Labor Economics

The WSJ Real Time Economics blog reviews a paper by Michael Lechner which finds that “sports-playing adults saw a boost in income of about 1,200 euros per year over 16 years when compared to their less active peers. That translates into a 5-10% rate of return on sports activities, roughly equal to the benefit of an extra year’s worth of education.” How can playing sports increase income?

The simplest mechanism is that playing sports increases one’s health level. Healthier people are less likely to get sick and more likely to be able to work to earn income. This health difference, however, only explains a portion of the income differential. Dr. Lechner claims that playing sports builds a social network which helps to increase pay (e.g., your friends are the ones who recommend you for jobs). In fact, Lechner finds that sports-playing men display a higher level of “social functioning” than did the less active men.

One worry of this study is that of reverse causation. If someone is very sick, they are not able to play sports. Further, if you are sick, you are probably less likely to engage in social activities. Thus, health–and not sports playing–may be a hidden, unobserved feature which may be driving these results.

• Hat tip to The Sports Economist
• The full paper is available here.

Nobel laureate James Heckman has a nice summary of how applied econometricians and policy researchers should define causality. Some of the more interesting points I have excerpted below.

On the source of randomness in a sample

“One reason why many statistical models are incomplete is that they do not specify the sources of randomness generating variability among agents, i.e., they do not specify why otherwise observationally identical people make different choices and have different outcomes given the same choice. They do not distinguish what is in the agent’s information set from what is in the observing statistician’s information set, although the distinction is fundamental in justifying the properties of any estimator for solving selection and evaluation problems. They do not distinguish uncertainty from the point of view of the agent whose behavior is being analyzed from variability as analyzed by the observing analyst. They are also incomplete because they are recursive. They do not allow for simultaneity in choices of outcomes of treatment that are at the heart of game theory and models of social interactions and contagion (see, e.g., Brock & Durlauf, 2001; Tamer, 2003).”

Unbundling a treatment

Researchers often say that a policy change will cause a change in some outcome measure. However, a policy change is often made up of many components. Which components of the policy change actually influenced the outcomes? In Heckman’s words:

“Many causal models in statistics are black-box devices designed to investigate the impact of “treatments”—often complex packages of interventions—on observed outcomes in a given environment. Unbundling the components of complex treatments is rarely done. Explicit scientific models go into the black box to explore the mechanism(s) producing the effects.”

Outcomes vs. Utilities

Most researchers pick an outcome variable of interest and if the outcome increases–assuming a beneficial outcome measure–than people are better off. This may not be the case however. For instance, Bill Clinton’s welfare reform act (PRWORA) may have increased employment rates and income for single mothers, but the mother’s utility may have decreased. The single mothers may (or may not) have valued spending time caring for their child more than working.

Problems with non-linearity

Issues such as “social interactions, contagion and general equilibrium effects” can complicate causal inference.

What are you measuring?

Let us assume that Y is the outcome variable of interest. Y depends on what state, s, you are in. For instance, in a treatment/no treatment world, Y(s) is the outcome if you would be treated and Y(s’) is the effect if you were not treated. D(s)=1 if you were actually treated and D(s)=0 if you did not receive treatment in the data. Thus, we can measure various things:

• Average Treatment Effect (ATE): E (Y s) − Y(s’)). This is equal to the average effect if all individuals moved from a untreated to a treated state.
• Treatment on the Treated (TT): E[(Y(s) − Y(s')) | D(s) = 1]. This looks at the average effect of treatment only on those who were treated. This is important if only certain individual select into the treatment group, or if the policy change is only relevant for certain individuals.
• Treatment on the Untreated (TUT): E[(Y(s) − Y(s')) | D(s) = 0]. It is also possible that treatment can affect those who are not treated. For instance, instituting a work training program for treated individuals may reduce community college enrollment and thus may affect untreated individuals (e.g., if the community college closes from lack of enrollment).
• Policy relevant treatment effect (PRTE):E_p[Y(s)] − E_p’ [Y(s)]. The estimator compares the average outcomes of two different policy choices.

Heckman, James (2008) “Economic Causality” NBER WP #13934.

Economists predict that longer life expectancy leads to more investment in education. For those who live a short time, sacrificing working years for education is not worthwhile if the payback period is short. For those with a longer life expectancy, an individual can reap the monetary rewards from education over a longer period of time.

An NBER working paper by Seema Jayachandran looks at a what happens to educational investment after there was a 70% reduction in maternal mortality risk in Sri Lanka. The decreased maternal mortality rate was due to a number of factors: an increase in the number of hospitals, clinics and health centers, an increase in the number of trained birth attendants, transportation improvements such as free ambulances, and increased adoption of western technologies such as sulfa drugs and penicillin. Further, there was significant success in eradicating malaria during this time. Also, most of the medical services were provided for free.

Jayachandran uses a difference-in-difference-in-difference (DDD) strategy. “The first difference is over time, since maternal mortality fell between 1946 and 1953. The second difference is across geographic areas; the magnitude of the MMR declines varied considerably across Sri Lanka’s 19 districts. The third difference is between genders; maternal mortality is quite unique among major causes of death in that it exclusively pertains to women.”

The results of the study are that the decline in “…maternal mortality risk over the sample period increased female life expectancy at age 15 by 4.1%, female literacy by 2.5%, and female years of education by 4.0%.”

• Seema Jayachandran (2008) “Life Expectancy and Human Capital Investments: Evidence From Maternal Mortality Declines” NBER WP #13947.

