R. Only in the special case that all fired workers fail to find reemployment in the survey period would WR = Rs. Hence R provides an upper bound on WR. The minimum bound on WR, is max (GD). That is, if 75 jobs were created economy- wide and 50 were destroyed, we know that at least 75 workers had to change employment status or place of employment. Imagine that all 50 workers whose jobs were destroyed found employment at the newly created jobs. Then 25 other workers would have had to be drawn in to fill the remaining 25 new jobs. Hence the 50 fired workers changed place of employment and the 25 new workers changed employment status, moving from unemployed to employed. 10.2 Job Creation and Destruction: Facts In this section we sketch briefly only the high points of the results in DHS. To answer the exercises at the back of this chapter, you will need to consult the text directly. Over the sample period 1973-1988, the net manufacturing job creation rate (a - d) aver- aged-1.1%. This basic fact obscures the variation of job creation and destruction over the business cycle, by industry and by plant characteristic. In this section, we hit some of the high points. The average annual rate of job destruction d, in manufacturing was 10.3%, and the average rate of job creation was slightly lower at 9.1%, so the average rate of job reallocation was 19.4%. The rate of job creation hit a peak of 13.3% in the recovery year of 1984, while the rate of job destruction peaked at 14.5% and 156% in the recession years of 1982-83. This points to the striking cyclical nature of job creation and destruction: in recessions job destruction spikes well above its mean, while job creation does not fall that much. Moreover, most of the job creation and destruction is concentrated in plants that open or close, rather than in plants that change size. When DHS look at gross job flows across industries, they find that rates of job reallocation are uniformly high, ie, all industries create and destroy lots of jobs. However, high-wage industries tend to have smaller gross job flows than low-wage industries. Finally, they find that the degree to which an industry faces competition from imports does not significantly affect job destruction. (For your information imports make up less than 13% of the market for 80% of U.S.industries.) Examining gross job flows by employer characteristics reveals that most jobs are not cre ated by small business but rather by large, old firms. The pervasive myth of small-business job creation is fed by bureaucratic self-interest and by some elementary statistical errors. Understanding these errors is an instructive exercise in its own right and one of the most interesting parts of the DHS book