Productivity

The NBER's Program on Productivity held its spring meeting in Cambridge on March 16. Organizers C. Lanier Benkard, NBER and Stanford University, and Iain M. Cockburn, NBER and Boston University, put together this program:

Pharmaceuticals are a special case of goods with some unobservable quality prior to consumption, which is called an "experience characteristic." Consumers learn about these experience characteristics from consumption and from advertising. Park proposes price indexes for pharmaceuticals that are innovation-adjusted and applies these indexes to the data for antidepressant drugs during 1980-95. He finds that the key source of innovation is the entry of new products, but that the effects of learning about experience characteristics also are significant. He finds too that the average annual growth rate of the focal price index declines by almost 9.5 percent, which suggests that the existing price indexes for pharmaceuticals may seriously overstate the rate of inflation in a rapidly growing market with the entry of innovative products.

Blow and Crawford suggest a method of valuing a new good that does not depend on the relationship between a household's economic welfare and the goods and services it consumes--only on the existence of such a relationship. They illustrate their technique with U.K. household budget survey data and calculate the welfare effects of the launch of the National Lottery in the United Kingdom in November 1994. The authors show that the increase in economic welfare associated with the arrival of the Lottery was greater for better-off households. They also show how measures of inflation over this period are affected by the inclusion of the new good, and they describe how the distributional effects of inflation are more strongly pro-rich when they allow for the new good.

Pakes considers the use of hedonic techniques to ameliorate new goods biases in price indexes. He provides a conceptual comparison of the hedonic to the traditional matched-model index, stressing the selection bias in the latter that arises from the fact that it does not make price comparisons for discontinued goods, and that the prices of discontinued goods tend to fall more than the average. He then provides a hedonic price index for desktop PCs and compares it to a traditional matched-model index calculated on the same data (the time period is 1995 to 1999). The hedonic price index is always negative, with an average rate of decline of about 15 percent, while the matched-model index is slightly positive and negatively correlated with the hedonic index. This negative correlation is expected because the selection bias in the traditional matched-model index is particularly large in years when many new products entered and made existing products obsolete. Those were precisely the periods when the hedonic index tells us that many product improvements were made.

Hendel and Nevo study intertemporal price discrimination in markets for nondurable but storable goods, such as groceries, with highly frequent sales. They consider a consumer's dynamic problem of being able to store a consumption good and facing uncertain future prices. To test the model, they use weekly store-level price and quantity scanner data on laundry detergents, as well as a household-level dataset. Their preliminary results suggest that duration from previous sale has a positive effect on the aggregate quantity purchased. Two indirect measures of storage are positively correlated with a household's tendency to buy on sale. One measure of inventory is negatively correlated with the quantity purchased (conditional on a purchase) and with the probability of buying, conditional on being in a store.

Stiroh examines the link between information technology (IT) and the U.S. productivity revival in the late 1990s. Industry-level data show a broad productivity resurgence that reflects both the production and the use of IT. The most IT-intensive industries experienced significantly larger productivity gains than other industries. A wide variety of econometric tests show a strong

correlation between IT capital accumulation and labor productivity. To quantify the aggregate impact of IT use and IT production, Stiroh presents a novel decomposition of aggregate labor productivity. He shows that virtually all of the aggregate productivity acceleration can be traced to the industries that either produce IT or use IT most intensively, with essentially no contribution from the remaining industries that are less involved in the IT revolution.

Kim and Marschke develop and test a model of the patenting and R and D decisions of a firm whose researcher-employees sometimes move to a competitor. In their model, a firm facing the prospect of a scientist leaving risks losing its innovations to the scientist's future employer. But a firm can mitigate this risk by moving quickly to patent its scientists' innovations. Thus, a firm's propensity to patent an innovation rises with the likelihood of a researcher's departure. Their model also shows that an increase in the probability of a scientist leaving is likely to reduce research expenditures, and therefore to raise the patent-R and D expenditures ratio. Using firm-level panel data on patenting and R and D and industry estimates of labor mobility, the authors show that firms in industries with higher job turnover rates generate more patents, consistent with firms using patenting to prevent employee misappropriation of intellectual property. Also consistent with their theory are the authors' findings that job turnover rates are negatively correlated with firm-level R and D outlays and positively correlated with the patent-R and D ratio. The evidence indicates that the increasing mobility of scientists may be driving part of the rapid rise in patenting since the mid-1980s.