Does an inflation targeting mandate of a central bank cause higher income inequality?

(1)

“Does an inflation targeting mandate

of a central bank cause higher income

inequality?”

Name: Roderik Nielen Student number: 10097163

(2)

This document is written by Student Roderik Nielen who declares to take full responsibility

for the contents of this document.

I declare that the text and the work presented in this document is 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)

Introduction

Income inequality has, arguably, been a problem for a long time and has even increased while countries were going through a period of economic and employment growth. Kuznets (1955) predicted that income inequality moves in an inverted U-curve. This is called the Kuznets-curve. Kuznets explains that as the economy of a country evolves and GDP increases, some people will adapt immediately to these changes. Early adopters of the changing economy will earn a higher wage because of their productivity advantage causing income inequality to increase. Over time, when the economy develops further, more people will adopt to the changes in the economy and their wages will also increase thereby decreasing income inequality. Looking at the data, income inequality has been decreasing from 1920, as predicted by Kuznets. However, contrary to Kuznets prediction, as of early 1980 income inequality is increasing again. The common understanding among economists as a source of rising inequality since early 1980 are skilled biased technological change (Bound & Johnson, 1992), increased global trade (Feenstra & Hanson, 2008) and changes in labour market institutions such as unionisation (Card, 2001).

Monetary policy is rarely mentioned as a cause of the rise in income inequality since the early 1980s. Monetary policy, however, has changed. Central banks abandoned exchange rate systems and instead turned their attention to fighting inflation (Rose, 2007). New Zealand was the first county to adopt an Inflation Targeting (IT) mandate and since then many developed and developing countries have followed their example. IT seemed to be working as the world entered “the Great Moderation”, a period of reduced macroeconomic volatility. The Financial crisis in 2008 originating in the subprime mortgage markets led to a recession, increasing macroeconomic volatility and to demonstrations such as the Occupy Wall Street movement. IT was put to a test and central banks had to take unconventional measures, trying to get inflation back on target with unclear consequences for income distribution. The financial crisis showed that central banks focus on inflation is not perfect and additional measures have to be taken to increase economic stability. Even when IT might change in its current form, IT will have had a lasting impact on monetary policy. Central bank transparency has increased and the use of forward guidance is now common even for countries that did not adopt IT (Walsh, 2009).

This paper will discuss whether an IT mandate affects inflation and inflation volatility and later on if an IT mandate causes higher income inequality. To study the effects of an IT mandate on inflation, inflation volatility and income inequality, panel data regressions will be conducted. This paper finds that an IT mandates can achieve lower inflation for both OECD and non-OECD countries. This paper further finds that an IT mandate given by a government to a central bank doesn’t have a lasting

(4)

effect on income inequality. The decreasing effect of an IT mandate on income inequality when inflation is brought down from hyperinflation levels is mostly cancelled because the central bank in hyperinflation countries lack credibility and commitment to their inflation target. Income inequality is also not increased by an IT mandate because for that to happen unexpected changes have to occur in interest rates and inflation. However unexpected changes are unlikely as an import tool of IT is forward guidance.

Literature review

The consequences of monetary policy on income inequality are a relatively new subject and, although growing, literature is limited. I will first discuss the common wisdom in literature on income inequality, after that I will examine which characteristics affect a country’s likelihood that they will adopt inflation targeting and whether this is really a regime switch. Thirdly, I will focus on the effects of inflation targeting (IT) on macroeconomic variables and conclude with the distributional consequences of monetary policy. The previous literature on the effects of IT on macroeconomic variables should help identify IT indicators which will be used as a proxy for the panel data research in the next section.

The view among economists is that the rise in income inequality since the 1980 is caused by an increase in global trade (Feenstra & Hanson, 2008), skilled-biased technological change (Berman, Bound & Machin, 1998) and changes in labour market institutions (Card, 2001). Feenstra and Hanson (2008) use a model of outsourcing and data on changes in industry behaviour, and find that foreign outsourcing is associated with skill upgrading and an increase in the unskilled-skilled wage gap. They argue that outsourced activities mostly use unskilled labour and moving these activities abroad reduces relative demand for unskilled labour. Berman et al. (1998) however argues that it is not increased global trade but skill-biased technological change that increased the unskilled-skilled wage gap. Skill-biased technological change dramatically decreased demand for unskilled workers and the most important contributions in skill upgrading are machinery (computer), electrical machinery and printing & publishing (Berman et al, 1998). Another main contributor of income inequality are changes in labour market institutions. Card (2001) uses Current Populations Survey micro data to study the effects of changing labour unions on wage inequality. Card (2001) concludes that unionization did not explain the rise in wage inequality among female workers, this is because the fraction of women belonging to unions was stable. However, for male workers, Card (2001) concluded that 15-20% of the rise in wage inequality is explained by a decrease in unionization. In the public sector an increase in unionism prevented a rise in wage inequality both for male and

(5)

female workers (Card, 2001). Also Mishel (2012) and Ghosh (2014) find evidence that de-unionization played a key role in the increase in wage inequality in the U.S. as it undermined the position of workers while improving that of employers.

Most empirical investigations on which characteristics affect a country’s decision on IT adoption, focus on inflation. Results are however mixed. For example Mukherjee and Singer (2008) find that higher inflation increases the likelihood that a country adopts IT as monetary strategy whereas Hu (2006) finds that low inflation increases the probability that countries choose for IT. Samaryna and De Haan (2011) did a more enhanced research and looked for economic, fiscal, external, financial and institutional characteristics that affect the IT adoption decision. With the use of a panel binary response transition model for 60 countries, Samaryna and De Haan (2011), find support that low inflation and GDP growth, high output and exchange rate volatility, a flexible exchange rate regime, fiscal discipline, less developed and a market-based financial system effect a countries decision on a IT regime. Their results differ however between non-OECD and OECD countries. No matter what the reason is for a country to adopt IT, all stayed with their choice and did not switch afterwards to another regime. Before investigating whether an IT mandate causes higher income inequality it must first be established that the adoption of IT represented a regime switch. With the use of an MSVAR technique Creel and Hubert (2008) find that in three different countries, Canada, Sweden and the UK, the adoption of IT has represented a regime switch. IT has been characterised by lower intercept and variances and policy responses have been more predictable (Creel & Hubert, 2008). Creel and Hubert (2008) also conclude that central banks with an IT mandate could achieve lower real interest rate compared with the same central banks when not under IT regime. Carare and Stone (2006) also looked at IT regimes and divided IT into three types, Full-Fledged-, eclectic- and lite-IT. Carare and Stone (2006) define this as:

“Full-Fledged inflation targeting (FFIT) countries have a medium to high level of credibility, clearly commit to their inflation target, and institutionalize this commitment in the form of a transparent monetary framework that fosters accountability of the central bank to the target”.

“Eclectic inflation targeting (EIT) countries have so much credibility that they can maintain low and stable inflation without full transparency and accountability with respect to an inflation target”. “Inflation targeting lite (ITL) countries announce a broad inflation objective but owing to relatively low credibility are not able to maintain inflation as the foremost policy objective”.

