Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania

Page created by Bernard Armstrong
 
CONTINUE READING
Minimum Wages and Employment:
                      A Case Study of the Fast-Food Industry
                         in New Jersey and Pennsylvania

         On April 1, 1992, New Jersey's minimum wage rose from $4.25 to $5.05 per
         hour. To evaluate the impact of the law we surveyed 410 fast-food restaurants in
         New Jersey and eastern Pennsylvania before and after the rise. Comparisons of
         employment growth at stores in New Jersey and Pennsylvania (where the
         minimum wage was constant) provide simple estimates of the effect of the higher
         minimum wage. We also compare employment changes at stores in New Jersey
         that were initially paying high wages (above $5) to the changes at lower-wage
         stores. We find no indication that the rise in the minimum wage reduced
         employment. (JEL 530, 523)

   How do employers in a low-wage labor                          cent studies that rely on a similar compara-
market respond to an increase in the mini-                       tive methodology have failed to detect a
mum wage? The prediction from conven-                            negative employment effect of higher mini-
tional economic theory is unambiguous: a                         mum wages. Analyses of the 1990-1991 in-
rise in the minimum wage leads perfectly                         creases in the federal minimum wage
competitive employers to cut employment                          (Lawrence F. Katz and Krueger, 1992; Card,
(George J. Stigler, 1946). Although studies                      1992a) and of an earlier increase in the
in the 1970's based on aggregate teenage                         minimum wage in California (Card, 1992b)
employment rates usually confirmed this                          find no adverse employment impact. A study
prediction,' earlier studies based on com-                       of minimum-wage floors in Britain (Stephen
parisons of employment at affected and un-                       Machin and Alan Manning, 1994) reaches a
affected establishments often did not (e.g.,                     similar conclusion.
Richard A. Lester, 1960, 1964). Several re-                         This paper presents new evidence on the
                                                                 effect of minimum wages on establishment-
                                                                 level employment outcomes. We analyze the
                                                                 experiences of 410 fast-food restaurants in
   *Department of Economics, Princeton University,               New Jersey and Pennsylvania following the
Princeton, NJ 08544. We are grateful to the Institute            increase in New Jersey's minimum wage
for Research on Poverty, University of Wisconsin, for            from $4.25 to $5.05 per hour. Comparisons
partial financial support. Thanks to Orley Ashenfelter,
Charles Brown, Richard Lester, Gary Solon, two
                                                                 of employment, wages, and prices at stores
anonymous referees, and seminar participants at                  in New Jersey and Pennsylvania before and
Princeton, Michigan State, Texas A&M, University of              after the rise offer a simple method for
Michigan, university of Pennsylvania, ~niversitJ of              evaluating the effects of the-minimum wage.
Chicago, and the NBER for comments and sugges-                   ~~~~~~i~~~~ within N~~ jerseybetween
tions. We also acknowledge the expert research assis-
tance of Susan Belden, Chris Burris, Geraldine Harris,                    high-wage                paying
and Jonathan Orszag.                                             than the new minimum rate prior to its
   'see Charles Brown et al. (1982,1983) for surveys of          effective date) and other stores provide an
this literature. A recent update (Allison J. Wellington,         alternative estimate of the impact of the
1991) concludes that the employment effects of the
minimum wage are negative but small: a 10-percent
                                                                 new lawe
increase in the minimum is estimated to lower teenage               In addition to the simplicity of our empir-
employment rates by 0.06 percentage points.                      ical methodology, several other features of
                                                           772
VOL. 84 NO. 4      CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT                                773

