• Selected
• Entire List

Abstract:
After documenting the long decline in private sector unionism over the last 50 years, we examine data on NLRB representation elections to determine if changes in the administration of the NLRA during the 1980s reduced the level of organizing activity and success. While organizing activity sharply declined in 1981 (just before President Reagan's showdown with the air traffic controllers' union, PATCO), we find little evidence that the changes in the administration of the NLRA later in the decade adversely affected the level of union organizing activity. We then present an accounting framework that decomposes the sharp decline in the private-sector union membership rate into components due to 1) differential growth rates in employment between the union and nonunion sectors, and 2) changes in the union new organization rate (through NLRB-supervised representation elections). We find that most of the decline in the union membership rate is due to differential employment growth rates and that changes in union organizing activity had relatively little effect. Given that the differential employment growth rates are due largely to broader market and regulatory forces, we conclude that the prospects are dim for a reversal of the downward spiral of labor unions based on increased organizing activity.

Abstract:
I examine changes in the incidence and consequences of job loss between 1981 and 2001 using data from the Displaced Workers Surveys (DWS) from 1984-2002. The overall rate of job loss has a strong counter-cyclical component, but the job loss rate was higher than might have been expected during the mid-1990's given the strong labor market during that period. While the job loss rate of more-educated workers increased, less-educated workers continue to have the highest rates of job loss overall. Displaced workers have a substantially reduced probability of employment and an increased probability of part-time employment subsequent to job loss. The more educated have higher post-displacement employment rates and are more likely to be employed full-time. The probabilities of employment and full-time employment among those reemployed subsequent to job loss increased substantially in the late 1990s, suggesting that the strong labor market eased the transition of displaced workers. Reemployment rates dropped sharply in the recession of 2001. Those re-employed, even full-time and regardless of education level earnings declines relative to what they earned before they were displaced. Additionally, foregone earnings growth (the growth in earnings that would have occurred had the workers not been displaced), is an important part of the cost of job loss for re-employed full-time job losers. There is no evidence of a decline during the tight labor market of the 1990s in the earnings loss of displaced workers who were reemployed full-time. In fact, earnings losses of displaced workers have been increasing since the mid 1990s.

Abstract:
We use a demand/supply framework to analyze 1) the decline in union membership since 1977 in the United States and 2) the difference in unionization rates between the Untied States and Canada. We extend earlier work on these problems by analyzing new data for 1991 from the General Social Survey and for 1992 from our own household survey on worker preferences for union representation. When combined with earlier data for 1977 from the Quality of Employment Survey and for 1984 from a survey conducted for the AFL-CIO, we are able to decompose changes in unionization into changes in demand and changes in supply. We also analyze data for 1990 from a survey conducted for the Canadian Federation of Labor on the preferences of Canadian workers union representation.

We find that virtually all of the decline in union membership in the United States between 1977 and 1991 is due to a decline in worker demand for union representation. There was almost no change over this period in the relative supply of union jobs. Additionally, very little of the decline in unionization in the U.S. can be accounted for by structural shifts in the composition of the labor force. Next, we find that all of the higher unionization rate in the U.S. public sector in 1984 can be accounted for by higher demand for unionization and that there is actually more frustrated demand for union representation in the public sector. Finally, we tentatively conclude that the difference in unionization rates between the U.S. and Canada is accounted for roughly in equal measure by differences in demand and in supply.

Abstract:
We study the reaction of stock prices to announcements of reductions in force (RIFs) using a sample of nearly 3878 such announcements in 1176 large firms during the 1970-97 period collected from the Wall Street Journal Index. We note that, although there has been a dramatic secular increase in news stories related to job loss, the total number of actual announcements fro the firms in our sample follows the business cycle quite closely. We then examine changes over time in standard summary statistics (means, meridians, fraction negative) of the distribution of stock market reactions as well as changes over time in kernel density estimates of this distribution. We find clear evidence that the distribution of stock market reactions has shifted to the right (became less negative) over time. One possible explanation for this change is that, over the last three decades, RIFs designed to improve efficiency have become more common relative to RIFs designed to cope with reductions in product demand. We find that, although this explanation shows some promise, most of the decline in the negative stock price reaction remains unexplained.

Abstract:
No abstract is available for this paper.