It is important to distinguish these types of IT since all have different implications and therefore might affect income inequality differently. FFIT’ers have a high level of credibility and are financially stable however they cannot maintain low inflation without strong commitment which causes them

(6)

to be less flexible for output stabilisation (Carare & Stone, 2006). EIT’ers already have low inflation and therefore do not have to show a strong commitment, also EIT’ers have the flexibility to smooth output. Their dual mandate however means they cannot be as transparent as FFIT (Carare & Stone, 2006). ITL’ers do not have enough credibility to maintain low inflation and are exposed to financial and cyclical shocks. ITL’ers therefore do not target inflation but want full flexibility to deal with shocks (Carare & Stone, 2006).

Now that it is established that an IT regime exists and that there are even different types of IT regimes, we can move forward to the effects IT has on macroeconomic variables. It is hard to quantify the effect IT has on macroeconomic variables because IT was implemented during an era of low and stable inflation combined with steady economic growth (Walsh, 2009). This period is referred to as the Great Moderation which according to some is due to monetary policy (e.g. Boivin & Giannoni, 2006) while others say it is simply good luck (e.g. Gali & Gambetti, 2009). What indicates that IT was responsible for the Great Moderation? First of all, a central banks announcement of an IT target can align public inflation expectations with the actual target (Walsh, 2009). This reduces the cost of achieving low inflation, IT should therefore be associated with low average inflation without an increase in volatility of output (Walsh, 2009). Secondly IT could be indicated by lower volatility of inflation when a central bank credibly manages inflation expectations (Walsh, 2009). Thirdly, by announcing a inflation target, IT can promote accountability and therefore lead to greater transparency.

Using cross-section regressions Ball and Sheridan (2004) find that inflation targeting has an insignificant beneficial effect on low average inflation. Similar results are found in Lin and Ye (2007) who use a propensity score matching methodology. With the propensity score matching methodology, Lin and Ye (2007) try to find what would have happened without IT by assessing what did happen in a non-IT country that they feel are similar. Among the countries Lin and Ye selected as non IT countries are the US and a large number of European countries who later formed the European Monetary union. Although these countries officially did not adopt inflation targeting they can be considered EIT and hence they are comparing IT countries to each other. Vega and Winkelried (2005) included developing countries and, also using a propensity score matching methodology, find that Inflation targeting has contributed to lower inflation volatility and a reduction in its average level. The lowering of inflation volatility works through managing inflation expectations. In a panel data study Johnson (2002) finds that IT leads to a significant drop in expected inflation, however he does not find a difference in the volatility of expected inflation across IT and non-IT. Gurkaynak et al. (2006 , 2007) find in two different studies that inflation expectations did not respond to economic news in Canada, Chile, Sweden and the UK. In a study across 36 developed and developing countries

(7)

Davis and Presno (2014) examined the response of inflation expectations following shocks to inflation, inflation expectations and oil prices. They find that in IT countries the response in the post-IT is much less significant whereas there was no difference in non-post-IT countries. Not only have central banks with IT managed to anchor inflation expectations, they are also more transparent. Eijffinger and Geraats (2006) made a comprehensive index for central bank transparency and find that IT central banks are the most transparent. Walsh (2009) concludes using the Dincer and Eichengreen index for central bank transparency that inflation targeters are more transparent and have the largest increase in transparency.

As said above, the literature for the distributional consequences of monetary policy is limited. Gornemann et al. (2014) built a dynamic stochastic general equilibrium model for the US and used this model to analyse how monetary policy affects different segments of the population. Gorneman et al. (2014) find that the rich gain when the policy rate increases. This is because the wealth-rich get most of their income from returns on financial assets. The wealth-poor get their income from labour which falls when an increase in the policy rate causes higher unemployment. Criticism on IT is that is focuses too much on inflation and thereby may sacrifice other objectives (Friedman, 2004). So when the policy rate needs to be adjusted upward to get inflation in line with the central bank’s target this could cause higher income inequality due to higher unemployment. Secondly, Gorneman et al (2014) find that the top 5 percent benefits from contractionary policy while the bottom 95 percent loses from such policy. Thirdly, they find that when monetary policy responds aggressively to a severe recession the bottom 80 percent gain significantly at the expense of the top 20 percent. This is because cuts in the policy rate leads to a lower real interest rate which induces labour firms to post more vacancies and capital firms to invest more. The increase in vacancies and investments by firm reduces short term profits and creates jobs (Gorneman et al., 2014). At last, Gorneman et al. (2014) find that countercyclical monetary policy reduces the labour market volatility which leads to a decrease in precautionary savings and a decline in the capital stock. This capital deccumulation will hurt all household in the long-run. Romer and Romer (1998) investigated the influence of monetary policy in general on poverty and inequality in the short- and long-run. With panel data regressions Romer and Romer (1998) show that cyclical boom created by expansionary monetary policy improve the well-being of the poor in the short run. For the long run they conclude that low inflation and stable aggregate demand improve the conditions of the poor. This could mean that IT actually causes lower income inequality because IT countries can achieve lower inflation rates compared to non-IT. Coibion et al. (2012) studied the effects of monetary policy shocks to consumption and income inequality in the US. They documented different channels through which monetary policy affects inequality. The first is the income composition channel. The bottom part of the population get their

(8)

income mostly from labour earnings while top part receives income form businesses and financial income. When monetary policy affects profits differently than wages it may have a distortionary effect on income inequality. With an IT mandate a central bank focuses on inflation and less on full-employment, therefore by acting more on inflation instead of unemployment central banks can influence income of labour earnings disproportionally to income for businesses and financial income. The second channel is the portfolio channel. “If low-income household hold more currency than high-income households, then inflationary actions on part of the central bank would represent a transfer from low-income household towards high-income households which tend to increase inequality” (Coibion et al, 2012). The third channel is savings redistribution channel. When there is an unexpected increase in interest rates or a decrease in inflation, this will be beneficial for savers and hurtful for borrowers. When IT reduces the volatility of inflation as in Vega and Winkelried (2005) this might reduce income inequality because there are less unexpected changes in interest rates and inflation. The last channel is the earning heterogeneity channel. Labour earnings may respond differently for low- and high-income households to monetary shocks. If unemployment mostly falls upon the low income households this could increase income inequality as suggested by Carpenter and Rogers (2004). With the use of these channels Coibion et al (2012) explain that monetary policy shocks significantly influence income and consumption inequality.

Data

The data used for the regressions all have yearly frequencies. The Gini coefficients and income quintiles originate from the World Income Inequality database (WIID). For central bank transparency the Dincer Eichengreen index is preferred over the De Haan index because the Dincer Eichengreen index better fits the sample period. Inflation data originates from the World Bank database. Similar to Rother (2004), inflation volatility is measured as the standard deviation over a calendar year of month on month inflation rates, this is done separately for each country. The data originates from Organisation for Economic Co-operation and Development (OECD) database and for Uruguay, New Zealand and Argentina from the central banks of those countries. Peru and Thailand are missing because for these countries monthly inflation levels are unavailable. As in Card (2001), the union density rate is used to indicate changes in labour market institutions and originate from Institutional Characteristics of Trade Unions Wage settings State Intervention and Social Pacts (ICTWSS). Trade is measured as a percentage of GDP and will be used to control for the effect increased global trade has on income inequality and comes from the OECD database. Countries are divided into 4 group, FFIT EIT ITL and no IT, this is done in a similar manner as Carare and Stone (2006). The date of when a

(9)

country is considered IT is when that country officially announced an IT target. Exceptions are the central bank of Denmark, ECB and FED because these institutions are not officially IT. The ECB can be considered IT and are in this paper IT since the start in 1999. The central bank of Denmark copies ECB behaviour and is therefore also IT as of 1999. For the FED it is a hard the give an precise date. This paper follows the IMF here and selected 2002 because as of this year the FED actively uses forward guidance as a monetary policy tool. The years that are taken into account are 1990 to 2010. A list of countries and which type of IT mandate their central bank has is given in table 25.The database on income quintiles is missing data for 9 countries and the regressions will therefore be short 9 countries compared to Gini regressions. The sample period which uses central bank transparency as an indicator of IT mandate will form 1998 to 2010 because the Dincer Eichengreen index is only available for these years.