Should employers provide health insurance to their employees? There are many reasons why they should. One is that employees are attracted to firms that offer health insurance, especially since their are tax and cost advantages to group health insurance purchased through an employer. Another reason is that if a worker becomes sick, that reduces productivity. But how much does it cost a firm when either 1) a worker is absent from work (absenteeism) or 2) the worker shows up for work but their productivity is impaired (presenteeism)?

A paper by Pauly et al. (2008) attempts to quantify these costs. Unlike most studies, they use manager estimates of lost productivity. While workers may have better information regarding how much productivity is lost during an illness, workers may have an incentive to answer strategically and further, managers will be the ones ultimately deciding how much health care preventing employee illness is truly worth.

The authors hypothesize that jobs with the following three characteristics will be more affected when a worker becomes ill:

• jobs with high values of team production,
• jobs with high requirements for timely output, and
• jobs with high difficulties of substitution for absent or impaired workers.

A firm of course, can invest in protection measures in the case of illness. For instance, it could cross-train workers, it could accumulate inventory to smooth out periods of down time, or might pay workers to work harder if an employee is missing. Nevertheless, the authors estimate the costs of absenteeism as follows:

The authors assume that the cost of being absent must be at least the employees wage if the labor market is competitive. We see that the cost of an illness is significantly higher than the wage. Also, jobs that were found to involve more teamwork, had times sensitive products, and for which workers were not substitutable had larger absentee multipliers.

For workers who are sick, but come to work, here is the cost to the firm as a percentage of the employees wage.

One problem with this analysis is that it assumes that this is a one-time illness. If these are long term illnesses, it may be more cost effective for the firm to fire the employee because their sickness is either 1) driving up insurance premiums or 2) causing them to miss too much work. Offering a generous health insurance benefit may help to prevent illness, but may also attract sicker people to the firm. Thus, despite the article’s demonstration that the cost of employees missing work is significantly higher than their wage and that there are large costs when workers come to work when they are sick, it still does not mean that employers will want to offer generous health plans.

• MV Pauly, S Nicholson, D Polsky, ML Berger, C Sharda (2008). “Valuing Reductions in On-the-job Illness: ‘Presenteeism’ from Managerial and Economic Perspectives.” Health Economics, vol. 17(4): 469-485.

Paul Levy, the president and CEO of Beth Israel Deaconess Medical Center in Boston made about $1 million dollars in 2005. Of this, $650,000 was base salary, $195,000 was made up of incentive bonus, and the balance was composed of compensation for health insurance, life insurance, and retirement.

How do I know these figures? Paul Levy told me. Well, he didn’t tell me directly, but he did post these figures on his Running a Hospital blog. Levy writes:

So, if you were on my board, how would you set an appropriate salary? You might look at the competition, and as the Globe notes, the salaries for most of the Boston-area hospital CEOs center around the same level. Would you look at salaries of people in for-profit companies, and, if so, how do you measure comparable size and complexity? Would you look at salaries of other types of non-profits, like universities and museums?

Does it matter that the average tenure of a hospital CEO is under six years? If that is roughly the tenure of a major league baseball player, should CEO salaries be in the same ballpark? Sorry, I couldn’t resist!

The New York Times recently posted an interactive page on executive pay. So the question remains, readers, how would you set Paul Levy’s salary?

Gruber (1994) shows that the costs of employer-provided health insurance benefits are passed on to employees through lower wages. But do employees with higher expected medical expenses have their wages reduced by a higher amount to reflect the additional medical costs to employers? This may not be the case if employers can not observe the health of employees when they hire them. On the other hand, for obese individuals, the fact that they are obese is easily observable. Papers such as Finkelstein, Fiebelkorn and Wang (2003) show that annual medical expenditures for obese individuals are $732 more than those of normal weight. Is this additional cost passed on to obese employees through lower wages?

This is what an NBER working paper by Bhattacharya and Bundorf (2005) aims to investigate. The authors compare the wages of obese and non-obese individuals in companies without health insurance. Then they compare the differences in wages of obese and non-obese employees in companies with health insurance. Since no health insurance costs will be passed on employees if no insurance is offered, the difference between the two wage gaps may be able to identify if higher health insurance premiums are passed on to employees through lower wages. If there are differences in wages between obese and non-obese workers (i.e.: due to discrimination, lower productivity, etc.) these differences are likely constant across firms with and without health insurance.

The authors find that “the incidence of obesity on wages for workers insured through their employers is -$1.68.” After controlling for a variety of covariates, this estimate lowers to -$1.44. This difference is mostly due to the fact that wages of obese workers with health insurance grew slower than thinner workers with health insurance. This may be due to the increasing price of medical care, the increasing severity of obesity–the BMI of individuals at the 95th body weight percentile has increased over time–or the aging of the population in the panel.

As a falsification test, the authors use a similar difference-in-difference estimation strategy comparing the obese vs. non-obese wage gap between employers with and without other fringe benefits (e.g.: life insurance, dental insurance, retirement benefits, child care, maternity leave, etc.). If obesity does not affect the cost of these fringe benefits, than we should see no difference in the obese/non-obese wage gap between employers who do and do not offer these benefits. The authors find that this difference-in-difference estimator is not statistically different from zero.

There are a few problems with this analysis. First the authors admit that “those with relatively low productivity due to health consequences of obesity may consume more medical care and, as a result, self select into firms offering health insurance.” Also, the data the authors have only reveals whether or not the individual has health insurance, and does not give the insurance premium paid by either the employer or the employee. Thus, these estimates are likely to be very imprecise.

