Hearing on "The U.S. Economic Outlook."


Prepared Statement of Mr. Alan B. Krueger
Professor of Economics
Princeton University

10:00 a.m., Tuesday, March 12, 2002 - Dirksen 538

Good morning, Mr. Chairman and distinguished members of the Senate Banking, Housing and Urban Affairs Committee. My name is Alan Krueger and I hold the Bendheim professorship in economics and public affairs at Princeton University. I appreciate the opportunity to share my views on recent economic developments, particularly as they relate to the labor market.

I. THE LABOR MARKET SITUATION AND SHORT-TERM OUTLOOK

Although some debate the exact meaning of the subjective definition commonly used to define a recession, there is little doubt that the labor market started to turn down in the beginning of 2001, and that March 2001 -- the official beginning of the recession according to the National Bureau of Economic Research -- marked a turning point. After reaching a 30-year low of 3.9 percent in April 2000, the unemployment rate fluctuated in a narrow range between 3.9 and 4.1 percent for the remainder of 2000, amid signs that economic growth was weakening. The unemployment rate increased from 4.3 percent in March 2001 to 4.9 percent in August 2001, and reached a recent peak of 5.8 percent in December 2001. The rate fell to 5.6 percent in January and fell again in February to 5.5 percent.

Because, other things being equal, the unemployment rate increases when discouraged workers decide to actively search for work, economists often prefer to examine employment growth from the establishment survey, and the employment-to-population rate from the household survey, in addition to the unemployment rate. These data tell a similar story. The employment-to-population rate reached an all-time high of 64.8 percent in April 2000, stood at 64.3 percent in March 2001 and fell to 63.4 percent as of August 2001. The employment rate continued to fall to 62.6 percent in January 2002, and increased to 63.0 percent in the latest employment report, which pertains to February 2002. Unlike the unemployment rate, the employment rate fell in January 2002, suggesting that the improvement in the unemployment rate that month resulted from labor force withdrawal rather than an increased rate of job finding. In February, the small decline in the unemployment rate and the rise in the employment-rate both pointed in the same direction.

Total payroll employment peaked at 132.7 million jobs in March 2001, and was down to 132.4 million in August 2001. It fell to 131.2 million in January 2002, and increased by 66,000 in February, an amount that is close to the average monthly absolute revision to the series. (The January and February figures are preliminary and subject to future revisions.) In the 11 months since March 2001, the month the recession began, total employment has fallen by 1.4 million jobs. Private sector employment is down by 1.8 million jobs in this period. By comparison, 11 months after the 1991 recession began, private sector employment was down 1.5 million jobs, and total employment was also down 1.5 million jobs. So, looking over a comparable interval, job destruction was somewhat greater in the private sector in the latest recession than in the previous one.

Employment continued to drift downward after the recovery began in March 1991, and reached bottom in February 1992, with private sector employment down a total of 1.8 million jobs from the peak and total employment down 1.6 million from the peak.

The latest GDP news suggests that the economy began to turn around late in 2001 and that the recession likely has ended. I think it will take more months of data before one can reach the conclusion that the labor market has reached bottom and is on the upswing, however. I also suspect that employment growth will remain sluggish for a time to come, especially for the less skilled. Employment and unemployment tend to be lagging indicators when the economy begins to improve. This point was made by Alan Greenspan in his prepared testimony before the House Committee on Financial Services on February 27th. "Even if the economy is on the road to recovery," he said, "the unemployment rate, in typical cyclical fashion, may resume its increase for a time."

Historically, the lingering effects of high unemployment in the first stages of a recovery tend to be concentrated among the less skilled and minorities. This seems to be the case despite the fact that recessions are becoming more egalitarian in terms of who they affect.

Such a pattern was clearly evident in the early 1990s. When the recession officially ended in March 1991 the unemployment rate was 6.8 percent. The rate continued to rise for another 15 months, however, and did not settle below 6.8 percent again until the end of 1993. Moreover, the unemployment rate rose from 12.3 to 13.5 percent for high school dropouts in the year after the recession ended, while for college graduates it held steady at 2.9 percent. The "jobless recovery" mainly involved the less skilled.

The current recession started out in a very unusual fashion. As the following table makes clear, from March to July of 2001, unemployment rose more for college graduates and those with some college education than it did for high school dropouts. Since July, however, unemployment has increased more for high school dropouts and high school graduates than for more highly educated workers, as is the usual pattern in a downturn. It is also worth noting that the unemployment rate ticked up for those with a high school degree or less last month, despite falling overall.

Table 1: Seasonally Adjusted Unemployment Rate by Education, Age 25 and Older
Education March 2001 July 2001 February 2002
Less than High School 6.8% 6.8% 8.3%
High School 3.8 4.1 5.3
Some College 2.7 3.1 4.1
BA or higher 1.9 2.2 2.9

Source: Bureau of Labor Statistics.