Abstract:
The central claim of a rapidly growing literature in international relations is that members of pairs of democratic states are much less likely to engage each other in war or in serious disputes short of war than are members of other pairs of states. Our analysis does not support this claim. Instead, we find that the dispute rate between democracies is lower than is that of other country pairs only after World War II. Before 1914 and between the World Wars, there is no difference between the war rates of members of democratic pairs of states and those of members of other pairs of states. We also find that there is a higher incidence of serious disputes short of war between democracies than between nondemocracies before 1914. We attribute this cross-temporal variation in dispute rates to changes in patterns of common and conflicting interests across time. We use alliances as an indicator of common interests to show that cross-temporal variation in dispute rates conforms to variations in interest patterns for two of the three time periods in our sample.

Abstract:
No abstract is available for this paper.

Abstract:
I model the labor supply of taxi drivers as the result of optimization based on an inter-temporal utility function. Since income effects in response to temporary fluctuations in daily earnings opportunities are likely to be small, cumulative hours will be much more important than cumulative income in the decision to stop work on a given day. However, if these income effects are large due to very high discount and interest rates, then labor supply functions could be backward bending, and, in the extreme case where the wage elasticity of daily labor supply is minus one, drivers could be target earners. Indeed, Camerer, Babcock, Lowenstein, and Thaler (1997) and Chou (2000) find that the daily wage elasticity of labor supply of New York City cab drivers is substantially negative and conclude that it is likely that cab drivers are target earners. I conclude from my empirical analysis, based on new data, of the stopping behavior of New York City cab drivers that, when accounting for earnings opportunities in a reduced form with measures of clock hours, day of the week, weather, and geographic location, cumulative hours worked on the shift is a primary determinant of the likelihood of stopping work while cumulative income earned on the shift is weakly related, at best, to the likelihood of stopping work. This is consistent with there being inter-temporal substitution and inconsistent with the hypothesis that taxi drivers are target earners.

Abstract:
New union members in the United States are typically gained through workplace elections. We find that the annual number of union elections fell by 50 per cent in the early 1980s. A formal model indicates that declining union election activity may be due to an unfavourable political climate which raises the costs of unionization, even though the union win-rate remains unaffected. We relate the timing of declining election activity to the air-traffic controllers' strike of 1981, and the appointment of the Reagan Labor Board in 1983. Empirical analysis shows that the fall in election activity preceded these developments.

Abstract:
New data on medical malpractice claims against a single hospital where a direct measure of the quality of medical care is available are used to address 1) the specific question of the role of the negligence rule in the dispute settlement process in medical malpractice, and 2) the general question of how the process of negotiation and dispute resolution in medical malpractice operates with regard to both the behavior of the parties and the outcome of the process. We find that the quality of medical care is an extremely important determinant of defendants' medical malpractice liability. More generally, we find that the data are consistent with a model where (1 the plaintiff is not well informed ex ante about the likelihood of negligence and 2) the ex ante expected value to the plaintiff of a suit is high relative to the costs of filing a suit and getting more information. Thus, suits are filed even where there is no concrete reason to believe there has been negligence, and virtually all suits are either dropped or settled based on the information gained after filing. We conclude that the filing of suits that appear, ex post, to be nuisance suits can be rational equilibrium behavior, ex ante, where there is incomplete information about care quality.

Abstract:
The public believes that job security has deteriorated dramatically in the United States. In this study, I examine job durations from eight supplements to the Current Population Survey (CPS) administered between 1973 and 1993 in order to determine if, in fact, there has been a systematic change in the likelihood of long- term employment. In order to measure changes in the distribution of job durations, I examine changes in selected quantiles (the median and the 0.9 quantile) of the distribution of duration of jobs in progress. I also examine selected points in the cumulative distribution function including the fraction of workers who have been with their employer 1) less than one year, 2) more than ten years, and 3) more than twenty years. The central findings are clear. By the measures I examine, there has been no systematic change in the overall distribution of job duration over the last two decades, but the distribution of long-term jobs across the population has changed in two ways. First, individuals, particularly men, with little education (less than twelve years) are substantially less likely to be in long jobs today than they were twenty years ago. Second, women with at least a high-school education are substantially more likely to be in long jobs today than they were twenty years ago.