Nevertheless, if this study’s results are true, then it would imply the following:

“If there are no externalities in these decisions, then “twinkie” taxes will only distort already optimal decisions. But if employer-provided insurance pools the health risk of the obese and non-obese, it will create an externality that reduces incentives to maintain a normal weight. Our evidence on the incidence of the obesity wage premium suggests that pooling of the obese and non-obese does not occur in the employer-sponsored insurance market; hence the externalities caused by health insurance on decisions about body weight are small.

• Jay Bhattacharya and M. Kate Bundorf (2005) “The Incidence of Healthcare Costs of Obesity,” NBER WP #11303.
• Finkelstein, Fiebelkorn and Wang (2003) “National Medical Spending Attributable To Overweight And Obesity: How Much, And Who’s Paying?,” Health Affairs, Web Exclusive, May 14, 2003.

Are smart people risk averse? Are dumb people impatient?

This is what Thomas Dohmen, Armin Falk, David Huffman, Uwe Sunde explore in their 2007 Discussion paper. Using data from a choice experiment of 1000 German adults, the authors tested for risk aversion using a Holt & Laury framework, and for impatience by varying the annual rate of return for a €100 investment. It is necessary to test the risk aversion and impatience parameter separately because in expected utility theory (EUT), a more concave utility function will cause more impatient choices, holding constant the discount rate. Cognitive ability was measured using questions similar to those on the Wechsler Adult Intelligence Scale (WAIS).

The authors found that individuals with higher cognitive abilities are less likely to be risk averse. Further, those who scored higher on the WAIS are significantly less impatient. This finding is true even after controlling for income, education, and credit constraint co-variates.

According to the authors:

“The paper also points to a different interpretation of reduced form models that have been estimated in the literature on cognitive ability and labor market outcomes. These models have typically included a measure of cognitive abilities, but not risk aversion or impatience, as explanatory variables (e.g., Cawley et al., 2001). Outcomes such as educational attainment or wages may by affected by risk aversion and impatience, and thus part of the impact of cognitive ability may reflect the correlation with these traits. In other words, our findings point to a potentially important source of omitted variable bias in this type of estimation.

Given that cognitive ability is known to be transmitted from parents to children, our findings could also be relevant for the literature on intergenerational transmission of preferences and socio-economic status.”

• Thomas Dohmen, Armin Falk, David Huffman, Uwe Sunde, 2007. “Are Risk Aversion and Impatience Related to Cognitive Ability?” IZA Discussion Paper No. 2735.

Do students who attend better schools preform better academically? This is tautologically correct, but not very informative. What would happen if we randomly moved students from low quality schools to high quality schools? Would they do better?

Using the results from Chicago Public Schools randomized lotteries of elementary studies, Julie Cullen–my dissertation adviser–attempts to answer this question in her latest NBER working paper. It is a follow-up to a similar Econometrica paper written with Brian Jacob and Steven Levitt (of Freakanomics fame) which studied a similar subject but used data on high school students. Below is the abstract.

In this paper, we examine whether expanded access to sought-after schools can improve academic achievement. The setting we study is the “open enrollment” system in the Chicago Public Schools (CPS). We use lottery data to avoid the critical issue of non-random selection of students into schools. Our analysis sample includes nearly 450 lotteries for kindergarten and first grade slots at 32 popular schools in 2000 and 2001. We track students for up to five years and examine outcomes such as standardized test scores, grade retention and special education placement. Comparing lottery winners and losers, we find that lottery winners attend higher quality schools as measured by both the average achievement level of peers in the school as well as by value-added indicators of the school’s contribution to student learning. Yet, we do not find that winning a lottery systematically confers any evident academic benefits. We explore several possible explanations for our findings, including the possibility that the typical student may be choosing schools for non-academic reasons (e.g., safety, proximity) and/or may experience benefits along dimensions we are unable to measure, but find little evidence in favor of such explanations. Moreover, we separately examine effects for a variety of demographic subgroups, and for students whose application behavior suggests a strong preference for academics, but again find no significant effects.

• Cullen JB, Jacob, BA (2007) “The Effect of School Choice on Participants: Evidence from Randomized Lotteries” NBER Working Paper #13443
• Cullen JB, Jacob BA, Levitt S (2006) “The Effect of School Choice on Participants: Evidence from Randomized Lotteries” Econometrica, 74(5) p. 1191-1230.

Most studies have found a strong correlation between a country’s unemployment benefit generosity–in terms of either duration, wage replacement rate, or eligibility–and the unemployment rate. Many economists contend that one reason why Europeans have higher unemployment rates is due to their more generous unemployment benefit package.

Yet a paper in Capitalism and Society by Howell, Baker, Glyn and Schmitt (2007) calls into question whether more generous unemployment benefits are really causing increased unemployment. Here are some of their more convincing points from the article.

• Take up. The authors claim that take-up rates of unemployment benefits is generally low. According to Hernanz et al. (2004), in the OECD countries, only 60-80 percent of eligible unemployed actually take-up their benefits. Thus not all people who are eligible for unemployment benefits actually take them up. However, this argument is not very convincing if the overwhelming majority of those eligible, take up benefits.
• Eligibility. Most countries have rules which require a certain amount of prior work experience to be eligible for unemployment. This means that the youth are less likely to be eligible for unemployment and thus will not respond to changes in unemployment benefit generosity changes. In most countries the employment rate for youths is much higher than those of working age. For instance, in the U.S. the unemployment rate among males aged 15-24 is 12.9%, but only 4.6% for males aged 25-54. Similarly in France, the UK and Spain the unemployment rate is much higher for youths (20.8%, 11.8% and 18.7% respectively) than for prime aged males (7.4%, 3.8% and 6.9% respectively). Since a large group of the employed are not eligible for unemployment benefits, unemployment benefit generosity may not have as big an impact on the unemployment rate as once thought.
• Policy lag. Changes in unemployment benefit generosity may take years to affect the unemployment rate. With an indeterminate lag between policy inaction and the policy result, spurious relationships may be found in the data.
• Policy Endogeneity. The most convincing argument is found with a Granger causality test. The authors find that high unemployment Granger causes an increase in unemployment benefits, but an increase in unemployment benefits does not Granger cause higher unemployment rates. This make sense if we are operating in the (likely) world of policy endogeneity. When politicians see a rise in unemployment, they may propose more generous benefits. Thus, economists will see a contemporaneous relationship between benefits and unemployment and may incorrectly attribute causation to the benefit rather than the unemployment rate.

