasing. These significant effects have also been found by Doerrenberg and Peichl, Cooper et al., Duncan and Sabirianova-Peter and Roine et al. This is probably due to the higher incomes getting relatively more money taken away from them than the poor, closing the income gap. The amount of low educated people is also highly significant, increasing inequality. As the theory explained, the lower educated part of the working population gets paid less, thus increasing income inequality. The amount of higher educated people is also significant and inequality increasing, but at the 10% level. This backs up research done by Iversen and Stephens, Calderón and Chong, Alderson and Nielsen, who all found significant effects. The aforementioned theory explains this as well, as higher educated people earn higher wages. Interestingly, union density and government intervention, variables often found to be significantly inequality decreasing, are not found to be significant here. This goes against claims from Doerrenberg and Peichl, Rueda, and Calderón and Chong, which all found significant inequality reducing effects from corporatist institutions. As stated in the theory, unions try to increase the wage of their members, and the type of members is important: if higher paid

(23)

23

workers are more organized than lower paid workers, it could increase inequality. Unions may still have a significant effect on inequality, but if membership of both higher paid workers and lower paid workers are about equal, the effects may cancel out. GDP growth is also not significant, as is trade openness. Significant effects of globalization have been found by Atkinson and Jaumotte, Lall, and Papageorgiou. The insignificance of trade openness might show that globalization does not have such a strong effect on income inequality as previously thought. However, a higher degree of trade openness has been linked to higher amounts of welfare redistribution and social programs to protect the losers of globalization. Higher openness is not correlated with higher percentages of social expenditure, though it is somewhat correlated with higher tax progressivity and government intervention. In subsequent research, using education combined with sectoral distribution and the openness of these sectors would be a better measure to measure the effects of globalization and education.

(24)

24

4.2 Links with pre-income tax inequality

Table 4 – Regression done on pre-tax income inequality

Variables Coef. Std. Err. P>t [95% Conf.Interval]

ECB Balance sheet _-0.5849 _{0.5677525 0.305} _-1.7086 _0.5388

ECB Policy Rate _0.0981 _{0.1728462 0.571} _-0.2439 _0.4403

GDP growth _-0.0609 _{0.0713829 0.395} _-0.2022 _0.0803 Unemployment _0.3519 _{0.0748687 0.000*** 0.2037} _0.5001 Social expenditure _0.0239 _{0.1950476 0.903} _-0.3621 _0.4099 Union density _-0.1802 _{0.1201966 0.136} _-0.4181 _0.0576 Trade openness _0.0036 _{0.0169667 0.829} _-0.0299 _0.0372 Government intervention -0.0789 0.285264 0.782 -0.6435 0.4856 Tax progressivity _-6.916 _2.557827 _{0.008*** -11.9791} _-1.8537 Low educated _-0.0479 _{0.0974826 0.624} _-0.2408 _0.1450 High educated _0.0444 _{0.1391696 0.750} _-0.2310 _0.3199 constant _58.5642 _12.99148 _{0.000*** 32.8504} _84.2779 N 147 sigma_u _3.0530 sigma_e _1.4947 rho _0.8066 R2_-within _0.4781 R2_-between _0.2009 R2_-overall _0.2178

(25)

25

The second regression is done on pre-tax income inequality with the results shown in table 4. Again, the variable for unconventional monetary policy is not significant, rejecting the second hypothesis, ’Expansionary unconventional monetary policy increases pre-tax income

inequality’. The ECB policy rate is also not significant, indicating that conventional monetary

policy might not have an effect. GDP growth is not significant as well, while unemployment is highly significant and welfare increasing, with a stronger effect than with the post-tax inequality. This could be explained due to the fact that in post-tax inequality, transfers are included, while in pre-tax inequality these are not included. Being unemployed thus has a stronger effect on pre-tax inequality than post-tax. Social expenditure does not have a significant effect, while it did in the post-tax analysis, now disputing claims by Doerrenberg and Peichl, Lee, Roine et al., and Calderón and Chong, but backing up results from García-Peñalosa and Checchi. Again, this is due to the effect of transfers, and these have not taken place when looking at pre-tax inequality. Union density, trade openness and government intervention are not significant in this analysis either, again going against research from Doerrenberg and Peichl, Rueda, and Calderón and Chong for corporatists institutions and Atkinson and Jaumotte, Lall, and Papageorgiou for effects from globalization. Tax progressivity is highly significant and inequality reducing, once more confirming research by Doerrenberg and Peichl, Cooper et al., Duncan and Sabirianova-Peter and Roine et al. This result may seem counterintuitive as transfers have not yet taken place, but having a more progressive tax system may make it less desirable for people to take higher paying and more intensive jobs or work more hours, as their work is relatively less rewarded. They may then forego this extra money to spend more time on leisure. Finally, both the education variables are not significant in this regression, now going against research from Iversen and Stephens, Calderón and Chong, Alderson and Nielsen. The constant in this regression is ten points above the mean value of the market Gini coefficient, which can be explained by the relatively large negative coefficients for the ECB balance sheet and unionization, both having large mean values.

