Inequality Midterm Discussion Questions

"Unions, Norms.." Western (2011): The authors suggest three ways that unions can affect non-union wages. What are they and in which direction do the go (i.e., do they widen inequality between union and non-union workers or reduce it)?

1. Threat effect: nonunion employers raise wages to the union level to avert the threat of unionization - less inequality 2. Spillover effect: when unions raise wages for their members, employers may cut workers, forcing them to find jobs in the nonunion sector, causing wages to fall there - more inequality 3. Moral economy: unions support norms of fairness and equity that extend beyond their membership: culturally, politically, and institutionally - less inequality

"Role of Demand Shift" Johnson 1998: What fact or change about the economy motivated the article?

A recent rise in earnings inequality motivated the article.

"A Grand Gender Convergence" Goldin 2014: What group of employers (e.g., technology, business, science, health, other) have the largest female wage penalty? Which have the smallest? What is Goldin's ultimate explanation for this and how does she show this is a reasonable explanation?

Business occupations have the highest female wage penalty while tech and science occupations have the lowest wage penalty. She says that technology may have the smallest because of selection, there are not many women in technology careers so this coefficient could be biased. People in technology roles in the technology industry specifically tend to earn considerably more than their counterparts in non-tech related roles.

"Statistical Discrimination or Prejudice: A Large Field Study" Ewens (2014) What kind of discrimination do the authors conclude is occurring in this market? What testable implication of that model of discrimination do the authors observe in the data that allow them to reach this conclusion?

Concluded that statistical discrimination is present because a negative signal was associated with an increasing racial gap (blacks with a negative signal less likely to get a repsonse than whites) A positive signal didn't have a significant effect on the response

"Racial Discrimination in Job Offers" Fryer 2001: What is the authors' main finding? How much of the raw black-white wage gap is due to discrimination according to the authors?

Differential treatment accounts for at least one-third of the black-white wage gap. Employers statistically discriminate upon the hiring of workers but adjust their wages upon learning more about their marginal product over time. Blacks' wages grow with tenure at a faster rate than whites.

"A Grand Gender Convergence" Goldin 2014: Fundamentally, this paper is about compensating differentials. What is the job amenity employees face a wage tradeoff for?

Employees want to trade flexibility for a lower wage.

"China Shock: Learning from Labor Market Adjustment to Changes in Trade" AUtor and Dorn (2016): What are the general effects the authors find (you do not need to remember magnitudes, just the basic directions)?

Employment has certainly fallen in US industries more exposed to import competition... BUT Overall employment has also fallen in the local labor markets within the same industries Input-output linkages between sectors appear to have magnified rather than dampened the employment effects of trade both within regions and nationally

"A Grand Gender Convergence" Goldin 2014: What empirical method does Goldin use to make the illustrative figures 2A to 2C? Why might such an approach be necessary?

Goldin uses comparisons in these three figures to show her point. Her regression includes log earnings equations using age, education dummies, race, log hours, and log weeks. Part A contains all full-time, full-year workers. Part B contains those who graduates college. Part C contains under 45 years old of those in part A. This is necessary because in order to see the larger picture of earnings, you must condition upon all of these variables while also taking different demographics.

"Marry Your Like" Greenwood (2014): The authors point out that even though there was assortative mating in 1960, random marriage in 1960 would not have decreased the GINI coefficient as it does today. How do they explain this odd fact?

In order for positive assortative mating to have an impact on income inequality, married women must work. Female labor force participation was much higher in 2005 than in 1960, therefore, their contribution to household labor income is significantly larger and this share rise with income percentile as well. The authors show that when matching is random, female labor force participation has a significant dampening effect on income inequality in 2005. Random matching would cause income inequality to drop because income is more diversified across husband and wife.

"Role of Demand Shift" Johnson 1998: What two explanations did the author put forward to help explain the changes discussed in the article? How would those changes affect relative demand?

Increased openness: an increase in globalization/trade increases the outsourcing of unskilled labor, unskilled labor in the US is not as demanded as we focus on skilled labor-intensive goods. Skill-biased technological change: technology increases productivity of skilled workers, skilled workers replace unskilled workers leading to increased demand of skilled laborers and decreased demand for unskilled workers

"Role of Demand Shift" Johnson 1998: What is the difference between extensive skill-biased technological change and intensive skill-biased technological change?

Intensive: skilled workers become more productive at the jobs they already perform Extensive: skilled workers become more efficient in jobs that were formerly done by unskilled workers

"Wage Differentials and Public Policy" Smith (1979): What are two types of data one could use to identify compensating differentials? What are some of the weaknesses of each?

Large, ongoing sample including individual characteristics and earnings and merge it with data on job characteristics and match them to individuals through industry codes. Weaknesses of this include matching average job characteristics to individuals who may not possess jobs with average characteristics. Self-reported data containing job and individual characteristics. Weaknesses of this include the subjectivity of self-reported measures.