While this paper makes some good points that unemployment benefits are likely not the only cause of high unemployment rates in Europe, it does not find evidence against the hypothesis that more generous unemployment benefits increase unemployment–only that this relationship has not yet been conclusively proven. In my mind, it is only sensible that being paid more not to work will increase an individual’s incentives not to work.

Policy makers must balance the work disincentive resulting from unemployment benefits with the social insurance efficiency gains from providing the insurance.

• Howell DR, Baker D, Glyn A, Schmitt J, (2007) “Are Protective Labor Market Institutions at the Root of Unemployment? A Critical Review of the Evidence” Capitalism and Society, vol. 2, issue 1.

When constructing policy, we are often faced by an equity-efficiency tradeoff. For instance, increasing welfare benefits will increase equity but may reduce a poor individual’s incentive to work if the welfare benefits are tied to having low income. A recent NBER working paper by the nobel laureate James Heckman and co-author Dimitriy Masterov (”…Investing in Young Children“) claims that programs which increase the standard of living of young children do not face this tradeoff; the programs both increase equity and are economically beneficial for society as a whole.

Heckman documents three of the major social problems in the U.S.

1. Low skills: “The U.S. has a thick lower tail of essentially illiterate and innumerate persons, who are a drag on productivity and a source of social and economic problems.�?
2. High crime rate: Compared to international standards of other developed countries, the U.S. has a high crime rate. A paper by Anderson (J Law Econ 1999) finds that crime creates huge expenses in the U.S.: $4818 per capita per year in 2004 dollars. This expense is calculated to include the cost of incarcerating criminals, the cost of police and private security, forgone wages of the incarcerated and additional injury/death risks of high crime levels.
3. Single Parenthood: The decrease of two-parent children has been widely documented. In 2003, 68% of children lived in a two-parent household while in 1980 that number was 77%. On average, single parents have lower education than married parents; children of parents with more education have more cognitive and emotional stimulation [Armour 2003].

According to the 1966 Coleman Report, the family environment is more important for school success than school spending, teacher quality, class ratios, etc. “Successful schools build on the efforts of successful families.�? One way of improving the home environment is separating the child from a ‘bad’ parent, but generally these programs have had disastrous results. Another option is to have school-based early life programs. Three programs have used this later philosophy to try to improve outcomes for underprivileged children by interventions early in these children’s lives.

• The Abecedarian Project: This randomized control trial used full-time year round classes for children from infancy through preschool. There was no affect on crime rates and a mild increase in IQ levels.
• The Perry Preschool Program: This randomized trial involved weekly home visits with parents, and intensive, high quality preschool services for 1-2 years. The program cost $19,000 per child per year but arrests by age 27 were 2.3 in the treatment group but 4.6 in the control. Only 7% of individuals in the treatment group were arrested more than 5 times compared to 35% in the control group. Test scores and literacy rates were consistently higher for treatment group.
• Chicago Parent Child Centers: This was a non-experimental, large-scale program. “…the center provided halfday preschool program for 3- and 4-year-olds during the 9 months that they were in school. The program provided an array of services, including health and social services, and free meals. It also sought to include the parents, including helping the parents complete school, home visits and field trips.”

While the Abecedarian project was not very successful, the Perry and Chicago programs had benefit-to-cost ratios of 9.11 and 7.77 respectively. The benefits include increased earnings, decreased crime, lower welfare payments. The authors, however, only use a 3% discount rate when discounting the future benefits which may be low.

The main conclusions to take from this paper are that generally intervention programs early in the child’s life are the most beneficial in terms of rate of return. Also the paper makes a point of the following:

“Schools can only work with what families give them. Successful schools are those that teach children from functioning families.�?

• Heckman; Masterov (2007) “The Productivity Argument for Investing in Young Children” NBER Working Paper #13016.

America has been characterized as “the best poor man’s country.” In the nineteenth century land was cheap and available and farming provided relatively high living standards. During the twentieth century, however, change has come. By 1920, income equality had risen to all-time highs. Over the next century, income inequality and the returns to skill (i.e.: schooling) have displayed a few broad trends. The chart below explains these trends.

A working paper (”The Race…“) by Claudia Goldin and Lawrence Katz seeks to explain these empirical findings using a structural model to estimate both relative labor supply and labor demand. One of the focuses on the paper is the wage premium individuals with higher schooling receive. In the short-run, year-to-year variation in the wage premium is almost entirely explained by the supply of college workers relative to other workers.

In the long-run, the authors find that the college wage premium and high school wage premium roses together between 1890 and 1920, but then “collapsed” between 1915 and 1950. In the period beginning in the late 1970s, the college wage premium began to rise, while the high school wage premium was anemic. There are a few explanations for this. The first is that in the early part of the century, high school graduates were considered skilled workers, workers much more qualified than high school dropouts. In the 21st century, high school graduates and dropouts have much more substitutability between them and only college graduates are now seen as skilled workers.  The authors note that the pace of the increase in the number of college graduates slowed after 1980. Since firms need more and more skilled workers to be able to take advantages of technological advances in their field, the wage premium for college graduates has increased.

