Growth and Inequality (Macro Perspectives)
This article provides a review of studies that examine the relationship between economic growth and inequality. These studies are divided into two groups. The first emphasizes the channels through which inequality affects growth, while the second emphasizes the opposite channel, where economic growth affects inequality. Although several empirical studies find a significant correlation between inequality and growth, it is still an open question as to whether the correlation is driven by the first or the second channel.
KeywordsConstitutions, economic approach to Creative destruction Distortionary taxes Economic growth Education Endogenous growth theory Entrepreneurship Equity–efficiency trade-off Expropriation Growth and inequality Human capital Income smoothing Increasing returns Inequality Innovation Kuznets, S. Policy reform Redistribution of income and wealth Skill-biased technical change Wage inequality
JEL ClassificationsD4 D10
The study of ‘growth and inequality’ has a long tradition. One well-known relationship in economic development is the Kuznets curve. Kuznets observed that in the early stage of human development, when agriculture was the main economic activity, inequality in the distribution of income was relatively low. As the economy industrialized and the workforce moved towards industry and away from agriculture, the distribution of income tended to widen. At some critical point in the economy’s development, this tendency reversed. Although more recent evidence does not support the Kuznets curve hypothesis (see, for example, the widening income inequality observed in the United States starting in the early 1980s), the original empirical finding of Kuznets (1955) stimulated a large body of research activity.
Other evidence of the relationship between inequality and growth comes from more recent cross-country observations. Data collected during the 1980s and 1990s show that there is a great deal of variation across countries in the degree of income inequality and economic growth. Are the different growth rates related to the degree of inequality of each individual country? Several studies find, indeed, that there is a cross-country negative correlation between inequality and growth, that is, countries with greater income inequality tend to experience slower growth; see Benabou (1996) and Perotti (1996) for a review of the empirical studies. But correlation does not imply causation, and there are good reasons to think that the causation can go in both directions. In other words, slow growth could generate greater inequality and equality could lead to faster growth.
Inequality Affecting Growth
One of the channels through which inequality affects growth is through the political and institutional system. A new series of studies in the 1980s, pioneered by Romer (1986) and Lucas (1988), developed a new class of models in which government policies could have a significant impact on the long-term growth of the economy (endogenous growth models). Given the importance of government policies for long-term growth, it becomes important to understand the forces and mechanisms underlying the choice of policies. This work stimulated a new series of studies in political economy. These studies start from the observation that, in a democratic society, the fundamental mechanism underlying the choice of policies is the electoral system. Therefore, in order to understand how policies are selected, we need to study the policy preferences of the population and how these preferences are translated into voting preferences.
Many factors affect the voting preferences of a society. However, for policies that have a clear redistributive content, the position of the voter in the distribution of income or wealth plays an important role. If a person is poor, his or her tax payments are smaller than the benefits he or she receives from government expenditures. Consequently, his attitude towards redistributive policies is more favourable than someone at the top of the income distribution (he or she has to pay more taxes than the received benefits). If the distribution of income is very unequal, then there will be many voters favouring larger governments. Of course this would not be a problem for efficiency if taxes were not distortionary. But, in a standard endogenous growth model, taxes have a negative impact on investment and growth. Therefore, the main conclusion of this literature is that inequality impairs the economic potential of a country because voters will demand more redistribution through distortionary taxes; Persson and Tabellini (1994), Alesina and Rodrik (1994), Krusell and Rìos-Rull (1996), Krusell et al. (1996) are some examples.
These studies also demonstrate the importance of the institutional system. Although greater inequality implies a greater demand for redistributive policies, the way political preferences are aggregated and the way policies are ultimately chosen depend on the particular institutional framework. For example, whether the representative democracy works through a parliamentary or a presidential system could lead to different sizes of government and, through distortionary taxes, to different levels of economic growth; see Persson and Tabellini (2005) for an analysis of the economic effects of constitutions.
The predictions of the politico-economic literature are consistent with several empirical studies as they find a negative relation between inequality and growth. However, a deeper empirical investigation of this channel poses some doubts. More specifically, the politico-economic channel can be divided into two sub-channels: a positive relation between ‘inequality’ and ‘redistributive policies’ and a negative relation between ‘redistributive policies’ and ‘growth’. Perotti (1996) shows that the negative effect of redistributive policies on growth is not a robust feature of the data. On the contrary, redistributive policies may even be positively associated with economic growth. How is this possible?
Several theories envision a beneficial effect of redistributive taxes. The key ingredient is the presence of financial constraints. Let us take the Shumpeterian view that entrepreneurship is central to economic growth. However, due to financial constraints and the lack of insurance markets, entrepreneurial investment is suboptimal. Under these conditions, redistribution may provide extra resources to constrained entrepreneurs and could facilitate more investments in growth-enhancing activities. At the same time, a redistributive system provides an implicit system of income smoothing (a person pays high taxes when he or she earns high profits but receives payments in case of losses), and therefore, it provides insurance. If entrepreneurs are risk averse, this encourages more investment. The issue of whether redistributive taxes increase or decrease entrepreneurial investment is still an open area of research.
