نابرابری درآمد و چرخه کسب و کار : روش هم انباشتگی حد آستانه

This paper investigates the impact of various socio-economic variables on various cohorts of the income distribution. We use asymmetric cointegration tests to show that unemployment and immigration shocks have real impacts on income inequality. In addition, using threshold test results we are able to show that positive and negative shocks to the economy do not have symmetric effects nor do the impacts of these shocks impact income quintiles uniformly.

Beginning with the seminal work of Kuznets (1955) many researchers have endeavored to investigate the nature of the relationship between economic growth and income inequality. The Kuznets hypothesis posited that the functional relationship between inequality and economic development had an inverted “U” shape. Kuznets speculated that inequality would initially be positively correlated with economic development but that the relationship between economic growth and inequality would become negative at higher levels of development. Results supporting this hypothesis typically come from the use of cross-sectional country-specific data. Some recent researchers dispute the Kuznets hypothesis such as Bruno et al. (1998). Still others, such as Blinder and Esaki (1978), have found results that support the basic premise of the Kuznets hypothesis but expand on how inflation and unemployment factor into inequality. What makes this area of research so inviting is that key to the debate over the relationship between growth and inequality is the question of what impact growth has on citizens throughout the entire income distribution. Most research in this area investigates the impact that economic development, namely GDP, has on some standard measure of income inequality, such as the Gini coefficient. What cannot be gleaned from these endeavors is whether a growing economy is helping all segments of society or is just helping certain subsets of society. If economic growth is found to be positively correlated with income inequality it is not clear if the increase in income inequality is being caused by an increase in the incomes of those in the highest quintile of the income distribution while those in the bottom quintile have incomes that are stagnant or falling. It could also be the case that while incomes are growing for all income quintiles, the income growth of the top quintile is greater than the income growth for other quintiles. These are empirical questions that cannot be answered by simply regressing a standard Gini coefficient against GDP. Our paper will go beyond what has been done previously in that we will be able to investigate the various degrees of impact that economic growth, unemployment, and to a lesser extent, immigration have on different cohorts of the income distribution. This paper has the advantage of using a time series that has nearly 60 years of reliable U.S. data. Specifically, we consider whether economic upturns have a different impact on income inequality than economic downturns. Asymmetric behavior over the business cycle has attracted considerable attention in the last decades. Neftci (1984) showed that several measures of U.S. unemployment display asymmetric adjustment over the course of the business cycle. Focusing on the asymmetric behavior of unemployment rates over the business cycle, Rothman (1991) showed that the primary source of asymmetry is the cyclical behavior of the unemployment rate in the manufacturing sector. Acemoglu and Scott (1994) have also shown asymmetries in the cyclical behavior of UK labor markets. Harris and Silverstone (2001) and Silvapulle et al. (2004) tested asymmetric adjustment in specifications of Okun's law. They found that the short-run effects of positive cyclical output on cyclical unemployment are quantitatively different from those of negative ones; as such, the relationship between labor market indicators and aggregate income is asymmetric. 1.1. Previous research and our model Typically there have been two approaches to this type of research. One approach uses a cross-section of countries with varying levels of economic growth and income inequality to investigate the impact of economic growth on country level inequality. Another approach is to take a consistent time series for one country (typically the United States) to analyze how growth over time has impacted income inequality without regard to redistributive policy measures such as taxes or social programs. What we propose in this paper is more akin to the latter methodology. Our model is similar to that of Blinder and Esaki (1978) and Bishop et al. (1994) with some key differences. Both papers use an extensive time series, although the one used in our paper is the longest, by far, and covers the period from 1948 to 2003. (See Perman and Stern (2003) for an excellent review of the issues and techniques related to this type of investigation.) Blinder and Esaki (1978) used income quintiles as left hand side variables with a time series of unemployment and inflation as explanatory variables. Their finding that inflation is progressive along the income distribution has been supported others such as Blank and Blinder (1986). Later work by Bishop et al. (1994) includes several controls for demographic, structural, and economic variables that might also affect movements along the income distribution. They also recognized that the possibility of random walks and cointegration should be addressed. The time series they used is relatively short and they did not include a measure of income inequality, such as the Gini coefficient to test the validity of the Kuznets hypothesis. Research, such as that done by Chen (2003), Gomez and Foot (2002), and Adams (2003), does test the Kuznets hypothesis with fairly reliable country-level cross-sectional data. Their results are complementary to Kuznets but fail to account for the variance in responses of the different quintiles along the income distribution. Income is being impacted by growth in their model but they cannot say whether those in the higher quintiles are being impacted more, less, or the same as