temporary layoffs and unemployment in canada

9
Temporary Layoffs and Unemployment in Canada MATTHEW ROBERTSON* The importance of tempora ry layoffs in measured unemployment has not been solidly established empirically. Clark and Summers (1 979) maintain that these separations account for only a small fraction of U.S. unemployment. Administrative data analyzed in this paper show that temporay layoffs are an important feature of the Canadian labor market. Approximately 35 per cent of the variation in the Canadian unemployment rate over time can be accounted for by changes in temporary layoffs and about 45 per cent by variations in all temporary employment separations. TEMPORARY LAYOFFS (rehires, recalls) have been recognized for over a decade as an integral and significant feature of the U.S. labor market (see Feldstein, 1975, 1976; Lilien, 1980). Both Feldstein and Lilien derive estimates of rehires from Bureau of Labor Statistics (BLS) establishment data. Clark and Summers (1979), on the other hand, use Current Population Survey (CPS) data to argue that temporary layoffs do not represent an important proportion of U.S. unemployment. They claim that only a small fraction of unemployment is due to those grouped in the official temporary layoff category, and that since temporary layoffs are of very short duration, only a small proportion of changes in the unemployment rate can be accounted for by firms recalling workers. This paper uses Canadian data to evaluate the following: (a) the proportion of total layoffs and other employment separations that are temporary as measured by ex ante household surveys and the proportion observed from ex post administrative data; (b) the quantitative importance of temporary separations among the unemployed; and (c) the significance of temporary ' Labour Market Studies Division, Employment and Immigration Canada, Ottawa. The author is grateful to Daniel Boothby and two anonymous referees for their comments. The views expressed in this study are those of the author and not necessarily those of Employment and Immigration Canada. INDUSTRIAL RELATIONS, Vol. 28, No. 1 (Winter 1989). 0 1989 Regents of the University of California 0019/8676/89/2 15/82/$10.00 82

Upload: matthew-robertson

Post on 30-Sep-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Temporary Layoffs and Unemployment in Canada

Temporary Layoffs and Unemployment in Canada

MATTHEW ROBERTSON*

The importance of tempora ry layoffs in measured unemployment has not been solidly established empirically. Clark and Summers (1 979) maintain that these separations account for only a small fraction of U.S. unemployment. Administrative data analyzed in this paper show that temporay layoffs are an important feature of the Canadian labor market. Approximately 35 per cent of the variation in the Canadian unemployment rate over time can be accounted for by changes in temporary layoffs and about 45 per cent by variations in all temporary employment separations.

TEMPORARY LAYOFFS (rehires, recalls) have been recognized for over a decade as an integral and significant feature of the U.S. labor market (see Feldstein, 1975, 1976; Lilien, 1980). Both Feldstein and Lilien derive estimates of rehires from Bureau of Labor Statistics (BLS) establishment data. Clark and Summers (1979), on the other hand, use Current Population Survey (CPS) data to argue that temporary layoffs do not represent an important proportion of U.S. unemployment. They claim that only a small fraction of unemployment is due to those grouped in the official temporary layoff category, and that since temporary layoffs are of very short duration, only a small proportion of changes in the unemployment rate can be accounted for by firms recalling workers.

This paper uses Canadian data to evaluate the following: (a) the proportion of total layoffs and other employment separations that are temporary as measured by ex ante household surveys and the proportion observed from ex post administrative data; (b) the quantitative importance of temporary separations among the unemployed; and (c) the significance of temporary

' Labour Market Studies Division, Employment and Immigration Canada, Ottawa. The author is grateful to Daniel Boothby and two anonymous referees for their comments.

The views expressed in this study are those of the author and not necessarily those of Employment and Immigration Canada.