Abstract:
I examine changes in the incidence and consequences of job loss by reported cause between 1981 and 1993 using data from Displaced Workers Surveys (DWS), conducted as part of the Current Population Survey (CPS) in even years since 1984. The overall rate of job loss is up somewhat in the 1990s. The increase in job loss is larger for older and more educated workers, but younger and less-educated workers continue to have the highest rates of job loss. Some significant changes are also found in the rate of job loss by reported reason. Next I examine the consequences of displacement for several post- displacement labor market outcomes, including the probability of employment, full-time/part-time status, the change in earnings, job stability, and self-employment status. The adverse consequences of job loss, which have always been substantial, do not appear to have changed systematically over time. More educated workers suffer less economic loss relative to income due to displacement than do the less educated. Self-employment appears to be an important response to displacement, and older workers and the more educated are more likely to turn to self-employment.

Abstract:
A central feature of the litigation process that affects case outcomes is the selection of cases for litigation. In this study, we present a theoretical framework for understanding the operation of this suit selection process and its relationship to the underlying distribution of potential claims and claimants. We implement the model empirically by assuming that individuals vary more in their litigiousness (inverse costs of litigation) than do corporations. This assumption, coupled with the case selection process we present, yields clear predictions on trial rates as a function of whether the plaintiff and defendant were individuals or corporations. The model also yields a prediction on the plaintiff's win rate in lawsuits as a function of the plaintiff's identity. Our empirical analysis, using data on over 200,000 federal civil litigations, yields results that are generally consistent with the theory. Lawsuits where the plaintiff is an individual are found to have higher trial rates and lower plaintiff win rates.

Abstract:
The process through which cases are selected for litigation cannot be ignored because it yields a set of lawsuits and plaintiffs that is far from a random selection either of potential claims or of potential claimants. We present a theoretical framework for understanding the operation of this suit-selection process and its relationship to the underlying distribution of potential claims and claimants. The model has implications for the trial rate and the plaintiff win rate at trial. Our empirical analysis, using data on over 200,000 federal civil litigations, yields results that are strongly consistent with the theory.

Abstract:
No abstract is available for this paper.

Abstract:
I establish four facts regarding the pattern of NLRB supervised representation election activity over the past 45 years: 1) the quantity of election activity has fallen sharply and discontinuously since the mid-70's after increasing between the mid-1950's and the mid-1970's; 2) union success in elections held has declined less sharply, though continuously, over the entire period; 3) it has always been the case that unions have been less likely to win NLRB-supervised representation elections in large units than in small units; and 4) the size-gap in union success rates has widened substantially over the last forty years. I develop a simple optimizing model of the union decision to hold a representation election that can account for the first three facts. I provide a pair of competing explanations for the fourth fact: one based on differential behavior by employers of different sizes and one purely statistical. I then develop and estimate three empirical models of election outcomes using data on NLRB elections over the 1952-98 time period in order to determine whether the simple statistical model can account for the size pattern of union win rates over time. I conclude that systematic union selection of targets for organization combined with the purely statistical factors can largely account for observed patterns.

Abstract:
I establish four facts regarding the pattern of NLRB supervised representation election activity over the past 45 years: 1) the quantity of election activity has fallen sharply and discontinuously since the mid-70's after increasing between the mid-1950's and the mid-1970's; 2) union success in elections held has declined less sharply, though continuously, over the entire period; 3) it has always been the case that unions have been less likely to win NLRB-supervised representation elections in large units than in small units; and 4) the size-gap in union success rates has widened substantially over the last forty years. I develop a simple optimizing model of the union decision to hold a representation election that can account for the first three facts. I provide a pair of competing explanations for the fourth fact: one based on differential behavior by employers of different sizes and one purely statistical. I then develop and estimate three empirical models of election outcomes using data on NLRB elections over the 1952-98 time period in order to determine whether the simple statistical model can account for the size pattern of union win rates over time. I conclude that systematic union selection of targets for organization combined with the purely statistical factors can largely account for observed patterns.

Abstract:
No abstract is available for this paper.

