The Competitiveness of Nations in a Global Knowledge-Based Economy

Morris Altman

Chapter 6:  (extracts)

A Behavioral Model of Endogenous Economic Growth

Worker Satisfaction and Economic Performance

M.E. Sharpe, Armonk, 2001, 119-136.




The Conventional Growth Model and Revisions

High and Low Wages and the Path and Pattern of Economic Growth

Conclusion: The Low Wage Economy and Market Failure



Conventional economic theory tends to support the view that there is only one sustainable wage path to economic development.  In the long run, one expects economies to converge to like levels of real per capita gross domestic product (GDP) with like mean factor prices, of which the wage rate is, of course, one.  There can be no such thing as a high-wage path to economic growth and development since convergence occurs through the process of interregional and international trade and factor mobility and is facilitated by the unfettered working of the marketplace.  A caveat to this argument is related to the work of Paul Romer (1986, 1990, 1994), who, following the earlier work of Arrow (1962) and Young (1928), introduces the notion of increasing returns and positive externalities into the basic neoclassical model pioneered by Solow (1956, 1957; see also Abramovitz 1952) by incorporating knowledge, in the form of learning by doing or research and development, in the production function.  Gene Grossman and Elhanen Helpman (1994) add endogenous innovations to the basic Romer framework (see also Aghion and Howitt 1998). [1]  Regions that begin first have a first-mover advantage over those that are followers in the investment and development process.  Economic growth then takes on a sustained and cumulative form.  In this case, convergence need not take place through the free market forces of supply and demand.

Off the mainstream, Gunnar Myrdal’s (1963) work on cumulative causation argues in favor of persistent divergence of per capita GDP and factor


prices as a consequence of the unfettered working of the market economy.  The law of supply and demand results in the rich getting richer and the poor getting poorer.  Once a region or country is ahead of the economic pack, it becomes profitable for investment to be biased toward the leaders.  The advantaged areas have, in effect, a first-mover advantage.  Only government intervention can break what becomes a vicious cycle.  Nicholas Kaldor (1957) also develops an endogenous theory of economic growth whereby the rate of technical change is a function of the growth in capital stock.  But like the neoclassical production function, Kaldor’s enhanced production function is characterized by diminishing returns.  He also constructs a model, based upon the work of Myrdal and P. J. Verdoon, wherein exports, exports of manufactured products in particular, cause differential levels of per capita output and growth rates through the process of cumulative causation since it is assumed that manufacturing is the most dynamic sector in an economy and is characterized by increasing returns (Kaldor 1970). [2]  However, like the basic neoclassical growth model, the alternative growth models do not specify why one region should be ahead of another to begin with, nor why and when regions should begin the process of convergence or what variables actually initiate the process of divergence.  The causal process in all of these models takes on a random form or, alternatively, becomes a function of events exogenous to the model.

The empirical evidence is heavily weighted in favor of the argument that convergence has not taken place internationally over time (Altman 1999c; Baumol 1986; Baumol and Wolff 1988; De Long 1988; Dowrick and Gemmell 1991; Pritchett 1997).  This is not to say that no convergence whatsoever has taken place.  Over the last one hundred-odd years the presently developed economies have converged.  Some less developed countries are in the process of converging.  Nevertheless, the majority of the less developed countries are characterized by laggard economic performances.  In point of fact, over the long haul, low-income and high-income economies have persisted side by side.  The coexistence of these two types of economies is an important stylized fact and should be explained by growth theory.