A similar picture holds by race. From March to July 2001 the unemployment rate increased from 3.7 to 4.1 percent for whites, and, uncharacteristically, fell from 8.4 to 8.1 percent for blacks. From July 2001 to February 2002, however, the rate increased from 8.1 to 9.6 percent for blacks, and increased more moderately, from 4.1 to 4.9 percent, for whites. The unemployment rate for Hispanics was also uncharacteristically stable in the beginning of the recession (standing at 6.2 percent in March and July), and then increased sharply to 8.1 percent by January 2002, before quixotically falling by a percentage point in February.

The broad nature of the first phase of the latest recession came as a surprise, but is consistent with the more-than-usual egalitarian tilt to the early 1990s recession, and probably resulted from the plunge in capital investment that apparently precipitated the downturn, the implosion of many dot-coms, and the fact that the cyclically sensitive manufacturing sector is much more skill intensive than it was 20 or 30 years ago. In any event, it is likely that the egalitarian phase is over.

There are many theoretical reasons to suspect that job growth would be slow at the beginning of the recovery. First, at the beginning of a recovery employers are not sure if improved conditions will persist, so they expand work hours rather than hire new employees. Second, many employers also "hoard" skilled workers (i.e., particularly those with specific training) during a recession because if they let them go it would be costly to hire and train replacements when conditions improve. Neither of these reasons, however, accounts for why job growth is particularly sluggish for the less skilled when the economy begins to turnaround, which seems to regularly occur. In fact, because of labor hoarding of skilled workers one might expect that employers are relatively "overstaffed" with skilled workers when the recession ends, and would therefore be less likely to hire skilled workers.

Third, as Melvin Reder suggested in a 1955 article, in a downturn many employers raise skill requirements for a given job, rather than cut pay. "Upskilling" of positions is common in a recession and probably at the beginning of a recovery as well. Consequently, the less skilled find their job options even more limited until demand picks up smartly, while skilled workers take positions further down the job ladder.

Finally, Lawrence Katz of Harvard suggests another reason: "Think of a recession as a time when firms reorganize." Reorganization tends to increase demand for skilled workers, who are more flexible, over less skilled workers. Furthermore, when companies introduce new technology as part of a reorganization they tend to hire skilled workers to operate the equipment and release unskilled workers whose jobs are made redundant.

To some extent, the lingering pattern of unemployment, especially among the less skilled, after growth resumes is probably inadvertently reinforced by interest rate cuts by the Federal Reserve. With a lag, rate cuts stimulate demand for new capital and consumer durables. But this is a two-edge sword for workers. On the one hand, a general rise in economic activity increases the demand for all factors of production, including workers. On the other hand, because machinery is cheaper than it used to be, in many industries companies replace some workers with machines, or hire fewer workers than they otherwise would have, because the machines can do the work at lower cost.

Also notice that capital – especially high-tech equipment -- and skilled labor are generally considered complementary inputs in production, while capital and unskilled workers are substitutes. In other words, high skilled workers are hired to operate and service the new machines, while less skilled workers are let go because the machines can do their work.

This leads me to the conclusion that the best way to avoid another "jobless recovery" is by stimulating demand for less skilled workers and by raising the skills of the unemployed. I think the interest rate cuts and the recently passed accelerated depreciation allowance will stimulate demand for more highly skilled workers. In looking forward, I would recommend policies that would increase employment of less skilled labor, such as job training.

Another important aspect of the labor market concerns wages. Table 2 (at the end of this testimony) reports real hourly wages by decile of the wage distribution each year since 1973 based on Current Population Survey Data. As is well known, real wages fell considerably for lower paid workers from 1979 to the mid 1990s. (This table uses the BLS’s new research series CPI to deflate wages, so the decline in real wages in the 1980s was not as great as it is with the conventional CPI deflator.) Real wage growth was very strong after 1996, however. Notably, most of the ground that was lost for those at the bottom in the 1979-95 period was regained in the last 5 years. The weakest wage growth in the last decade was for those in the middle of the distribution, a phenomenon that I have previously called "the sagging middle."

Most research finds that real wages have moved slightly procyclically since 1970, although I agree with Katharine Abraham and John Haltiwanger that "the cyclicality of real wages is not likely to be stable over time." The exceptionally low unemployment in the late 1990s, combined with two minimum wage increases, spurred the impressive wage growth in the second half of the 1990s, especially for the least paid workers. In the latest downturn, nominal wage growth slowed down, but inflation slowed even more, so real wages continued to grow. This factor has probably bolstered consumer spending, which was surprisingly robust during the downturn.

