Wage inequality in Brazil is exacerbated by high firing costs—a form of worker protection that may actually worsen the lot of low-income employees
Published February 6, 2018
The ability of job markets to respond flexibly to variations in labor demand is central to economic vitality. A company experiencing higher demand for its products can fulfill it quickly only if workers are available and the costs of hiring and training are relatively low. By the same token, a company suffering decreased demand can survive only by slowing output and trimming production costs. The latter may entail cutting labor—a painful but essential part of a functioning economy.
But several realities reduce job market flexibility. The difficulty of matching workers to positions—so-called job search—is one. Another is the cost of firing a worker. Severance pay, for instance, can make an employer less likely to lay off employees because the cost of doing so is higher than it would be without such a job benefit. Golden parachutes for workers can be iron handcuffs for bosses.
Such costs can induce firms to hire fewer workers, and research on firing costs suggests that severance pay requirements decrease employment.
In a recent Opportunity & Inclusive Growth Institute working paper (IWP5), Brazilian economist Ana Luisa Pessoa de Araujo, a visiting scholar at the Minneapolis Fed’s Institute, looks at firing costs from a new direction. She examines the impact of firing costs not on employment levels, but on wage inequality. Her conclusion is striking: Welfare for less-educated, 25- to 55-year-old Brazilian males working in the private sector would increase significantly if firing costs were abolished. By eliminating such costs, wage inequality among these workers would drop 26 percent, as measured by wage variance.
High turnover? Low wages
Why do firing costs have such a large impact? “High-turnover jobs are also lower-wage jobs,” writes Pessoa de Araujo. Think fast-food employees, customer service reps and retail store clerks. Firing costs depress wages for high-turnover trades because employers anticipate paying such costs frequently. High-wage jobs, by contrast, have low firing rates, so firing costs cause very little wage reduction for well-compensated employees. “Therefore, the gap between low-wage and high-wage workers increases with the firing cost.”
To arrive at her findings, Pessoa de Araujo focuses on the nation she knows best. “Brazil is a country that boasts high levels of wage inequality as well as high levels of firing costs,” she writes. Average firing costs in Brazil are equal to roughly 30 percent of one year’s wages.
But before focusing on these empirical realities, she creates a model of job search and matching that includes a vital feature: heterogeneous layoff rates.
Some firms and industries fire workers more often than others, but few researchers have examined job stability—let alone firing costs—as a potential source of wage dispersion. By including layoff rates in her model, Pessoa de Araujo is able to pinpoint the role that job stability—or, by proxy, firing costs—plays in determining wage inequality.
The model includes workers, who may be employed or unemployed, and companies that hire and fire. Workers can search for jobs; those who are unemployed do so more efficiently than employees. Firms recruit based on production demands and hiring/firing costs. The model includes other technical elements and refinements such as “continuation rates” and “destruction values.” In short, it’s a rich and fairly complex job search model.
To catalyze her model, she hypothesizes that the government mandates a firing cost in the form of a severance payment to employees and a government tax to fund unemployment insurance. “If the firm fires the worker (which occurs with the probability δ), then the firm must pay a lump-sum transfer to the worker … and a tax to the government.”
A model fit
The model works well. Once calibrated to key characteristics of Brazil’s economy, it generates a very close fit to actual Brazilian data on wages, employment and job stability. In one key chart, Pessoa de Araujo shows that her mathematical model replicates the negative empirical relationship between firing rates and wages. (See the accompanying figure.) The model also matches data on employment distribution at low-firing and high-firing rate companies.
Having confirmed that her model works well, she conducts her central experiment. What would happen to wages, employment and worker well-being if the government eliminated the hypothetical firing tax?
For the population analyzed—25- to 55-year-old males with no more than a high school education working in the private sector—the answer is clear. If firing costs were repealed, wages would increase disproportionately more for workers in low-wage, unstable jobs than for higher-wage, more securely employed workers. “Therefore, by taking into account layoff heterogeneity,” Pessoa de Araujo concludes, “my model shows how firing costs amplify wage inequality.”
The result is true for a range of measures of wage inequality, including variance in wages and wage ratios of high- and low-wage percentiles. Eliminating firing costs also increases employment a bit and decreases market tightness. And she finds that these results are true across the range of both tax reduction percentages and firing rates. For instance, if the firing cost is cut in half, companies that pay the lowest wages increase their pay by nearly 11 percent, while high-wage companies increase theirs by less than 2 percent. Worker welfare also increases, more for low-wage workers than for those getting higher pay.
In sum, the price of severance payments may be lower wages and greater job insecurity. Eliminating such government mandates could actually improve the lot of the less fortunate.