4.3 Robustness

The same regressions have been done with the Gini’s from the OECD (OECD, 2017). The results can be viewed in appendix A for pre-tax inequality and appendix B for post-tax inequality. This inequality measure is not preferable over the one used for the main regressions, since there are far less observations and more gaps in the data. For pre-tax inequality, only

(26)

26

unemployment and tax progressivity are significant, differing from the main regression. For post-tax income inequality, there are more differences. Tax progressivity is no longer significant, while the policy rate, government intervention and percentage of the population which has a tertiary education does become significant. Doerrenberg and Peichl (2016) show that different inequality datasets yield different results, which may explain the differences in significance for variables here. For the main variable of interest, unconventional monetary policy, the results do not differ however, since this is again never significant. Taking a one year lag for tax progressivity for the main regression makes GDP growth significant at the 10% level for post-tax inequality, but does not alter the significance of any other variables for both measures. Lagging unemployment by one year makes GDP growth significant at the 5% level for post-tax inequality, but does not change the significance of other variables. For pre-tax inequality, the ECB rate and unionization become significant at the 10% level. The variable for unconventional monetary policy is never significant.

(27)

27

5. Conclusion

This thesis researches if unconventional monetary policy has an effect on income inequality. A dataset consisting of twelve euro countries has been used, with data from 1999 to 2013. Two regressions have been done, one for pre-tax and one for post-tax inequality. Both regressions do not give a significant correlation between unconventional monetary policy and income inequality. This goas against research from Saiki and Frost and Davtyan which found significant inequality increasing effects. In addition, several other variables have been tested. Conventional monetary policy, measured by the ECB policy rate, is not significant in both analyses. It should be noted that this is just by small margin for post-tax inequality. Unemployment is highly significant and inequality increasing in both regressions. Tax progressivity is highly significant and inequality reducing for both analyses, bolstering claims from Doerrenberg and Peichl, Cooper et al., Duncan and Sabirianova-Peter and Roine et al. This is probably due to the effect of people substituting time worked for leisure, as well as redistribution for post-tax inequality. Social expenditure is significant and inequality reducing for post-tax inequality only. The percentage of persons with primary education only is also significant for post-tax inequality and inequality increasing. Interestingly, union density, government intervention and trade openness are never significant, while significant effects have been found in other studies. For robustness, the Gini coefficients from the OECD database are used. These yield different significances for some variables, but unconventional monetary policy is unaffected. Lagging the tax variable or the unemployment variable by one year also does not alter the significance of unconventional monetary policy. As for policy implications regarding this subject, the results seem to indicate that the expansive unconventional monetary of the ECB did not affect income inequality. However, some caution has to be taken in for this conclusion. First, while the results may be robust, different inequality datasets and different econometric techniques may yield different results, as does having different operationalizations for certain variables. Second, the time period used here is still relatively short and getting more data points may change the results. Third, as stated by Bivens who researched the counterfactual outcome, no unconventional monetary policy may increase inequality. Unconventional monetary policy may thus indeed have an effect compared to doing nothing. If policy has to be made for reducing income inequality, one could better look at ways to reduce unemployment, increase tax progressivity or increase social expenditure. Additionally, policies to increase the amount of people having at least a secondary education would also reduce inequality.

(28)

28

References

Albanesi, S. (2007). Inflation and inequality. Journal of Monetary Economics, 54, 1088–1114 Alderson, A. & Nielsen, F. (2002). Globalization and the Great U-Turn Income Inequality

Trends in 16 OECD Countries. American Journal of Sociology, 107, 1–56.

Amaral, P. (2017). Monetary Policy. Economic Commentary, 2017-01, Federal Reserve bank of Cleveland

Atkinson, A. (2003). Income Inequality in OECD Countries: Data and Explanations. CESifo

Working Paper. 881.

Arnold, J. (2008). Do Tax Structures Affect Aggregate Economic Growth?: Empirical Evidence from a Panel of OECD Countries, OECD Economics Department Working Papers, 643. Auclert, A. (2016). Monetary Policy and the Redistribution Channel, unpublished manuscript,

Stanford University.

Bivens, J. (2015). Gauging the Impact of the Fed on Inequality during the Great Recession. at

Brookings. (Hutchins Center Working Paper 12). Brookings: Hutchins Center on Fiscal and

Monetary Policy.