the New Jersey law and our data set are          and 1991 to measures of the relative mini-
also significant. First, the rise in the mini-   mum wage in each state.
mum wage occurred during a recession. The
increase had been legislated two years ear-                   I. The New Jersey Law
lier when the state economy was relatively
healthy. By the time of the actual increase,       A bill signed into law in November 1989
the unemployment rate in New Jersey had          raised the federal minimum wage from $3.35
risen substantially and last-minute political    per hour to $3.80 effective April 1, 1990,
action almost succeeded in reducing the          with a further increase to $4.25 per hour on
minimum-wage increase. It is unlikely that       April 1, 1991. In early 1990 the New Jersey
the effects of the higher minimum wage           legislature went one step further, enacting
were obscured by a rising tide of general        parallel increases in the state minimum wage
economic conditions.                             for 1990 and 1991 and an increase to $5.05
    Second, New Jersey is a relatively small     per hour effective April 1, 1992. The sched-
state with an economy that is closely linked     uled 1992 increase gave New Jersey the
to nearby states. We believe that a control      highest state minimum wage in the country
group of fast-food stores in eastern Pennsyl-    and was strongly opposed by business lead-
vania forms a natural basis for comparison       ers in the state (see Bureau of National
with the experiences of restaurants in New       Affairs, Daily Labor Report, 5 May 1990).
Jersey. Wage variation across stores in New         In the two years between passage of the
Jersey, however, allows us to compare the        $5.05 minimum wage and its effective date,
experiences of high-wage and low-wage            New Jersey's economy slipped into reces-
stores within New Jersey and to test the         sion. Concerned with the potentially ad-
validity of the Pennsylvania control group.      verse impact of a higher minimum wage, the
Moreover, since seasonal patterns of em-         state legislature voted in March 1992 to
 ployment are similar in New Jersey and          phase in the 80-cent increase over two years.
 eastern Pennsylvania, as well as across         The vote fell just short of the margin re-
 high- and low-wage stores within New Jer-       quired to override a gubernatorial veto, and
 sey, our comparative methodology effec-         the Governor allowed the $5.05 rate to go
 tively "differences out" any. seasonal em-      into effect on April 1 before vetoing the
 ployment effects.                               two-step legislation. Faced with the prospect
    Third, we successfully followed nearly 100   of having to roll back wages for minimum-
percent of stores from a first wave of inter-    wage earners, the legislature dropped the
views conducted just before the rise in the      issue. Despite a strong last-minute chal-
 minimum wage (in February and March             lenge, the $5.05 minimum rate took effect
 1992) to a second wave conducted 7-8            as originally planned.
 months after (in November and December
 1992). We have complete information on                 11. Sample Design and Evaluation
 store closings and take account of employ-
 ment changes at the closed stores in our          Early in 1992 we decided to evaluate the
 analyses. We therefore measure the overall      impending increase in the New Jersey mini-
 effect of the minimum wage on average           mum wage by surveying fast-food restau-
 employment, and not simply its effect on        rants in New Jersey and eastern Pennsylva-
 surviving establishments.                       niae2 Our choice of the fast-food industry
   -Our analysis of employment trends at         was driven by several factors. First, fast-food
 stores that were open for business before       stores are a leading employer of low-wage
 the increase in the minimum wage ignores        workers: in 1987, franchised restaurants em-
 any potential effect of minimum wages on
 the rate of new store openings. To assess
 the likely magnitude of this effect we relate
 state-specific growth rates in the number of       2At the time we were uncertain whether the $5.05
 McDonald's fast-food outlets between 1986       rate would go into effect or be overridden.
THE AMERICAN ECONOMIC REVIEW                                   SEPTEMBER 1994

                                                                                       Stores in:
                                                                   A1l          NJ             PA
                  Waue I, February 15-March 4, 1992:

                    Number of stores in sample frame:a            473          364            109
                    Number of refusals:                            63           33             30
                    Number interviewed:                           410          331             79
                    Response rate (percentage):                    86.7         90.9           72.5

                  Wace 2, Nocember 5 - December 31, 1992:

                    Number of stores in sample frame:             410          331             79
                    Number closed:                                  6            5              1
                    Number under rennovation:                       2            2              0
                    Number temporarily closed:'                     2            2              0
                    Number of refusals:                             1            1              0
                    Number i n t e r v i e ~ e d : ~              399          321             78
                  aStores with working phone numbers only; 29 stores in original sample frame had
              disconnected phone numbers.
                 '~ncludes one store closed because of highway construction and one store closed
              because of a fire.
                 'Includes 371 phone interviews and 28 personal interviews of stores that refused an
              initial request for a phone interview.

ployed 25 percent of all workers in the                     rants in New Jersey and eastern Pennsylva-
restaurant industry (see U.S. Department of                 nia from the Burger King, KFC, Wendy's,
Commerce, 1990 table 13). Second, fast-food                 and Roy Rogers chain^.^ The first wave of
restaurants comply with minimum-wage reg-                   the survey was conducted by telephone in
ulations and would be expected to raise                     late February and early March 1992, a little
wages in response to a rise in the minimum                  over a month before the scheduled increase
wage. Third, the job requirements and                       in New Jersey's minimum wage. The survey
products of fast-food restaurants are rela-                 included questions on employment, starting
tively homogeneous, making it easier to ob-                 wages, prices, and other store characteris-
tain reliable measures of employment,                       tic~.~
wages, and product prices. The absence of                      Table 1 shows that 473 stores in our sam-
tips greatly simplifies the measurement of                  ple frame had working telephone numbers
wages in the industry. Fourth, it is relatively             when we tried to reach them in February-
easy to construct a sample frame of fran-                   March 1992. Restaurants were called as
chised restaurants. Finally, past experience                many as nine times to elicit a response. We
(Katz and Krueger, 1992) suggested that                     obtained completed interviews (with some
fast-food restaurants have high response                    item nonresponse) from 410 of the restau-
rates to telephone survey^.^                                rants, for an overall response rate of 87
   Based on these considerations we con-                    percent. The response rate was higher in
structed a sample frame of fast-food restau-                New Jersey (91 percent) than in Pennsylva-