INDUSTRIAL RELATIONS, Vol. 28, No. 1 (Winter 1989). 0 1989 Regents of the University of California 0019/8676/89/2 15/82/$10.00

82

Page 2: Temporary Layoffs and Unemployment in Canada

Research Note I 83

layoffs in adjusting the workforce to changes in business conditions.2 Our estimates generally support the Feldstein-Lilien findings that temporary layoffs are an important aspect of unemployment. Our estimates also indicate, however, that temporary layoffs comprise a significantly higher proportion of total unemployment spells than of total time unemployed during a year, due to the fact that temporary layoffs have a shorter average duration than all unemployment spells.

Measuring Temporary Layoffs Data from the Canadian Labour Force Survey (LFS), which is very similar

in design to the CPS, suggest that temporary layoffs constitute a small percentage of layoffs and unemployment. The LFS is a monthly sample survey of about 50,000 households. The sample is designed to represent the noninstitutional civilian population, 15 years of age and over. Each month, detailed questions are asked about the activities of individuals in the sample of households. Respondents are classified as employed, unemployed, or not in the labor force.

To be considered as a temporary layoff, an individual must respond negatively to the following question: “Last week, did ... do any work at a job or business?”; and then respond positively to the question: “Last week, did ... have a job or business at which he/she did not work?” If the answer to the latter question is yes, then the respondent is asked: “Why was.. .absent from work last week?” If the answer is “layoff, expects to return” (one of several possible responses3), the individual is classified as unemployed. This is the measure of temporary layoffs in the LFS. It accounts for only about 5 per cent of the unemployed (see Table 1). Below, administrative un- employment insurance (UI) data are used to argue that the LFS measure substantially underestimates unemployment due to temporary layoffs.

Data from the CPS reinforce this view. In the CPS, individuals are asked, “Did ... have a job or business from which he/she was temporarily absent or on layoff last week?” If yes, then the respondent (or proxy) is asked, “Why was ... absent from work last week?”, of which temporary layoff (under 30 days) and indefinite layoff (30 days or more or no definite recall date) are two possible responses. These two components accounted for 12.7 per cent

The points made in this paper must be modified appropriately if, as Clark and Summers argue, total unemployment is underestimated because individuals are classified as out of the labor force who ought to be considered as unemployed.

“Other Reasons” include illness or disability, personal or family responsibilities, bad weather, labor dispute (strike or lockout), new job to start in the future, vacation, or seasonal business (excluding paid workers).

Page 3: Temporary Layoffs and Unemployment in Canada

84 / MATTHEW ROBERTSON

TABLE 1

FORCE SURVEY DATA”

Individuals Per cent “on layoff’ “on layoff’

not looking

TOTAL UNEMPLOYED ACCOUNTED FOR BY INDIVIDUALS “ON LAYOFF”, 1976-1987 LABOUR

as a per cent of total unemployed for work

1976 9.2 82.0 1977 7.9 80.6 1978 6.7 80.3 1979 6.7 80.3 1980 8.2 76.1 198 1 8.2 77.0 1982 8.8 74.1 1983 5.7 73.5 1984 4.6 76.6 1985 4.2 76.8 1986 4.9 75.4 1987 5.0 77.6

Source: Statistics Canada, The Labour Force, 71-001 (1982). * According to Statistics Canada, persons are classified as “on layoff’ only if they are expected to return to the job or business from which they were laid off.

of the unemployment in 1987, much higher than the 5 per cent figure for Canada.

The magnitude of temporary layoffs and other employment separations in Canada can be determined directly from Record of Employment (ROE) data. By law, employers in Canada must issue an ROE for individuals who have an interruption of earnings (layoffs, quits, maternity, retirement, return to school, e t ~ . ) . ~ The ROE specifies the reason for separation, earnings, occupation, weeks worked, date of issue, and other information. The distribution of employment separations for 1984 is shown in Table 2. Layoffs (lack of work) comprise the largest proportion of total employment separations (48.3 per cent in 1984). Other quantitatively important groups include voluntary quits (19.2 per cent) and other separations (21.0 per cent). The latter group includes predominantly discharges, casual (seasonal) labor , fixed-

An interruption of earnings occurs if, following a period of employment, there is a layoff or separation from that employment and a period of seven consecutive days when an employee performs no work and has no earnings from that employer; and when persons insured under the unemployment insurance act stop working because of illness, injury, quarantine, pregnancy, or adoption, and their weekly earnings (for example, paid sick leave) fall below 60 per cent of the normal insurable earnings from that employment.