Borjas, G.J. (1994). “The Economics of Immigration.” Journal of Economic Literature. 32, 1667-1717.

Calderón, C., & Chong, A. (2009). Labor market institutions and income inequality: an empirical exploration. Public Choice. 138, 65 –81.

Carpenter, S., & Rodgers III W., (2004). The Disparate Labor Market Impacts of Monetary Policy, Journal of Policy Analysis and Management, 23(4), 813–30.

Cohan, W. (2014). How Quantitative Easing Contributed to the Nation’s Inequality Problem, Dealbook, New York Times, October 22.

Coibion, Gorodnichenko, Silvia & Kueng, (2012). Innocent bystanders? Monetary policy and

inequality in the U.S.. (NBER working paper No. 18170). Cambridge: National Bureau of

Economic Research

Cooper, D. H., Lutz, B. F., & Palumbo, M. G. (2012). Quantifying the role of federal and state taxes in mitigating income inequality. Finance and Economics Discussion Series.

Davtyan, K. (2016). Income Inequality and Monetary Policy: An Analysis on the Long Run

Relation. (Working paper 2016-04). Barcelona: Research Institute of Applied Economics.

Diener, E., Kesebir, S., & Shigehiro I. (2011). Income Inequality and Happiness. Psychological

Science, 22(9), 1095-1100.

Dobbs, R., Lund, S., Koller, T., & Shwayder, A. (2013). QE and Ultra-low Interest Rates:

(29)

29

Doepke, M., & Schneider, M. (2006). Inflation and the Redistribution of Wealth, Journal of

Political Economy, 114(6), 1069–97.

Doerrenberg, P., & Peichl, A. (2014). The impact of redistributive polices on inequality in OECD countries. Applied Economics, 46(17), 2066-2086.

Domanski, D., Scatinga, M., & Zabai., A. (2016). Wealth inequality and monetary policy, BIS

Quarterly Review, March 2016, 45-64.

Duncan, D., & Sabirianova-Peter, K. (2008). Tax progressivity and income inequality. Andrew

Young School of Policy Studies Research Paper Series.

Erosa, A., & Ventura, G. (2002). On Inflation as a Regressive Consumption Tax, Journal of

Monetary Economics, 49(4), 761–795.

Gagnon, J., M. Raskin, J. Remache, and B. Sack, 2011. “The Financial Market Effects of the Federal Reserve’s Large-Scale Asset Purchases,” International Journal of Central

Banking, 7(1):3–43.

García-Peñalosa, C., & Checchi, D. (2008). Labour market institutions and income inequality.

Economic Policy. 23, 601–49.

Gold R., Kawachi, I., Kennedy, B.P., Lynch, J.W., & Connell, F.A. (2001). Ecological analysis of teen birth rates: association with community income and income inequality, Maternal

Child Health Journal, 5, 161–67.

Gornemann N., Kuester, K., & Nakajima, M. (2012). Monetary Policy with Heterogeneous

Agents. (Working Paper no. 12-21). Philadelphia: Federal Reserve Bank of Philadelphia

Heathcote, J., Perri, F., & Violante, G., (2009). Unequal We Stand: An Empirical Analysis of

Economic Inequality in the United States, 1967–2006. (Research Department Staff Report

no. 436). Minneapolis: Federal Reserve Bank of Minneapolis.

Hsieh, C-C., & Pugh, M.D. (1993). Poverty, income inequality, and violent crime: a meta-analysis of recent aggregate data studies. Criminal Justice Review, 18, 182–202.

Iversen, T. & Stephens, J.D. (2008). Partisan Politics, the Welfare State, and Three Wolds of Human Capital Formation. Comparative Political Studies. 41, 600-637.

Jaumotte, F., Lall, S. & Papageorgiou, C. (2008). Rising Income Inequality: Technology, or Trade and Financial Globalization?. IMF Working Paper, IMF WP/08/185.

Joyce, M, Miles, D., Scott, A. and Vayanos, D. (2012). Quantitative easing and unconventional monetary policy – an introduction. The Economic Journal, 122, 271-288.

Kaplan G.A., Pamuk, E.R., Lynch, J.W., Cohen, R.D., & Balfour, J.L. (1996). Inequality in income and mortality in the United States: analysis of mortality and potential pathways.

(30)

30

Kennedy, B.P., Kawachi, I., & Prothrow-Stith, D. (1996). Income distribution and mortality: cross sectional ecological study of the Robin Hood index in the United States. BMJ, 312, 1004–7.

Lee, C.-S. (2005). Income Inequality, Democracy, and Public Sector Size. American

Sociological Review, 70(1), 158-181.