Abstract:
We examine whether the decline in the availability of employer-provided health insurance is a phenomenon common to all jobs or is concentrated only on certain jobs. In particular, we investigate the extent to which employers have continued to provide health insurance on what we term reducing the availability of health insurance on jobs. We consider two dimensions on which jobs may be considered peripheral: if they are new (tenure less than one year) or part-time. We consider three outcomes whose product is the health insurance coverage rate: 1) the fraction of workers who are in firms that offer health insurance to at least some workers (the offer rate); 2) the fraction of workers who are eligible for health insurance, conditional on being in a firm where it is offered (the eligibility rate); and 3) the fraction of workers who enroll in health insurance when they are eligible for it (the takeup rate). We find that declines in own-employer insurance coverage over the 1988-1997 period are driven primarily by declines in takeup for core workers and declines in eligibility for peripheral workers. We also look at trends by workers' education level and see how much of the decline is offset by an increase in coverage through a spouse's policy. Our findings are consistent with the view that employers are continuing to make health insurance available to their core long-term employees but are restricting access to health insurance by their peripheral short-term and part-time employees.

Abstract:
Central to understanding the effect of arbitration schemes on the process of collective bargaining is understanding the process by which arbitrators make decisions. A model of arbitrator behavior inconventional arbitration is developed that allows the arbitration award to be a function of both the offers of the parties and the(exogenous) facts of the case. The weight that the arbitrator puts on the facts relative to the offers is hypothesized to be a function of the quality of the offers as measured by the difference between the offers. Two special cases of this model are derived: 1) the arbitrator bases the award strictly on the offers of the parties(split-the-difference) and 2) the arbitrator bases the award strictly on the facts of the case.The model is implemented empirically using data gathered from practicing arbitrators regarding their decisions in twenty-five hypothetical cases. These data have the advantage that they allow causal inference regarding the effect on the arbitration award of the facts relative to the offers. On the basis of the estimates, both of the special case models are strongly rejected. The arbitration awards are found to be influenced by both the offers of the parties and the facts of the case. In addition, the weight put on the facts of the case relative to the offers is found to vary significantly with the quality of the offers. When the offers are of low quality (far apart)the arbitrator weights the facts more heavily and the offers less heavily.These results suggest that the naive split-the difference view of arbitrator behavior, which is the basis of the critique of conventional arbitration that has led to the adoption of final-offer arbitration, is no correct in its extreme view. On the other hand,the awards are affected by the offers so that the parties can manipulate the outcome to some extent by manipulating their offers. However, the scope for this sort of influence is limited by the finding that the offers are weighted less heavily as their quality deteriorates.

Abstract:
We develop a model of the plaintiff's decision to file a law suit that has implications for how differences between the federal government and private litigants and litigation translate into differences in trial rates and plaintiff win rates at trial. Our case selection model generates a set of predictions for relative trial rates and plaintiff win rates depending on the type of case and whether the government is defendant or plaintiff. In order to test the model, we use data on about 350,000 cases filed in federal district court between 1979 and 1997 in the areas of personal injury and job discrimination where the federal government and private parties work under roughly similar legal rules. We find broad support for the predictions of the model.

Abstract:
It is a well known fact that the extent of unionization is lower in states with Right-to-Work (RTW) laws. A framework is developed for determining whether RTW laws actually cause a decrease in the extent of unionization or whether they simply mirror preexisting tastes of workers against unions. A set of empirical tests is proposed that can distinguish between these explanations based on differences between RTW and non-RTW states in the demand for union representation, the supply of union jobs relative to that demand, and the observed union-nonunion wage differential. Data from the Quality of Employment Survey and from the Current Population Survey are utilized to implement the tests.The results indicate that the demand for union representation is significantly lower in states with RTW laws.At the same time no significant difference is found on the basis of RTW laws in the supply of union jobs relative to demand. It is also found that the observed union-nonunion wage differential is slightly larger in RTW states.This pattern is consistent with the hypothesis that RTW laws simply mirror preexisting preferences against union representation. In its entirety it is not consistent with the hypothesis that RTW laws cause a decrease in the extent of unionization.A final interesting result is that it is found that the extent of unionization in the south is lower even after controlling for the presence of RTW laws in many of the states in that region. Further, it is determined that this is due to a supply of union jobs in the south that is more constrained relative to demand than elsewhere. This suggests that there exist a set of institutional or economic factors in the souththat makes union organizing more difficult and expensive independent of the existence of RTW laws.