Page 4: Temporary Layoffs and Unemployment in Canada

Research Note I 85

TABLE 2 EMPLOYMENT SEPARATIONS BY REASON FOR SEPARATION, 1984

Reaon for separation

Per cent returning to former employer

Number Per cent within (thousands) of total two years

Layoff Labor dispute Return to school Injury/illness Quit Maternity Retired Work-sharing Apprenticeship training Other reasons Total separations

2,510 10

185 167 996 132 43 43 21

1,091

5,198

48.3 0.2 3.6 3.2

19.2 2.5 0.8 0.8 0.4

21.0 100.00

61.0 63.2 50.2 58.2 18.3 56.4 22.5 92.9 81.9 41.1 53.3

Source: Employment and Immigration Canada, Administrative Record of Employment (ROE) data.

term contracts or projects, and employment terminations associated with business failureklosure.

Since the administrative data cover all employment separations over time (since 1974), it is possible to search ROES and other administrative data sequentially to determine whether or not an individual returned to hislher former empl~ye r .~ In excess of 50 per cent of all employment separations in 1984 ended with the individual returning to the same employer within two years (see Table 2). For layoffs, this figure is about 60 per cent (1984). When broken down by sector, temporary layoffs represent about three- quarters of all layoffs in the manufacturing sector, a figure which is similar to Feldstein’s (1975) estimate of 80 per cent for the U.S. labor market.

Each firm is issued a Revenue Canada employer identification number which is entered on the Record of Employment when issued to the individual upon an employment separation. To determine whether or not an individual returned subsequently to the same firm, the next Record of Employment is consulted. A temporary layoff or separation is defined as a situation where the two identification numbers are identical. If the individual did not have another ROE subsequently, T4s data (drawn from forms issued annually by firms to individuals for income tax purposes) are used. As with the ROE, the T4s contains the Revenue Canada employer identification number. There is a small attrition factor associated with those who never had subsequent labor market activity (an ROE or a T4s). If there is no ROE or T4s within two years after separation, an individual is considered to have withdrawn from the labor force.

Page 5: Temporary Layoffs and Unemployment in Canada

86 / MATTHEW ROBERTSON

The difference between the Labour Force Survey data, in which relatively few respondents report themselves to be on temporary layoff (see Table 1) and the administrative data estimates (Table 2) is likely due to the phrasing of the sequence of questions which determine whether an individual is classified as being on temporary layoff. Recall that, to be classified as having a job but not at work, the LFS respondent (who may be a proxy) must answer that the individual in question “[had] a job at which he/she did not work.” I suspect that the answer to this question will be “no” if there is any doubt as to whether the individual will ultimately be recalled, or when the temporary layoff is for more than a few weeks. If the answer is no, the individual cannot be classified as being on temporary layoff. Moreover, some individuals who eventually return to the same employer may not have expected to do so at some time during their layoff.

Temporary Layoffs as a Component of the Current Unemployed

Temporary and permanent layoffs and other employment separations constitute the job losers and job leavers component of the current unemployed. The current unemployed also contains entrants and re-entrants to the labor force. In 1986, the latter two groups accounted for about 25 per cent of the unemployed in Canada. Since very few entrants and re- entrants have an employer to which they can return, the fraction of flows into unemployment accounted for by temporary layoffs is less than the fraction of persons separating from a job who eventually return to their original employer. It should also be noted that temporary layoffs are the major but not the only source of temporary employment separations. Some separations are accounted for by persons who return to their former employer after leaving for reasons other than layoff. Thus, it is important to focus on the return and nonreturn behavior of all the unemployed, not just the job losers/leavers component.