                                                               4 ~ h seample was derived from white-pages tele-
   3 ~ an pilot survey Katz and Krueger (1992) obtained     phone listings for New Jersey and Pennsylvania as of
very low response rates from McDonald's restaurants.        February 1992.
For this reason, McDonald's restaurants were excluded          'copies of the questionnaires used in both waves of
from Katz and Krueger's and our sample frames.              the survey are available from the authors upon request.
VOL. 84 NO. 4             C A m AND KRUEGER: MINIiiMUM WAGE AND EMPLOYMENT                                             775

nia (72.5 percent) because our interviewer                      nently closed stores but is treated as missing
made fewer call-backs to nonrespondents in                      for the temporarily closed stores. (Full-
Penn~ylvania.~  In the analysis below we in-                    time-equivalent [FTE] employment was cal-
vestigate possible biases associated with the                   culated as the number of full-time workers
degree of difficulty in obtaining the first-                    [including managers] plus 0.5 times the
wave interview.                                                 number of part-time workers.)' Means are
   The second wave of the survey was con-                       presented separately for stores in New Jer-
ducted in November and December 1992,                           sey and Pennsylvania, along with t statistics
about eight months after the minimum-wage                       for the null hypothesis that the means are
increase. Only the 410 stores that re-                          equal in the two states.
sponded in the first wave were contacted in                        Rows la-e show the distribution of stores
the second round of interviews. We success-                     by chain and ownership status (company-
fully interviewed 371 (90 percent) of these                     owned versus franchisee-owned). The
stores by phone in November 1992. Because                       Burger King, Roy Rogers, and Wendy's
of a concern that nonresponding restaurants                     stores in our sample have similar average
might have closed, we hired an interviewer                      food prices, store hours, and employment
to drive to each of the 39 nonrespondents                       levels. The KFC stores are smaller and are
and determine whether the store was still                       open for fewer hours. They also offer a
open, and to conduct a personal interview if                    more expensive main course than stores in
possible. The interviewer discovered that six                   the other chains (chicken vs, hamburgers).
restaurants were permanently closed, two                           In wave 1, average employment was 23.3
were temporarily closed (one because of a                       full-time equivalent workers per store in
fire, one because of road construction), and                    Pennsylvania, compared with an average of
two were under renovation.' Of the 29 stores                    20.4 in New Jersey. Starting wages were
open for business, all but one granted a                        very similar among stores in the two states,
request for a personal interview. As a re-                      although the average price of a "full meal"
sult, we have second-wave interview data                        (medium soda, small fries, and an entree)
for 99.8 percent of the restaurants that re-                    was significantly higher in New Jersey. There
sponded in the first wave of the survey, and                    were no significant cross-state differences in
information on closure status for 100 per-                      average hours of operation, the fraction of
cent of the sample.                                             full-time workers, or the prevalence of bonus
   Table 2 presents the means for several                       programs to recruit new worker^.^
key variables in our data set, averaged over                       The average starting wage at fast-food
the subset of nonmissing responses for each                     restaurants in New Jersey increased by 10
variable. In constructing the means, employ-                    percent following the rise in the minimum
ment in wave 2 is set to 0 for the perma-                       wage. Further insight into this change is
                                                                provided in Figure 1, which shows the dis-
                                                                tributions of starting wages in the two states
                                                                before and after the rise. In wave 1, the
                                                                distributions in New Jersey and Pennsylva-
   6 ~ e s p o n s erates per call-back were almost identical   nia were very similar. By wave 2 virtually all
in the two states. Among New Jersey stores, 44.5
percent responded on the first call, and 72.0 percent
responded after at most two call-backs. Among Penn-
sylvania stores 42.2 percent responded on the first call,          ' w e discuss the sensitivity of our results to alterna-
and 71.6 percent responded after at most two call-              tive assumptions on the measurement of employment
backs.                                                          in Section 111-C.
   7 ~ ofs April 1993 the store closed because of road             ' ~ h e s e programs offer current employees a cash
construction and one of the stores closed for renova-           "bounty" for recruiting any new employee who stays
tion had reopened. The store closed by fire was open            on the job for a minimum period of time. Typical
when our telephone interviewer called in November               bounties are $50-$75. Recruiting programs that award
1992 but refused the interview. By the time of the              the recruiter with an "employee of the month" desig-
follow-up personal interview a mall fire had closed the         nation or other noncash bonuses are excluded from our
store.                                                          tabulations.
THE AMERICAN ECONOMIC REVIEW                         SEPTEMBER 1994

                                                                    Stores in:
           Variable                                         NJ                    PA       ta

           1. Distribution of Store Types (percentages):

             a.   Burger King
             b.   KFC
             c.   Roy Rogers
             d.   Wendy's
             e.   Company-owned

