Measuring Payoffs from Information-Technology Investments: New Evidence from Sector-Level Data on Developed and Developing Countries

The payoffs from investments in information technology are investigated at the international level by analyzing a database containing six years of output, employment and investment data for six industry sectors in 36 countries. The Solow endogenousgrowth model is employed to derive parameters of a Cobb-Douglas production function relating aggregate GDP to levels of three inputs: IT capital, non-IT capital and labor. For the full set of countries we find the coefficients on IT capital in the linear regression model to be significant and positive, indicating productivity payoffs from IT investment during the period. Using the database to estimate actual factor shares, we find that there was a substantial underinvestment in IT compared to the levels predicted by applying neoclassical assumptions. Furthermore, estimates of the marginal products of IT and nonIT capital indicate that IT investments were five to eight times more productive at the margin than non-IT investments. We found that the same overall pattern holds for the subset of the database consisting only of developed countries, but no evidence of similar relationships were found among in the developing-country subset. These results confirm the main findings of earlier country-level investigations into the so-called productivity paradox, while diverging from earlier papers in terms of the relative elasticities of IT and non-IT capital. By applying the Solow model to sectors as the unit of analysis, we have avoided the need to estimate capital-stock variables from capital-flow variables and to account for investments in human capital. We speculate that this methodology contribution may account for the differences between our elasticity estimates and those of earlier studies. Measuring Payoffs from IT Investments page 2 R.K. Plice and K.L. Kraemer