Innovation and income inequality12

Preparing to load PDF file. please wait...

0 of 0
100%
Innovation and income inequality12

Transcript Of Innovation and income inequality12

Innovation and income inequality12
Cristiano Antonelli* Agnieszka Gehringer**
*Dipartimento di Economia, Università di Torino & BRICK Collegio Carlo Alberto. **Volkswirtschaftliches Seminar, Lehrstuhl für Wirtschaftspolitik, Georg-August-Universität Göttingen
Abstract. The paper articulates and tests the hypothesis that innovation is a major factor in the reduction of income inequalities. The relationship between the pace of technological change and the dynamics of income inequalities has been first suggested by Kuznets (1955), but found little elaboration and empirical investigation in the subsequent literature. The evidence of a large data set including advanced countries, such as the US, Canada and the members of the European Union, as well as the newly industrializing BRIC members, in the years 1995-2011, confirms the virtuous circle between technological change and income inequalities.
JEL classification: N30, O33
Keywords: Income inequality; Innovation; Trade openness; Economic growth.
1 The authors acknowledge the financial support of the European Union D.G. Research with the Grant number 266959 to the research project ‘Policy Incentives for the Creation of Knowledge: Methods and Evidence’ (PICK-ME), within the context of the Cooperation Program / Theme 8 / Socio-economic Sciences and Humanities (SSH), in progress at the Collegio Carlo Alberto and the University of Torino.

1. Introduction The paper elaborates and tests the hypothesis that technological change is a powerful factor that reduces
income inequality. Technological change helps reducing income inequalities for two reasons. First, because it magnifies the rates of economic growth and hence the increase of wage levels complementing the traditional hypothesis that economic growth reduces income inequalities. Second, due to the dynamics of market rivalry based upon innovation, the higher the price competition of factor and product markets and the lower the accumulation of rents and hence the increase of income inequalities. At the same time, however, it should be kept in mind that income distribution affects the rates of technological change: lower levels of income inequality increase the incentives and opportunities of increasing human capital and hence a new pool of knowledge externalities sustaining innovative activity is freed up for the economic system.
The analysis of the causes and effects of the distribution of income and specifically of income inequality is a long discussed issue in economics. Large empirical evidence discusses episodes of increasing income inequality both in advanced and industrializing countries since the end of the XX century (Aghion, Caroli, García-Peñalosa, 1999; Kaplan and Rauh, 2010). The raising levels of income inequality have called increasing attention especially to try and understand its determinants. Among other determinants, the past literature has focused on the direct effects of the output growth. Quite surprisingly little attention has been paid to appreciating the role of indirect growth effects and most importantly, of technological change, as the main source of economic progress.
The rest of the paper is structured as it follows. Section 2 explores the relations between technological change and income distribution and articulates the hypothesis that the rates of introduction of technological and organizational innovations have significant impact in reducing income inequality. Section 3 presents the empirical evidence and the results of econometric investigations that confirm the negative relationship between the rate of technological change and the levels of income inequality in an inclusive data set comprising all the EU countries, USA, Canada, Japan, Turkey, Croatia, Island, Norway, Switzerland and the BRIC countries (Brazil, Korea, China, India). The conclusions section summarizes the results and elaborates policy implications.
2. Technological change and income distribution
2.1 The standard view over economic growth and income inequality
After more than fifty years, the pathbreaking contribution of Simon Kuznets (1955, 1963) remains the basic reference in the economics of income inequalities, for many reasons. First and most important reason has to do with the identification of the issue: income distribution is a relevant aspect of the economic structure of every economic system. Second, it is not static, but intrinsically dynamic as it keeps changing simultaneously throughout economic history and across countries (Aghion, Caroli, García-Peñalosa, 1999). Third, it is not exogenous but, quite on the contrary, the endogenous consequence of the structure and dynamics of the economic system. Fourth, and more specifically, income distribution stems from the distribution and remuneration of three well distinct factors: capital, labor and skills.
2