Since it is not possible to accurately measure temporary employment separations from Labour Force Survey data, UI data for individuals are used. Regular UI benefit weeks should be a good proxy for unemployment. About 95 per cent of all paid workers in Canada are insured under the UI Act. About 80 per cent of the unemployed in the LFS draw regular UI benefits.6 Regular UI beneficiaries in turn account for approximately 90 per cent of total beneficiaries. The remaining 10 per cent of beneficiaries

In 1986, 78.6 per cent of the unemployed were drawing regular UI benefits. Those not receiving benefits are largely entrants and re-entrants and individuals who have exhausted their UI benefits and are still unemployed.

Page 6: Temporary Layoffs and Unemployment in Canada

Research Note / 87

essentially are not in the labor force (they are recipients of maternity, injury/ illness, or retirement benefits).

A 1 per cent random sample of regular UI beneficiaries was drawn from UI administrative data for each year in the 1974-1984 period to proxy the current unemployed as measured by the Labour Force Survey. Using the Record of Employment, UI beneficiaries were identified in terms of the reasons for employment separation (layoff, quit, other) and whether or not they returned to their former employer.’ The number of weeks that each group drew UI benefits during a year was also determined. This procedure makes the analysis analytically comparable to the Clark and Summers (1979) study. Information on those who drew UI is obtained from the Status- Vector file, which is a derivative of the administrative “Benefits and Overpayments” (BNOP) file of the universe of UI claimantsg

The analysis focuses on a random sample of regular UI beneficiaries for the year 1984. About 2.5 million individuals drew regular UI benefits in 1984 (out of a labor force of approximately 12 million). The year 1984 was chosen in order to allow a subsequent period for identifying workers’ return or nonreturn status. Table 3 shows that 53.3 per cent of these UI recipients returned to their former employer, while 61.5 per cent of those who had been laid off returned subsequently. Those who had been laid off constitute about 70 per cent of all UI return employment. The table also shows the number of weeks drawing UI benefits in 1984 for those who return to their former employer and those who do not, by reason for separation. Regular UI beneficiaries who returned to their former employer account for 43.8 per cent of total UI weeks. UI beneficiaries who had been laid off and who returned to their former employer account for 31.3 per cent of total regular UI benefits. The average duration of a return layoff for 1984 was 21 weeks. For nonreturn layoffs, the average duration was 30 weeks.

When these figures are adjusted for the fact that those drawing regular UI benefits account for about 80 per cent of the unemployed, it can be seen that approximately one-third of weeks unemployed can be accounted for by persons who eventually will return to their former employer. For the

’ A completed copy of a Record of Employment is required if an individual wishes to draw UI benefits.

The Status-Vector File contains detailed week-by-week information on the activities of UI claimants during the time they reported to the Unemployment Insurance Commission (UIC). The file is based on a 10 per cent sample of the universe. The data represent an extensive re-processing of the UIC administrative files that makes the information more amenable to research and policy analysis. For detailed information on the administrative data, see Department of Employment and Immigration, 1971-1986. In addition to UI data, Record of Employment, and T4s data are -Is0 part of the Longitudinal Labour Force File. The period covered is 1971-1986, except for the ROE, which begins in 1974.

Page 7: Temporary Layoffs and Unemployment in Canada

88 / MATTHEW ROBERTSON

TABLE 3

RECIPIENTS OF REGULAR UI BENEFITS, WHO RETURNED AND DID NOT RETURN TO FORMER EMPLOYER, 1984

Per cent of Per cent of Per cent of Reason for UI beneficiaries total total UI weeks employment Non- Non- Non- separation Return return Return return Return return Total

Layoff 61.5 38.5 71.3 50.8 31.3 28.0 59.3 Quit 18.3 81.7 3.4 17.4 2.0 9.9 11.9 Other reasons 47.6 52.4 25.3 31.8 10.5 18.3 28.8

All separations 53.3 46.7 100.0 100.0 43.8 56.2 100.0

Source: Employment and Immigration Canada, Administrative Record of Employment (ROE) data.

unemployed who had been laid off, the proportion is about 25 per cent of all weeks, after adjustment.

