Are unemployment figures meaningless?

Are unemployment figures meaningless?

ARE U N E M P L O Y M E N T FIGURES MEANINGLESS? KENNETH W. CLARKSON AND ROGER E. MEINERS Peter Drucker has charged that the traditional u n e m p l ...

472KB Sizes 1 Downloads 122 Views

ARE U N E M P L O Y M E N T FIGURES MEANINGLESS? KENNETH W. CLARKSON AND ROGER E. MEINERS

Peter Drucker has charged that the traditional u n e m p l o y m e n t index is "meaningless and misleading." Policy makers are misled by official u n e m p l o y m e n t statistics because they are still politically accepted; economically they are, as Professor Drucker says, "an abomination, an Alice-in-Wonderland stew of apples, oranges and red herrings. ''1 In the twenty-five years before 1972, official u n e m p l o y m e n t averaged 4.7 percent. In the five years after 1972, u n e m p l o y m e n t averaged over 6.7 percent, an increase of over 40 percent in measured unemployment. It is our hypothesis that the measured high rate of u n e m p l o y m e n t , especially in recent years, can be explained in large part by federal programs that directly and indirectly cause beneficiaries of the programs to report themselves as u n e m p l o y e d when they are not seeking work, 1. Peter F. Drucker, "Meaningful Unemployment Figures," Wall Street Journal, November 3, 1978: 18.

K E N N E T H W. C L A R K S O N teaches at the Law and E c o n o m i c s C e n t e r at the University o f Miami. R O G E R E. M E I N E R S is a m e m b e r of the managem e n t faculty at Texas A&M University.

or to engage in activities prolonging the period of u n e m p l o y m e n t ) These individuals, who were previously not included in the labor force, are now counted in the official unemployment statistics.

WORK REGISTRATION REQUIREMENTS Since July l, 1972, work registration requirements have been in effect for the two largest federal income transfer programs, food stamps and AFDC (Aid to Families with Dependent Children). By the terms of 1971 amendments to both programs, certain beneficiaries must register with the employment service for work. Essentially, participants in these programs, who are physically and mentally fit, and between 18 and 60 years of age, must register for work, conduct job-search activ2. For a comprehensive presentation of our hypothesis, see Kenneth W. Clarkson and Roger E. Meiners, "Institutional Changes, Reported Unemployment, and Induced Institutional Changes," Journal of Monetary Economics, forthcoming.~ Our original findings were reported in Inflated Unemployment Statistics (Coral Gables, Florida: Law and Economics Center, 1977).

FEBRUARY 1979

47

K E N N E T H W. C L A R K S O N AND R O G E R E. MEINERS

ities, and accept certain offers of employment, or they lose their welfare benefits. Exempted from registration are persons who must take care of children or incapacitated persons, some students, and participants in drug addiction or alcoholic treatment programs.

Food-stamp and AFDC recipients who have registered for work are required to follow the rules of the e m p l o y m e n t service in the same manner as all other applicants. Hence, they may be called in by the employment service for testing, counseling, referral to training programs for employment, or obtaining additional information. Individuals who refuse to register or to comply with e m p l o y m e n t service rules are denied welfare benefits. About 6 percent of the food stamp registrants have lost their benefits because of their failure to comply with these rules.

48

MEASURING UNEMPLOYMENT Each m o n t h the Bureau of Labor Statistics analyzes and publishes information on population, labor force, and unemployment. The u n e m p l o y m e n t statistics are derived from the Current Population Survey conducted b y the Bureau of the Census for the Bureau of Labor Statistics. Three key elements determine whether or not an individual surveyed by the Bureau of the Census is counted as unemployed: not working, available for work, and looking for work. Since the first two elements are satisfied by registering for work as part of the eligibility requirements for the particular program, it is important to concentrate on what constitutes "looking for work." The Interviewer's Reference Manual used for the Current Population Survey explicitly states that "registration in a public or private employment office" constitutes looking for work. 3 3. U.S. D e p a r t m e n t of C o m m e r c e , Bureau of the Census,

Current Population Survey: Interviewers Reference Manual.. CPS-250, rev. August 1976.

BUSINESS HORIZONS

Large numbers of individuals must register for work and, as a result, must be counted in the official u n e m p l o y m e n t statistics. If these individuals generally prefer not to work at existing wage rates and welfare benefits, or are largely unemployable, the work registration requirements will permanently increase the measured rate of unemployment. This means that the u n e m p l o y m e n t data collected since the implementation of the work registration programs are not comparable to the data collected before that time. Most important, the data collected are invalid for public policy purposes since they are now based on incorrect notions of what the u n e m p l o y m e n t figures represent.