Methodology

To study the effects of an IT mandate on income inequality a panel data regression will be conducted. A panel data study is chosen for multiple reasons. Firstly the number of observations of years with target is increased, secondly individual years are isolated and thirdly, a comparison is possible across countries and also between early and late years of targeting (Johnson, 2002). Before looking at the effects of an IT mandate on income inequality it must first be established what can be used to indicate an IT mandate. The macro variables this paper wants to use as an indicator of an IT mandate are central bank transparency and inflation expectations. For robustness a dummy variable is used to indicate and IT mandate. The macro variables are however preferred over a dummy as an indicator of an IT mandate because a dummy variable can indicate anything and by using macro variables this is narrowed down. As discussed above, the literature about the effects of an IT mandate on inflation expectations and central bank transparency is clear and therefore these variables can be used as an indicator of an IT mandate. The Dincer-Eichengreen index, which is used for central bank transparency, only ranches from 1998 to 2014, since some countries started with IT in 1990, this does not cover the entire sample period.

Before looking at the effects of an IT mandate on income inequality, this paper first discusses the effect an IT mandate has on inflation and inflation volatility because it is expected that the effects of an IT mandate on income inequality run via inflation and inflation volatility. The literature on the effects of an IT mandate on inflation and inflation volatility is somewhat mixed. To make sure an IT mandate has an effect on inflation and inflation volatility regressions (1) and (2) are conducted. These are necessary because an IT mandate is expected to have an effect on income inequality via

(10)

inflation and inflation volatility through multiple transmission channels. The first transmission channel is the portfolio channel which states that inflationary actions of a central bank affect those who hold relatively more currency (Coibion et al, 2012). Erosa and Ventura (2002) find that low-income households hold relatively more currency than high-low-income households so when a central bank successfully effects inflation this would increase income inequality. Another transmission channel is the savings redistribution channel. This channel states that household who save will benefit, and household that borrow will lose, when the interest rate unexpectedly increases or when inflation unexpectedly decreases (Doepke & Schneider, 2006). Savers tend to be wealthier than borrowers therefore this could increase income inequality. The saving redistribution and portfolio require that an IT mandate affects inflation and inflation volatility and therefore this paper will look for the effects an IT mandate has on these macro variables. If an IT mandate fails to have an effect on inflation and inflation volatility this might lead to the conclusion that an IT mandate has no effect on income inequality.

In line with Johnson (2002), who studied the effect of an IT mandate on inflation expectations, this paper will also use a panel data regression to study the effects an IT mandate has on inflation and inflation volatility. Both inflation and inflation volatility are expected to decrease when a central bank switches to an IT mandate. A panel data study is chosen for similar reasons as listed above. The regression will use inflation and inflation volatility as dependent variables and the independent variable is central bank transparency or inflation expectations which are used to indicate an IT mandate. The regression will control for unemployment which is added as a lag to avoid the simultaneity problem and lagged inflation (1) or lagged inflation volatility (2). The regressions will take the following form:

(1) 𝐼𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛𝑖𝑡= 𝛽𝑖,0𝐶𝑖 + 𝛾𝑖𝑌𝑡 + 𝛿(𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 𝑜𝑓 𝐼𝑇 𝑚𝑎𝑛𝑑𝑎𝑡𝑒𝑖𝑡) + 𝜃(𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛𝑡−1) +

𝜑(𝑢𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑡−1)

(2) 𝑖𝑛𝑓𝑙. 𝑣𝑜𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦𝑖𝑡 = 𝛽𝑖,0𝐶𝑖 + 𝛾𝑖𝑌𝑡+ 𝛿(𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 𝑜𝑓 𝐼𝑇 𝑚𝑎𝑛𝑑𝑎𝑡𝑒𝑖𝑡) +

𝜃(𝑖𝑛𝑓𝑙. 𝑣𝑜𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦𝑡−1) + 𝜑(𝑢𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑡−1)

In (1) and (2) 𝛽𝑖,0𝐶𝑖 represents the effect of factors within a specific country i that do not vary over

the sample. The term 𝛾𝑖 𝑌𝑡 in (1) and (2) measures common effects of inflation (1) or inflation

volatility (2) in all countries in year t. The different types of indicators of an IT mandate are discussed above and will measure the effect an IT mandate has on inflation and inflation volatility.

To study the effects of an IT mandate on income inequality, this paper will regress a proxy of income inequality on a proxy of IT mandate and several control variables. This paper uses several proxies to measure income inequality and indicators of an IT mandate in order to check the robustness. Three

(11)

proxies are selected for inequality: the Gini index, the incomes for the top 10% over the bottom 10% (S90/10) and income for the top 20% over the bottom 20% (S80/20). The share of the different income quintiles are also used by Romer and Romer (1998), and gives a different measurement than the commonly used Gini index. The sample with the different income quintiles will be short 9 countries compared with Gini because data on income quintiles is missing for some countries. The proxies indicating an IT mandate are inflation expectations, central bank transparency. The regression will control for indicators for labour market institutions and global trade, these will be the union density rate and trade measured as the sum of exports and imports of goods and services as a share of GDP. The regressions will take the following form:

(3) 𝐺𝑖𝑛𝑖𝑖,𝑡= 𝛽𝑖,0𝐶𝑖+ 𝛾𝑖𝑌𝑡+ 𝛿(𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 𝑜𝑓 𝐼𝑇 𝑚𝑎𝑛𝑑𝑎𝑡𝑒𝑖𝑡) + 𝜑(𝑢𝑛𝑖𝑜𝑛 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 𝑟𝑎𝑡𝑒𝑖𝑡) + 𝜌(𝑡𝑟𝑎𝑑𝑒𝑖𝑡) (4) 𝑆90/10𝑖,𝑡 = 𝛽𝑖,0𝐶𝑖+ 𝛾𝑖𝑌𝑡+ 𝛿(𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 𝑜𝑓 𝐼𝑇 𝑚𝑎𝑛𝑑𝑎𝑡𝑒𝑖𝑡) + 𝜑(𝑢𝑛𝑖𝑜𝑛 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 𝑟𝑎𝑡𝑒𝑖𝑡) + 𝜌(𝑡𝑟𝑎𝑑𝑒𝑖𝑡) (5) 𝑠80/10𝑖,𝑡= 𝛽𝑖,0𝐶𝑖+ 𝛾𝑖𝑌𝑡+ 𝛿(𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 𝑜𝑓 𝐼𝑇 𝑚𝑎𝑛𝑑𝑎𝑡𝑒𝑖𝑡) + 𝜑(𝑢𝑛𝑖𝑜𝑛 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 𝑟𝑎𝑡𝑒𝑖𝑡) + 𝜌(𝑡𝑟𝑎𝑑𝑒𝑖𝑡)

In (3) (4) (5) ) 𝛽𝑖,0𝐶𝑖 represents the effect of factors within a specific country i that do not vary over

the sample. The term 𝛾𝑖 𝑌𝑡 in (3), (4) and (5) measures common effects of income inequality in all

countries in year t. The different types of indicators of an IT mandate are discussed above and will measure the effect an IT mandate has on income inequality, this is the key variable for this paper. The term 𝜑(𝑢𝑛𝑖𝑜𝑛 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 𝑟𝑎𝑡𝑒𝑖𝑡) and 𝜌(𝑡𝑟𝑎𝑑𝑒𝑖𝑡) control for changes in labour market institutions

and increasing global trade which are considered main contributors of increasing income inequality (Card, 2001 and Feenstra & Hanson, 2008).