Abstract:
I investigate how the the threat of union organization affects the wage paid to nonunion workers. I start by outlining the standard model of wage determination by a nonunion employer when faced with the threat of union organization. The model suggests that the nonunion wage will be directly related and the union wage gap will be inversely related to the threat. I use repeated cross-section data from the CPS from 1977-2002 to develop a measure of the threat as the predicted probability of union membership. I use this measure to estimate earnings functions that use several sources of variation in the likelihood of union membership to identify the threat effect in a manner that reduces the likelihood of omitted variable bias. Finally, I investigate two cases where there has arguably been a change in the likelihood of union organization that is not correlated with changes in the demand for labor. These include the wage changes surrounding the introduction of right-to-work (RTW) laws in two states during the period studied and wage changes surrounding deregulation of key industries in the late 1970s and early 1980s. The results are mixed. The preferred estimates from the analysis using predicted probability of unionization as the threat measure, imply very little relationship between either nonunion wages or the union wage gap and the threat. The estimates that rely on the introduction of RTW laws show a significant relationship between nonunion wages and the introduction of RTW laws in one of the two states. Stronger evidence of threat effects is found in the experience of deregulated industries, where regulation was a central factor in union strength.

Abstract:
No abstract is available for this paper.

Abstract:
A model of the determination of the union status of workers is developed that incorporates the separate decisions of workers and potential union employers in a framework which recognizes the possibility of an excess supply of workers for existing union jobs.This theoretical framework results in an empirical problem of partial observability because information on union status is not sufficient to determine whether nonunion workers are nonunion because they do not desire union representation or because they were not hired by union employers despite a preference for union representation.The problem is solved by using data from the Quality of Employment Survey that have a unique piece of information on worker preferences which allows identification and estimation of the model.The empirical results yield some interesting insights into the process of union status determination that cannot be gained from a simple logit or probit analysis of unionization. Chief among these relate to the unioniza-tion of nonwhites and southerners.The well-known fact that nonwhites are more likely to be unionized than otherwise equivalent whites is found largelyto be due to a greater demand for union representation on the part of non-white workers. The equally well-known lower propensity to be unionized among southern workers is found to be due to a combination of a lower demand for union representation on the part of southern workers and a supply of union jobs which is more constrained relative to demand than in the North.

Abstract:
We develop and estimate a model of the union's optimal extent of
organizing activity that accounts for the decision of employers
regarding resistence to union organizing. The central exogenous
variable in the analysis is the quantity of quasi-rents per
worker available to be split between unions and employers.

We measure available quasi-rents per workers as the difference
per worker between total industry revenues net of raw materials
costs and labor costs evaluated at the opportunity cost of the
workers. Using two-digit industry level data for thirty-five
U.S. industries for the period 1955 through 1986, we find that
both organizing activity and employer resistance to
unionization are positively related to available quasi-rents
per workers. However, there is still a strong negative trend
in union organizing activity and a strong positive trend in
employer resistance after controlling for quasi-rents per worker.
Thus, the explanation for the decline in union organizing
activity and the increase in employer resistance to
unionization since the mid 1970's lies elsewhere.

Abstract:
In 1955 only a few states had laws governing collective bargaining by public employees. By 1984 only a few states were without such laws. The emergence of these policies coincides with a dramatic increase in unionization among public employees, and an important puzzle is the direction of causality between the laws and public employee unionization. A key piece of the solution is understanding the evolution of the public policy in this area, and this is the focus of the analysis in this study. A Markov model of the evolution of these laws is developed based on the idea that states will change their existing policy if and only if their preferences deviate from the existing policy by more than the cost of a change in policy. The key underlying constructs are 1) the intensity of state preferences for or against public sector collective bargaining and 2) the cost of changing an existing policy or enacting a new policy. The model is implemented empirically using state level data on policy for each year from 1955 to 1984. The results suggest that state preferences for a pro-bargaining policy are positively related to 1) the COPE score (a measure of pro-union congressional voting behavior on labor issues), 2) income per capita, and 3) the size of the public sector and negatively related to southern region. The costs of policy change were hypothesized to be a function of structural measures of the legislative process, but no support for this was found in the data. Use of the estimates to predict probabilities of states having laws of various kinds suggests that the model can predict the aggregate distribution of laws relatively well but that the model does less well at distinguishing between states with laws of different types.

Abstract:
One of the most prominent features of U.S. unionism is the key role played by seniority. However, in cross-sectional data, the positive association between seniority and earnings is typically much stronger for nonunion workers than for union workers. This finding has puzzled previous researchers, since it seems inconsistent with the generalization that seniority is more important in the union sector than in the nonunion sector. We show that standard estimates of the return to seniority are likely to be biased upward and argue that the bias is likely to be larger in the nonunion sector than in the union sector. Corrected estimates imply that the return to seniority is, in fact, larger in the union sector than in the nonunion sector.

