Raicho Bojilov, Gylfi Zoega, and Hian Teck Hoon
The first part of the book has three chapters in which the rate of innovation in different countries is estimated and a distinction is made between indigenous and Schumpeterian innovation. Chapter 1 examines recent calculations of the time series of total factor productivity drawn from historical data by the Banque de France, Chapter 2 studies the transmission of innovations between countries, and Chapter 3 focuses on innovation and its transmission during the period of the IT revolution.
This first chapter presents the data used throughout the book in a cross-country comparison of productivity and innovation. The results of the analysis are a set of stylized facts that motivate the rest of our analysis:
In the early 20th century, there were at least two countries, the UK and the US, that were generating innovation, which afterward spread around the world. Yet most existing empirical analysis of macro- and industry-level productivity trends assume away the possibility of multiple innovation centers.
We propose a general framework for the study of the propagation of innovation (or TFP shocks) through the world and its evolution over time. We then apply the framework to the data and present the empirical results.
We distinguish between the adoption of innovation originating and imported from abroad and indigenous innovation, which is innovation generated by a given country on its own. Our distinction, therefore, allows us to narrow down our focus on the part of TFP in each country that cannot be traced back to innovation somewhere else.
Our results show that before WWII there was no dominant global leader in innovation. In the years before WWI, the main generators of indigenous innovation were the UK, the US, France, and to a lesser extent Germany. In contrast, during the interwar years, indigenous innovation originated mainly from the US and France. The results for continental Europe support the hypothesis that, in the 30 years after the end of WWII, war-torn Europe caught up very successfully and quickly with the technological leader at the time, the US.
Our key contribution in this context is the qualification that the observed rapid TFP growth in continental Europe in the period 1945–1972 is primarily due to catching up to the world technological frontier rather than indigenous innovation.
The estimates present evidence for a global slowdown in TFP growth by almost 2 percent in the period 1972–2011 relative to the preceding period, 1950–1972. After an initial sharp decline during the 1970s, indigenous innovation made a partial recovery in the US, the UK, and, especially, the Scandinavian countries. The decline in indigenous innovation was particularly dramatic in continental Europe, where it never recovered after the initial drop in the 1970s and the 1980s. A striking result, which justifies our emphasis on indigenous innovation, is that the exogenous innovation attributed to scientific discoveries is not of major quantitative importance.
This chapter focuses on indigenous innovation and, in turn, productivity growth over the period of the IT revolution. First, we compare the average growth rate of indigenous innovation for the period across countries and then examine the country-specific dynamics for the US, the UK, France, Germany, and Japan. Our results show that there has been no structural change in the transmission network for the period after 1990 relative to the preceding post-WWII decades. Moreover, we see that the annual rates of indigenous innovation in the US and the UK have made only a partial recovery during the IT revolution: while higher than the rates for the period 1970–1990, they are still lower relative to the rates witnessed in the postwar years until the late 1960s. Also, our estimates show that the continental European economies and Japan have not experienced even such a moderate recovery. Interestingly, we find that unlike in the glorious 30 years after WWII, these economies have been rather slow in adopting the (IT) innovations originating in the US and the UK.
This part has four chapters on the roots of innovation in the system of values found in each country and the benefits measured in terms of job satisfaction and labor force participation. Chapter 4 is a case study of successful innovators in Iceland. The objective of the interviews with these innovators is to establish a set of their shared values. In Chapters 5 and 6, we try to relate these values and innovation in a cross section of countries to test the hypothesis that nations that share these values turn out to be innovative. The last chapter in this section, Chapter 7, shows the effect of innovation on job satisfaction and labor force participation.
In this chapter, we establish a common set of values and attitudes that characterize the founders of four successful innovating companies. Each company is founded on the basis of an indigenous innovation, not merely the application of new technology. The values detected and determined are then used in subsequent studies on the relationship between values and innovation at the national level.
