Twenty-seven percent of the Union Army prisoners captured July 1863 or later died in captivity. At Andersonville, the death rate may have been as high as 40 percent. How did men survive such horrific conditions? Using two independent datasets, we find that friends had a statistically significant positive effect on survival probabilities and that the closer the ties between friends as measured by such identifiers as ethnicity, kinship, and the same hometown, the bigger was the impact of friends on survival probabilities.

The conventional wisdom is that African-Americans, Hispanics, and American Indians are underrepresented among faculty in postsecondary institutions because they are underrepresented among Ph.D. recipients. Thus, the putative solution to the problem of minority faculty underrepresentation is to increase the supply of minority Ph.D.’s. In our book, Faculty of Color in Academe: Bittersweet Success (Turner and Myers, 2000) we point out that the supply-side argument has several flaws. In this and a companion paper (Myers and Turner, 2003) we replicate and update the analysis performed in Chapter 7 of our book using more recent census data and a larger sample for 1990. Once again, we demonstrate that an autonomous increase in the Ph.D. supply, uniform across all groups, would leave the representation of African-American and Hispanic faculty largely unchanged. This conclusion challenges the view that the underrepresentation of minority faculty is solely a supply-side phenomenon that can be addressed primarily by increasing the pipeline for new minority Ph.D.’s. Although a strong case can be made for increasing the minority pipeline, the pipeline itself does not appear to be the central cause of the continued underrepresentation of minority faculty. I. The Problem: At every point in the educational pipeline from the Bachelor’s degree to the doctoral degree, African-Americans, Hispanics, and American Indians are substantially underrepresented.

Nonmarket goods such as unpaid household labor, leisure, health and longevity, and the environment are important components of the standard of living. They represent a large fraction of all activities. Prime-aged men and women spend 17 percent of their day in leisure activities, and 5 and 13 percent of their time, respectively, in unpaid housework compared to 23 and 13 percent of their time, respectively, in paid work. The quantity of nonmarket goods is rising over time. In the United States, life expectancy at birth is now 77 years, having risen by 29 years since 1900. In Los Angeles County between 1980 and 1998, average annual daily exceedences of the national smog standard declined by 60 days from 71 to 11. Falling big-city murder rates merit national news headlines. Studies of living standards have focused on the tremendous change in the quantity of nonmarket goods but have assumed that the value of nonmarket goods, except for unpaid labor, has remained constant. This assumption underlies most health studies (e.g., David Cutler and Elizabeth Richardson, 1997; Kevin M. Murphy and Robert H. Topel, 2003; William T. Nordhaus, 2003). Nordhaus (2003) valued declines in mortality since 1900 using a constant value of life. Even the Boskin CPI Commission (Michael J. Boskin et al., 1998) discussed trends in the quantity of nonmarket goods (i.e., pollution and crime progress) without mentioning incorporating such goods’ implicit prices into a broader CPI measure. There is no reason to think that implicit prices or the willingness to pay for nonmarket goods has remained constant. Rising real and shadow wages have made both leisure and unpaid household labor more expensive. Rising incomes have also made such normal goods as safety, health, a temperate climate, and the environment more valuable. We document the price dynamics of nonmarket goods by estimating repeat cross-sectional hedonic regressions. We focus on two important and measurable nonmarket goods: job fatality risk and climate. In both cases we find that both price and quantity have been rising. This evidence is consistent with rising valuation. We use our estimates of job-risk compensating differentials to construct new evidence on long-run trends in value of life. Accounting for price changes affects how we view the retrospective and prospective benefits of medical innovations. A rising value of life implies that marginal improvements in safety and in longevity are becoming more valuable. We report evidence that the price of living in a temperate winter and summer climate has significantly increased over time.

The U.S. and U.K. Great Depressions Through the Lens of Neoclassical Growth Theory by Harold L. Cole and Lee E. Ohanian. Published in volume 92, issue 2, pages 28-32 of American Economic Review, May 2002

The authors study the implications for shareholder wealth of interfirm antitrust litigation and how the costs of the dispute affect the propensity to settle. Upon filing, defendants experience significant wealth losses that are ten million dollars larger than the wealth gains of plaintiffs. Financial distress, behavioral constraints, and follow-on suits are sources of wealth leakage and influence settlement behavior. Since the threat of a monetary transfer has little power to explain either wealth effects or the likelihood of settlement, the central concern of defendants may be the potential prohibition of profitable business practices.