A similar story applies to investment in education or human capital. If education is important for economic growth, but because of financial constraints households choose sub-optimal levels of education, then government transfers may allow for greater investment and growth. A more direct effect could be generated by financing public education, as in Glomm and Ravikumar (1992). Examples of studies that emphasize the importance of inequality for growth in the presence of financial constraints are Galor and Zeira (1993), Banerjee and Newman (1993) and Aghion and Bolton (1997).
Another group of studies emphasizes social conflict and expropriation. Greater inequality means that a larger group of individuals is at the bottom of the distribution and faces poor economic conditions compared to the rest of the population. Faced with poor economic conditions, people have strong incentives to expropriate either by ‘stealing’ or through ‘revolutions’. The risk of expropriation has two negative effects. First, it acts as an investment tax that discourages investment. Second, more resources are devoted to protect property rights, which detracts from resources devoted to productive and growth enhancing activities. An example of this kind of theory is Benhabib and Rustichini (1996).
Another theory of inequality affecting growth is that developed in Murphy et al. (1989). This theory assumes that there are technologies with increasing returns. These technologies become profitable only if the domestic market is sufficiently large, that is, there is a large demand for the goods produced with the new technologies. If wealth is highly concentrated the domestic market remains small (since there are not enough consumers who can afford these goods). As a result, these growth-enhancing technologies will not be implemented. However, the theory finds weak support in the data (see Benabou 1996).
Growth Affecting Inequality
If we take the view that growth requires innovative risky activities and these activities cannot be easily insured, we would expect that faster growth is associated with greater ex post inequality. At the same time, a faster rate of innovation implies greater destructions of monopoly positions (creative destruction). This would generate lower inequality because the monopoly positions, which are the source of high-income revenues, last for a shorter period of time. Therefore, it is not obvious whether faster innovation and growth create greater inequality. However, within this environment, faster growth generates higher mobility due to a higher turnover in the holding of monopoly positions. Therefore, even if growth leads to greater inequality, it also creates a healthier social environment. Long-term growth requires technological innovation and there is no doubt that new technologies affect different groups in different ways. Therefore, growth and inequality are intrinsically related. Since 1980, wage inequality among different education groups has been widening in almost all industrialized countries. Katz and Murphy (1992) show that this increase is due to a raising demand of skilled labour. Krusell et al. (2000) propose an explanation for the increasing demand of skilled labour based on the introduction and development of new technologies that are more complementary to skilled labour (skill-biased technologies).
Suppose that there are two types of workers, skilled and unskilled. The stocks of skilled and unskilled workers change slowly over time. Now suppose that there is the introduction of skill-biased technologies, that is, technologies that require more skilled labour than unskilled labour. This will lead to an increase in the demand for skilled workers. Given the limited increase in the supply, the wages of skilled workers will increase. On the other hand, the demand for unskilled workers will decline, which leads to a fall in the wages of these workers.
This is a compelling explanation for the increasing wage premium started at the beginning of the 1980s. However, it raises the question of why the ratio of skilled versus unskilled workers has not increased that much during this period, certainly not as much as we would expect given the size of the wage premium change.
The technological innovations introduced in the 1970s seem to have affected the economy in other respects. Greenwood and Jovanovic (1999) and Hobijn and Jovanovic (2001) believe that new information technologies required a level of restructuring that incumbent firms could not face. As a result, their stock market value dropped. This is another form of redistribution in the sense that the owners of incumbent firms lose market value to the owners of the new firms.
Whether we concentrate on the first channel of causation – in which growth affects inequality – or to the second – in which inequality affects growth – there are no obvious policy recommendations. If we think that inequality has a negative impact on growth because society demands more redistributive policies (as in the standard political economy literature), then the constitutional system of electoral representation becomes central. Changing the constitutional system could lead to different political outcomes. However, changing the constitutional system is not easy. We could also think of reallocating resources once and for all to change the initial distribution. Although this is possible in theory, it is difficult from a political point of view.
If we concentrate on the opposite channel, in which growth impacts on inequality, and we are concerned about having an excessively unequal distribution of incomes, then we may consider possible redistributive policies. However, these policies may also have undesired effects on efficiency. If the tax system keeps the after-tax skill premium low, the incentive to acquire skills will be lower. But, because of skilled-biased technologies, more skills are required. This could also discourage the introduction of these technologies, which would impact negatively on growth. The equity–efficiency trade-off becomes central to the analysis. The positioning of a society in this trade-off will depend on society preferences about the degree of inequality that is socially acceptable. These preferences are based on individual beliefs that are likely to depend on individual experiences and they change very slowly over time. The relationship between personal experience and beliefs is formalized in Piketty (1995); see also Quadrini (1999).
- Benabou, R. 1996. Inequality and growth. In NBER macroeconomics annual, ed. B.S. Bernanke and J. Rotemberg. Cambridge, MA: MIT Press.Google Scholar
- Kuznets, S. 1955. Economic growth and income inequality. American Economic Review 45: 1–28.Google Scholar
- Persson, T., and G. Tabellini. 1994. Is inequality harmful for growth? American Economic Review 84: 600–621.Google Scholar
- Persson, T., and G. Tabellini. 2005. The economic effects of constitutions. Cambridge, MA: MIT Press.Google Scholar