Sachs and Shatz (SSh) (1994) follow the LS methodology with two minor changes and one major change. First, they time the decade of the 1980s as 1978 through 1989. Second, they use a slightly different dependent variable: they share of production employment in total industry employment. Their major difference from LS is the treatment of the computer industry (SIC 357 at the three-digit level, SIC 3573 at the four-digit level). They argue (p. 37) that LS “should have separated the effects of computer prices from the other sectors.” The reason is that the computer-industry price data are exceptionally mismeasured: “The relative prices of computers fell sharply during the decade, matching extraordinary productivity increases. The exact measurement of these price and productivity changes is highly problematic, so that it is important that these changes do not overwhelm the message in the rest of the data.” Thus, using both import prices and the full sample of three-digit SIC domestic prices, SSh run the following regression using OLS. payday loans az
(SSh) Pj*1980s = a+ b (PW/(PW+NPW))j1980 + bc(Dcomputers) + ej
The main results from these regressions (Table 16, p. 38) are b is estimated to be negative (insignificantly so for import prices, significantly (with a t-statistic of -1.98) for domestic prices) while bc is also estimated to be negative (very significantly so (t-statistic of -13.40) for domestic prices). Thus, SSh find among the non-computer sample that industries employing a larger share of production workers had lower relative-price increases over the 1980s. SSh conclude this supports the hypothesis that international trade contributed to rising U.S. wage inequality by raising the relative price of skilled-labor-intensive products.