5.3Output Growth Contribution of IT Capital, by State

This analysis is concerned with the relative contributions of IT to output growth among states, and not their absolute levels. IT capital might have a low absolute contribution, and still a high contribution relative to a state’s share of national IT stock. The idea is that many states have seen positive effects of IT on productivity but these states account for a small share of the national IT capital stock. Only 10% of the states own more than 50% of the total U.S. stock of IT capital. Hence, aggregating the states’ contributions may show a small overall contribution of IT to productivity in the United States. Indeed, as stated in chapter 2, more than 50% of the U.S. total IT capital stock is located in 8 states: California, New York, Texas, Illinois, Florida, Pennsylvania, New Jersey and Ohio. More than 80% of this stock is located mainly in the service sector (and more specifically in transportation, F.I.R.E, services and retail trade, respectively) and manufacturing accounts for 15%. Thus, the question of IT and productivity should be analyzed with a closer look at what happened in these eight states and specifically in their transportation, F.I.R.E, services, manufacturing and wholesale trade industries.

Following the method of Oliner and Sichel (1994), I assume the output growth contribution of IT capital (OGCKIT) can be computed as the product of IT capital’s income share and its growth rate. The income share is defined as the product of the return to IT capital (r) times the ratio of its stock with total output (KIT/Y):

OGCKIT, t = [r * (KIT/Y)t-1] * gr(KIT)t(5.5)

where t denotes the year and r is the return to IT capital net of depreciation. I computed the average yearly output growth contribution for each state. Results appear in Table 5.8. States are sorted in descending order by their respective growth contribution. This contribution is also expressed as a percentage of a state’s output growth rate. The average growth rates of output and IT capital stock are also given for each state. The average share of a state IT capital stock over the total national stock and the state IT ratio also appear in Table 5.8. According to these results, the contribution of IT capital to output growth ranges from less than 5 to more than 14 average yearly percentage points.

Oliner and Sichel (1994) found a comparable average contribution of 16 percentage points for the period 1970-1992 at the national level. However, this absolute value depends on the methodology adopted, which varies greatly among authors. Here, the focus is on the relative contribution by state. The main result is that there are some important variations in the output growth contribution of IT capital among states. Furthermore, according to the theory of convergence, as capital accumulates, the speed of convergence is reduced. In other words, it is possible that the output growth contribution of IT capital is lower in states that own a larger share of the national IT capital stock. As stated earlier, eight states own more than half of the IT capital stock of the United States. If output growth contributions are lower in these states, the overall national aggregated contribution of IT capital to output growth would also be lower. This would also partly explain the productivity paradox. Looking at Table 5.8, it seems that these eight states do not indeed present the highest growth contribution of IT capital: California, New York, Texas, Illinois, Florida, Pennsylvania, New Jersey and Ohio are ranked 26th, 49th, 9th, 32nd, 11th, 44th, 28th and 41st, respectively. The average ranking for those eight states is 30th. Although not significant, the correlation between the contribution to growth and the share of national IT capital stock is estimated at -0.07. Hence, IT capital may make an important contribution to growth for many states and a less important one in the few states that account for most of the national IT capital stock. Thus, the productivity paradox may be true at the national level, not at the level of individual states.