           2. Means in Wave I:

             a. FTE employment                             20.4
                                                           (0.51)
             b. Percentage full-time employees             32.8
                                                           (1.3)
             c. Starting wage                               4.61
                                                           (0.02)
             d. Wage = $4.25 (percentage)                  30.5
                                                           (2.5)
             e. Price of full meal

             f. Hours open (weekday)

             g. Recruiting bonus

           3. Means in Ware 2:

             a. FTE employment                             21.0                  21.2     - 0.2
                                                           (0.52)                (0.94)
             b. Percentage full-time employees             35.9                  30.4       1.8
                                                           (1.4)                 (2.8)
             c. Starting wage                               5.08                  4.62    10.8
                                                           (0.01)                (0.04)
             d. Wage = $4.25 (percentage)                   0.0                  25.3      -
                                                                                 (4.9)
             e. Wage = $5.05 (percentage)                  85.2                   1.3     36.1
                                                           (2.0)                 (1.3)
             f. Price of full meal                          3.41                  3.03     5.0
                                                           (0.04)                (0.07)
             g. Hours open (weekday)                       14.4                  14.7     - 0.8
                                                           (0.2)                 (0.3)
             h. Recruiting bonus                           20.3                  23.4     - 0.6
                                                           (2.3)                 (4.9)
           Notes: See text for definitions. Standard errors are given in parentheses.
             aTest of equality of means in New Jersey and Pennsylvania.

restaurants in New Jersey that had been                      Despite the increase in wages, full-time-
paying less than $5.05 per hour reported a                 equivalent employment increased in New
starting wage equal to the new rate. Inter-                Jersey relative to Pennsylvania. Whereas
estingly, the minimum-wage increase had no                 New Jersey stores were initially smaller,
apparent "spillover" on higher-wage restau-                employment gains in New Jersey coupled
rants in the state: the mean percentage wage               with losses in Pennsylvania led to a small
change for these stores was - 3.1 percent.                 and statistically insignificant interstate
VOL. 84 NO. 4   CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT

                                February 1 9 9 2

                                    Wage Range

                               November 1 9 9 2

                                    Wage Range

                              New Jersey           Pennsylvania
                    FIGURE
                         1. DISTRIBUTION
                                     OF STARTING
                                               WAGERATES
778                                    THE AMERICAN ECONOMIC REVIEW                        SEPTEMBER I994

difference in wave 2. Only two other vari-                    our survey. We present data by state in
ables show a relative change between waves                    columns (i) and (ii), and for stores in New
1 and 2: the fraction of full-time employees                  Jersey classified by whether the starting
and the price of a meal. Both variables                       wage in wave 1 was exactly $4.25 per hour
increased in New Jersey relative to Pennsyl-                  [column (iv)] between $4.26 and $4.99 per
vania.                                                        hour [column (v)] or $5.00 or more per hour
   We can assess the reliability of our survey                [column (vi)]. We also show the differences
questionnaire by comparing the responses                      in average employment between New Jersey
of 11 stores that were inadvertently inter-                   and Pennsylvania stores [column (iii)] and
viewed twice in the first wave of the survey.10               between stores in the various wage ranges
Assuming that measurement errors in the                       in New Jersey [columns (viil-(viii)].
two interviews are independent of each                           Row 3 of the table presents the changes
other and independent of the true variable,                   in average employment between waves 1
the correlation between responses gives an                    and 2. These entries are simply the differ-
estimate of the "reliability ratio" (the ratio                ences between the averages for the two
of the variance of the signal to the com-                     waves (i.e., row 2 minus row 1). An alterna-
bined variance of the signal and noise). The                  tive estimate of the change is presented in
estimated reliability ratios are fairly high,                 row 4: here we have computed the change
ranging from 0.70 for full-time equivalent                    in employment over the subsample of stores
employment to 0.98 for the price of a meal."                  that reported valid employment data in both
   We have also checked whether stores with                   waves. We refer to this group of stores as
missing data for any key variables are dif-                   the balanced subsample. Finally, row 5 pre-
ferent from restaurants with complete re-                     sents the average change in employment in
sponses. We find that stores with missing                     the balanced subsample, treating wave-2
data on employment, wages, or prices are                      employment at the four temporarily closed
similar in other respects to stores with com-                 stores as zero, rather than as missing.
plete data. There is a significant size differ-                  As noted in Table 2, New Jersey stores
ential associated with the likelihood of the                  were initially smaller than their Pennsylva-
store closing after wave 1. The six stores                    nia counterparts but grew relative to Penn-
that closed were smaller than other stores                    sylvania stores after the rise in the mini-
(with an average employment of only 12.4                      mum wage. The relative gain (the "dif-
full-time-equivalent employees in wave 1).12                  ference in differences" of the changes in
                                                              employment) is 2.76 FTE employees (or 13
           111. Employment Effects of the                     percent), with a t statistic of 2.03. Inspec-
              Minimum-Wage Increase                           tion of the averages in rows 4 and 5 shows
                                                              that the relative change between New Jer-
          A. Differences in Differences                       sey and Pennsylvania stores is virtually iden-
                                                              tical when the analysis is restricted to the
  Table 3 summarizes the levels and                           balanced subsample, and it is only slightly
changes in average employment per store in                    smaller when wave-2 employment at the
                                                              temporarily closed stores is treated as zero.
                                                                 Within New Jersey, employment ex-
   10
      These restaurants were interviewed twice because        panded at the low-wage stores (those paying
their phone numbers appeared in more than one phone           $4.25 per hour in wave 1) and contracted at
book, and neither the interviewer nor the respondent
noticed that they were previously interviewed.                the high-wage stores (those paying $5.00 or
   11
      Similar reliability ratios for very similar questions   more per hour). Indeed, the average change
were obtained by Katz and Krueger (1992).                     in employment at the high-wage stores
   ''A probit analysis of the probability of closure          ( - 2.16 FTE employees) is almost identical
shows that the initial size of the store is a significant
predictor of closure. The level of starting wages has a
                                                              to the change among Pennsylvania stores
numerically small and statistically insignificant coeffi-     ( - 2.28 FTE employees). Since high-wage
cient in the probit model.                                    stores in New Jersey should have been
V O L . 84 NO. 4         CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT

largely unaffected by the new minimum                        els of the form:
wage, this comparison provides a specifica-
tion test of the validity of the Pennsylvania                (la)        AE,=a+bXi+cNJi+~,
control group. The test is clearly passed.
Regardless of whether the affected stores
are compared to stores in Pennsylvania or
high-wage stores in New Jersey, the esti-
                                                                                         +
                                                             ( l b ) AE, = a' + blXi clGAPi + E{
mated employment effect of the minimum                       where AE, is the change in employment
wage is similar.                                             from wave 1 to wave 2 at store i, Xi is a set
   The results in Table 3 suggest that em-                   of characteristics of store i, and NJ, is a
ployment contracted between February and                     dummy variable that equals 1 for stores in
November of 1992 at fast-food stores that                    New Jersey. GAP, is an alternative measure
were unaffected by the rise in the minimum                   of the impact of the minimum wage at store
wage (stores in Pennsylvania and stores in                   i based on the initial wage at that store
New Jersey paying $5.00 per hour or more                     (W,,):
in wave 1). We suspect that the reason for
this contraction was the continued worsen-                   GAP, = 0        for stores in Pennsylvania
ing of the economies of the middle-Atlantic
states during 1992.13 Unemployment rates                            =0       for stores in New Jersey with
in New Jersey, Pennsylvania, and New York
all trended upward between 1991 and 1993,
with a larger increase in New Jersey than
Pennsylvania during 1992. Since sales of
franchised fast-food restaurants are pro-                                    for other stores in New Jersey.
cyclical, the rise in unemployment would be
expected to lower fast-food employment in                    GAP, is the proportional increase in wages
the absence of other factors.14                              at store i necessary to meet the new mini-
                                                             mum rate. Variation in GAP, reflects both
         B. Regression-Adjusted Models                       the New Jersey-Pennsylvania contrast and
                                                             differences within New Jersey based on re-
   The comparisons in Table 3 make no                        ported starting wages in wave 1. Indeed, the
allowance for other sources of variation in                  value of GAP, is a strong predictor of the
employment growth, such as differences                       actual proportional wage change between
across chains. These are incorporated in the                 waves 1 and 2 (R* = 0.75), and conditional
estimates in Table 4. The entries in this                    on GAP, there is no difference in wage
table are regression coefficients from mod-                  behavior between stores in New Jersey and
                                                             Pennsylvania. l5
                                                                The estimate in column (i) of Table 4
                                                             is directly comparable to the simple
    13
      An alternative possibility is that seasonal factors
                                                             difference-in-differences of employment
produce higher employment at fast-food restaurants in        changes in column (iv), row 4 of Table 3.
February and March than in November and December.            T h e discrepancy between the two
An analysis of national employment data for food             estimates is due to the restricted sample in
preparation and service workers, however, shows higher       Table 4. In Table 4 and the remaining ta-
average employment in the fourth quarter than in the
first quarter.                                               bles in this section we restrict our analysis
    14
       To investigate the cyclicality of fast-food restau-   to the set of stores with available employ-
rant sales we regressed the year-to-year change in U.S.      ment and wage data in both waves of the
sales of the McDonald's restaurant chain from
1976-1991 on the corresponding change in the unem-
ployment rate. The regression results show that a
1-percentage-point increase in the unemployment rate            1 5 regression
                                                                     ~         of the proportional wage change be-
reduces sales by $257 million, with a t statistic of 3.0.    tween waves 1 and 2 on GAP, has a coefficient of 1.03.
THE AMERICAN ECONOMIC REVlEW                                      SEPTEMBER 1994

