5.4Labor Productivity Growth Contribution of IT Capital, by State

This section focuses on the contribution of IT capital stock to growth in labor productivity by sate. The equation estimated is:

gr(Ys/Ls)= α0 gr(KNITs/Ls)+ α1 gr(KITs/Ls) + TFP(5.6)

Labor productivity growth depends on IT and traditional capital deepening (KIT/L and KNIT/L), and total factor productivity (TFP), which includes labor quality in this study. First, output, capital and hours worked are aggregated across industries for each state each year between 1977 and 1997. In order to compute income shares for each state and years, the lagged ratio of input to output must be multiplied by the input’s marginal return. Average output elasticities, or marginal returns of IT and traditional capital, are estimated from equation 5.2 for each state at the detailed industry level (Table 5.4). Income shares are computed for each state each year. Growth rates of IT and traditional capital per hour worked (capital deepening) are computed for each state each year. All these values are averaged over the period 1977-1997. Then, the labor productivity growth contribution of IT and traditional capital deepening are computed by state, as the product of average income shares and average growth rates. Table 5.9 shows: the productivity growth contribution and percentage of average growth in state productivity for IT and traditional capital; the average growth rate of state productivity; IT and traditional capital deepening; and TFP, respectively. Across states, 6% of the average labor productivity growth was due to IT capital deepening, 18% was due to other non-residential capital deepening, and the remaining 76% was due to residential capital deepening, labor quality improvement and total factor productivity. However, these results vary by state. The contribution of IT capital deepening varies from 2.25% to 11.07% across states. Furthermore, some of the lowest contributions of IT capital deepening are observed in the states that own more than half of the country’s IT capital stock. Indeed, California, New York, Texas, Illinois, Florida, Pennsylvania, New Jersey and Ohio are ranked 34th, 44th, 17th, 37th, 7th, 42nd, 43rd and 38th (average ranking is 33rd). The correlation between a state’s share of national IT capital stock and the contribution of this stock to productivity growth is negative although not significant (-0.133). The productivity paradox may again be explained with the convergence theory: IT capital highly contributes to growth in productivity when states start to accumulate IT capital. The magnitude of this contribution is then reduced as states converge to their ideal level of IT capital stock. However, when IT capital stock is considered nationally, its contribution to productivity growth seems lower because it is actually lower in states that own the highest share of this capital stock. Thus, long learning lags are needed to allow benefits from IT capital, bur rent dissipation make the returns to IT capital diminish over time as capital accumulates.