Results

Multiple regression have been conducted to study the effects of an IT mandate on inflation, inflation volatility and income inequality. In this part the results of these regression will be analysed and see to what extend they match the hypotheses. The analysis will be split into 2 parts, first the effects of an IT mandate on inflation and inflation volatility will be analysed. In the second part the effects of an IT mandate on income inequality will be analysed on a global level, further the countries will be divided into OECD and non-OECD and into the different types of IT mandates, as proposed by Carare and stone (2006).

(12)

The first part of regressions conducted measures the effects of an IT mandate on inflation and inflation volatility. The results in table 1 show that having an IT mandate decreases inflation with 0.41% when central bank transparency is used to indicate an IT mandate. When inflation expectations are used to indicate an IT mandate an IT mandate decreases inflation with 0.86% (table 2). These results do not differ between OECD and non-OECD countries. In OECD countries an IT mandate lowers inflation by 0.19% (table 3) and 1.01% (table 4) when central bank transparency respectively inflation expectations are used to indicate an IT mandate. In non-OECD countries an IT mandate decreases inflation by 0.29% (table 5) and 0.71 (table 6) when central bank transparency and inflation expectations are used to indicate an IT mandate. Although inflation also decreased in non IT countries (Walsh, 2009) the results show that with an IT mandate a central bank can achieve lower inflation rates than other mandates. This is most likely due to a formal target which aligns inflation expectations of current and future target rates with the actual goal of a central bank (Walsh, 2009). Indeed, Inflation expectations became more aligned with actual inflation and dropped significantly due to IT (Johnson, 2002). It was expected that there would be a difference between OECD and non-OECD countries because of the different starting point in the level of inflation. Non-OECD countries experienced much higher levels of inflation which should allow inflation to drop more in non-OECD countries (Vega & Winkelried, 2005). That this effect did not showed up in the results is most likely due to the lack of credibility and commitment of the central banks in non-OECD countries (Carare & Stone, 2006). An IT mandate has a small effect on inflation volatility. The results in table 1 show that having an IT mandate decreases inflation volatility only by 0.08 points when central bank transparency is used to indicate an IT mandate. When inflation expectations are used to indicate an IT mandate, inflation volatility decreases with 0.25 points (table 2). Separating the countries into OECD and non-OECD has, similar as with inflation, no effect on the results. In both OECD and non-OECD countries an IT mandate has a small effect on inflation volatility (table 3, 4 ,5 and 6). It was expected that an IT mandate would lower inflation volatility due to the formal alignment of an inflation target (Walsh, 2009). It was also expected that there would be a difference between OECD and non-OECD countries due to the different starting points. Non-OECD countries experienced much higher levels of inflation volatility compared with OECD countries, this should allow inflation volatility to drop more in non-OECD countries (Vega & Winkelried, 2005). That an IT mandate only has a small effect on inflation volatility in non-OECD countries, is most likely due to lacking credibility of central banks in these countries (Carare & Stone, 2006).

Reasons for the decrease in the level of inflation and inflation volatility have not been fully identified (Gali & Gambetti, 2009) but the result of this paper are more in line with the good policy- than with the good luck hypothesis. The results are in line with Ball and Sheridan (2004), Lin and Ye (2007) who

(13)

also find that IT decreases inflation. Contrary to Vega and Winkelried (2005), this paper finds that an IT mandate has little effects inflation volatility in non-OECD countries.

The second part of regressions conducted measure the effect of an IT mandate on the Gini-index. For these regressions central bank transparency and inflation expectations will again be used to indicate an IT mandate. The results in table 7 and 8 show that the effect of an IT mandate on income inequality are negligible. When central bank transparency is used to indicate an IT mandate, the Gini-index decreases with 0.08 points (table 7) and when inflation expectations are used as an indicator of an IT mandate the Gini-index decrease with only 0.01 point (table 8). The results hint towards an decrease rather than an increase of income inequality due to an IT mandate. An IT mandate could lead to lower income inequality when a central bank cuts the policy rate, for example to aggressively fight to keep inflation on target. The low policy rates induces firms to create more investments and jobs, this reduces firm’s profits in the short run and thereby income of rich household who mostly own the firms (Gornemann et al, 2014). In the long run firm’s revenue likely increases due to the investments and business owners will benefit however, when labour income increases more than profits income inequality will decrease. It was expected that an IT mandate would affect income inequality through multiple transmission channels. The portfolio channel states that inflationary actions of a central bank affect those who hold relatively more currency (Coibion et al, 2012). Erosa and Ventura (2002) find that low-income households hold relatively more currency than high-income households so when a central bank successfully effects inflation this would increase income inequality. Another transmission channel is the savings redistribution channel. This channel states that household who save will benefit, and household that borrow will lose, when the interest rate unexpectedly increases or when inflation unexpectedly decreases (Doepke & Schneider, 2006). Savers tend to be wealthier than borrowers therefore this could increase income inequality. For the savings redistribution channel to have effect, unexpected changes have to occur, however an important tool of an IT mandates is forward guidance which makes monetary policy more predictable. Not only central banks with an IT mandates make more use of forward guidance also central banks without an IT mandate increased the use this tool, this could explain why an IT mandate does not have an effect on income inequality.

Separating the countries into OECD and non-OECD doesn’t change the results, similar as for inflation and inflation volatility. The results in table 10, 11, 13 and 14 show again that an IT mandate only slightly decreases income inequality in OECD- and non-OECD countries. For OECD countries the Gini-index decreases with 0.16 (table 10) and 0.01 (table 11) when central bank transparency and inflation expectations are used to indicate an IT mandate. In non-OECD countries the decrease in the Gini-index is only 0.08 (table 13) and 0.02 (table 14) when central bank transparency and inflation

(14)

expectations are used as indicator of an IT mandate. It was expected that the effect of an IT mandate on income inequality in non-OECD countries would have been bigger because of the different starting points in the level of inflation. Non-OECD countries are more likely to start from hyperinflation levels and Bulíř (2001) finds that reducing inflation from hyperinflation levels significantly lowers income inequality however further reduction toward low levels of inflation only has a negligible effect on income inequality. Although some of these countries experienced hyperinflation levels which are brought down by an IT mandate, this decrease didn’t lead to lower income inequality. It is likely that the inflation channel is partly cancelled out by the saving distribution channel. The lack of credibility and commitment of these central banks make unexpected changes in inflation and interest rates likely causing an increase in income inequality. Therefore there is no difference between OECD and non-OECD countries as is confirmed by the results in table 10, 11, 13 and 14.