                     TABLE3-AVERAGE EMPLOYMENT  PER STOREBEFOREAND                       I   ~   E   THE
                                                                                                      R RISE
                                     IN NEW JERSEYMINIMUM WAGE

                                         Stores by state                Stores in New Jersey a      Differences within N J ~
                                                   Difference,   Wage =        Wage =      Wage r Low-        Midrange-
                                  PA       NJ        NJ-PA        $4.25      $4.26-$4.99      $5.00   high       high
 Variable                          (i)     (ii)        (iii)       (iv)           (v)          (vi)   (vii)      (viii)
 1. FTE employment before,
    all available observations
 2. FTE employment after,
    all available observations
 3. Change in mean FTE
    employment
 4. Change in mean FTE
    employment, balanced
    sample of storesC
 5. Change in mean FTE
    employment, setting
    FTE at temporarily
    closed stores to Od
Notes: Standard errors are shown in parentheses. The sample consists of all stores with available data on employment. FTE
(full-time-equivalent) employment counts each part-time worker as half a full-time worker. Employment at six closed stores
is set to zero. Employment at four temporarily closed stores is treated as missing.
   astares in New Jersey were classified by whether starting wage in wave 1 equals $4.25 per hour ( N = 101), is between
$4.26 and $4.99 per hour ( N = 140), or is $5.00 per hour or higher ( N = 73).
   b ~ i f f e r e n c ein employment between low-wage ($4.25 per hour) and high-wage ( 2$5.00 per hour) stores; and difference
in employment between midrange ($4.26-$4.99 per hour) and high-wage stores.
   'Subset of stores with available employment data in wave 1 and wave 2.
         this row only, wave-2 employment at four temporarily closed stores is set to 0. Employment changes are based on the
subset of stores with available employment data in wave 1 and wave 2.

                           TABLE4-REDUCED-FORM MODELSFOR CHANGEIN EMPLOYMENT

                                                                                Model
                Independent variable                      (i)        (ii)        (iii)            (iv)     (v)
                1. New Jersey dummy                      2.33        2.30         -               -        -
                                                        (1.19)      (1.20)
                2. Initial wage gapa                      -           -         15.65            14.92    11.91
                                                                                (6.08)           (6.21)   (7.39)
                3. Controls for chain and                 no         yes          no              yes      yes
                   ownershipb
                4. Controls for regionC
                5. Standard error of regression
                6. Probability value for controlsd

                Notes: Standard errors a r e given in parentheses. T h e sample consists of 357 stores
                with available data o n employment and starting wages in waves 1 and 2. T h e
                dependent variable in all models is change in F T E employment. T h e mean and
                standard deviation of the dependent variable are -0.237 and 8.825, respectively. All
                models include a n unrestricted constant (not reported).
                   aProportional increase in starting wage necessary to raise starting wage t o new
                minimum rate. For stores in Pennsylvania the wage gap is 0.
                   b ~ h r e de ummy variables for chain type and whether or not the store is company-
                owned are included.
                   'Dummy variables for two regions of New Jersey and two regions of eastern
                Pennsylvania are included.
                   d ~ r o b a b i l i t yv alue of joint F test for exclusion of all control variables.
VOL. 84 NO. 4       CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT                                     781