A model of endogenous growth is developed here that complements both orthodox and nonorthodox growth theories.  The main concern of this chapter is a causal one.  I attempt to provide an explanation for sustainable “equilibrium” differences in real per capita GDP, as well as for either transitory or “permanent” differences in growth rates that ultimately cause differences in real per capita GDP.  I therefore go beyond a discussion of the immediate sources of growth, such as the traditional factor inputs, research and development, and human capital, to an analysis of the motivational structure that might help explain some of these important differences. [3]  Using the more


realistic behavioral assumptions embedded in efficiency wage and x-efficiency theory, I argue that there are at least two sustainable paths to economic growth.  At one extreme is the low-wage path, and at the other is a high-wage path.  The latter is associated with a higher level of equilibrium per capita real GDP and the former with a lower level.  Convergence in real per capita output need not take place through the workings of the free market.  Moreover, competitive pressures (or market forces in general) need not result in the convergence of wage rates.  In a word, what I show here is that both a low-wage and a high-wage economy can be consistent with a competitive economic regime in the long run.  But a high-wage economy potentially yields a higher level of material well-being for all its members.  I further maintain that in the short run and under certain conditions in the longer run, the high-wage economy may also be characterized by a higher rate of growth.

High wages serve to pressure firms to become more efficient and innovative, whereas low wages allow firms to engage in less efficient economic behavior, all the while remaining competitive and profitable.  For this reason, high wages are viewed as a potential boon to the process of economic growth and development, whereas low wages are viewed as a potential drag.  This is contrary to the view typically held by both neoclassical and nonconventional economists wherein high wages are considered to be anathema to the growth process.  In the model presented here, economic agents can choose among different available growth paths.  There is no one road to growth.  The choice made, however, affects the long-run equilibrium level of real per capita GDP and the rate of economic growth.  It is evident that owners and managers of business enterprises, the decision makers, when free to choose, are to an increasing extent choosing the low-wage path to economic growth and profitability (Bluestone and Harrison 1990; Perelman 1993).  In my model, this choice would have predictable negative consequences for the level of material well-being for the majority of a region’s population.  To the extent that the purpose of growth is to allow for the maximization of the consumption basket of the typical consumer, this “preferred” path to growth generates inferior or suboptimal welfare outcomes.  Only if the low-wage path is deemed the only path to growth can it and all that it entails be considered in any way optimal or welfare maximizing.


The Conventional Growth Model and Revisions

Conventional neoclassical growth theory does not rule out the possibility of nonconvergence.  In particular, to the extent that technical change or innovations are not transferable interregionally or internationally, there is no reason for per capita GDP to converge.  The same holds true if there are significant


and persistent barriers to trade.  In fact, the classic neoclassical growth model as developed by Solow is not concerned with the question of international or interregional differences in the level or rate of growth in real per capita GDP.  Instead, it is largely concerned with establishing the necessary conditions for the long-run sustainability and stability of a given rate of growth.  Solow’s model was developed in response to Harrod’s (1939) work on modern growth theory, which suggests the likelihood of capitalist growth being characterized by considerable long-run instability.

In Solow’s analytical framework the long-run equilibrium level of per capita output is determined by an exogenously given propensity to save and rate of technical change.  The long-run equilibrium growth rate is given by an exogenously determined labor force growth rate.  Long-run stability is achieved through variations in the capital to output ratio.  In the basic model, real output is a function of capital and labor inputs, where the production function is characterized by constant returns to scale while the individual inputs are subject to diminishing returns.  Perfect competition is also assumed, as is full employment in the long run, with Keynesian macroeconomic policy in play in the background.  In addition, savings are identically equal to investment so that the growth in capital stock is identically equal to savings.  It is further assumed that production is x-efficient.  Therefore, all economic agents are working as hard and as well as they can…


123 [not displayed]

If one introduces technical change into the argument, this too has the effect of temporarily increasing the growth rate as the production function is shifted upward, only to return to the equilibrium growth rate... Nevertheless, as is true for an increase in the propensity to save, the economy subject to technical change ends up with a higher level of output per worker.  It is important to note that, for Solow, technical change is a shorthand classification for any factor that shifts the production function.  The term “technical change” embodies any and all variables, such as learning by doing, human capital formation, innovations, research and development, and improved efficiency, that contribute to enhancing the productivity of factor inputs.  Unless all regions and nations are characterized by an identical propensity to save and rates of technical change, which is what conventional wisdom suggests to be true, convergence in per capita GDP is not expected.

