3.2.6Small Share of Computer Capital

Oliner and Sichel (1994) have shown that the small share of IT capital to total capital (IT ratio) was a reason why effects from IT on productivity have been so negligible. The authors estimated the contribution of computer equipment capital to output growth, using the neoclassical framework previously described in section 3.1.1. As stated in equation 3.4, the contribution of any input to output growth is measured as the product of this input’s nominal income share and the growth rate of its nominal net stock.

First, computer income share is estimated using equation 3.5. Sichel (1997) considered the following values for 1992:

• r COMP, the competitive rate of return to all nonresidential equipment and structures, is 12% as calculated by the Bureau of Labor Statistics

• d, the rate of depreciation for computers is estimated at 25%, based on data from the Bureau of Economic Analysis.

• KIT, the nominal stock of computers and peripheral equipment was $95.9 billion in 1992.

• Y, total nominal income in 1992 was $4,494.4 billion.

Hence, computers had a very small income share of 0.8 % in 1992. Finally, price decline in computer equipment may have lower marginal returns to IT investment, thus reducing the income share of IT capital. As a matter of fact, the drop in computer price was spectacular over the last three decades, as noted by Brynjolfsson and Yang (1996):

‘The price of computing has dropped by half every 2-3 years. If progress in the rest of the economy had matched progress in the computer sector, a Cadillac would cost $4.98, while ten minutes of labor would buy a year’s worth of groceries. ’

As the price of computers falls, firms will buy them and use them for tasks with lower payoffs that were not profitable with the higher price. In other words, firms will invest in computers until the marginal product from an additional computer just equals its marginal cost. If the price of computers drops, companies can then invest in additional computers that have a lower marginal product. Pioneer computers of the 1960s cost millions of dollars, but they were used for high-level applications such as space programs, whereas today’s computers are used for more trivial tasks such as scheduling meetings, surfing the Internet or typing reports. Therefore Sichel (1997) argued:

‘Thus, when one considers how much more computing power a dollar buys today than some years ago, one must remember that today’s marginal computer dollar may be going to a lower payoff activity and to a machine that is less heavily utilized.’

After reviewing the different theoretical explanations of the productivity paradox, the literature survey continues with the description of some of the main empirical studies measuring the effects of IT capital on productivity.