survey. This restriction results in a slightly   null hypothesis of a zero employment effect
smaller estimate of the relative increase in     of the minimum wage. One explanation for
employment in New Jersey.                        this attenuation is the presence of measure-
   The model in column (ii) introduces a         ment error in the starting wage. Even if
set of four control variables: dummies for       employment growth has no regional compo-
three of the chains and another dummy for        nent, the addition of region dummies will
company-owned stores. As shown by the            lead to some attenuation of the estimated
probability values in row 6, these covariates    GAP coefficient if some of the true varia-
add little to the model and have no effect       tion in GAP is explained by region. Indeed,
on the size of the estimated New Jersey          calculations based on the estimated reliabil-
dummy.                                           ity of the GAP variable (from the set of 11
   The specifications in columns (iiil-(v) use   double interviews) suggest that the fall in
the GAP variable to measure the effect of        the estimated GAP coefficient from column
the minimum wage. This variable gives a          (iv) to column (v) is just equal to the ex-
slightly better fit than the simple New Jer-     pected change attributable to measurement
sey dummy, although its implications for the     error.16
New Jersey-Pennsylvania comparison are              We have also estimated the models in
similar. The mean value of GAPi among            Table 4 using as a dependent variable the
New Jersey stores is 0.11. Thus the estimate     proportional change in employment at each
in column (iii) implies a 1.72 increase in       store.17 The estimated coefficients of the
FTE employment in New Jersey relative to         New Jersey dummy and the GAP variable
Pennsylvania.                                    are uniformly positive in these models but
   Since GAP, varies within New Jersey, it is    insignificantly different from 0 at conven-
possible to add both GAP, and NJ, to the         tional levels. The implied employment ef-
employment model. The estimated coeffi-          fects of the minimum wage are also smaller
cient of the New Jersey dummy then pro-          when the dependent variable is expressed in
vides a test of the Pennsylvania control         proportional terms. For example, the GAP
group. When we estimate these models, the        coefficient in column (iii) of Table 4 implies
coefficient of the New Jersey dummy is in-       that the increase in minimum wages raised
significant (with t ratios of 0.3-0.7), imply-   employment at New Jersey stores that were
ing that inferences about the effect of the      initially paying $4.25 per hour by 14 per-
minimum wage are similar whether the             cent. The estimated GAP coefficient from a
comparison is made across states or across       corresponding proportional model implies
stores in New Jersey with higher and lower       an effect of only 7 percent. The difference is
initial wages.                                   attributable to heterogeneity in the effect of
   An even stronger test is provided in col-     the minimum wage at larger and smaller
umn (v), where we have added dummies             stores. Weighted versions of the propor-
 representing three regions of New Jersey        tional-change models (using initial employ-
(North, Central, and South) and two regions      ment as a weight) give rise to wage elastici-
 of eastern Pennsylvania (Allentown-Easton
 and the northern suburbs of Philadelphia).
 These dummies control for any region-
                                                     16
 s~ecificdemand shocks and identifv the ef-             In a regression model without other controls the
 feet of the minimum wage by                     expected attenuation of the GAP coefficient due to
                                                 measurement error is the reliability ratio of GAP (yo),
 employment changes at higher- and lower-        which we estimate at 0.70. The expected attenuation
wage          within the same region of New      factor when region dummies are added to the model is
Jersey. The probability value in row 6 shows     y l = (Yo- ~ 2 ) / ( 1 - ~ 2 ) where
                                                                                ,      ~2  is the R-square
 no evidence of regional components in em-       statistic of a regression of GAP on region effects (equal
 ployment growth. The addition of the re-        to 0.30). Thus, we expect the estimated GAP coeffi-
                                                 cient to fall by a factor of Y I / Y O = 0.8 when region
 gion dummies attenuates the GAP coeffi-         dummies are added to a regression model.
 cient and raises its standard error, however,       " ~ h e s e specifications are reported in table 4 of
 making it no longer possible to reject the      Card and Krueger (1993).
782                                   THE AMERICAN ECONOMIC REVIEW                             SEPTEMBER I994

ties similar to the elasticities implied by the             little effect on the models for the level of
estimates in Table 4 (see below).                           employment but yield slightly smaller point
                                                            estimates in the proportional-employment-
              C. Specification Tests                        change models.
                                                               In row 6 we present estimates obtained
   The results in Tables 3 and 4 seem to                    from a subsample that excludes 35 stores in
contradict the standard prediction that a                   towns along the New Jersey shore. The ex-
rise in the minimum wage will reduce em-                    clusion of these stores, which may have a
ployment. Table 5 presents some alternative                 different seasonal pattern than other stores
specifications that probe the robustness of                 in our sample, leads to slightly larger mini-
this conclusion. For completeness, we re-                   mum-wage effects. A similar finding emerges
port estimates of models for the change in                  in row 7 when we add a set of dummy
employment [columns (i) and (ii)] and esti-                 variables that indicate the week of the
mates of models for the proportional change                 wave-2 inter vie^.^'
in employment [columns (iii) and (iv)].18 The                  As noted earlier, we made an extra effort
first row of the table reproduces the "base                 to obtain responses from New Jersey stores
specification" from columns (ii) and (iv) of                in the first wave of our survey. The fraction
Table 4. (Note that these models include                    of stores called three or more times to ob-
chain dummies and a dummy for company-                      tain an interview was higher in New Jersey
owned stores). Row 2 presents an alterna-                   than in Pennsylvania. To check the sensitiv-
tive set of estimates when we set wave-2                    ity of our results to this sampling feature,
employment at the temporarily closed stores                 we reestimated our models on a subsample
to 0 (expanding our sample size by 4). This                 that excludes any stores that were called
change has a small attenuating effect on the                back more than twice. The results, in row 8,
coefficient of the New Jersey dummy (since                  are very similar to the base specification.
all four stores are in New Jersey) but less                    Row 9 presents weighted estimation re-
effect on the GAP coefficient (since the size               sults for the proportional-employment-
of GAP is uncorrelated with the probability                 change models, using as weights the initial
of a temporary closure within New Jersey).                  levels of employment in each store. Since
   Rows 3-5 present estimation results us-                  the proportional change in average employ-
ing alternative measures of full-time-equiv-                ment is an employment-weighted average of
alent employment. In row 3, employment is                   the proportional changes at each store, a
redefined to exclude management employ-                     weighted version of the proportional-change
ees. This change has no effect relative to                  model should give rise to elasticities that
the base specification. In rows 4 and 5, we                 are similar to the implied elasticities arising
include managers in FTE employment but                      from the levels models. Consistent with this
reweight part-time workers as either 40 per-                expectation, the weighted estimates are
cent or 60 percent of full-time workers (in-                larger than the unweighted estimates, and
stead of 50 percent).19 These changes have                  significantly different from 0 at conventional
                                                            levels. The weighted estimate of the New
                                                            Jersey dummy (0.13) implies a 13-percent
                                                            relative increase in New Jersey employment
   18
      The proportional change in employment is de-          -the same proportional employment effect
fined as the change in employment divided by the            implied by the simple difference-in-dif-
average level of employment in waves 1 and 2. This          ferences in Table 3. Similarly, the weighted
results in very similar coefficients but smaller standard
errors than the alternative of dividing by wave-1 em-
                                                            estimate of the GAP coefficient in the
ployment. For closed stores we set the proportional         proportional-change model (0.81) is close to
change in employment to - 1.
   19
     Analysis of the 1991 Current Population Survey
reveals that part-time workers in the restaurant indus-
try work about 46 percent as many hours as full-time
                                                               20
workers. Katz and Krueger (1992) report that the ratio            We also added dummies for the interview dates
of part-time workers' hours to full-time workers' hours     for the wave-1 survey, but these were insignificant and
in the fast-food industry is 0.57.                          did not change the estimated minimum-wage effects.
VOL. 84 NO. 4            CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT                                           783