When the countries are separated into the different forms of IT as proposed by Carare and Stone (2006), a couple of minor changes occur. The effects of the purest form of IT, FFIT, on income inequality are similar as the results above. Having an FFIT mandate causes income inequality to decrease however the effect is again negligible. The Gini-index decreases with 0.12 (table 16) and 0.01 (table 17) points when central bank transparency and inflation expectations are used to indicate FFIT. Countries with an FFIT mandate clearly commit to their inflation target (Carare & Stone, 2006). A clear commitment to their inflation target makes it unlikely that unexpected changes occur in interest rate or inflation, which are necessary for the saving distribution channel to increase income inequality. It is likely that the inflation channel causes the small decrease in the Gini-index for FFIT countries. The decrease is however negligible as countries with an FFIT mandate did not start from hyperinflation levels. The results change a little for EIT and ITL mandates as due to these mandates income inequality increases slightly. An EIT mandate increases income inequality measured by the Gini-index with 0.01 (table 19) and 0.5 (table 20) points when central bank transparency and inflation expectations are used to indicate the mandate. having an ITL mandate leads to opposites effects for the indicators of IT mandate, central bank transparency increases the Gini-index with 0.61 (table 22) points and inflation expectations decreases the Gini-index with 0.01 (table 23) points. ITL mandates differ from FFIT mandates because they lack credibility and commitment to keep inflation on their target. This makes it more likely that there are unexpected changes in interest rate and inflation which cause income inequality to increase. ITL is performed by non-OECD countries and are likely to have experienced hyperinflation levels. The inflation channel than states that these countries should have decreasing income inequality when hyperinflation levels are brought down, as concluded above inflation is brought down by an IT mandate. However this effect is cancelled out by the saving

(15)

distribution channel. Central banks in ITL countries lack credibility and commitment making unexpected changes in interest rate and inflation likely which increases income inequality. An EIT mandate is performed only in OECD countries and are characterised by high credibility without full transparency and stable inflation. OECD countries did not suffer from hyperinflation levels. This makes it likely that the increase in income inequality is caused by the saving redistribution channel. EIT countries have stable inflation and so much credibility that they have the flexibility to pursue other objectives without full transparency (Carare & Stone, 2006). The flexibility to pursue other objectives could lead to unexpected changes in interest rates which could increase income inequality.

Robustness check

For a robustness check some minor changes have been made for the regressions. For the robustness check on the effects of an IT mandate on inflation and inflation volatility the indicators, central bank transparency and inflation expectations are replaced by a dummy variable. The regressions take the following form:

(6) 𝐼𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛𝑖𝑡= 𝛽𝑖,0𝐶𝑖 + 𝛾𝑖𝑌𝑡 + 𝛿(𝑇𝑖,𝑡𝐴𝑙𝑙) + 𝜃(𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛−1) + 𝜑(𝑢𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑡−1)

(7) 𝑖𝑛𝑓𝑙. 𝑣𝑜𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦𝑖𝑡 = 𝛽𝑖,0𝐶𝑖 + 𝛾𝑖𝑌𝑡+ 𝛿(𝑇𝑖,𝑡𝐴𝑙𝑙) + 𝜃(𝑖𝑛𝑓𝑙. 𝑣𝑜𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦−1) +

𝜑(𝑢𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑡−1)

In (6) and (7) 𝛽𝑖,0𝐶𝑖 represents the effect of factors within a specific country i that do not vary over

the sample. The term 𝛾𝑖 𝑌𝑡 in (6) and (7) measures common effects of inflation (6) or inflation

volatility (7) in all countries in year t. The term 𝛿(𝑇_𝑖,𝑡𝐴𝑙𝑙) will measure the effect an IT mandate has on inflation and inflation volatility. The dummy will equal one in all years a country has an IT mandate and will be zero otherwise. Lagged unemployment and lagged inflation are added as control variables. The results of these regressions are different from the ones were the indicator, central bank transparency and inflation expectations are used. This difference is mostly explained by the effect of an IT mandate on inflation in non-OECD countries. Table 28 shows that an IT mandate increases inflation with 3.48%, so the dummy variable picks up an effect that increases inflation that is mist by the indicators inflation expectations and central bank transparency. The results for OECD countries is however similar as an IT mandate decreases inflation with 0.49% (table 27).

To check for robustness in the effects of an IT mandate on income inequality two different proxies are used for income inequality and a dummy variable is added to indicate an IT mandate. The

(16)

regressions are described in the methodology. The dummy variable used to indicate an IT mandate will equal one in all years a central bank has an IT mandate and will be zero otherwise. In non-OECD countries it is found, when using s90/10 as a proxy for income inequality, that there is a somewhat larger decreasing effect of an IT mandate on income inequality (Table 13, 14 and 15). This might indicate that the inflation channel is present. The inflation channel states that the effect of an IT mandate on income inequality is bigger when inflation is brought down form hyperinflation levels. Non-OECD countries have more likely suffered from hyperinflation. The results in table 16 to 24 show that the effects for the different types of IT mandate are similar and thereby confirm that the effect of an IT mandate on income inequality are negligible.

Conclusion

As of the early 1980’s income inequality has been increasing again. Known causes are increased global trade (Feensta & Hanson, 2008), skilled-biased technological change (Berman, Bound & Machin, 1998) and changes in labour market institutions (Card, 2001). Monetary policy, especially an IT mandate, is however rarely mentioned. This paper first looked at the effects of an IT mandate on inflation and inflation volatility because the transmission channels which could cause an IT mandate to affect income inequality work through inflation and inflation volatility. With the use of a panel data regression this paper finds that an IT mandate decreases inflation in OECD and non-OECD countries. Although inflation also decreased in non IT countries (Walsh, 2009) the results show that with an IT mandate a central bank can achieve lower inflation rates than other mandates. This is most likely due to a formal target which aligns inflation expectations of current and future target rates with the actual goal of a central bank. These results are in line with Ball and Sheridan (2004) and Lin & Ye (2007) who also find that IT has a beneficial effect on inflation.

In the second part, this paper looked at the effects an IT mandate has on income inequality. Again, panel data regressions are used to study the effects of an IT mandate on income inequality. The results hint towards an decrease of income inequality rather than an increase which could happen when a central bank cuts its policy rate to, for example aggressively fight inflation. It was expected that an IT mandate would increase income inequality through the portfolio and savings redistribution channel. For the saving redistribution channel to have effect unexpected changes have to occur however an important tool of an IT mandate is forward guidance which makes monetary policy more predictable and thus lead to a negligible effect of an IT mandate on income inequality. The results are similar between OECD- and non-OECD countries. The reduction from hyperinflation levels, of which some non-OECD countries suffered, could lead to an decrease in income inequality. It is likely that

(17)

the inflation channel is partly cancelled out by the savings redistribution channel as central banks of non-OECD countries lack credibility and commitment which makes unexpected changes in interest rates and inflation likely causing an increase in income inequality.