A rising tide continues to lift all boats, and the late 1990s provides ample evidence that strong economic growth greatly helps all segments of society. But the effect of a rising tide on employment does not appear to occur immediately. In the early 1990s, weak employment growth lingered long after the national economic tide began to rise. Moreover, research suggests that a given change in economic conditions has a more gradual effect on labor demand in a recovery than in a recession. As Edward F. McKelvey, a senior econmist at Goldman, Sachs observed after the latest unemployment report, "It would be premature to say that there is going to be heavy net hiring soon."

II. UNEMPLOYMENT INSURANCE

Because I expect unemployment to linger at relatively high levels in the beginning of the recovery, I think it is important and appropriate that last week both Houses of Congress passed, and the President signed, a bill to extend unemployment insurance (UI) benefits for an additional 13 weeks.

But I think recent history highlights the importance of making additional reforms to make unemployment insurance a more efficient and more effective automatic stabilizer. First, the automatic triggers that temporarily turn on extended benefits without Congressional action should be set at more realistic levels. The state triggers are connected to the insured unemployment rate; that is, the fraction of covered workers who receive benefits. The insured unemployment rate must exceed 5 percent for extended benefits to be provided, and must be 120 percent above the rate in the corresponding period in each of the prior two calendar years. Because insured unemployment has drifted down relative to the BLS’s total unemployment rate, and because the natural rate of unemployment has declined, it is very unlikely that a state will automatically trigger extended benefits. In practice, the automatic triggers have become beyond reach, and we rely on Congress to vote for extended benefits during a downturn.

It should not be necessary for Congress to have to agree to ad hoc extended UI benefits when it is clear that the economy has deteriorated in a specific region. Realistic automatic triggers would be much more expedient and more efficient. Funds would be saved if extended benefits were more closely targeted to specific states experiencing severe economic distress, rather than applied nationwide. Furthermore, if extended benefits turned on more quickly in contracting areas, consumption would be smoothed and the downturn would be less severe.

Second, the financing of UI could do more to stabilize the economy and discourage layoffs. To pay for benefits, the UI system builds up reserves during prosperous times and draws them down during slack times. A common measure of the health of trust funds is the reserve ratio: the ratio of accumulated trust fund balances to annual payroll. A higher reserve ratio provides more protection in an economic downturn.

Unfortunately, the UI reserve fund in several states – most notably, New York and Texas – were quite low even before September 11th. Phillip B. Levine, an economist at Wellesley College, calculates that to remain solvent through a severe recession, like the one experienced in the early 1980's, unemployment insurance funds would require a reserve ratio of at least 1.25 percent. Using this standard, 16 states were at risk of insolvency in a severe recession based on their reserve funds as of the first quarter of 2001. In New York the reserve ratio was 0.28 percent and in Texas it was 0.22 percent.

This predicament arose because many states did not build up their funds during the 1990s, and because experience rating – that is, the extent to which a business’s payments increase with its past record of laying off workers – is poorly implemented. If the funds become insolvent, they will borrow from the federal government at close to market rates, and probably tighten eligibility standards to stem the shortfall. I would recommend considering that the states be required to implement real experience rating and maintain ample fund balances within three years (i.e., after the economy improves sufficiently). This would shore up the long-run financing of the state programs. In addition, a study by Phillip B. Levine and David Card of U.C. Berkeley estimates that the unemployment rate would decline by six-tenths of a percentage point if industries were fully experience rated – that is, if employers in an industry were required to pay the full additional costs of unemployment benefits for layoffs in that industry.

The federal government sets minimum standards for state unemployment insurance programs and has a history of encouraging experience rating. This is a unique aspect of the American system of UI, and may in part help to account for the relatively low unemployment in the U.S. compared to other economically advanced countries. Better experience rating could be accomplished by increasing the 5.4 percent maximum tax rate on high-layoff employers, and by requiring the states to have at least 10 different rates. Some states currently only have two rates: 0 or 5.4 percent. In addition, I would recommend that the per employee taxable earnings cap – which range from $7,000 to $10,000 in most states – be raised, which would allow better experience rating at lower tax rates and make the financing of the program less regressive. Raising the caps and lowering the rates would also increase demand for less skilled workers. Improved experience rating would discourage employers from laying off workers, and help to internalize the externalities layoffs impose on society.

Third, unemployed workers who are otherwise eligible for UI but are searching for a part-time job (e.g., because of family obligations) are ineligible for benefits in most states. These workers pay into the system, but they are prevented from receiving benefits. States could be required to expand eligibility. Workers who would be made eligible for UI benefits as a result of this reform would be primarily single-parent, female, and low-income workers.

I realize that Congress is likely to be reluctant to make additional changes to UI having just voted to expand benefit payments, but perhaps a commission could be established to study longer term issues in UI, including the automatic triggers, financing, and eligibility requirements of the state programs.