Another similar explanation is skill biased technological change (see Berman, Bound and Machin QJE 1998 for more on this topic). Between 1920 and 1950, technological advancements—particularly in manufacturing—decreased the need for skilled labor since a machine took the place of craftsmen. Technological advance in the later part of the 20th century—such as personal computers—required skilled workers in order to be able to maximize the technology’s marginal product. Thus, the wage premium for individuals who could work with technologies such as computers (i.e.: college graduates) increased and wages of those without these skills (i.e.: high school graduates and dropouts) decreased.

Goldin and Katz also look at whether immigration has played a role in the changing wage premium. “[I]mmigration had but a minor impact on the growth in the relative supply of the college educated and a moderate effect on the supply of high school graduate workers relative to dropouts for the 1980 to 2005 period. Consequently immigration played only a modest role in the surge in the skill premium during those years.” Immigration restrictions, such as the National Origins Act of 1924, were also found to have little impact on inequality or skill premia. “In 2005 17 percent of the foreign born population had fewer than nine years of education whereas less than 1 percent of native-born Americans did,” but were also more likely to have a post-graduate degree.  Thus, it is possible that immigration did reduce the wages of individuals on the lowest end of the skill/earnings spectrum.

Goldin and Katz’s conclusion is worth repeating:

“Technological change is the engine of economic growth. Yet, it also has a potentially dark side…technological change creates winners and losers and can sometimes have adverse distributional consequences that may foment social tension. Such distributional problems are more likely when technological change is skill biased, that is when new technologies increase the relative demand for more skilled and more advantaged workers.

A nation’s economy will grow more as technology advances, but the earnings of some may advance considerably more than the earnings of others. If workers have flexible skills and if the educational infrastructure expands sufficiently, then the supply of skills will increase as demand increases for them. Growth and the premium to skill will be balanced and the race between technology and education will not be won by either side and prosperity will be widely shared.”

A few weeks ago I was able to attend Till von Wachter’s presentation on his working paper “Mortality, Mass-Layoffs, and Career Outcomes: An Analysis using Administrative Data,” co-written with Daniel Sullivan. The authors investigated whether or not losing one’s job affects mortality rates.

There could be many reasons why losing one’s job may affect mortality. The first of which is that losing one’s job may reduce either short- or long-term income and it has been found that those with lower incomes have higher mortality. Secondly, the emotional stress which appears among those who have recently lost a job could adverse impact health. Finally, losing one’s job can lead to a loss of health insurance which may lead to a reduction in the quantity of medical care consumed by each individual.

The paper uses earnings records from Unemployment Insurance records in Pennsylvania between 1974 and 1991 which include data on the age, sex, earnings, and industry of the individual as well as firm identifiers. The researchers link this data with mortality information from the Social Security administration between 1974 and 2002. Only those workers who were fired due to a ‘mass-layoff’ (classified as “workers who left their employer at the same time that the employer experienced a 30% or larger decline in employment”) are examined in order to focus on job loss situations which are not due to individual on-the-job performance.

The authors find that job loss leads to a 15%-20% increase in the hazard rate of dying in the twenty years following the layoff. In other words, for the average displaced middle aged worker, the loss in life expectancy is about 1.5 years. The authors claim that two-thirds of this increase in mortality is directly due to a loss in income and the other one-third of the mortality increase is due to to other factors such as stress from losing the job or other career outcome factors.

This paper does have a few problems. First, is that it only examines individuals with at least six years of tenure at their employer, when they were fired. Limiting the sample in this way reduces the generalizability of the study. Also, since the data has no information on health insurance, the researchers can not test whether the loss of health insurance had an impact on health (since COBRA was passed in 1985, many of the individuals in the study likely lost their health insurance after the layoff). Further, the analysis takes place in a partial equilibrium setting. It is likely that mass layoffs in one firm will lead to either: 1) increases in hiring in other firms in the same industry, or 2) increases in hiring at firms in other industries. Thus, in a more general equilibrium analysis, job losses hurt some workers, but benefit others . Those workers who are hired at the other firms may experience a decrease in mortality. Long run evidence tends to show that Schumpeter’s concept of “creative destruction” has decreased average mortality for individuals all over the world due to rising living standards.

• von Wachter, Till; Sullivan, Daniel; “Mortality, Mass-Layoffs, and Career Outcomes: An Analysis using Administrative Data“

If taxation reduces the amount of hours worked, why does Scandinavia—which has some of the world’s highest tax rates—have labor force participation (LFP) rates similar to the U.S? This is the question addressed by Richard Rogerson in the NBER working paper “Taxation and market work: Is Scandinavia an outlier?�?

The study looks at three groups of countries: the U.S., the EU (Belgium, France, Germany and Italy), and Scandinavia (Denmark, Finland, Norway, and Sweden). Europeans and Scandinavians work approximately 1500 hours per employed person per year compared to the American figure of about 1850. In the year 2000, the employment to population ratio in Scandinavia is approximately 0.70, which is the same as in the U.S. Europe has an employment to population ratio of only 0.60 in 2000. We can see Scandinavia’s similarity to the U.S. workforce occurs mostly on the extensive margin. Yet, Scandinavia has the highest tax rates while the United States has the lowest. How can this be explained if taxes provide a disincentive to work?

Rogerson uses OECD and GGDC (Gronigen Growth and Development Center) data and claims that where the government revenue is spent can explain the differential. Government spending is divided into four categories as shown below.