The New Jobs Tax Credit was one of the four programs in the 1977 economic stimulus package. This program, although viewed primarily as a countercyclical measure, may also alter the equilibrium unemployment rate, UN.' This paper presents our preliminary analysis of the Department of Labor survey, conducted by the Bureau of the Census, in which firms described their responses to this employment tax credit (ETC). To date, our results indicate the potential for a large employment effect. Ordinary least squares estimates suggest that firms which knew about the program increased employment 3 percent faster than other firms. A second analysis which uses multinomial logit techniques indicates that the ETC shifted the entire distribution of employment growth to the right: slowly growing firms increased employment to capture the credit. Since the firms which knew about the program, however, were not randomly drawn, our results may overstate the program's employment effect. Due to the nature of the survey data, we can only focus on direct employment effects.2 It is useful, however, to at least mention the other potential effects of the program. First, unlike Comprehensive Employment Training Act (CETA) programs which increase public employment, the ETC should increase employment in the private sector. Within the private sector, the rules of the current ETC program provide an additional stimulus to the growing industries and, to a lesser extent, to small establishments. Second, the long-run structural effects of this two-year program are probably small. A permanent ETC, however, may be able to lower UN of disadvantaged workers.

William Bomberger in his comment on our recent article raises three points. All are related to our long-run tradeoff between unemployment and inflation. His first point is that our C-shaped Phillips curve is not dependent on noncompetitive forces, as we claim to be the case, but is related to our rather curious use of the planning horizon. His second point is that our longrun model faces serious stability problems. His third point is that our analysis does not yield results which are congruent with the economy's experiences. Bomberger's first point is incorrect. He seems to believe that the size of the markup variable VI captures all of the noncompetitive elements in our model. This is not the case. Although the 4, variable is unique to the noncompetitive sector, measuring the safety factor or the risk aversion of the noncompetitive firm, it is not the only factor which distinguishes the noncompetitive sector. Rather, the main impact of the noncompetitive sector is in introducing the timing or planning period problem. Whereas competitive firms may be viewed as setting wages and prices continuously, the noncompetitive firm acts discontinuously. This creates a planning period over which the noncompetitive firm maintains a fixed wage and price strategy. We also make the important assumption that the firm is more concerned about a customer queue than a labor queue. The result is that it will set a wage premium so as to have a labor queue in the bulk of the planning period. At the extreme where it views a customer queue as prQhibitive, this will mean setting a wage premium sufficient to insure a labor queue throughout the period. The identical result can be obtained by assuming that the elasticity of the supply of labor to the noncomDetitive firm is larger than the elasticity of the derived demand for labor. Both assumptions-that the firm is more concerned with a customer than with a labor queue and that the supply elasticity is greater than the demand elasticity-seem to us to be theoretically viable and well established in the empirical literature. Since Bomberger views 4, as representing the noncompetitive elements in the model, he does not appreciate the role played by the fixed planning period. This leads him to suggest that a more straightforward way of dealing with the planning period is for the firm to consider its average market position. This, however, would eliminate one of the basic assumptions of our analysis. Since the noncompetitive firm is more concerned with a customer than a worker queue, it looks toward the end of the planning period in setting wages and prices. Bomberger observes that when firms look to the middle of the period, setting a wage premium sufficient to have a labor queue in the first half of the period and a customer queue in the second, the numerical elasticity of the long-run Phillips curve is quite low. As he verifies in equation (5) by using a period average (ignoring discounting), the wage markup is, in fact, exactly independent of the inflation rate. This is just what one would expect. We analyzed the midpoint case precisely because it gives a lower bound to the slope of the long-run tradeoff. Our point was and still remains, however, that noncompetitive firms aim beyond the midpoint. As for the numerical elasticity, in the extreme where only a labor queue is tolerable, the noncompetitive firm sets E=er and the aggregate wage premium of our equation (15) is proportional to