III. CONCLUSION

Since the summer of 2001 the downturn has looked more like a typical downturn, with the labor market softening more for the less skilled and minorities than for highly educated, white workers. Recessions typically last longer and are more severe for the less skilled and minorities. Job growth early in a recovery is typically weaker for these groups as well. My guess is that the typical pattern will continue in the near future, but I have to confess a great deal of uncertainty as the recession initially was unusual in terms of the breadth of groups of workers affected. Additionally, the temporary help sector is much larger than it was in the early 1990s, and it is possible that employment adjustment over the business cycles will be quicker because the option of hiring from temporary help firms enhances labor market flexibility. If this is the case, employment growth on the upswing may not lag economic growth as much as it has in the past. Nevertheless, it is probably more likely than not that higher unemployment will linger for less skilled and minority workers in the beginning of the recovery -- and such a process has already begun if the recession ended in the fourth quarter of 2001.

To have a balanced recovery I would argue that policy has to be balanced as well. Fiscal and monetary policies are in place to lower the costs of capital and stimulate growth. Because of the phenomenon of capital-skill complementarity, this will likely increase demand for employment of skilled workers and reduce demand for employment of less skilled workers in the future. Policies are also in place (e.g., extended UI benefits) to maintain consumption. To the extent that further policy initiatives are sought to stimulate job growth, I would expect that policies geared to stimulate demand for hiring less skilled workers would be most effective. Such policies could include job training, the Targeted Jobs Tax Credit, and a temporary reduction of payroll taxes. But I think it is also important to recognize that there will be pressures on the federal budget as the baby boom cohort retires because of Medicare and Social Security commitments, so policies to address short-run cyclical adjustments should be careful not to weaken the long-run budget outlook.



Table 2: Real Hourly Wages by Decile of the Wage Distribution (2000 Dollars)

Year

1

2

3

4

5

6

7

8

9

1973

$6.03

$7.28

$8.65

$10.06

$11.53

$13.23

$15.36

$17.57

$22.07

1974

5.96

7.15

8.49

9.83

11.26

12.96

15.02

17.31

21.84

1975

5.80

7.09

8.38

9.72

11.27

13.13

14.86

17.31

21.86

1976

6.25

7.26

8.44

9.76

11.34

13.20

15.13

17.59

22.11

1977

6.18

7.16

8.36

9.72

11.40

13.09

15.19

17.95

22.02

1978

6.27

7.40

8.62

10.05

11.77

13.50

15.59

18.42

23.09

1979

6.55

7.47

8.77

10.32

11.68

13.53

16.00

18.65

22.90

1980

6.19

7.32

8.59

9.97

11.55

13.43

15.59

18.47

22.62

1981

6.29

7.17

8.54

9.76

11.26

13.31

15.52

18.30

22.66

1982

6.06

7.06

8.44

9.82

11.44

13.35

15.72

18.52

22.93

1983

5.87

6.93

8.27

9.69

11.38

13.27

15.89

18.41

23.29

1984

5.74

6.97

8.22

9.67

11.45

13.31

15.78

18.78

23.64

1985

5.65

7.04

8.28

9.70

11.56

13.47

15.69

18.92

23.44

1986

5.64

7.21

8.52

10.05

11.78

13.83

16.19

19.13

24.17

1987

5.62

7.15

8.55

10.12

11.78

14.01

16.13

19.17

24.62

1988

5.63

7.10

8.50

10.12

11.71

13.92

16.22

19.35

24.81

1989

5.62

6.99

8.38

10.02

11.64

13.60

16.18

19.35

24.44

1990

5.70

7.06

8.47

9.98

11.63

13.48

15.99

19.24

24.55

1991

5.81

7.12

8.49

9.93

11.69

13.52

15.92

19.11

24.69

1992

5.79

7.05

8.43

9.85

11.78

13.45

15.93

19.24

24.33

1993

5.76

7.03

8.42

9.89

11.68

13.67

16.09

19.41

24.70

1994

5.70

6.96

8.27

9.70

11.47

13.51

16.05

19.54

25.12

1995

5.68

6.95

8.30

9.77

11.37

13.46

16.04

19.42

25.09

1996

5.65

7.00

8.38

9.78

11.31

13.36

16.09

19.45

25.16

1997

5.84

7.21

8.50

9.90

11.58

13.58

16.14

19.66

25.57

1998

6.16

7.44

8.71

10.36

11.91

14.00

16.65

20.07

26.31

1999

6.25

7.60

9.01

10.44

12.27

14.40

17.01

20.60

26.92

2000

6.31

7.77

9.08

10.51

12.26

14.51

17.19

20.91

27.50

2001

6.51

7.85

9.36

10.73

12.52

14.65

17.46

21.14

28.25

Source: Economic Policy Institute analysis of CPS data.
Note: Hourly wages were deflated by the BLS Research Series CPI deflator.



Home | Menu | Links | Info | Chairman's Page