The results do change a little between different types of an IT mandate. An FFIT mandate has a small decreasing effect on income inequality, this effect occurs most likely through the inflation channel. The effect are however minimal because FFIT countries did not suffer from hyperinflation levels. Countries with an FFIT mandate clearly commit to their inflation target making it unlikely that unexpected changes occur (Carare & Stone, 2006). Unexpected changes have to occur for the saving redistribution channel to have an effect. The results of an ITL mandate on income inequality are mixed between the indicators but the effect is small. ITL mandates lack credibility and commitment to keep inflation on their target, making it likely that unexpected changes occur. The saving distribution channel would than lead to a small increase. An EIT mandate has a small increasing effect on income inequality. This effect is most likely caused by the saving distribution channel as EIT mandates have so much credibility that they don’t have to be fully transparent. Without full transparency unexpected changes in policy are likely which is necessary for the saving distribution channel.

To summaries, an IT mandate given by a government to a central bank doesn’t have a lasting effect on income inequality. The decreasing effect of an IT mandate on income inequality when inflation is brought down from hyperinflation levels is mostly cancelled because the central bank in these countries lack credibility and commitment to their inflation target. Income inequality is also not increased by an IT mandate because for that to happen unexpected changes have to occur in interest rates and inflation. However unexpected changes are unlikely as an import tool of IT is forward guidance.

(18)

References

Ball, Laurence M., and Niamh Sheridan. "Does inflation targeting matter?." The Inflation-Targeting Debate. University of Chicago Press, 2004. 249-282.

Berman, Eli, John Bound and Stephen Machin. “Implications of Skill-Biased Technological Change: International Evidence”, Quarterly Journal of Economics, Vol. 113, no. 4 (November 1998): 1245-1280 Boivin, Jean, and Marc P. Giannoni. "Has monetary policy become more effective?." The Review of Economics and Statistics 88.3 (2006): 445-462.

Bulíř, A. (2001). Income inequality: does inflation matter?. IMF Staff papers, 48(1), 139-159. Carare, Alina, and Mark R. Stone. "Inflation targeting regimes." European Economic Review 50.5 (2006): 1297-1315.

Card, David. "The effect of unions on wage inequality in the US labor market." Industrial & Labor Relations Review 54.2 (2001): 296-315.

Carpenter, Seth B., and William M. Rodgers III. "The disparate labor market impacts of monetary policy." Journal of Policy Analysis and Management 23.4 (2004): 813-830.

Coibion, Olivier, et al. “Innocent bystanders? Monetary policy and inequality in the US”. No. w18170. National Bureau of Economic Research, 2012.

Creel, Jerome, and Paul Hubert. "Has the Adoption of Inflation Targeting Represented a Regime Switch? Empirical evidence from Canada, Sweden and the UK." Empirical Evidence from Canada, Sweden and the UK (November 17, 2008) (2008).

Davis, J. Scott, and Ignacio Presno. “Inflation targeting and the anchoring of inflation expectations: cross-country evidence from Consensus” Forecasts. Vol. 174. Federal Reserve Bank of Dallas, 2014.

Doepke, M., & Schneider, M. (2006). “Inflation and the redistribution of nominal wealth”, Journal of Political Economy, 114(6), 1069-1097.

Eijffinger, Sylvester CW, and Petra M. Geraats. "How transparent are central banks?." European Journal of Political Economy 22.1 (2006): 1-21

.

Erosa, A., & Ventura, G. (2002). “On inflation as a regressive consumption tax”. Journal of Monetary Economics, 49(4), 761-795.

Feenstra, Robert, and Gordon Hanson. “Global production sharing and rising inequality: A survey of trade and wages”. No. w8372. National Bureau of Economic Research, 2001.

Friedman, Benjamin M. "Why the Federal Reserve should not adopt inflation targeting." International Finance 7.1 (2004): 129-136.

Galí, Jordi, and Luca Gambetti. "On the sources of the great moderation." American Economic Journal: Macroeconomics 1.1 (2009): 26-57.

Ghosh, Pallab Kumar. "The Role of Unionization on the US Wage Inequality During the Period 1990-2010." Labour Economics 28 (2014): 58-69.

(19)

Gornemann, Nils, Keith Kuester, and Makoto Nakajima. "Doves for the rich, hawks for the poor? Distributional consequences of monetary policy." Federal Reserve Bank of Philadelphia (2014).

Gurkaynak, Refet S., Andrew T. Levin, and Eric T. Swanson. "Does inflation targeting anchor long-run inflation expectations? Evidence from long-term bond yields in the US, UK and Sweden." Evidence from Long-Term Bond Yields in the US, UK And Sweden (March 2006) (2006).

Gürkaynak, Refet S., et al. "Inflation Targeting and the Anchoring of Inflation Expectations in the Western Hemisphere*." Economic Review-Federal Reserve Bank of San Francisco (2007): 25.

Hu, Yifan. "The choice of inflation targeting—an empirical investigation." International Economics and Economic Policy 3.1 (2006): 27-42.

Johnson, David R. "The effect of inflation targeting on the behavior of expected inflation: evidence from an 11 country panel." Journal of Monetary Economics 49.8 (2002): 1521-1538.

Kuznets, S. (1955). Economic growth and income inequality. The American economic review, 45(1), 1-28.

Lin, Shu, and Haichun Ye. "Does inflation targeting really make a difference? Evaluating the treatment effect of inflation targeting in seven industrial countries." Journal of Monetary Economics 54.8 (2007): 2521-2533.

Mishel, Lawrence

. "

Unions, inequality, and faltering middle-class wages."Economic Policy Institute Issue Brief 342 (2012): 1-12.

Mukherjee, Bumba, and David Andrew Singer. "Monetary institutions, partisanship, and inflation targeting." International Organization 62.02 (2008): 323-358.

Romer, Christina D., and David H. Romer. Monetary policy and the well-being of the poor. No. w6793. National Bureau of Economic Research, 1998.

Rose, A. K. (2007), ‘A Stable International Monetary System Emerges: Bretton Woods, Reversed’, Journal of International Money and Finance, 26(5), 663–81.

Rother, P. (2004). Fiscal policy and inflation volatility. Working paper series no. 317 / March 2014

Samaryna, Hanna, and Jakob de Haan. "Right on target: Exploring the determinants of inflation targeting adoption." De Nederlandsche Bank working paper no.321, (2011).

Vega, Marco, and Diego Winkelried. "Inflation targeting and inflation behavior: a successful story?." International Journal of Central Banking 1.3 (2005): 153-175.

Walsh, Carl E. "Inflation Targeting: What Have We Learned?*." International Finance 12.2 (2009): 195-233.

(20)

Appendix

Table 1: The effects of an inflation target mandate on inflation and inflation volatility, central bank transparency used to indicate an IT mandate.

Coefficient (standard errors) Inflation (FE) Inflation volatility (FE) Constant 4.15 (1.83) 1.36 (0.50)

Central bank transparency -0.41 (0.16)* -0.08 (0.02)*

Inflation_1 0.61 (0.03)* x Inflation volatility_1 x 0.48 (0.03)* Unemployment_1 -0.02 (0.07) 0.05 (0.03)* R_sq 0.69 0.80 F- statistic 34.71 35.79 Number of groups 43 41 number of obs. 664 692

Balanced panel Yes Yes

* Indicates coefficient is statistically significant at 5 percent.

Table 2 The effect of an IT mandate on inflation and inflation volatility, inflation expectations used to indicate an IT mandate

Coefficient (standard errors)

Inflation (FE) Inflation volatility (FE) Constant 1.66 (1.35) 1.82 (0.42) Inflation expectations -0.86 (0.01)* -0.25 (0.06)* Inflation_1 0.10 (0.01)* X Inflation volatility_1 x 0.03 (0.01)* Unemployment_1 0.14 (0.03)* 0.18 (0.03)* R_sq 0.71 0.70 F- statistic 477.03 70.68 Number of groups 42 41 number of obs. 910 1009

(21)

Table 3: The effect of an IT mandate on inflation and inflation volatility. Central bank transparency used to indicate an IT mandate. OECD countries.