Niehues, J. (2010). Social spending generosity and income inequality: A dynamic panel approach. SOEPpapers on Multidisciplinary Panel Data Research, 336.

O’Farrell, R., L. Rawdanowicz, and K. Inaba, (2016). Monetary Policy and Inequality. OECD

Economics Department Working Paper, 1281.

Pickett K.E., Kelly, S., Brunner, E., Lobstein, T., & Wilkinson, R.G. (2005a). Wider income gaps, wider waistbands? An ecological study of obesity and income inequality. Journal of

Epidemiology and Community Health, 59, 670–74.

Pickett, K.E., Mookherjee, J., & Wilkinson, R.G. (2005b). Adolescent birth rates, total homicides, and income inequality in rich countries. American Journal of Public Health, 95, 1181–83.

Pickett, K.E., & Wilkinson, R.G. (2007). Child wellbeing and income inequality in rich societies: ecological cross sectional study, BMJ, 335, 1080–87.

Roine, J., Vlachos, J. & Waldenström, D. (2009). The long-run determinants of inequality: What can we learn from top income data? Journal of Public Economics, 93(7-8), 974–988.

Rosa, C., 2012. “How ‘Unconventional’ Are Large-scale Asset Purchases? The Impact of Monetary Policy on Asset Prices,” Federal Reserve Bank of New York, Staff Report no. 560.

Rueda, D. (2008). Left Government, Policy, and Corporatism: Explaining the Influence of Partisanship on Inequality. World Politics. 60(3), 349-389.

Saiki, A. & Frost, J. (2014). Does unconventional monetary policy affect inequality? Evidence from Japan. Applied Economics, 46(36), 4445–4454.

Smeeding, T., & Brandolini, A. (2007). Inequality patterns in Western-Type Democracies.

Working paper.

Solt, F. (2016). The Standardized World Income Inequality Database. Social Science Quarterly, 97(5), 1267-1281.

Subramanian, S.V., & Kawachi, I. (2006). Whose health is affected by income inequality? A multilevel interaction analysis of contemporaneous and lagged effects of state income inequality on individual self-rated health in the United States. Health Place, 12, 141–56.

(31)

31

Appendices

Appendix A – Regression done on OECD Gini for pre-tax income inequality

ECB Policy Rate _0.0580 _0.1119859 _0.605 _-0.1643 _0.2805

GDP growth _0.0004 _0.0439985 _0.992 _-0.0869 _0.0878 Unemployment _0.3216 _0.0522269 _0.000*** _0.2179 _0.4254 Social expenditure _0.2086 _0.1252308 _0.099* _-0.0401 _0.4574 Union density _-0.0899 _0.0879795 _0.309 _-0.2647 _0.0848 Trade openness _0.0093 _0.0108523 _0.391 _-0.0122 _0.0309 Government intervention -0.0623 0.1882528 0.741 -0.4362 0.3116 Tax progressivity _0.9737 _1.578736 _0.539 _-2.1626 _4.1101 Low educated _-0.0058 _0.0357859 _0.870 _-0.0769 _0.0652 High educated _0.0233 _0.0713021 _0.744 _-0.1182 _0.1650 constant _40.35284 _6.959797 _0.000*** _26.52599 _54.1797 N 113 sigma_u _3.1862 sigma_e _0.8755 rho _0.9297 R2_-within _0.7607 R2_-between _0.0311 R2_-overall _0.1424

(32)

32

Appendix B – Regression done on OECD Gini for post-tax income inequality

ECB Policy Rate _-0.3474 _0.0939766 _0.000*** _-0.5338 _-0.1610

GDP growth _0.0015 _0.0396883 _0.970 _-0.0772 _0.0802 Unemployment _0.1174 _0.0450521 _0.010** _0.0281 _0.2068 Social expenditure _-0.2625 _0.1063124 _0.015** _-0.4734 _-0.0516 Union density _0.0139 _0.072817 _0.848 _-0.1304 _0.1584 Trade openness _-0.0118 _0.0095736 _0.219 _-0.0308 _0.0071 Government intervention -0.3000 0.1576422 0.060* -0.6127 0.0125 Tax progressivity _-0.1957 _1.392929 _0.889 _-2.9585 _2.5671 Low educated _0.1000 _0.0323777 _0.003*** _0.0358 _0.1642 High educated _0.1078 _0.0602479 _0.076* _-0.0116 _0.2273 constant _34.489 _5.721925 _0.000*** _23.13965 _45.83848 N 125 sigma_u _1.87535 sigma_e _0.79818 rho _0.84663 R2_-within _0.2508 R2_-between _0.6398 R2_-overall _0.5726

The effect of unconventional monetary policy on income inequality in the Eurozone