"Marry Your Like" Greenwood (2014): Describe the regression methodology the authors' use to identify any increase in assortative mating?

Regress education of wife on education of husband and time variables (see doc for more info)

"Wage Differentials and Public Policy" Smith (1979): What disagreeable characteristic does Smith report had been most researched back in 1979? What are some of the others considered? What are the main results?

Risk of injury or death is the most researched. Results are ambiguous because it is hard to differentiate between safe and dangerous jobs on a 0, 1 scale. In another study, risk of death has a positive coefficient meaning they have higher wage (by 3-4% on average). Results with respect to risk of injury are less clear- cut. Job insecurity and hard/stressful work are other two characteristics that come the closest to being rated unpleasant. Tests overall are inconclusive except for "risk of death".

"The Growth of Low-Skill Service Jobs and Polarization of the U.S. Labor Market" Autor and Dorn (2013): What kinds of jobs do the authors hypothesize are replaced by technology? What kind of jobs cannot be replaced easily both on the high-skill and low-skill end?

Routine jobs suffer with the advances in technology, high-skill and low-skill jobs alike. See question 3 for low-skill jobs that can't be replaced. On the high-skill end, roles such as doctors, lawyers and business managers cannot be easily replaced because technology has only enhanced their abilities. We have not yet developed robotics or artificial intelligence that can surpass or even match the general, broad intelligence of a human.

"The Growth of Low-Skill Service Jobs and Polarization of the U.S. Labor Market" Autor and Dorn (2013): What jobs that are typically thought of as "low-skill" jobs have actually seen some increases in wages?

Service-related jobs are typically low-skill (do not require higher education) jobs that involve helping others and have not suffered from an increase in technology. Examples of these jobs are food service, child care, security guards, gardeners and hairdressers.

"Role of Demand Shift" Johnson 1998: Although the author explains that changes in labor demand may explain increases in the wage ratio between college educated and high school educated individuals, what trend mitigated this increase?

Some positions that were previously staffed by unskilled workers are now staffed by skill workers so the marginal skilled worker will receive less in wages and the marginal unskilled worker will receive more so the wage ratio will fall.

"Role of Demand Shift" Johnson 1998: What general economic theory you learned about as early as Principles of Microeconomics was used to broadly explain why the changes may have occurred question 1?

Supply and demand: there was a sharp increase in the labor demand of skilled workers. Supply of skilled workers also increased.

"Power of the Pill" Goldin and Katz (2002): Name at least two effects identified by Goldin and Katz that result from the advent of the invention from question 1?

The accessibility of the pill allows young people to put off marriage without putting off sex, which could lead to higher-quality matches Makes career investments cheaper and women with greater career ability become more attractive marriage partners

"Unions, Norms.." Western (2011): What empirical approach do the authors use? How do they identify the direct effect of union membership on members versus the indirect effect on non-union members?

The authors argue that declining union membership plays a role in increasing income inequality. To determine this, they utilize the variance of wages as a measure of inequality. Their key predictors are an indicator for union membership and a continuous variable that records for each respondent the unionization rate for the industry and region in which they work. They measure between group wage inequality using variance of conditional means and within group inequality by using the residual variance. They hold unionization constant at the 1973 level to calculate adjusted variances, then add threat and norms by holding unionization rate fixed at the 1973 level, and finally fixing education coefficients to 1973 values. These adjusted values can be seen as measures of inequality.

"Single Moms and Deadbeat Dads" Skira (2016): What mechanism is the most important in terms of explaining the differences in outcomes between blacks and whites?

The black and white earnings gap and the heterogeneity of preferences explain a substantial portion of racial differences. DIfferences in preferences between whites and blacks was the most important.

"A Grand Gender Convergence" Goldin 2014: What is Goldin's argument for why some employers are willing to pay more for workers willing to sacrifice this amenity?

The idea is that when someone works more hours in a job, the employer will offer a higher pay and greater reward for doing so. When there is less flexibility, that means that the worker is sacrificing more of their life and time potentially for the job and the potential inconveniences. It makes sense that an employer will pay lower amounts of money to women who demand a more flexible schedule or fewer hours.

"Power of the Pill" Goldin and Katz (2002): What is the "direct cost" reduction that Katz and Goldin point to for women seeking professional education that results from the invention identified in question 1? What is the "indirect cost" reduction?

The pill lowered the cost of engaging in long-term career investments because women could feel more secure in their ability to get pregnant when they were ready for a child The indirect cost reduction was the reduced marriage market cost because individuals could delay marriage and not pay a large penalty for it since they wouldn't lose out on the best available bachelors to women who were not as focused on their career.

"Why has CEO pay increased so much?" Gabaix Landier (2008): The authors list several other explanations for the rise in CEO pay from the literature. What are they?