Temporary Layoffs and the Unemployment Rate

The significance of temporary employment separations in changes in the unemployment rate is examined briefly here. Although temporary layoffs could be an important factor in the adjustment of the workforce to cyclical fluctuations in demand, this issue has not received any consideration in Canada.

The unemployment rate in a given year can be defined as the number of weeks of unemployment as a proportion of the number of weeks of labor market activity. Variations in the unemployment rate are determined largely by changes in total weeks of unemployment, since relative changes in weeks of labor force participation vary considerably less. lo Again, total weeks of unemployment is proxied by weeks drawing regular UI benefits.

Table 4 shows the number of weeks drawing UI benefits for a random sample of UI beneficiaries. The years 1979 to 1982 were selected because this was a period in which the unemployment rate rose significantly; regular

Lilien (1980) has conducted an empirical examination for the U.S. economy. lo For example, largely as a result of the 1981-1982 recession, total weeks of unemployment

rose by 90 per cent from 1978 to 1982. Persons in the labor force at some time during the year increased for the same period by 16 per cent. On the other hand, total weeks of labor market activity (employment plus unemployment) rose by only 9 per cent. Total weeks of unemployment is the number of people unemployed at some time during the year times the average number of weeks unemployed per person unemployed at some time during the year. Calculations are based on data from Statistics Canada (1978, 1982).

Page 8: Temporary Layoffs and Unemployment in Canada

Research Note I 89

TABLE 4 PER CENT OF THE TOTAL INCREASE IN WEEKS DRAWING UI BENEFITS ACCOUNTED FOR BY

TEMPORARY AND PERMANENT EMPLOYMENT SEPARATIONS, 1979-1982

Temporary Permanent employment employment separations separations (per cent) (per cent)

Layoff Quit Other reasons

37.3 0.4

10.3

34.5 4.6

12.9 Total 48.0 52.0

Source: Employment and Immigration Canada, Administrative Record of Employment (ROE) data.

UI benefit payments about doubled over these years. Then the data were used to determine whether individuals returned to their former employer.

Calculations based on data presented in Table 4 indicate that 48 per cent of the total rise in UI weeks was associated with temporary employment separations and about 37 per cent with temporary layoffs. The estimate for temporary layoffs is roughly similar to Lilien’s (1980) estimate for the U.S. economy.

Conclusion Canadian administrative data (1984) indicate that about 50 per cent of all

employment separations and 60 per cent of layoffs end with individuals returning to their former employer. In manufacturing, about 75 per cent of layoffs end in recall. Clark and Summers’ (1979) conclusion that temporary employment separations in the U.S. play an insignificant role in the analysis of unemployment does not apply to Canadian labor market behavior. Estimates for Canada indicate that about 35 per cent of the current unemployed will return to their former employer. In addition to comprising a significant proportion of the unemployed, temporary employment separ- ations are an important factor in variations in the unemployment rate.

REFERENCES Clark, Kim B. and Lawrence H. Summers. “Labor Market Dynamics and Unemployment:

Department of Employment and Immigration. “Longitudinal Labour Force File Document-

Feldstein, Martin S. “The Importance of Temporary Layoffs: An Empirical Analysis,”

A Reconsideration,” Brookings Papers on Economic Activity, I: 1979, 13-60.

ation,” No. 1, 6th ed. Ottawa: 1971-1986.

Page 9: Temporary Layoffs and Unemployment in Canada

90 / MATTHEW ROBERTSON

- . “Temporary Layoffs in the Theory of Unemployment,” Journal of Political Economy,

Lilien, David M. “The Cyclical Pattern of Temporary Layoffs in United States Manufacturing,”

Statistics Canada. Annual Work Patterns Survey, 71-531, 71-001. Ottawa: 1978 and 1982.

Brookings Papers on Economic Activity, 111: 1975, 725-744.

LXXXIV (October, 1976), 937-957.

R e v i m of Economics and Statistics, LXII (February, 1980), 24-31.