Income distribution, in other words, should be investigated over two dimensions, the functional one and the personal one. The first one concerns the distribution of income between the main factors of production, i.e. capital and labour. It depends upon the distribution of profits and dividends paid to capital owners as well as of wages and wage premiums paid to the levels of human capital embedded into skilled labor. The second one pertains the distribution of income among households (or individuals), irrespective of the source of income.
Building upon these pillars, Simon Kuznets elaborates an interpretative framework, where income distribution reflects the changing distribution of wealth and skills and their changing prices with respect to standard labor that takes place along economic growth. While the methodological foundations of the analysis are still valid, the specific contents of his analysis are quite dated by now. Kuznets, indeed, paid much attention to the changing structure of the economic systems at the time of industrialization, with the rapid shift away from an agricultural and rural economy into an urban and industrial one.
On such a historical background he noted that, in the early phases of industrializations, income inequality increases because of the large differences in factor productivity between rural and urban activities. Afterwards, however it eventually declines with the completion of the industrial transformation. Once the full system has been able to complete the industrial transformation the standard dynamics of economic growth favors the reduction of income inequality along the following chain of factors: 1) savings increase the supply of capital 2) and decrease the levels of interest rates; 3) capital intensity increases, and 4) labor productivity increases (with the appropriate supply of complementary skilled workforce), 5) leading to higher wages that make possible a larger supply of savings; 6) moreover, in an advanced industrial economy, tighter competition in product and factor markets makes it possible to minimize monopolistic profits. Figure 1 synthesizes the working of the mechanism that relates economic growth to the reduction of income inequality. Following this chain of factors it appears clearly that the standard mechanism of economic growth reduces income inequality by means of the decrease of interest rates, the reduction of monopoly profits and the increase of wages. Hence, the famous inverted U-shaped relationship between income inequality and revenue per capita, initially suggested by Kuznets (1963), acquires a fresh understanding.
3

labour productivity ↑ real wages ↑

income inequality ↓ income per capita ↑

relative capital intensity ↑

savings ↑

real interest rates ↑
asymmetry in wealth distribution ↓
income inequality ↓

capital ↑

Figure 1 Economic growth and income inequality in the standard view.
The empirical literature provides contrasting evidence on the issue. The inverted U-shaped relationship between the stage of development and income inequality finds only partial support and seems to be quite sensitive to the composition of the datasets and the time periods analyzed (Adelman and Robinson, 1989; Atkinson and Piketty, 2007 and 2009). Dollar and Kraay (2002) conclude their inclusive study, extended to all the countries of the Penn World Tables, that there is no systematic evidence confirming economic effects on the income distribution. There is, instead, converging evidence about the positive effects of income growth on the reduction of income inequality, especially in developing countries (Adams, 2002; Chen and Ravallion, 2001; Ravallion, 1995). The evidence about advanced countries is on the opposite mixed, especially, when the last decade of the XX century is considered (Atkinson and Pikkerty, 2009). As a matter of fact, Kaplan and Rauh (2010) show that income inequality has increased in the recent past at a time of fast economic growth.
The theoretical literature has shared the basic intuitions that support Kuznets analysis. By articulating and expanding his basic hypothesis, past theoretical contributions show consensus on the fact that in the long-run the economic growth process reduces income inequality. An emphasis here was put towards the appreciation of the role of equilibrium conditions. Accordingly, the closer are the working of product and
4

