Specialization

Conclusion

This paper has shown that the specialization of capital is endogenous labor- augmenting and capital-augmenting organizational or technological change. It may be implemented as either shiftwork within the firm or as different sets of individuals time-sharing the same set of rental capital services in a rental market. In addition, this organizational change may be “embodied” in the measured capital inputs using the Griliches-Jorgenson [Griliches and Jorgenson (1966)) and Jorgenson and Griliches (1967)] measurement methods. The specialization of capital is costly to implement. In our model, the specialization of capital is limited by coordination/ transaction costs; and, as per-capita wealth increases, an increased amount of the specialization of capital is demanded. This paper has thus shown that the continuous specialization of capital is a mechanism for endogenous technological change in the neoclassical-growth model.

There are many issues left for future research. For example, we have not studied under which conditions rental capital services in rental markets are chosen instead of shiftwork within the firm. Insights from earlier works [see Coase (1937)

and Williamson (1975, 1985) for more discussions] suggest that, in order to do this, we need to model coordination costs in greater detail. We also need to do more empirical studies of rental services and shiftwork. Some statistics from the World Bank's World Development Report 1997 suggest that rental services are becoming more important. According to that report, the services sector accounts for 66% (32% for manufacturing) of the GDP of the high-income economies in 1995; and that the services sector has grown at the average annual rate of 3.4% (3.2% for manufacturing) from 1980 to 1990 and at 2.3% (0.7% for manufacturing) from 1990 to 1995 in the high-income economies. Indeed, since this paper was first written, the “sharing economy” has grown tremendously in many industries.

It is also useful to study capital sharing in urban economies, where the US Bureau of the Census (1992) shows that services are concentrated. The results in this paper imply that the shift in the total productivity parameter across cities found in the data by Segal (1975) may go away if the Griliches-Jorgenson adjustments are used in measuring capital inputs.
A service industry in which coordination and coordination costs may be studied in greater detail is the financial services/banking industry. The capital- augmenting effects of the borrowing and lending of specialized capital goods at the same time — e.g. I use your set of specialized capital in return for using your set of specialized capital — can be viewed as a “production multiplier effect” not unlike the money multiplier effect in the money supply process. Empirical studies [e.g. King and Levine (1993) and Jayaratne and Strahan (1996)] and theoretical works [e.g. Greenwood and Jovanovic (1990) and Bencivenga and Smith (1991)] have studied the role of financial markets and institutions in economic growth. See Levine (1997) for a survey of this literature.

The results in this paper have some general policy implications. Our results imply that better organized rental markets for capital services and better coordinated shiftwork within the firm may be important for economic growth. Therefore, when investing in either physical or human capital, we have to make sure that sound coordination and institutions are there to keep labor productivity and capital utilization up at the same time. Otherwise, a country which has acquired more machines and more engineers may simply have more idle machines and more idle engineers. These implications are consistent with Winston's (1971) finding that capital was idle two-thirds of the time in a typical underdeveloped country. They are also consistent with the general observations by North and Thomas (1973) and North (1990) that market institutions are important for economic growth. Coordination policies and institutions may be specific to each industry — for example, institutions in the banking industry may be different from those in the ride-sharing industry.