With the introduction of technical change the Solow growth equation is transformed into… a permanent feature of the growth equation insofar as technical change is a permanent and persistent feature of the economy.  More generally, with technical change the growth in real GDP is a function of the growth in factor inputs plus the rate of technical change, and the growth in labor productivity becomes a function of the growth in the capital to labor ratio (which is subject to diminishing returns) and the rate of technical change.  


… The key to sustained growth in the Solow model is, therefore, technical change writ large.  Increased savings can, in the long run, affect only the level of per capita output.  And there is a limit to which a society can increase the average propensity to save.  Moreover, to the extent that the assumption of diminishing returns to capital is reasonable, there is a limit to the increase in output that can be evoked by increases to the capital stock.  Since technical change is so much of a catchall phrase in the neoclassical model and it is assumed to be exogenously determined, this model has been criticized for leaving most of the growth process unexplained.  But as Solow argues, the assumption of exogenous technical change makes his model more methodologically flexible: “To say that the rate of technological progress is exogenous is not to say that it is either constant, or utterly erratic, or always mysterious.  One could expect the rate of technological progress to increase or decrease from time to time.  Such an event has no explanation within the model, and may have no apparent explanation at all.  Or else it might be reasonably understandable in some reasonable but after-the-fact way, only not as a systematic part of the model” (1994, 48).  This, of course, leaves us with no clear causal mechanism to explain the process of technical change and economic growth.


To address this potential problem, the new growth theory, centered on the work of Paul Romer (1986, 1990; also Aghion and Howitt 1998; Grossman and Helpman 1994; Helpman 1992; Lucas 1988), has evolved wherein growth is explained endogenously.  In the new growth models increasing returns to scale are assumed to characterize the aggregate production function.  In particular, increasing returns are assumed to characterize investment in knowledge as distinct from physical capital.  Therefore, increasing investment in human capital or research and development has the potential effect of increasing labor productivity without bound (Solow 1994, 50). Thus the production function becomes convex so that the rate of change in labor productivity increases as investment in the knowledge sector(s) of the economy increases.  As Grossman and Helpman put it with respect to their version of the endogenous growth model: “The endogenous learning here - like the exogenous technological progress of the neoclassical model - prevents the marginal product of capital from falling to the point where investment ceases to be profitable.  Innovation sustains both capital accumulation and growth” (1994, 35).  With investments in knowledge, sustained and persistent growth is possible above and beyond the rate of growth in the labor force…, and technological change… becomes endogenous to the model.  Once the growth process is triggered through investments in knowledge, it becomes self-sustaining.  Moreover, since investment in knowledge is assumed to be characterized by positive externalities to the economic agent or the firm, it is quite possible for an economy to underinvest in knowledge and for the free market to fail in generating the “optimal” level of investment and therefore the “optimal” rate of economic growth and level of per capita output.

The new growth theory, however, does not provide us with a framework within which to understand why convergence does not take place or, indeed, why divergence may take place over time or why leading economies might eventually become laggards - Britain being a classic case in point.  One is left with an understanding that persistent differences in per capita GDP among economies and persistent differences in rates in per capita economic growth are a product of a variety of ad hoc factors, among which are different types of economic policy or historical accident.  For this reason, the new growth theory leaves us no further ahead of the analytical game than Solow’s theory of economic growth.  Moreover, empirical work on the new growth theory does not provide strong support for the hypothesis that investment in knowledge is a fundamental cause for sustained economic growth and differentials in growth and per capita output among economies (Fagerberg 1994; Mankiw, Romer, and Weil 1992; Pack 1994; Parente and Prescott 2000; Solow 1994).