Rogerson contends that the EU has more government spending in the ’subsidy to leisure’ category while Scandinavia spends more on ’subsidy to work’ programs. The paper states that Scandinavia spends 8% of government spending on child and elderly care compared to only 2% in Europe. Further, government employment is highest in Scandinavia (i.e.: 15% in the U.S., 18% in Europe, and 28% in Scandinavia). Of course, government spending on employing individuals will increase the labor participation rates. Increased taxes decrease the incentive to work, but increased spending on programs which decrease the cost of going to work will increase labor force participation.

Rogerson then creates a model where households gain utility from consumption, leisure, and family services (e.g.: child care). In the model, taxes due decrease work incentives, but programs which decrease the price of child care increase the labor supply curve. The paper then argues that the differential in service employment explains the overall employment differential because the service sector is a proxy for the amount of child care. More and less expensive child care services make it easier for parents to enter the labor force.

This argument is very interesting and believable, but the evidence given is weak. The fact that Scandinavia has a higher service sector LFP rate does not necessarily mean that more child care is being given since the service sector includes everyone from economists, lawyers, technology consultants, fast food workers as well as child care specialists. More convincing evidence would include empirical data which showed that after a government increased child and elderly care spending, labor force participation also rose. Still, it seems obvious to Public Economists that Ricardian Equivalence will not hold in reality; where the government spends its money has significant effects on all markets.

Daniel Polsky and Sean Nicholson have two papers which aim to look at employer health insurance offerings.  The first [Polsky, Nicholson (2004)] tries to estimate the factors driving the cost differences between HMO plans and non-HMO (eg: PPO, indemnity) plans.  The authors deconstruct the cost differences into three factors:

1. Utilization Effect: this occurs if individuals enrolled in an HMO plan use fewer services than those in a non-HMO plan.
2. Risk Selection Effect: this phenomenon appears if HMO enrollees are generally healthier than non-HMO enrollees.
3. Reimbursement Effect: this is apparent if HMOs negotiate lower payments to providers–for instance by offering higher volume–than do non-HMO plans.

The authors have an ingenious way of estimating this.  They use the 1996/1997 Community Tracking Study (CTS) to estimate the following equation:

• M=b_1*H + b_2*X + v

The medical resource use (’M‘) is a function of whether or not the individual is in an HMO (’H‘) as well as observable characteristics (’X‘).  The error term ‘v’ is composed of the sum of a random element (’e‘) as well as the unobservable characteristics (b_3*Z)  It is possible to estimate b_1 and b_2 consistently only if Cov(H,Z)=0.  The authors accomplish this by estimating the above equation for individuals whose employer offers them no choice of health plan–thus risk selection should not be a problem.  In order to determine the reimbursement effect, the authors estimate prices for each service using the MEPS.  Below are the formulas for each effect.  The coefficients are all estimated from the subsample with no choice (C=0) for the employee.

1. Utilization Effect: b_1
2. Risk Selection Effect:
1. Observable: b_2*X{H=1,C=0} -b_2*X{H=0,C=0}
2. Unobservable: b_2*X{H=1,C=1} -b_2*X{H=0,C=1}
3. Reimbursement Effect: [p{H=1}S{H=1}-p{H=0}S{H=1}] - [p{H=0}S{H=1}-p{H=0}S{H=0}].  This is done for each service type and summed over all services. 

The results of the study are that HMO expenditures per person were $188 lower (9.3% lower) than non-HMO expenditures per person.  Utilization Rates and Risk Selection were similar for HMO and non-HMO plans.  The difference was made up from different reimbursement rates for providers.

This paper is very clever econometrically and the CTS gives a wide variety of useful variables for the study.  The key assumption to this paper is that employees do not choose their work based on the health insurance that it offers (which Bhattacharya and Vogt would disagree with).  The authors claim if unhealthy employees choose jobs with non-HMO health insurance, then the utilization effect would be biased downward and the unobserved risk selection estimate would be biased upward.  Thus, one cannot even sign the direction of this problem. 

A second paper by Polsky and Nicholson, et al. (2005) uses the CTS data set as well.  It examines the factors which determine whether a worker takes up the health insurance offered by their employer.  A worker can choose to take-up the insurance of their employer, choose an ‘alternative insurance’ (such as government provided insurance or go on a spouse’s health insurance plan), or elect to remain uninsured.  A multinomial logit regression finds that single individuals generally take up an HMO plan if it is the only one offered, but married individuals often refuse to take up the HMO plan if it is the only one offered.  Married individuals have a choice to be on their spouses insurance while single individuals are left with a choice of either state health insurance or expensive individual insurance. 

The authors also use the CTS employer survey to estimate the cost of the plans offered to employees.  They estimate the cost to the employer (premium) and the cost of the plan to the employee (net premium).  They find that a higher net premium reduces employee take-up but a higher premium does not affect take-up rates.  Unsurprisingly, younger, less educated and poorer families are more likely to forgo health insurance. 

Some problems with this paper are the authors assume that the employer does not offer health insurance strategically.  This is unlikely.  The authors showed that when a firm offers only a HMO plan, a married worker is likely to take-up insurance with their spouse.  Thus, offering parsimonious coverage will reduce a firms costs by causing many of its employees to move off their roles.  The authors also assume that employees do not select jobs according to the generosity of their health insurance, and that the net premium workers pay is also exogenous.  Both assumptions are unlikely to be proven true in reality. 

Polsky, Daniel; Nicholson, Sean; (2004) “Why are managed care plans less expensive: risk selection, utilization or reimbursement?” The Journal of Risk and Insurance, pp. 21-40.