EMPIRICAL ESTIMATES A preliminary estimate of the impact of the work registration requirements on measured u n e m p l o y m e n t is straightforward. 4 Using data available from the Department of Labor and official Bureau of Labor Statistics publications, corrected rates of u n e m p l o y m e n t are determined. Table 1 gives these estimates as well as the official measured unemployment. In each case the corrected number of unemployed persons was determined by subtracting the average total active food stamp and AFDC work registrants from the average official number of unemployed persons. The number of work registrants is estimated by using the active file for those on the food stamp program and the mandatory file for those receiving AFDC benefits. Both of these files represent less than the total number of welfare recipients who must register for work prior to receiving benefits. For example, in September 1976 there were 3.7 million initial food stamp registrants, but only 1.2 million active food stamp work registrants. The average number of food stamp and AFDC regis4. Other empirical evidence in support of the effect of work registration is discussed in detail in Clarkson and Meiners, "Institutional C h a n g e s . . . "

Are Unemployment Figures Meaningless?

TABLE 1 Estimates of the Effects of Food Stamp and AFDC Work Registration Requirements on the U n e m p l o y m e n t Rate, 1974-78

Item

1978 (lst half)

1974

1975

1976

1977

91,011

92,613

94,773

97,401

99,705

Average official unemployment (thousands)

5,076

7,830

7,288

6,855

6,058

Average official unemployment rate

5.6%

8.5%

7.770

7.0%

6.1%

Average food stamp work registration-active employment service applicants and AFDC (WIN) mandatory registrants (thousands)

1,527*

2,081

2,065

1,994

1,839t

Corrected average unemployment (thousands)

3,549

5,749

5,223

4,861

4,235

Corrected civilian labor force (thousands)

89,484

90,532

92,708

95,407

97,851

Corrected unemployment rate

4.0°70

6.4%

5.6%

5.1%

4.3%

Average civilian labor force (thousands)

*Based on a monthlyaveragefor June throughDecember.

49

tBased on a five-monthaverage,excludingApril. SOURCES: Bureau of Labor Statistics, Employment and Earnings, Table A-l; Department of Labor, Employment Series Automated Reporting Systems (unpublished statistics). trants was subtracted from the average civilian labor force to determine the corrected average civilian labor force. The ratio of the corrected average number of unemployed persons to the corrected average civilian labor force yielded the corrected u n e m p l o y m e n t rate. For example, in 1976 the average number of work r e g i s t r a n t s was 2,065,000. Subtracting this figure from the average official unemployment figure of 7,288,000 yields the corrected u n e m p l o y m e n t figure of 5,223,000. The corrected average civilian labor force is determined by subtracting the average number of work registrants from the official average civilian labor force of 94,773,000, which yields a corrected average civilian labor force of 92,708,000. Thus the corrected rate of u n e m p l o y m e n t is 5.6 percent (5,223,000 + 92,708,000 = 0.0563).

THE EMPLOYMENT "'PARADOX" As Professor Drucker and other scholars have noted, the official u n e m p l o y m e n t statistics do not make sense in light of the record high rates of e m p l o y m e n t in recent years. For that reason, it may be that employment, which is more accurately counted, may be a better measure of the state of the e c o n o m y in terms of the status of workers. Because of welfare work registration requirements, unemployment compensation, minimum wages, and other federal programs, as well as the changing demographic composition of the labor force, measured u n e m p l o y m e n t appears to be increasingly unreliable over time. Official u n e m p l o y m e n t statistics do not seem comparable to u n e m p l o y m e n t rates in earlier years. In Table 2, periods of relatively

FEBRUARY 1979

©

N

O

*Based on data from July--December 1972. N.A. = Not applicable.

