6.2 Job Search and Frictional Unemployment


One reason for unemployment is that it takes time to match workers and jobs. The equilibrium model of the aggregate labour market discussed in Chapter 3 assumes that all workers and all jobs are identical, and therefore that all workers are equally well suited for all jobs. If this were really true and the labour market were in equilibrium, then a job loss would not cause unemployment: a laid-off worker would immediately find a new job at the market wage.

In fact, workers have different preferences and abilities, and jobs have different attributes. Furthermore, the flow of information about job candidates and job vacancies is imperfect, and the geographic mobility of workers is not instantaneous. For all these reasons, searching for an appropriate job takes time and effort, and this tends to reduce the rate of job finding. Indeed, because different jobs require different skills and pay different wages, unemployed workers may not accept the first job offer they receive. The unemployment caused by the time it takes workers to search for a job is called frictional unemployment.

Some frictional unemployment is inevitable in a changing economy. For many reasons, the types of goods that firms and households demand vary over time. As the demand for goods shifts, so does the demand for the labour that produces those goods. The invention of the personal computer, for example, reduced the demand for typewriters and, as a result, the demand for labour by typewriter manufacturers. At the same time, it increased the demand for labour in the electronics industry. Similarly, because different regions produce different goods, the demand for labour may be rising in one part of the country while it is falling in another. An increase in the price of oil may cause the demand for labour to rise in an oil-producing province such as Alberta, but because expensive oil makes driving more expensive, it decreases the demand for labour in an auto-producing province, such as Ontario. Economists call a change in the composition of demand among industries or regions a sectoral shift. Because sectoral shifts are always occurring, and because it takes time for workers to change sectors, there is always frictional unemployment.

Sectoral shifts are not the only cause of job separation and frictional unemployment. In addition, workers find themselves unexpectedly out of work when their firm fails, when their job performance is deemed unacceptable, or when their particular skills are no longer needed. Workers also may quit their jobs to change careers or to move to different parts of the country. Regardless of the cause of the job separation, it will take time and effort for the worker to find a new job. As long as the supply and demand for labour among firms is changing, frictional unemployment is unavoidable.

Public Policy and Frictional Unemployment

Many public policies seek to decrease the natural rate of unemployment by reducing frictional unemployment. Government employment agencies disseminate information about job vacancies with a view to matching jobs and workers more efficiently. Publicly funded retraining programs are designed to ease the transition of workers from declining to growing industries. If these programs succeed at increasing the rate of job finding, they decrease the natural rate of unemployment.


Other government programs inadvertently increase the amount of frictional unemployment. One of these is employment insurance (EI). Under this program, unemployed workers can collect a fraction of their wages for a certain period after losing their jobs. Although the precise terms of the program differ from year to year and from province to province, a typical worker covered by employment insurance in Canada has received about 50 percent of his or her former wages for about half a year. Before change in the EI program in the mid-1990s, Canada’s insurance system was one of the most generous in the world.

By softening the economic hardship of unemployment, employment insurance increases the amount of frictional unemployment and raises the natural rate. The unemployed who receive employment-insurance benefits are less pressed to search for new employment and are more likely to turn down unattractive job offers. Both of these changes in behaviour reduce the rate of job finding. In addition, because workers know that their incomes are partially protected by employment insurance, they are less likely to seek jobs with stable employment prospects and are less likely to bargain for guarantees of job security. These behavioural changes raise the rate of job separation.

That employment insurance raises the natural rate of unemployment does not necessarily imply that the policy is ill advised. The program has the benefit of reducing workers’ uncertainty about their incomes. Moreover, inducing workers to reject unattractive job offers may lead to a better matching between workers and jobs. Evaluating the costs and benefits of different systems of employment insurance is a difficult task that continues to be a topic of much research.

Economists often propose reforms to the EI system that would reduce the amount of unemployment. One common proposal is to require a firm that lays off a worker to bear the full cost of that worker’s employment benefits. Such a system is called 100 percent experience rated, because the rate that each firm pays into the employment-insurance system fully reflects the unemployment experience of its own workers. The programs in many countries are partially experience rated. Under this system, when a firm lays off a worker, it is charged for only part of the worker’s employment benefits; the remainder comes from the program’s general revenue. Because a firm pays only a fraction of the cost of the unemployment it causes, it has an incentive to lay off workers when its demand for labour is temporarily low. Canada’s system has no experience rating, so this incentive problem is acute. By reducing that incentive, the proposed reform may reduce the prevalence of temporary layoffs.