We interviewed the innovators in order to establish a set of personal and cultural traits that are conducive to innovation. We can summarize them as follows.
This chapter studies the relation between values, institutions, and innovation for 20 OECD countries.
We employ the method of canonical correlations to map the relationship between values and innovation. The method makes sense of the cross-covariance matrices of two multidimensional variables. To perform the canonical correlation, we gather together some observed measures into two different variable sets, X and Y, which represent the two multidimensional components of the latent variables, henceforth known as the canonical variables X and Y. The variable X is our measure of values, and the variable Y is our measure of innovation.
Next, we assign weights to the variables within X and Y in order to create two linear combinations, X* and Y*, one for each variable set, which maximize the bivariate correlation between the canonical variables. The set of linear combinations, called canonical functions, is chosen to maximize the canonical correlation between the two latent canonical variables X* and Y*. Several uncorrelated components or functions can be determined, as in principal components analysis.
We find a strong relationship between a latent variable that captures both values and institutions, on the one hand, and a latent variable measuring economic performance, on the other hand. In particular, we find that trust, the willingness to take the initiative, the desire to achieve on the job, teaching children to be independent, and the acceptance of competition contribute positively to economic performance. Teaching children to be obedient may, in contrast, reduce economic performance. A measure of economic freedom—that captures the paucity of regulations in goods, capital, and labor markets—is positively related to performance. Economic performance is measured by indigenous TFP growth, imported TFP growth, job satisfaction, male labor force participation, and employment. Fertility is also included in this variable as a measure of optimism about the future.
In this chapter, we relate the annual average growth rates of TFP for the period 1993–2013 for a set of developed economies to an index of modernism, and we find a strong positive correlation between our measure of modernism and economic performance as measured by TFP growth. Naturally, this hardly constitutes a proof of causality, although our empirical strategy is designed to address some of the obvious concerns about endogeneity and reversed causality.
As a first step, we consider individual-level data from a set of developed economies on beliefs, attitudes, and social norms about economic life, which come from the second wave of the World Values Survey. Using these data, we design an index of modernism and an index of traditionalism. Next, we relate the indexes to the annual average TFP growth rates, and to indigenous and imported innovation for our set of developed economies. We document a strong positive association between our index of modernism and our measures of productivity growth and innovation.
This chapter is about the consequences of low rates of innovation in a sample of 16 European countries. In particular, we study the relationship between these low rates of job satisfaction and the labor force participation of men. Job satisfaction has fallen in the EU over this period, as has the labor force participation of men.
We have found that low levels of indigenous innovation are accompanied by low and falling levels of job satisfaction, falling male labor force participation, and less measured happiness.
In a sample that also includes some non-European countries, we find that performance—measured by the rate of indigenous and transmitted innovation, TFP, and the proportion declaring themselves to be very happy—is lower in the continental European countries than in the United States, the UK, Sweden, the Netherlands, Australia, Canada, and Finland.
We conclude that low levels of innovation may be adversely affecting the level of job satisfaction and happiness in many European countries.
In this last part of the book, the focus is turned to the introduction of robots into the labor force. We distinguish between robots that replace humans and robots that augment their productivity, then compare and contrast the effects of these two applications on wages and employment. We also compare the effects of robots when capital is constant and when it moves slowly over time. Finally, we explore the effect of robots on the rate of innovation.
We address the issue of how to model both the creation and the effects of robots in the context of the neoclassical growth model. Our approach is to model capital as malleable and flexible in its use either as conventional machines—that is, as traditional capital—or as robots. Given its malleability, the current stock of conventional machines can be retrofitted to function as robots once the robotic technology is available. We model two types of robots: additive and multiplicative.
Additive robots can perform the same functions as a human worker and, thus, are perfectly substitutable for a human worker. Denoting the quantity of additive robots by RA and human labor by H, the total labor force (both robotic and human) is equal to RA + H. The development of artificial intelligence and machine learning has also produced a second kind of robot, which we call multiplicative, that can be described as augmenting labor power. Such AI-enabled robots, or “multiplicative robots,” are denoted by RM. With the introduction of multiplicative robots, total (robotic and human) labor is given by (1 + RM)H.