Polsky; Stein; Nicholson; Bundorf ;(2005) “Employer health insurance offerings and employee enrollment decisions” Health Services Research, pp. 1260-1278

According to the Kaiser Family Foundation, 160 million Americans receive their health insurance from their employers.  That figure represents three out of five non-elderly individuals.  Many experts argue that using employer provided health insurance eliminates the problem of adverse selection by forming an insurance pool around a non-medical issue (employment).  Jayanta Bahattacharya and William Vogt are not sure this is the case.  In their 2006 NBER working paper, the authors aim to test whether or not healthy individuals are more likely to receive employer provided insurance at their jobs. 

Model

The model Bhattacharya and Vogt set up is a two period model.  In the first period i) employer set wage and benefit levels, ii) workers are informed if they are health or sick, iii) workers choose their employer and then iv) a health shock occurs and the employee consumes the needed medical care.   In the second period: i) the workers see a new health state, ii) there is an involuntary turnover rate ‘T‘ where these workers must seek new employment, iii) workers in the ‘(1-T)‘ group can elect to switch employers, and iv) a second health shock occurs and the employee consumes the needed medical care.

The individual’s utility function is:

• u(Y-m)+v[H+f(m,e)]

‘Y‘ is one’s income and ‘m‘ is out of pocket medical expenses.  The separable v function is utility gained from health capital.  ‘H‘ is the initial health capital level and f is an increased or decreased level of health depending on health spending ‘m‘ and a random shock ‘e‘.  The shock variable is distributed according to the cdf ‘F‘ which depends on whether the individual is sick or healthy. F_sick first order stochastically dominates F_well.  Because of this assumption, we can prove (mathematically) the following three statements:

• The cost of care for the sick is higher than the cost of care for the well
• The Utility of the insured who are sick is lower than the utility of the insured who are well.
• The Utility of the uninsured who are sick is lower than the utility of the uninsured who are well.

There are four possible indirect utility functions in the first period. 

• U_SU: E_{F_s} [U(Y-m,H+f(m,e))]
• U_SI: E_{F_s} [U(Y,H+f(m,e))]
• U_WU: E_{F_w} [U(Y-m,H+f(m,e))]
• U_WI: E_{F_w} [U(Y,H+f(m,e))]

In the second period, however, the author assumes that there is a switching cost of ‘c‘ utils if the employee decides to change jobs.  This can be justified as a psychic cost to the worker or the loss of job-specific human capital.  The author assumes U_WU(W-p)-c<U_WU(W) meaning that one will reduce overall utility if the person leaves a job and purchases private health insurance rather than staying at a job and using the employer provided insurance. 

Equilibrium

The authors seek a symmetric subgame perfect Nash equilibrium where 1) both sick and well workers choose an employer offerreing insurance in period one and 2) neither sick nor well workers voluntarily turn over to change insurance status in period two.  Working out the mathematics we find the following conclusions:

1. Among people who do not receive employer provided insurance, the sick benefit the most from purchasing individual insurance: U_SI(W-p)-U_SU(W) > U_WI(W-p)-U_WU(W).  Thus the only people left uninsured will be the healthy ones.
2. A pooling equilibrium is more likely if there are high switching costs ‘c‘. 
3. A pooling equilibrium is more likely with a low exogenous turnover rate ‘T‘. 
4. A pooling equilibrium is more likely if the probability that a well person will become sick is high and the probability a sick person will become well is also high.  Algebraically, the author say this means that ‘P_ww‘ is low.

Empirical Tests

The authors use data from the 1995-2005 March CPS as well as the Occupational Information Network and the Census Public Use Microdata Sample.  They aim to test conclusions 2-4 above as well as seeing if there is any evidence of adverse selection.  The data also show that industries with a high level of job specific human capital–which in this case the authors use as a proxy for ‘c‘ –have workers which are more likely to be covered by their employers.  The coefficient on ‘P_ww‘ is not significantly different from zero, but the ‘P_ww‘ variable is likely not measure with accuracy.  A high turnover rate ‘T‘ actually has the opposite effect than the one predicted by the authors.  The measure ‘T‘ however is problematic to measure in the data since turnover empirically is a mix of exogenous and self selected turnover (quits) and thus the authors disregard any results for the variable. 

Analysis

The model of this paper is elegant and produces sensible conclusions.  While the homogeneity of employers greatly simplifies the model, it makes the mathematical conclusions less robust.  The empirical work gives some suggestive evidence that adverse selection in employer choice of workers may be a problem.  Using such a large data set adds precision to the author’s estimates, but it may be difficult to decompose the countervailing forces within each industry without having more detailed knowledge of each sector.  Still, the combination of theory and empirical work is good; using other data sets to substantiate the authors claims would make their conclusions more robust.

Bhattacharya, Vogt (2006); “Employment and adverse selection in health insurance,” NBER Working Paper No. 12430.

Recently, we have seen a shift towards the left in Latin America. Socialistic presidents Evo Morales of Bolivia (who just nationalized Bolivia’s national gas industry) and Hugo Chavez of Venezuela both enjoy much popular support. Brazilian president Lula Da Silva has a background as a union leader and Nestor Kirchner of Argentina also rose to power by bashing neo-liberal economics. The World Bank’s Private Sector Development blog gives links to a variety of articles arguing why this is the case.