factor markets- including financial markets and all considered in the national and global dimension - to competitive equilibrium and the lower the income inequalities. When economic growth is associated with an increase of market imperfections, income asymmetry would actually increase.
These results stem directly from the appreciation of the factors that account for economic growth in the standard framework. According to traditional growth theory, in fact, economic growth is engendered by the increased availability of capital via the accumulation of savings and the consequences in terms of lower interest rates and increased capital intensity of production processes. All imperfections of financial markets have strong negative effects on the correct allocation of resources favoring inefficiency that discriminate income distribution in favor of the wealthy (Aghion and Bolton, 1992). All measures that remove frictions from the financial markets and, hence, improve the availability and efficiency of capital, for given levels of savings, are likely to favor the reduction of interest rates, and finally, the increase of capital intensity. The consequences for income distribution in terms of reduction of income inequality are straightforward and strongly complementary because the reduction of interest rates shrinks the effects of the possible -actually frequent- asymmetries in the distribution of wealth that are, in relative terms further reduced by the increase in the wage levels (Beck, Demirgüç-Kunt, Levine, 2007).
Within the same interpretative framework it is clear that the increase in competition both in domestic and international product markets is likely to affect positively the rates of economic growth. This works through the positive effects in terms of both more rational allocation of resources, improved division of labor, as well as specialization.
The exposure to international competition is especially effective to reduce and sometimes overcome the barriers to entry and to mobility. Such barriers limit competition in domestic markets, and consequently, the overall growth dynamics. Greater foreign competition, visible through the increase in imports, makes markups decline and, thus, also profit margins fall (Chen et al., 2009). The relationship has been found effective both for exports and imports. Large empirical evidence confirms the positive relationship between the share of imports to GDP and competition (eg. Mac Donald, 1996). For the same token, the larger are the shares of exports to GDP and the higher the levels of international competiveness and the closer the conditions of product markets to the standards of workable, if not perfect competition. It is consequently clear that the larger are levels of the openness to trade - as measured in terms of the share of imports and exports to GDP - the closer the levels of prices to minimum average costs and the lower the levels of markups and quasi rents. Low levels of mark-ups and quasi-rents insure that the distribution of income is close to competitive levels, with capital and labour remunerated on their marginal productivities. Firms, and consequently firm owners, cannot accumulate profits. In financial markets, interest rates are less inflated by profit margins. It becomes evident - following this chain of arguments- that the larger are levels of openness to international trade and the lower are the income inequalities (Wood, 1994; Roine, Vlachos, Waldenstrom, 2009).
Careful analysis of international economics, however, provides an opposite argument. Openness to trade in capital abundant countries may have a positive effect on income inequality. This might come as a
5

consequence of the increase of imports from labor-abundant countries, where more capital intensive techniques are applied, in turn, increasing the derived demand of capital and help interest rates to rise. Ultimately, asymmetric effects in the income distribution would stem from the increasing share of the wealthy citizen of capital abundant countries (Manasse and Turrini, 2001).
The empirical evidence on the effects of openness to trade on income inequality confirms that the effects on income inequality are prevalently negative when trade takes place horizontally among advanced countries (reference). The evidence about the effects of horizontal international trade confirms the basic intuition that extra-profits and quasi-rents play a central role in increasing income inequality. The possible explanation for this is that the horizontal trade flows free up capacities and permit to take advantage from economies of scale. As a consequence, increasing market shares lead to an increase in market power and, finally, higher extra-profits in the hands of few. The closer are the conditions of product and factor markets to competitive equilibrium and the lower are the chances that the accumulation of profits may help increasing income inequality.
It is important to note that the ultimate result of the different strands of the literature that analyze the relations between economic growth and income inequality confirm the key role of the reduction of market imperfections and specifically of monopolistic rents in reducing the levels of income inequalities. In the static framework of analysis, shared by a large part of this literature, there is a strong and clear causality between the levels of imperfections of product and factor markets, the levels of profits, hence, the asymmetric accumulation of rents into long-lasting wealth and ultimately income inequalities.
Much broader framework, however, is necessary as soon as we acknowledge the role of technological change as an intrinsic component of economic dynamics (Crenshaw, 1992).
2.2 Innovation dynamics and income distribution
After years of neglect, the effects of technological change on income distribution have been recently taken into consideration. The first step has been when the direction of technological change - defined in terms of changes in the output elasticity of production factors - determined by the introduction of biased technological change, has been considered. This took place in the context of the skill-biased technological change hypothesis. In this context, the skill-bias has been found as an important factor responsible for the increase of income inequalities, observed in the new century. According to this literature, the introduction of new technologies, strongly biased in favor of skilled labor, might have determined an increase in income asymmetry because of the increasing divergence of wages between skilled and un-skilled labor (Vanhoudt, 2000).
This approach, as most of the skill-bias hypothesis, however, misses to appreciate the strong capital saving characteristics of the skill-biased direction of technological change. This second, contemporary and complementary bias of the new direction of technological change should reduce the derived demand for fixed capital and hence, for given amount of available capital, interest rates should fall. Consequently, the reduction of income stemming from wealth would contribute to diminishing income asymmetries between
6