With the creation of additive robots, we obtain the stark result that real wages are permanently depressed and the (human) labor share of national income asymptotically tends toward zero, although total (human and robotic) labor share of national income tends toward a positive constant. In contrast, we find that with the arrival of multiplicative robots, while the immediate impact is to cause the stock of conventional machines to fall when it is profitable to adopt robots, the real wage need not fall because of an offsetting labor-augmenting effect coming from the multiplicative nature of the robot. The real wage, even if it drops initially, continues to steadily increase along a balanced-growth path as nonhuman wealth grows even in the absence of steady technological progress. The (human) labor share in national income is also a positive constant in the long run.
In contrast to the assumption of Chapter 8 that conventional machines can be instantaneously retrofitted to function as additive or multiplicative robots, this chapter assumes that the cooperating factor is a slow-adjusting asset like factory buildings and physical structures that take time to build. The adjustment cost of such investments is large. We discover that the adoption of robots leads to the stimulation of investment in slow-adjusting nonmalleable capital that complements total human and robotic labor. Such complementary investment, we show, exerts an upward pull on wage levels.
When additive robots are introduced, we show that when it is profitable for firms to adopt the robots in the production process, the real wage gradually declines to reach a permanently lower level even as the price of physical structures jumps up and gradually increases to reach a permanently higher level. Thus, this polar case of additive robots confirms the fear people have that the introduction of robots hurts workers’ wage earnings and enriches the owners of nonhuman wealth. In the long run, however, the stock of factory buildings and physical structures increases, thus restoring the real wage.
However, when multiplicative robots are introduced in the other polar case, we show that there are two offsetting effects with a fixed stock of factory buildings and physical structures: a factor-intensity effect (with more effective labor working on a fixed supply of physical structures in the short run), which tends to reduce wages, and a labor-augmenting effect (with multiplicative robots making each worker more effective), which raises hourly pay. We establish the result that if the share of factory buildings and physical structures is less than the elasticity of substitution between physical structures and effective labor, then introducing robots must raise the whole path of real wages. In the long run, the growth of real wages is further propelled by the accumulation of capital taking the form of factory buildings and physical structures.
In a two-sector model, we formalize the idea that when an influx of additive robots in the capital goods industry causes the relative price of the capital good to fall, economic incentives are created to make indigenous innovation in the consumer goods sector more profitable. This causes some workers in the economy to shift from participating in production to participating in innovative activity. Consequently, the new wage path must be ultimately rising, as long as innovation in the consumer goods industries continues undiminished, exerting its upward pull on wage rates.
We set up a two-sector model with the characteristic that there is a sector that uses only human labor to produce a capital good before the creation of robots. The capital good is used to produce a pure consumption good. With zero technological progress and zero population growth, the model exhibits a stationary equilibrium with constant wages, consumption per capita, constant relative price of the capital good, and constant stock of conventional machines. With the creation of additive robots, we obtain the result that capital accumulation is stimulated, consumption per capita grows, and the real wage declines to a permanently lower level as the relative price of the capital good permanently declines to a lower level.
We also study a two-sector model where the consumption goods sector produces a good derived from conventional machines working along with a continuum of intermediate inputs produced by imperfectly competitive firms that also use conventional machines as inputs. Indigenous innovation takes the form of improving the quality of these intermediate inputs. We derive two main results: (1) a drop in the relative price of the capital good stimulates indigenous innovation, which acts to boost real wages; and (2) the creation of additive robots spurs investment in conventional machines, which also stimulates indigenous innovation. Thus, once we depart from the first two-sector model to allow for indigenous innovation that raises the productivity of the consumption goods sector, the arrival of additive robots is good for wage growth.