High and Low Wages and the Path and Pattern of Economic Growth

A critical assumption of all standard models of economic growth is that wage rates play no positive role in the growth process.  At most, relatively high wage rates, by inducing higher capital to labor ratios, yield a relatively higher level of per capita output, although at the price of higher unit costs.  But economic agents are assumed to be x-efficient in behavior.  For this reason, economic agents cannot be made to work harder or better, and all firms are assumed to be cost minimizers in the sense of maximizing output per unit of input.  Once the assumption of x-efficient behavior is dropped, however, the predictions of the conventional neoclassical growth model are dramatically affected.  In a word, by only marginally modifying the simplifying assumptions of the Solow model, a model of economic growth is developed that can help explain persistent differences in per capita GDP and rates of economic growth, as well the movement of regions from leaders to laggards in the growth process. [4]

There is now a significant literature suggesting that typically economic agents do not work as hard or as well as they might.  In other words, effort is a variable in the process of production (chapter 9; Akerlof and Yellen 1986, 1990; Altman 1992; Blinder, ed. 1990; Frantz 1988, 1997; Leibenstein 1966).  If effort is variable, then so is labor productivity, and if effort is not maximized, neither is productivity.  Effort and therefore productivity become discretionary variables.  Effort variability itself is a product of incomplete contracts (both informal and formal) due to transaction costs in drawing up, monitoring, and enforcing them.  Moreover, the behavior of the firm hierarchy is more in line with utility maximization as opposed to profit maximization so that if the marginal utility costs of increasing the hierarchy’s own effort or that of its workers exceed the marginal utility benefits, effort per unit of time will not be maximized.  Nevertheless, the bottom line for the firm hierarchy remains the competitiveness of the firm, at least in the short run.

An important tenet of the behavioral model presented in this book is that in the absence of an ideal work environment, x-efficient behavior will be the exception to the rule.  X-efficiency is defined as that level of output that can be achieved in the ideal, relatively cooperative work environment.  It is possible for x-inefficient firms to survive even with perfect product markets if wage rates differ among firms and if the different wage rates and the work cultures, for which they may be a proxy, affect labor productivity by affecting the effort inputted into the process of production.  The rate of labor compensation is therefore positively related to effort per unit of time and thereby to the level of labor productivity.  Under these circumstances, a low-wage


regime goes hand in hand with a low-productivity regime and a high-wage regime goes hand in hand with a high-productivity regime, and the two wage regimes need not necessarily yield any differences in unit costs or even in rates of return…

Under these conditions, one can postulate a given level of average and marginal costs and rate of return to capital that correspond to an array of rates of labor compensation.  There need not be any unique wage rate or labor compensation package that minimizes costs or maximizes the level of profits (chapters 1 and 2; Altman 1998; Akerlof and Yellen 1990; Stiglitz I987). [5]  In this case, high-wage firms can compete with low-wage firms by becoming more productive, and output is a function of effort per unit of labor input…, as well as of labor, capital, and technical change…

In this scenario, low-wage-low-productivity and high-wage-high-productivity firms and, more generally, low-wage-low productivity and high-wage-high-productivity economies can exist simultaneously and persist over time since there is no mechanism in place within the structure of the free market (perfect product market competition) to drive the wage rates and the levels of labor productivity toward convergence: low-wage x-inefficient firms do not have a competitive advantage over high-wage firms that can ultimately force convergence to take place.  Indeed, low wages serve to shelter x-inefficient firms from competitive pressures.  High wages are compensated for by higher productivity.  On the other hand, if relatively high rates of labor compensation are not followed by sufficient increases in labor productivity, the high-wage firms will, ceteris paribus, fall by the wayside in the process of interfirm or international competition.  One can expect this to occur once, given technology, firms become x-efficient and effort becomes perfectly inelastic to increases in wages or improvements in overall working conditions.  It is theoretically possible for wage rates to converge in the long run if perfect mobil-


ity exists in the labor market.  I assume that labor market mobility is far from perfect.  In fact, labor does not move freely, nor does it costlessly flow from one region or country to another.  Apart from this, many economies, especially the high-wage economies, have institutional mechanisms in place (the right of workers to organize and a social welfare safety net) that can maintain wages at relatively high levels even in the face of significant labor mobility.  As long as relatively high rates of labor compensation are accompanied by relatively high rates of productivity, the product market cannot generate the forces necessary to bring about the convergence in per capita GDP.