Inflation (FE)

Inflation volatility (FE)

Constant 3.63 (1.19) 1.65 (0.36)

Central bank transparency -0.19 (0.09)* -0.14 (0.04)*

Inflation_1 0.80 (0.02)* X Inflation volatility_1 x 0.51 (0.01)* Unemployment_1 -0.15 (0.04) -0.01 (0.02) R_sq 0.87 0.77 F- statistic 172 13.80 Number of groups 32 32 number of obs. 502 544

Table 4: The effects of an inflation target mandate on inflation and inflation volatility. Inflation expectations used to indicate an IT mandate. OECD countries.

Coefficient (standard errors) Inflation (FE) Inflation volatility (FE) Constant 1.66 (0.33) 1.14 (0.10) Inflation expectations -1.01 (0.01)* -0.25 (0.01)* Inflation_1 -0.02 (0.00)* X Inflation volatility_1 x 0.03 (0.01)* Unemployment_1 -0.00 (0.00) -0.00 (0.00) R_sq 0.68 0.70 F- statistic 34.59 73.19 Number of groups 32 32 number of obs. 748 838

(22)

Table 5: The effect of an IT mandate on inflation and inflation volatility. Central bank transparency used to indicate an IT mandate. Non-OECD countries.

Coefficient (standard errors) Inflation (FE) Inflation volatility (FE) Constant 1.54 (0.91) 1.49 (0.85)

central bank transparency -0.29 (0.49) -0.00 (0.22)

Inflation_1 0.24 (0.08) x Inflation volatility_1 x 0.45 (0.08)* Unemployment_1 0.51 (0.26)* 0.24 (0.11)* R_sq 0.52 0.53 F- statistic 1.51 3.67 Number of groups 11 9 number of obs. 162 148

Table 6: The effect of an IT mandate on inflation and inflation volatility. Inflation expectations used to indicate an IT mandate. Non-OECD countries.

Coefficient (standard errors) Inflation (FE) Inflation volatility (FE) Constant 4.22 (0.50) 1.48 (0.81) inflation expectations -0.71 (0.04) -0.23 (0.03)* Inflation_1 0.07 (0.04) X Inflation volatility_1 x 0.21 (0.01)* Unemployment_1 0.59 (0.14)* 0.39 (0.09)* R_sq 0.48 0.57 F- statistic 17.88 15.68 Number of groups 10 9 number of obs. 162 171

* Indicates coefficient on this dummy variable is statistically significant at 5 percent.

Table 7: The effects of an IT mandate on income inequality, central bank transparency as indicator of an IT mandate. Coefficient (standard errors) Gini (FE) S80/20 (FE) S90/10 (FE) Constant 37.14 (1.93) 4.93 (0.92) 14.36 (3.17)

Central bank transparency -0.08 (0.16) 0.04 (0.07) -0.03 (0.18)

Trade 0.01 (0.01) 0.01 (0.00)* 0.03 (0.02)**

union density rate -0.07 (0.04)* 0.05 (0.02) -0.04 (0.06)

R_sq. 0.14 0.12 0.12

F- statistic 0.99 3.29 2.24

Number of groups 43 43 42

number of obs. 566 498 443

Balanced panel Yes Yes Yes

(23)

Table 8: The effects of an IT mandate on income inequality, inflation expectations used as an indicator of an IT mandate. Coefficient (standard errors) Gini (FE) S80/20 (FE) S90/10 (FE) Constant 31.21 (1.83) 6.07 (1.06) 14.38 (7.94) Inflation expectations -0.01 (0.01) -0.00 (0.00) 0.00 (0.01) Trade 0.02 (0.01) 0.01 (0.00)* 0.05 (0.02)

union density rate -0.01 (0.02) 0.01 (0.01) -0.09 (0.07)

R_sq 0.11 0.18 0.31

F- statistic 1.03 2.08 6.51

Table 9: The effects of an IT mandate on income inequality, dummy used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (Fixed effects) Constant 34.41 (1.30) 6.32 (0.74) 19.81 (3.89) 𝑇𝑖,𝑡𝐴𝑙𝑙 -0.56 (0.43) -0.81 (0.23)* -2.29 (1.01)* Trade 0.02 (0.01)* 0.01 (0.00)* 0.03 (0.02)**

union density rate -0.04 (0.02)* -0.00 (0.01) -0.16 (0.06)*

R_sq 0.06 0.11 0.16

F- statistic 1.63 2.11 3.13

(24)

Table 10: The effects of an IT mandate on income inequality OECD countries, central bank transparency used to indicate an IT mandate.

Gini

(fixed effects) S80/20 S90/10

Constant 35.03 (3.84) 6.64 (1.85) 11.51 (3.92)

central bank transparency -0.16 (0.23) 0.02 (0.10) 0.20 (0.19)

Trade 0.01 (0.02) -0.00 (0.01) 0.01 (0.02)

union density rate -0.07 (0.07) -0.01 (0.03) -0.08 (0.07)

R_sq 0.09 0.17 0.09

F- statistic 0.70 1.16 1.57

Table 11: The effects of an IT mandate on income inequality OECD countries, inflation expectations used to indicate an IT mandate.

Coefficient (standard errors) Gini (FE) S80/20 (FE) S90/10 (FE) Constant 29.69 (2.50) 6.08 (1.42) 9.42 (8.55) Inflation expectations -0.01 (0.01) -0.00 (0.00) 0.01 (0.01) Trade 0.01 (0.01) 0.01 (0.01) 0.05 (0.03)

R_sq 0.06 0.11 0.36

F- statistic 0.76 1.60 6.43

Table 12: The effects of an IT mandate on income inequality OECD countries, dummy used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (fixed effects) S90/10 (fixed effects) Constant 30.97 (1.89) 6.04 (1.01) 15.26 (5.02) 𝑇𝑖,𝑡𝐴𝑙𝑙 0.82 (0.58) -0.15 (0.29) -0.62 (1.30) Trade 0.01 (0.01) 0.01 (0.01) 0.04 (0.03)

union density rate 0.01 (0.03) -0.00 (0.01) -0.12 (0.07)**

R_sq 0.15 0.07 0.19

F- statistic 0.71 1.10 3.14

(25)

Table 13: The effects of an IT mandate on income inequality non-OECD countries, central bank transparency used as an indicator of an IT mandate.

Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) Constant 44.67 (2.03) 10.24 (1.31) 39.19 (6.29)

central bank transparency -0.08 (0.25) -0.13 (0.15) -1.15 (0.44)*

Trade 0.01 (0.01) 0.02 (0.01)* 0.10 (0.05)*

union density rate -0.10 (0.05) 0.02 (0.03) -0.32 (0.16)*

R_sq 0.29 0.23 0.42

F- statistic 1.18 2.81 2.16

Table 14: The effect of an IT mandate on income inequality non-OECD countries, inflation expectations used to indicate an IT mandate.

Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) Constant 38.74 (3.77) 11.19 (2.71) 19.28 (10.82) Inflation expectations -0.02 (0.03) -0.02 (0.02) 0.00 (0.08) Trade 0.01 (0.01) 0.02 (0.01)* 0.13 (0.06)*

union density rate -0.08 (0.04)** -0.02 (0.03) -046 (0.17)*

R_sq 0.19 0.18 0.36

F- statistic 1.30 1.41 2.01

** Indicates coefficient is statistically significant at 10 percent

Table 15: The effects of an IT mandate on income inequality non-OECD countries, dummy used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) Constant 41.42 (2.00) 9.07 (1.45) 20.67 (6.69) 𝑇_𝑖,𝑡𝐴𝑙𝑙 -1.20 (0.85) -1.50 (0.61)* -4.08 (1.96)* Trade 0.02 (0.01) 0.01 (0.01) 0.07 (0.04)

union density rate -0.10 (0.03) -0.02 (0.02) -0.30 (0.11)*

R_sq 0.19 0.28 0.26

F- statistic 2.87 1.81 1.21

(26)

Table 16: The effects of an FFIT mandate on income inequality, central bank transparency used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) Constant 43.53 (3.82) 15.01 (3.94) 21.36 (9.49)

central bank transparency -0.12 (0.21) 0.11 (0.15) -0.16 (0.36)

Trade -0.01 (0.03) -0.02 (0.02) 0.02 (0.06)

union density rate -0.14 (0.07) -0.13 (0.07) -0.46 (0.17)*

R_sq 0.11 0.19 0.17

F- statistic 1.60 1.69 2.92

Table 17: The effects of an FFIT mandate on income inequality, inflation expectations used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) Constant 34.53 (2.59) 13.85 (2.91) 19.32 (15.66) inflation expectations -0.01 (0.01)* -0.00 (0.01) 0.00 (0.02) Trade 0.01 (0.01) 0.00 (0.02) 0.04 (0.08)

union density rate -0.02 (0.03) -0.09 (0.04)* -0.24 (0.18)

R_sq 0.15 0.21 0.16

F- statistic 1.39 1.63 4.65

(27)

Table 18: The effects of an FFIT mandate on income inequality, dummy used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) constant 34.67 (2.30) 11.88 (2.54) 21.07 (13.68) 𝑇𝑖,𝑡𝐴𝑙𝑙 0.69 (0.66) -0.14 (0.61) -0.16 (2.65) Trade 0.01 (0.02) 0.01 (0.02) -0.01 (0.09)

R_sq 0.14 0.17 0.21

F- statistic 1.38 1.41 2.93

Table 19: The effects of an EIT mandate on income inequality, central bank transparency used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) constant 43.53 (3.82) 3.45 (2.51) 7.51 (4.66)

Central bank transparency 0.01 (0.57) -0.15 (0.16) 0.07 (0.32)

Trade -0.00 (0.03) 0.01 (0.01) 0.01 (0.01)

union density rate -0.01 (0.20) 0.09 (0.05) 0.35 (0.09)*

R_sq 0.14 0.17 0.15

F- statistic 0.48 1.42 1.48

(28)

Table 20: The effects of an EIT mandate on income inequality, inflation expectations used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) constant 27.56 (7.62) 2.98 (2.12) 9.10 (3.79) Inflation expectations 0.50 (0.02)* 0.04 (0.07) 0.23 (0.15) Trade -0.03 (0.02) 0.00 (0.01) 0.01 (0.01)

union density rate -0.13 (0.13) 0.01 (0.04) 0.09 (0.07)

R_sq 0.19 0.22 0.21

F- statistic 0.81 1.59 1.68

Table 21: The effects of an EIT mandate on income inequality, dummy used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) constant 28.63 (4.14) 3.85 (1.17) 7.26 (2.55) 𝑇𝑖,𝑡𝐴𝑙𝑙 0.36 (1.23) -0.43 (0.33) -0.16 (0.74) Trade 0.00 (0.02) 0.00 (0.00) 0.01 (0.01)

union density rate 0.10 (0.08) 0.05 (0.02)* 0.07 (00.01)**

R_sq 0.19 0.18 0.19

F- statistic 0.76 1.66 1.48

(29)

Table 22: The effects of an ITL mandate on income inequality, central bank transparency used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) constant 37.01 (6.85) 7.13 (3.61) 24.34 (9.17)

Central bank transparency 0.61 (0.42) 0.03 (0.20) -0.08 (0.50)

Trade -0.05 (0.06) 0.01 (0.03) -0.10 (0.08)

R_sq 0.14 0.27 0.25

F- statistic 0.95 0.94 0.64

Table 23: The effect of an ITL mandate on income inequality, inflation expectations used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) constant 38.86 (6.78) 5.31 (2.61) 23.04 (6.91) Inflation expectations -0.01 (0.08) -0.09 (0.03)* -0.18 (0.08)* Trade -0.04 (0.07) 0.05 (0.03) -0.04 (0.09)

R_sq 0.34 0.47 0.41

F- statistic 0.97 1.75 1.16

Table 24: The effect of an ITL mandate on income inequality, dummy used to indicate an IT mandate. Coefficient (standard errors) Gini (fixed effects) S80/20 (Fixed effects) S90/10 (fixed effects) constant 41.80 (7.44) 4.32 (3.64) 15.30 (9.60) 𝑇_𝑖,𝑡𝐴𝑙𝑙 1.48 (2.16) 1.10 (1.03) -0.77 (2.71) Trade -0.02 (0.05) 0.02 (0.03) -0.07 (0.06)

R_sq 0.20 0.32 0.33

F- statistic 1.61 1.25 0.98

(30)

Table 25: List of countries used in the data.

# country name IT Type OECD

start of IT

1 Argentina NO No -

2 Australia * FFIT Yes 1993

3 Belgium EIT Yes 1999

4 Brazil FFIT No 1999

5 Bulgaria NO No -

6 Canada * FFIT Yes 1991

7 Chile * FFIT Yes 1999

8 China * NO No -

9 Colombia FFIT No 1999

10 Czech Republic FFIT Yes 1998

11 Denmark EIT Yes 1999

12 Estonia EIT Yes 2011

13 Finland EIT Yes 1993

14 France EIT Yes 1999

15 Germany EIT Yes 1999

16 Hungary FFIT Yes 2001

17 Iceland FFIT Yes 2001

18 Indonesia * ITL No 2005

19 Ireland EIT Yes 1999

20 Israel * FFIT Yes 1997

21 Italy EIT Yes 1999

22 Japan EIT Yes 1998

23 Korea * FFIT Yes 1998

24 Latvia EIT Yes 2014

25 Lithuania EIT No 2015

26 Luxembourg EIT Yes 1999

27 Mexico FFIT Yes 2001

28 Netherlands EIT Yes 1999

29 New Zealand FFIT Yes 1990

30 Norway FFIT Yes 2001

31 Peru ITL No 2002

32 Poland FFIT Yes 1999

33 Portugal EIT Yes 1999

34 Russian Federation ITL No 2006

35 Slovenia ITL Yes 2007

36 Spain EIT Yes 1995

37 Sweden FFIT Yes 1993

38 Switzerland * EIT Yes 2000

39 Taiwan NO No -

40 Thailand FFIT No 2000

41 Turkey * FFIT Yes 2006

42 UK FFIT Yes 1992

43 Uruguay ITL No 2007

44 US EIT Yes 2003