The best article of the bunch is from a blog by famed economists Gary Becker and Richard Posner where the authors claim that crony capitalism is the reason why Latin American voters have been trending to the left. Becker states:

One legitimate reason for the opposition to capitalism in Latin America is that it frequently has been “crony capitalism” as opposed to the competitive capitalism that produces desirable social outcomes. Crony capitalism is a system where companies with close connections to the government gain economic power not by competing better, but by using the government to get favored and protected positions. These favors include monopolies over telecommunications, exclusive licenses to import different goods, and other sizeable economic advantages. Some cronyism is found in all countries, but Mexico and other Latin countries have often taken the influence of political connections to extremes.

In essence, crony capitalism often creates private monopolies that hurt consumers compared to their welfare under competition. The excesses of cronyism have provided ammunition to parties of the left that are openly hostile to capitalism and neo-liberal policies. Yet when these parties come to power they usually do not reduce the importance of political influence but shift power to groups that support them.

Another possible explanation for Latin America’s rejection of capitalism is the increased inequality in their societies.  The Economist (”Inequality in Latin America“) reports on this phenomenon:

Latin America is much more unequal than other regions. The reason is history. European colonisation set a pattern of exploitation of indigenous Indians. Its legacy, and that of slavery, live on. This helps to explain the ethnic character of some of the region’s inequality. The Bank finds that in Guatemala one in five “white�? men have a car, compared with only one in 20 men of indigenous blood. In Bolivia, 84% of “white�? women have access to electricity, compared with 64% of Indian women.

Skill biased technological change (SBTC) may be the cause of this increased inequality.   In Berman, Bound and Machin (QJE 1998), the authors argue that the increased inequality in the wages of workers in developed countries is due to the fact that returns to high skill (education) jobs has been increasing, while the return to low skill jobs has been flat or decreasing over the past decades.  While the authors do not specifically examine this phenomenon in the case of Latin American (OECD economies are used), it is very likely that SBTC could be taking place in Latin America as well.  If this is the case, the median voter would be more likely to have experience flat or reduced wages over time and thus a vote for socialism would be rational.

Economists typically assume that the majority of additional costs employers incur from hiring a worker are reflected in a lower compensation package for the employee.  For instance, employees owe payroll taxes for Social Security and Medicare on 7.65% (Social Security taxes are limited to earnings below $94,000 in 2006, but Medicare taxes are applied on all earnings).  Employers also pay a tax of 7.65% of workers wages to the government.  Economists believe that the true tax to workers is 15.3% since firms pass on the added costs to employees.

Increasing health insurance costs is another arena where employers may pass the costs on to employees.  A study by Goldman, Sood and Leibowitz (2005) analyzes how wages and benefits change in response to rising health insurance costs at one specific benefits consulting firm with a ‘cafeteria plan’.  The firm pays workers a wage and gives them a credit towards the purchase of a variety of benefits.  The credit can be used towards acquiring health insurance, as well as other benefits such a pension, accident insurance, life insurance, long term disability, etc.  The credit the firm grants to workers is based on their salary and their tenure at the firm.  Workers who wish to receive more benefits above their credit allocation can pay for the difference out of their salary.

Using a fixed-effects model with data between 1989-91, the authors estimate how changes in the price of health insurance affected the workers choice of benefits.  They find that when the price of health insurance increases by $1, workers finance the increase with a 52 cent reduction health insurance expenditures, a 37 cent reduction in take home wages, and a 17 cent reduction in other benefits.  This result shows the demand for health insurance is inelastic since a price increase leads to increased expenditures–48 cents in this example–on the good.

A problem with the study is the narrowness of its scope.  Since it only deals with one white collar firm, this result may not be generalizable to other sectors or the economy as a whole.  Also, the authors claim that as employees reduce the amount of other benefits as health care prices rise, an employee may be more vulnerable to health, mortality, disability and other risks.  We do not know, however, whether or not the employees decided to increase their savings rate in the face of decreased benefits in order to ’self insure’ against future risks.  Thus the magnitude and sign of the change in an employee’s vulnerability to future risk is indeterminate.

Source: Dana P. Goldman, Neeraj Sood, and Arleen Leibowitz (2005) “Wage and Benefit Changes in Response to Rising Health Insurance”, Forum for Health Economics & Policy, Forum: Frontiers in Health Policy Research, Volume 8: Article 3. http://www.bepress.com/fhep/8/3

I have recently completed a paper titled Job Lock: A Literature Review, which has been posted in the ‘Papers by HC Economist‘ page. Here is a brief excerpt from the beginning of the paper.

“Three in 10 Americans say they or someone in their household have at some time stayed in a job they wanted to leave mainly to keep the health benefits, according to a New York Times/CBS News Poll. The survey provides some of the strongest evidence yet of pervasive concern about the costs of medical insurance and care.”

“We will strengthen health savings accounts — making sure individuals and small business employees can buy insurance with the same advantages that people working for big businesses now get. We will do more to make this coverage portable, so workers can switch jobs without having to worry about losing their health insurance.”
- President George W. Bush,
- State of the Union Address, 2006

INTRODUCTION

Health insurance is a major issue in the United States. Nearly
everyday, residents can pick up a newspaper and read a story
expounding on the worsening `health care crisis.’ Many workers
fear losing their job, not simply due to the loss in wages, but
even more due to the loss in health insurance coverage. In fact
many workers decide to stay at a job whose compensation is less
than their marginal product because they fear the loss of their
health insurance. This phenomenon is known as `job lock.’ The
term has received much publicity both in the economics literature
and in the popular press. The economics literature seeks to
answer the following three questions:

1. Does job lock exists?
2. If job lock does exist, what is its impact on social welfare?
3. Does current regulation aimed at combatting job lock improve social welfare?

In the following literature review, I will show how the field of
Economics has attempted to answer these questions.