those income earners in the lowest quintile. There has been some research on the impact of growth on after-tax inequality. Hayes et al. (1991) examines impacts on the income distribution after policy initiatives, such as redistributive taxes, have been implemented. More recent work by Lundberg and Squire (2003) points out that the impact of policy initiatives on inequality and on growth should share some common characteristics and as such should be studied simultaneously. They report that by doing so it is shown that the determinants of growth and inequality are not mutually exclusive. Theoretical work on the relationship between business cycles and income inequality fails to account for the increase in inequality and wealth concentration in the United States. Castañeda et al. (1998) examined the extent to which unemployment spells and cyclically moving factor shares account for the behavior of income inequality over the business cycle. While their model somewhat accounts for income inequality business cycle dynamics, it does not account for wealth concentration. Dolmas et al. (2000), and Albanesi (2007) model inflation in a public choice framework whereby political conflicts lead to inflation; as such, agent heterogeneity allows for the modeling of monetary policy as a function of inequality. The results will depend on the political powers of agents. Heer and Sussmuth (2003) analyze the effects of a permanent change in inflation on the distribution of wealth and find a significant relationship between inequality and inflation for the U.S. economy. Specifically, higher inflation leads to higher nominal interest rates and a higher real tax burden on interest income. An increase in inflation results in a lower stock market participation rate and an increase in wealth inequality. Focusing on cross-country correlations between inflation unemployment and income inequality, Romer and Romer (1999) find a strong relationship between unemployment and poverty, and no clear relationship between inflation and poverty. Even though the literature is suggestive of potential variables that are associated with worsening of income inequality, the dynamic interaction between movement in the income distribution and the business cycle have not been explored in the literature. The primary objective of this paper is to investigate the dynamic interactions between income inequality and measures of the business cycle. We also consider a host of socio-economic variables such as immigration, imports from less developed countries, annual inflation, female headed households, labor force participation rate by females, total transfers, the share of services in GDP, service employment as share of total employment, and productivity. After identifying significant variables, we test whether the relationship between business cycles and income inequality is asymmetric. This issue is germane since the relationship between labor market indicators and aggregate income has been found to be asymmetric. Since the principal sources of income, particularly for persons in the lower quintiles of the income distribution, are wages and salaries, it is natural to examine whether the relationship between business cycles and the income distribution is asymmetric. Using threshold and momentum models of cointegration developed by Enders and Granger (1998) and Enders and Siklos (2001), we test whether economic expansions have a different impact on the income distribution than economic contractions. Since income inequality also can affect the business cycle via the propensity to spend at the lower quintiles of the income distribution, there can be feedback effects between the income distribution and the business cycle. Impulse response functions based on the Vector Error Correction Model are suitable to analyze such feedback effects. To preview our results, we find asymmetric adjustment between the quintiles comprising the U.S. income distribution and business cycle measures over the last 50 years. Particularly, increases in unemployment cause increases in income inequality, but negative shocks to unemployment have only short-lived positive benefits on income inequality. In what follows, we spell out our methodology and data. Section 3 presents empirical results and Section 4 concludes the paper with a discussion of policy implications.

The interaction of the business cycle and income inequality is of great importance to academics and policy makers alike. If it is true that “a rising tide lifts all boats” then policies designed to grow the economy as fast and as vigorously as possible may be viewed as desirable. This type of policy ensures that all segments of the income distribution are enhanced. However, if different cohorts along the income distribution do not behave uniformly to shocks to economy it may be necessary to derive a more extensive policy instrument. In this paper we show that asymmetries do exist among the quintiles comprising the U.S. income distribution over the last 50 years. Of particular note, this paper finds that increases in unemployment causes increases in income inequality but that negative shocks to unemployment have only short-lived positive benefits to income inequality. This is a new and important finding. In addition, this work shows that shocks to unemployment have impacts that are not uniform to cohorts along the income distribution. In particular, those individuals with the lowest mean family income (sorted by quintile) are most adversely affected by shocks to unemployment but are quickest to return to the steady state. The paper also shows that immigration is an important factor in explaining changes in inequality. Increases in immigration lead to significant increases inequality but it should be noted that we do not distinguish between country of origin of incoming groups. What is principally important to note in this work is that policies designed to increase the well-being of the entire range along the income distribution should be tailored and exact since it is clear that persons in different income quintiles have distinctly different reactions to business cycle changes.