income holders. The appreciation of this second effect should mitigate, if not overcome, the claim that skillbiased technological change increases income inequality. The net effect stems, in fact, from the balance between the asymmetry-increasing effects stemming from the increased remuneration of skills, and the asymmetry-decreasing effects stemming from the reduction of interest rates.
In sum, following the distinction of Aghion, Caroli, García-Peñalosa (1999), the skill-biasedtechnological-change hypothesis may increase wage inequalities rather than income inequalities.
Most importantly, however, not only the direction, but also and most importantly the rate of technological change does play a role in determining economic growth. Consequently, it directly might have an impact on income distribution. Economic growth is not determined only by the well-known virtuous relationship between labor productivity, income per capita, savings, increase in the amount of capital, reduction of interest rates, increase of capital intensity, increase of wages and increase of labor productivity. The introduction of technological innovations is the second and actually the most important determinant of economic growth. Kuznets was well aware of the crucial and endogenous role of technological change both in fostering economic growth and in reducing income inequality. The effect of technological change is strong as it works both indirectly via the positive effects in terms of increase of labor productivity and directly as crucial factor in reducing market imperfections.

skills ↑

technological change

TFP ↑

labour productivity ↑ real wages ↑

income inequality ↓ income per capita ↑

relative capital intensity ↑

savings ↑

real interest rates ↑
asymmetry in wealth distribution ↓

capital ↑

income inequality ↓
Figure 2 Innovation dynamics and income inequality. 7

Nevertheless, the literature after Simon Kuznets has been mute on this relationship. This is quite surprising if one recalls the importance given by Kuznets to technological change. According to Kuznets, technological change affects the composition of the stock of wealth reducing the share of incumbents and increasing that of newcomers: “In such a society, technological change is rampant and property asset that originated in older industries almost inevitably have a diminishing proportional weight in the total because of the more rapid growth of younger industries. Unless the descendants of a high-income group manage to shift their accumulating assets into new fields and participate with new entrepreneurs in the growing share of the new and more profitable industries, the long-range returns on their property holdings are likely to be significant lower than those of the more recent entrants into the class of substantial asset holder” (Kuznets, 1955: 10).
2.3 Schumpeterian view on inequality reducing technological change
The analysis of the effects of technological change on income distribution in terms of reduction of income asymmetries can be reviewed and strengthened by the basic contributions of Schumpeter (1934 and 1942). According to Schumpeter, innovation is not an occasional characteristic of the evolution of an economic system but an intrinsic component and the key determinant of its dynamics. The introduction of innovation is at the basic imperfections in product markets. Innovators command quasi-rents that last as long as the entry of imitators is impeded by barriers to entry and cost differences between incumbents and potential entrants. Increasing returns based upon learning economies and economies of scale, in fact, provide incumbents with increasing cost advantages that delay market entry based upon imitation. The strategic pricing of incumbents can stretch the duration of monopolistic rents with the reduction of prices below the costs of potential entrants and imitators. This could eventually favor the increase of demand and the diffusion of product innovations. In this context, only the introduction of new technological innovations by potential competitors may reduce the duration of monopolistic quasi-rents. New products and processes make possible the creative destruction with the substitution of new quasi-monopolies to the old ones. In this analytical framework, only the introduction of innovations on a regular basis, and more generally, the positive rates of technological change may impede the formation of long-lasting quasi rents based on previous vintages of technological innovations. Only the rate of technological change can increase the actual levels of rivalry, as distinct and opposed to competition.
According to Schumpeter (1947) firms caught in out-of-equilibrium by unexpected changes in factor and product markets try and react. Their reaction will be merely adaptive if appropriate levels of skills and knowledge externalities are not available. When instead the levels of human capital and skills of manpower are large enough and support high levels knowledge externalities, the attempts of firms to react to unexpected changes can be actually creative. Adaptive reactions enable firms to change their techniques i.e. to move on the existing map of isoquants. Creative reaction consists in the actual introduction of new superior technologies that enable to increase the dynamic efficiency of the production process.
8