The relationship between the size of a CEO's firm and the average firm size in that country gives insight into why there has been a rise in CEO pay. Another reason is sensitivity to talent across firms. Newer tech firms that come to the market are in high demand for an active, excellent CEO which will drive their compensation up due to the higher responsibility. Another possibility is misconception of the cost of compensation. Board members in the past have mistaken the value of giving CEOs stock options and has therefore allowed their wealth to increase more than anticipated. They become fooled by the idea that giving out stock is not like giving out cash.

"Power of the Pill" Goldin and Katz (2002): What medical invention is the subject of this paper? What is the variation in this invention's availability that Katz and Goldin utilize in the econometric portion of this paper?

The subject of this paper is how the changes in the availability of the pill altered women's career and marriage decisions. In 1970, the pill became easier for young, single women to gain access to because "mature minors" were able to gain access without needing parental permission, and the "age of majority" (being a legal adult) was lowered to 18

"Single Moms and Deadbeat Dads" Skira (2016): What are the three major mechanisms put forward?

The three mechanisms are: -Lower educated males and black males tend to make less money and be less likely to be employed than well-educated/white males, limiting the value of marriage -There tend to be fewer black men than black women, limiting the potential for matches to be made. -Blacks and whites may have different preferences over marriage.

"Single Moms and Deadbeat Dads" Skira (2016): What are the three primary questions the paper tries to answer?

The three questions are: Why do some men father children without providing support? Why do women have children outside of marriage when they receive little support from fathers? Why is this behavior more common among blacks than whites?

"Racial Discrimination in Job Offers" Fryer 2001:What variable to the authors use try to control for the market valuation of an individual's typically unobservable "skill bundles, non-cognitive skills, and similar variables?" Under what assumptions does this variable work?

The variable they used was a person's most recent wage right before a spell of unemployment. This variable works under two assumptions: -Blacks and whites draw job offers from comparable sets of firms in similar markets -Previous wage does not systematically overstate blacks' productivity relative to whites'

"Racial Discrimination in Job Offers" Fryer 2001: In general, why is discrimination hard to identify in a regression approach?

There are a lot of variables that go into determining someone's wage and they are difficult to control for using a regression approach ie ability, education, experience, geographic location, etc.

"Statistical Discrimination or Prejudice: A Large Field Study" Ewens (2014) What is the basic setup of the experiment?

They are emailing relators on Craigslist for 1 bedroom or studio apartments where they either have a positive signal, negative signal or no signal, and use names that are either clearly white or black.

"Statistical Discrimination or Prejudice: A Large Field Study" Ewens (2014) What two models of discrimination to the authors seek to test? What market (e.g., job market, marriage market, housing market) are they looking at?

They are testing for taste-based discrimination produced by racial prejudice and statistical discrimination (when imperfect information causes agents to form expectations based on limited signals that correlate with race) They are testing for these forms of discrimination in the apartment rental market

"Why has CEO pay increased so much?" Gabaix Landier (2008): In this paper, the authors start with Superstar Theory and argue one major characteristic of firms has changed in the U.S. over the last several decades and led to the rapid increase in CEO pay. What is that characteristic and why could it lead to large increases in pay according to Superstar Theory?

Those who are skilled in a way that influences many people in a way where quality is not completely sacrificed are part of the top percentile. Firms have gotten larger over the past several decades in the U.S. which allows CEOs and similar management to have a more far reaching role in the company. This can be attributable to technological advancements in communication and connectivity. The CEO has the opportunity to more efficiently manage more people which is always a positive according to superstar theory.

"Unions, Norms.." Western (2011): What is the primary finding of the article? Do the authors' offer any caveats?

Union decline is associated with about ⅓ of the rise in wage inequality for men and ⅕ for women. It is possible that deunionization is not directly implicated in the process of growing inequality, it is just a byproduct of new labor market norms that have affected inequality as well, like computerization, deregulation, and globalization, etc.

"The Growth of Low-Skill Service Jobs and Polarization of the U.S. Labor Market" Autor and Dorn (2013): What patterns in wages and employment at the bottom, middle, and top of the income distribution that motivates their hypothesis?

We are experiencing job polarization, where the middle of the income distribution has been carved out because... The wage gap between college educated and high school educated individuals has increased Labor force participation dropped more for high school graduates than for college graduates The largest gains in employment have gone to those in high-skill occupations, yet earnings growth has been greater in low-skill jobs and less in high-skill jobs

"Why has CEO pay increased so much?" Gabaix Landier (2008): Does the authors' theory help explain the fact (noted in class) that other countries have not seen such a rapid rise in CEO pay?

Yes, it is possible that firm size has only increased by a significant margin in the U.S.

"Marry Your Like" Greenwood (2014): Do the authors find any increase? How much lower would the GINI coefficient be today if people got married randomly?

Yes, the authors find that there is a stronger correlation between husband and wife education in 2005 than in 1960. The GINI coefficient would have been .34 instead of .43 (in 2005) if people married randomly. This is a 20% increase.