1974 1977

1969

High Employment Periods

1956 1968 1972 Average 1976

Medium Employment Periods

1950 1955 1966 1971 Average 1975

Low Employment Periods

Year

57.11

56.52 56.98

56.06 56.00 56.05 56.04 56.06

55.25% 55.15 55.57 55.49 55.37 55.25

Civilian E m p l o y m e n t as a Percentage o f Total Noninstitutional Population, Age 16 and Over

Employment and Unemployment: A Comparison

TABLE 2

3.5 5.6 7.0

4.1 3.6 5.6 4.4 7.7

5.3% 4.4 3.8 5.9 4.9 8.5

Measured U n e m p l o y m e n t as a Percentage o f Civilian Labor Force

5.1%

N.A. 4.2%

N.A. N.A. 5.1%* N.A. 5.7%

N.A. N.A. N.A. N.A. N.A. 6.6%

Corrected UnemploTment as a Percentage o f Civilian Labor Force

0

0 ~Z >

f3 >

Are Unemployment FiguresMeaningless?

high, medium, and low employment (the ratio of civilian employment to total population) between 1950 and 1976 are shown together with the corresponding rate of unemployment. During periods of low employment (55.15 to 55.57 percent), measured unemployment averaged 4.9 percent of the total noninstitutional population. Yet in 1975, a similar rate of employment (55.25 percent) was associated with a measured unemployment rate of 8.5 percent. Medium employment periods, in which unemployment averaged 4.4 percent, have been associated with civilian employment averaging 56.0 percent of the population. But in 1976, a 56.1 percent employment rate yielded a 7.7 percent rate of unemployment, almost double the historical average. Finally, the highest period of employment after World War II, but before welfare work registration, occurred in 1969, when employment of 56.5 percent was associated with a 3.5 percent rate of unemployment. However, the highest employment rate

in the past 30 years, 57.1 percent in 1977, yielded a higher-than-average measured rate of unemployment of 7.0 percent. Table 2 clearly shows that the 1974-77 official unemployment rates do not fit previous historical patterns, which have been stable for many years. However, the corrected unemployment rates yield unemployment levels more consistent with earlier periods, before welfare work registration requirements.

UNEMPLOYMENT

AND PUBLIC POLICY

The use and misuse of unemployment statistics by politicians is a well-established common practice. Recently, however, legislators have been explicitly tying federal spending to measured unemployment. As noted by the Commissioner of the Bureau of Labor Statistics: "In recent years, another use of' statistical indicators has emerged: large sums of 51

, , r r//1~

fFr,,,,

_

/

D

FEBRUARYI979

KENNETH W. CLARKSON AND ROGER E. MEINERS

"There has been a permanent increase in the number of individuals included in the unemployment statistics who represent those excluded from the labor force, those who find work less attractive, and those not seeking work but required to register for work as a condition for receiving welfare benefits. Thus we find the circle is complete: Changes in our institutions-namely, the work registration requirements-have permanently raised official unemployment statistics, and in turn can be linked to new changes in institutions-public programs to reduce unemployment."

52

money are being transferred among different groups of Americans on the basis of movements of these statistics. Thus in 1977 alone, over $16 billion has been allocated to states and communities on the basis of unemployment statistics. "s This expenditure may represent the lower boundary of total federal spending to lower measured unemployment to "traditional" levels. For instance, in 1975 unemployment was reported as 8.5 percent, the highest since the Depression. This figure spurred arguments for new federal programs to combat unemployment, and for a general increase in federal spending to stimulate economic activity. The billions of dollars spent by the federal government because of the unusually high measured unemployment figures during this period are impossible to estimate. Higher levels of unemployment may have led to the passage of the Humphrey-Hawkins Full Employment and Balanced Growth Act. Federal legislation now tied to official unemployment rates includes the Public Works and Economic Development Act of 1965; the Comprehensive Employment and Training Act (CETA) of 1973, which includes Comprehensive Manpower Services and Public Employment Programs; the Emergency Jobs and Unemployment Assistance Act of 1974; and the Public Works Employment Act of 5. Julius Shiskin, "A New Role for Economic Indicators," Monthly Labor Review, November 1977: 4.

BUSINESS HORIZONS

1976. 6 Most of these programs were passed in response to high measured unemployment. Spending is directly tied to the official rate of unemployment in an effort to combat a problem whose magnitude has been seriously distorted. Although some of these programs were supposed to be temporary, most appear to have become permanent, and projected budget outlays will increase in future years.

~]][~l]~ Future research on unemployment data will surely yield more accurate estimates of errors resulting from work registration requirements and other federally instituted programs. However, it seems unlikely that our general conclusions will be altered. There has been a permanent increase in the number of individuals included in the unemployment statistics who represent those excluded from the labor force, those who find work less attractive, and those not seeking work but required to register for work as a condition for receiving welfare benefits. Thus we find the circle is complete: Changes in our institutions-namely, the work registration requirements-have permanently raised official unemployment statistics, and in turn can be linked to new changes in institutions-public programs to reduce unemployment. [Z3 6. Details of these programs are discussed in Clarkson and Meiners, "Institutional C h a n g e s . . . "