In the Schumpeterian framework, the faster is the rate of technological change and the faster the reduction of income inequality. With a slow pace of innovation, monopoly rents, reinforced by the barriers to entry and to imitation, are long lasting and market prices decline towards actual production costs only in the very long term. The transfer of the benefits of technological change, in terms of the overall increased efficiency of the production processes and higher quality of the products - especially if measured in terms of hedonic prices so as to include the effects of product innovations- to consumers is very slow. Innovators can retain for themselves large shares of these benefits and increase income inequality both via the increase of their current incomes and the related increase of their wealth with long lasting effects on the income inequalities in the future.
When instead the introduction of product and process innovations is fast, the monopolies and related quasi-rents impinging upon the previous vintages of innovations, last much less. The transfer of the benefits to consumers is much faster. The duration of the accumulation of quasi-rents is much smaller. The lower levels of income inequalities favor the accumulation of human capital that reinforces the rates of technological change. The Schumpeterian framework of analysis confirms the hypothesis put forward by Kuznets: the faster are the rates of introduction of innovations and the lower are income inequalities.
Only the introduction of a new vintage of technological innovation can overcome the barriers to entry and destroy the competitive advantage of incumbents. Fast rates of technological change make it possible the working of the creative destruction and the transfer to consumers of all the advantages of the new technology. Slow rates of introduction of technological innovations, on the opposite, may engender long lasting barrier to entry and, hence, impede the working of price competition with the consequent accumulation of quasi-rents and ultimately the increase of income inequalities both in terms of current profits and increase rents, stemming from accumulated wealth.
Technological change contributes to the reduction of income inequalities not only because its rates trim the duration of transient quasi-rents associated to previous technological vintages, but also because it is the main source of increase of economic growth with the increase of the general efficiency of production processes. Technological change is at the origin of the increase of total factor productivity that, in turn, leads to higher levels of labor productivity and, consequently, to higher levels of savings. The latter make possible the increase of the stock of financial resources available for investment, the consequent reduction of interest rates, the increase of wages and, hence, the reduction of income from wealth that is at the main origin of income inequalities.
The causal relationship between technological change and income distribution works both ways. So far we have seen why and how technological change helps reducing income inequalities. Now, following Schumpeter (1947) we can appreciate how and why income inequality slows the rates of technological change. As soon as we stop considering technological change as an exogenous event and we fully apply the Schumpeterian approach - according to which technological change is endogenous to economic dynamics as it is at the same time the cause and the consequence of technological change - we can see that the relationship between technological change and income inequality works both ways.
9

The amount of skills and the levels of human capital embedded in the working population play a crucial role in the Schumpeterian framework of explanation of the endogenous origin of technological change. Following Aghion, Caroli, García-Peñalosa (1999) it seems clear that income inequality prevents, delays and reduces the levels of human capital in a given economic system. The concentration of income impedes the access of large shares of the population to training and reduces the chances to identify agents with high potential levels.
Economic systems with low levels of income inequality have larger chances to improve the levels of human capital and learning. Higher abilities of the workforce provide firms - that try and react to unexpected shocks in factor and product markets - with larger and better knowledge externalities. Hence, the chances that the reaction of firms becomes creative, is clearly related to the levels of human capital. In sum, it seems now evident that faster rates of introduction of innovations lead to lower levels of income inequalities and lower levels of income inequalities lead to faster rates of introduction of innovations.
2.4. The hypothesis
We can now spell clearly our hypothesis: technological change plays a crucial role in shaping the distribution of income. The rate of technological change and income inequalities, in fact, are tightly intertwined by three well distinct processes: first, innovation helps sustaining the rates of economic growth, also through increasing labor productivity and, thus, the increase of wage levels. This complements the traditional hypothesis that economic growth reduces income inequalities. Second, the dynamics of rivalry based upon innovation complements and widens the hypothesis that the higher the competition in factor and product markets and the lower the accumulation of rents and, hence, the increase of income inequalities. Thirdly, income inequality slows down the rates of innovation. Income inequality, in fact, reduces the opportunities to increase the levels of human capital and the accompanying knowledge externalities. Larger stocks of human capital favor the accumulation of knowledge and ultimately the introduction of innovations. Technological change and income inequality are strongly related by a virtuous cycle where both elements are part of an endogenous self-reinforcing dynamic process.
3. Empirical evidence
3.1 Econometric strategy
The previous discussion on the link between income inequality and growth suggests that there are numerous conceptual and methodological caveats to be taken into account. First, from the conceptual point of view, the link between inequality and innovation is shaped by a two-way causality that challenges the estimation strategy with obvious endogeneity problems. Consequently, the choice of the right methodology should be at the center of our analysis.
Second, it seems not an obvious choice to apply an adequate measure of technological change. Most importantly, due to the complex and manifold nature of technological change, we recognize that there does
10
ChangeIncome InequalityLevelsGrowthIncome Inequalities