Abstract:
No abstract is available for this paper.

Abstract:
I examine the extent to which workers who lose jobs find work in alternative employment arrangements including temporary work and independent contracting and find part-time work, both voluntary and involuntary. The analysis is based on data from the Displaced Worker Supplements (DWS) and the February 1994 and 1996 Current Population Surveys (CPS) which I match to the Contingent and Alternative Employment Arrangements Supplements (CAEAS) to the February CPSs in the subsequent years (1995 and 1997 respectively). I find that job losers are significantly more likely than non-losers to be in temporary jobs (including on-call work and contract work). I also find evidence that the likelihood of temporary employment falls with time since job loss. With regard to part-time employment, I find that involuntary part-time employment is an important part of the employment experience subsequent to job loss and that the likelihood of involuntary part-time employment falls with time since job loss, particularly for full-time job losers. Thus that temporary and involuntary part-time jobs are part of a transitional process subsequent to job loss leading to regular full-time employment.

Abstract:
No abstract is available for this paper.

Abstract:
One prominent explanation for disagreement in bargaining is that the parties have divergent and relatively optimistic expectations regarding the ultimate outcome if they fail to agree. The fact that settlement rates are much higher where final-offer arbitration is the dispute settlement procedure than where conventional arbitration is the dispute settlement procedure is used as the basis of a test of the role of divergent expectations in causing disagreement in negotiations. Calculations of identical-expectations contract zones using existing estimates of models of arbitrator behavior yield larger identical-expectations contract zones in conventional arbitration than in final-offer arbitration. This evidence clearly suggests that divergent expectations alone are not an adequate explanation of disagreement in labor-management negotiations. A number of alternative explanations for disagreement are suggested and evaluated.

Abstract:
No abstract is available for this paper.

Abstract:
The labor supply of taxi drivers is consistent with the existence of intertemporal substitution. My analysis of the stopping behavior of New York City cabdrivers shows that daily income effects are small and that the decision to stop work at a particular point on a given day is primarily related to cumulative daily hours to that point. This is in contrast to the analysis of Camerer et al., who find that the daily wage elasticity of labor supply of New York City cabdrivers is substantially negative, implying large daily income effects. This difference in findings is due to important differences in empirical methods and to problems with the conception and measurement of the daily wage rate used by Camerer et al.

Abstract:
Despite a record of sustained growth in employment in the United States, there is long-standing concern that the new jobs are of poor quality, implying that the quality of the stock of jobs in the economy is deteriorating. In fact, it is difficult to define a new job, much less identify such jobs and evaluate their quality. In this study, I define new jobs operationally as worker-firm matches that have begun within the last year (i.e., tenure less than one year), and I investigate the extent to which the quality of new (low tenure) jobs relative to old (higher tenure) jobs has changed over the period from 1979-1996. I consider three dimensions of quality: real wages, the rate of part-time employment, and the rate of coverage by employer-provided health insurance. The empirical analysis relies on data from eight mobility and benefit supplements to the Current Population Survey over the period studied. The results are clear-cut. Real hourly wages on new jobs have deteriorated slightly relative to wages on older jobs, but the general patterns of the wage distribution on new jobs have changed in ways similar to the overall wage distribution (a large increase in the return to education driven by a large decline in wages for less-skilled workers). There is no evidence of an increase in the rate of part-time employment on either new jobs or old jobs so that the quality of new jobs has not changed absolutely or relative to the quality of older jobs in this dimension. The quality of new jobs has deteriorated substantially for some workers in the provision of employer-provided health insurance. Less-educated workers have become substantially less likely to be covered by or offered health insurance by their employer, and the decline is particularly large on new jobs. There has been a smaller decline in employer-provided health insurance coverage for more-educated workers on both old and new jobs, but the decline is no larger on new jobs than on old jobs. On balance, there has been a decline in the quality of jobs for less-skilled workers, as measured by the availability of a key fringe benefit, that is especially severe for new jobs and that reinforces the well known deterioration of the labor market for less-skilled workers more generally. There has been relatively little change in the quality of jobs available to more highly skilled workers, either on new or on old jobs.