The notion that relatively high wages can positively affect labor productivity and thereby cause long-run international differentials in real per capita GDP is not a new one.  Such a scenario was suggested by H.J. Habbakkuk (1962) in his classic comparative study of nineteenth-century American and British economic development.  He argues that America’s advantage lay in the relatively high rates of labor compensation that characterized its industrial sector.  Habbakkuk maintains that in order to remain competitive and profitable, American firms were forced to make more efficient use of their factor inputs (to be more x-efficient), and American entrepreneurs were induced into becoming more creative and innovative in developing new technology, as well as being more apt to adopt known technology.  Thus in Habbakkuk’s interpretation of American-British comparative economic development, America’s relatively high-priced labor was the primary cause of two key events.  First, it affected the x-efficiency of production.  Second, it affected both the extent and the rate of technological progress.  All of this would have the effect of shifting the production function upward… yielding a higher level of per capita output in British as compared to American firms.

The critical adjustment to the conventional neoclassical growth model suggested in this chapter is that given the possibility of x-inefficiency, relative differences in wage rates in like industries affect the differences in x-inefficiency among industries and economies and the extent and pace of technological progress.  This, in turn, affects equilibrium differences in real GDP per capita among economies, as well as, at a minimum, the short-run growth rate in per capita real GDP.  Output becomes a function not only of capital and labor, but also of relative rates of labor compensation and other labor costs, which, in turn, affects the shift or technical change parameter through its effect on x-inefficiency and technological progress…


Output is now only an indirect function of effort since effort per unit of time is assumed to be the product of relative rates of labor compensation among economies... To simplify the discussion, I assume that the output of two (or more) economies is identical.

In this model, an increase in the relative price of labor has the immediate and short-term effect of increasing the cost of production and reducing the rate and amount of profit.  If the firm is to survive, however, ceteris paribus, an increase in the relative rate of labor compensation must be accompanied by a sufficient increase in labor productivity.  This wage effect on productivity can be broken down into three components.  The x-efficiency effect, the technological progress effect, and the savings effect will be discussed in turn.

1. The x-efficiency effect.  High-wage firms are pressured into becoming more x-efficient.  The decisionmakers of the firm cannot be expected to increase x-efficiency on their own if the benefits of the increase accrue largely to the employees and if the process incurs increased disutility to members of the firm hierarchy in terms of increased effort, loss of status, and even loss of jobs to some (chapter 9).  Increasing the level of x-efficiency serves to shift the production function outward.  On the other hand, the relatively low-wage firms remain relatively x-inefficient as a result of low wages.

2. The technological progress effect.  High-wage firms are forced to adopt already available technologies that are new to these firms.  These firms may also find it worthwhile to develop new technologies (to innovate).  The new technologies might also require the investment in on-the-job training and formal education for employees and the firm.  Technological progress shifts the production function upward.  Low-wage firms need not adopt new technology or innovate if they can effectively compete on the basis of low wages.  Moreover, the low-wage regime may result in labor being too x-inefficient for the new technology to be cost-effective…


131[not displayed]

3. The savings effect  Under pressure to adopt new technology or new forms of organization by higher rates of labor compensation, high-wage firms may be forced to increase the propensity to save so as to increase the rate of investment.  This may result in an increase in the economy’s propensity to save.  Much depends on the extent to which increasing wages serve to pull down the economywide or average propensity to save (with workers being characterized by a relatively low propensity to save).  To the extent that the saving rate increases, the production function shifts upward.

These three wage effects serve to increase the level of real per capita GDP in the high-wage economy.  In the short run, at least, the rate of per capita growth is also augmented as the high-wage economy adjusts to the new higher level of per capita output.  It is quite possible that in the long run the growth rate in output will converge toward the growth rate of the labor force.  However, growth may be positively affected over the longer run to the extent that relatively higher and increasing rates of labor compensation affect the rate of improvement in x-efficiency and the rate of technological progress writ large.  But clearly the wage effects can be expected, at a minimum, to increase real per capita output and the short-run per capita growth rate.  Moreover, the consequences of higher rates of labor compensation should be stable insofar as they are consistent with Solow’s stability conditions for economic growth.  In addition, the low-wage economies have no competitive advantage over the high-wage economies.  In this sense, both the high- and low-wage path to economic growth and development are sustainable in the long run.