Canada’s unemployment rate was rising quickly during the 2009 recession, reaching 8.6 percent, up two and a half percentage points over a one-year period. Heated debate centred on the regional disparities and the incomplete coverage features of Canada’s EI system. The opposition Liberals made a call for reform on these issues their central attack on the governing Conservatives. An excellent assessment of these controversies is available in a special report published in April 2009 entitled Is Canada’s Employment Insurance Program Adequate? Andrew Coyne, National Editor of Maclean’s, Canada’s national news magazine, also wrote a thought-provoking article on EI reform at this time entitled “What Would Real Job Loss Insurance Look Like?” 2



Employment Insurance and the Rate of Job Finding

Between World War I and World War II, Britain experienced persistently high unemployment. From 1920 to 1938 the unemployment rate in Britain averaged 14 percent and never fell below 9 percent.

Economists Daniel Benjamin and Levis Kochin have suggested that Britain’s generous unemployment benefits can largely explain this high rate of unemployment. They cite three pieces of evidence to support their view. First, during this period, increases in British unemployment benefits coincided with increases in the economy’s unemployment rate. Second, teenagers, who received few or no unemployment benefits, had much lower unemployment rates than adults. Third, when the benefits for married women were reduced in 1932, their unemployment rate dropped significantly relative to that for men. All three pieces of evidence suggest a connection between unemployment benefits and unemployment rates.

This explanation of interwar British unemployment is controversial among economists who study this period. One difficulty in interpreting the evidence is that the data on unemployment benefits and unemployment rates may reflect two different relationships—one economic and one political. On the one hand, the higher the level of benefits, the more likely it is that an unemployed person will turn down an unattractive job offer, and the higher the level of frictional unemployment. On the other hand, the higher the rate of unemployment, the more pressing unemployment becomes as a political issue, and the higher the level of benefits the government chooses to offer. Hence, high unemployment rates may have caused high unemployment benefits, rather than the other way around. When we observe an empirical relationship between unemployment rates and unemployment benefits, we cannot tell whether we have identified an economic connection, a political connection, or some combination of the two.3 Given this problem, other studies have examined data on uemployed individuals, rather than data on economy-wide rates of unemployment. Individual data sometimes provide evidence that is less open to alternative interpretations.


One finding from individual data is that when unemployed workers become ineligible for employment insurance, the probability of their finding a new job rises markedly. Canadian evidence shows that the probability of an unemployed person finding employment varies—depending on how many weeks that person has been unemployed and how many weeks of employment-insurance benefits remain. In the first few weeks after becoming unemployed, the probability of being employed is about 15 percent. Then, if the person remains unemployed, the probability of a job falls to quite a low number—3 percent or 4 percent after 20–25 weeks of unemployment. As the unemployment spell lasts even longer, however, the probability of work rises again—to the 12–13 percent range—as the exhaustion of employment-insurance benefits is approached. Many economists interpret this evidence as verifying that at least some part of recorded unemployment is voluntary.4

More evidence in this vein comes from considering the regional dimensions of the Canadian unemployment system. The Canadian employment-insurance program has been substantially more generous in the high-unemployment regions of the country. These differences have been mainly due to differences in what is called the duration ratio—the maximum number of weeks one can collect benefits relative to the minimum number of weeks of work that generate the insurable earnings necessary to become elligible for any benefits. This ratio has varied between 26/14 in relatively low-unemployment provinces like Saskatchewan to 42/10 in relatively high-unemployment provinces like Newfoundland. Multiplying these duration ratios by the proportion of a worker’s wage that could be received through employment insurance (0.57 at the time these particular duration ratios were relevant) generates an estimated wage subsidy that varies between 1.06 and 2.39. These values indicate the number of weeks of insurance benefits and leisure (at full pay) to which Canadians were entitled for each week of work completed up to the minimum qualifying value (of 10 weeks or 14 weeks). Clearly, work disincentives were much greater in high-unemployment provinces. Many economists have concluded that the Canadian employment-insurance system has exacerbated regional unemployment disparities by discouraging the migration of labour from high- to low-unemployment rate regions.

Additional evidence on how economic incentives affect job search comes from an American experiment that the state of Illinois ran in 1985. Randomly selected new claimants for unemployment insurance were each offered a $500 bonus if they found employment within 11 weeks. The subsequent experience of this group was compared to that of a control group not offered the incentive. The average duration of unemployment for the group that was offered the $500 bonus was 17.0 weeks, compared to 18.3 weeks for the control group. Thus, the bonus reduced the average spell of unemployment by 7 percent, suggesting that more effort was devoted to job search. This experiment shows that the incentives provided by the employment-insurance system affect the rate of job finding.5 image