The high-wage regime might be more unstable than a low-wage regime, however, insofar as corporate leaders might have a preference for the latter and have the capacity to undermine the former by political means or by relo-


cating to low-wage regions.  This is particularly true if an economy approaches the neoclassical ideal of perfect competition, with the perfect mobility of capital and the absence of tariffs.  Footloose industries would be more prone to this particular problem.  High-wage industries tied to a region or country for geographical, economic (transportation costs, economies of scale), or political reasons would be more stable.

A high-wage regime would also be difficult to sustain if the necessary infrastructure to maintain a high level of productivity were not established.  This includes the educational, research and development, health, legal, and transportation and communication infrastructure, which Moses Abramovitz (1986) refers to as a nation’s “social capabilities.”  Social capabilities represent necessary conditions for sustained growth and development (see also Olson 1996).  High wages may induce governments to engineer the social capabilities necessary.  A low-wage regime may have the opposite effect on government.  On the other hand, adequate social capabilities are not sufficient to induce high rates of economic growth and high levels of real per capita output.  On a microeconomic level, high wages are necessary to induce firms to become more x-efficient and to adopt more productive technology. [7]

The key point this chapter seeks to make is that one can develop a reasonable endogenous theory of economic growth by introducing wage rates as a causal determinant of the extent and rate of x-inefficiency and technical change into the Solow growth model.  Thus the origins of important and stable differences in per capita GDP and growth rates become more comprehensible.  Moreover, it is now possible to more clearly explain how an economy can revert from leader to laggard in the growth and development process.  Whether or not convergence takes place depends on the state of the labor market and the extent to which an economy can sustain relatively high rates of labor compensation.  This is not to say that other factors, both endogenous and exogenous, are not important to explaining this process.  However, by focusing on the relative rate of labor compensation, one is isolating an easily identifiable and measurable as well as significant economic variable as a possible supply-side determinant of the growth process.  This variable has typically been ignored in the literature as an important determinant of x-efficiency and technical change.  As with all hypotheses, the one presented in this chapter must bear the test of empirical scrutiny.


Conclusion: The Low Wage Economy and Market Failure

To the extent that both high- and low-wage economic regimes are sustainable over time, the low-wage regime can be regarded as a product of a market failure.  The objective of economic growth is not growth itself; rather it is


to maximize a social welfare function.  One might argue that an important argument in such a welfare function is the basket of commodities consumed by members of an economy over a specified time horizon.  A low-wage-low-productivity economy cannot serve as the basis for maximizing the consumption basket of the typical consumer of such an economy.  Apart from this, such an economy produces a level of output that is below its potential where this potential is given by what is produced in the high-wage-high-productivity economy.  But the behavior of economic agents in both economies can be consistent with utility-maximizing behavior.  And this is the rub.  The low-wage economy might be preferred by the utility-maximizing members of the firm hierarchy and corporate leaders in general, as well as political leaders - that is, those individuals who ultimately determine the wage regime firms follow given the institutional, social, political, and economic constraints they face.  What is of some importance from the perspective of social welfare is whether the utility function or preferences of the decision makers take into account the preferences of the larger population.  In other words, do externalities exist when choosing the wage regime that an economy is to follow?

If a high-wage economic regime is adopted, this can be expected to increase the utility of workers in terms of the higher income generated.  These benefits are typically not part of the preference function of corporate leaders.  In this case, by maximizing their own utility, the corporate leadership will tend to “underinvest” in the development of a high-wage economy.  But why might a low-wage regime be favored by corporate decisionmakers when both regimes yield competitive and profitable economic structures?  It all depends on their utility function.  If utility is negatively related to more work and effort on the part of the firm hierarchy and these are required to develop a high-wage economy, the marginal costs of developing a more productive economy tend to outweigh the marginal benefits.  Such benefits might be zero, given our assumptions.  If creating a high-wage-high-productivity economy also involves investments that can only be covered over time, this will further encourage a preference for a low-wage economy.  This is especially true if members of the firm hierarchy are present-oriented, if they have a high rate of time preference.  Since choosing a low-wage regime ignores the benefits accruing to consumers in general, such a choice generates a market failure.

The preference for a low-wage regime on the part of members of the firm hierarchy cannot always be realized, however.  They are not always free to choose.  The state of the labor market acts as a powerful constraint on the behavior of corporate leaders.  Labor market conditions, which are influenced by supply and demand considerations as well as by institutional factors, can prevent the low-wage route to growth and development.  But once the labor mar-


ket becomes “flexible,” the low-wage regime becomes the chosen path of corporate leaders unless their utility functions undergo a significant transformation.  As Bluestone and Harrison remark, an environment has emerged “in which it was simply easier for many firms to attempt to contain their own labor costs than to seek enhanced profits through investments in expensive new plant and equipment.  For the most part, firms abandoned revenue-enhancing strategies to boost profits and turned sharply towards tactics that emphasized cost reduction instead” (1990, 369).  This comment on the American economy is applicable to an increasing number of developed high-wage economies.

To the extent that the choice of a low-wage route to growth represents a market failure, it is important to develop institutions that prevent such a failure.  At a minimum, labor market institutions need to be developed that minimize the probability of the low-wage path when the high-wage path is a competitive and profitable option.  Moreover, an institutional environment (social capabilities) must be developed that contributes toward making the high-wage path competitive and profitable.  This is particularly important since, according to the behavioral model of economic growth presented here, the low-wage path is viable and sustainable over time in spite of its being suboptimal from a social welfare perspective.  Market forces, left to their own devices, cannot easily push an economy from a low- to a high-wage path.  As a result, there are significant long-run regional and international differences in per capita output.



1. Many of these arguments, often referred to as the new growth theory, are predated by Abramovitz (1952), albeit in a less formal presentation.  Moreover, Schumpeter’s (1974) classic study on technological change, first published in 1934, has also informed much of the new growth theory literature, as well as the recent work on evolutionary growth (see Nelson 1995, 1998).

2. See Hodgson (1989) for a critical assessment of Kaldor’s theory of cumulative causation.  Kaldor’s (1970) rendition of “Verdoon’s Law” is that there exists a strong positive relationship between the rate of growth of productivity and efficiency and the rate of growth in the scale of economic activities.  Thus there is a strong positive correlation between the rate of growth of exports and the rate of growth of productivity and efficiency.

3. Nelson (1998) points out that both the basic Solow growth framework and the new growth theories fail to take us beyond a discussion of the immediate sources of economic growth.  He argues that variables that underlie the immediate sources of growth hold the secret to a better understanding of the persistence of growth and per capita income differentials across nations.  Nelson identifies the organizations that adopt technologies, the institutional framework in which they operate, and the process of technological change as key underlying variables that affect the process of economic growth.


4.Unlike the Solow model, the model developed in this chapter does not assume x-efficiency in production or perfect labor market mobility.  Also, as with the conventional growth models the demand side is not dealt with here, which is not to say that it is unimportant.  Rather, I focus on the supply-side variables that may affect the growth process.  See You (1994) for an alternative wage-led growth model that incorporates the demand side.

5. In the traditional efficiency wage literature there exists a unique wage that maximizes effort per unit of time and thereby minimizes the cost of labor per unit of time.  This unique wage is chosen by the profit- maximizing firm (Akerlof and Yellen 1986).

6. The argument presented here clearly fits in with the notion of induced technological change, wherein changes in relative factor prices affect the direction of technological change (Hayami and Ruttan 1971; Hicks 1932; Ruttan 1997).

7. Reviewing the history of technological change, Nelson and Wright (1992) find that the globalization of the world economy has resulted in common technologies across nations, easing the international transfer of technology when the appropriate social capabilities are available in the target nations.  Fagerberg (1994) finds that the most important determinant of technological change, given the existence of a sufficiently strong social capability, are the “intentional activities of private firms.”  The model presented in this chapter suggests that in a world of accessible technologies and adequate social capabilities differentials in wage regimes can have a significant impact on the “intentional activities of private firms.”