                                                                                          Proportional change
                                                   Change in employment                     in employment
                                               NJ dummy          Gap measure         NJ dummy          Gap measure
Specification                                      (i)               (ii)               (iii)              (iv)
 1. Base specification                             2.30              14.92
                                                  (1.19)             (6.21)
 2. Treat four temporarily closed stores
    as permanently closeda
 3. Exclude managers in employment
    countb
 4. Weight part-time as 0.4 x full-timec

 5. Weight part-time as 0.6 X full-timed

 6. Exclude stores in NJ shore areae

 7. Add controls for wave-2 interview
    dateE
 8. Exclude stores called more than twice
    in wave lg
 9. Weight by initial employmenth

10. Stores in towns around Newark'                  -                33.75
                                                                    (16.75)
11. Stores in towns around CamdenJ                  -                10.91
                                                                    (14.09)
12. Pennsylvania stores only                        -               - 0.30
                                                                    (22.00)

Notes: Standard errors are given in parentheses. Entries represent estimated coefficient of New Jersey dummy
[columns (i) and (iii)] or initial wage gap [columns (ii) and (iv)] in regression models for the change in employment
or the percentage change in employment. All models also include chain dummies and an indicator for company-
owned stores.
   aWave-2 employment at four temporarily closed stores is set to 0 (rather than missing).
   b~ull-timeequivalent employment excludes managers and assistant managers.
   CFull-time equivalent employment equals number of managers, assistant managers, and full-time nonmanage-
ment workers, plus 0.4 times the number of part-time nonmanagement workers.
   d~ull-timeequivalent employment equals number of managers, assistant managers, and full-time nonmanage-
ment workers, plus 0.6 times the number of part-time nonmanagement workers.
   eSample excludes 35 stores located in towns along the New Jersey shore.
   ' ~ o d e l sinclude three dummy variables identifying week of wave-2 interview in November-December 1992.
   gSample excludes 70 stores (69 in New Jersey) that were contacted three or more times before obtaining the
wave-1 interview.
   h~egressionmodel is estimated by weighted least squares, using employment in wave 1 as a weight.
   i
  . Subsample of 51 stores in towns around Newark.
     Subsample of 54 stores in town around Camden.
     Subsample of Pennsylvania stores only. Wage gap is defined as percentage increase in starting wage necessary
to raise starting wage to $5.05.
784                                 THE AMERICAN ECONOMIC REVIEW                       SEPTEMBER 1994

the implied elasticity of employment with                    We have also investigated whether the
respect to wages from the basic levels speci-             first-differenced specification used in our
fication in row 1, column (iiI2l These find-              employment models is appropriate. A
ings suggest that the proportional effect of              first-differenced model implies that the level
the rise in the minimum wage was concen-                  of employment in period t is related to the
trated among larger stores.                               lagged level of employment with a coeffi-
   One explanation for our finding that a                 cient of 1. If short-run employment fluctua-
rise in the minimum wage has a positive                   tions are smoothed, however, the true co-
employment effect is that unobserved de-                  efficient of lagged employment may be less
mand shocks within New Jersey outweighed                  than 1. Imposing the assumption of a unit
the negative employment effect of the mini-               coefficient may then lead to biases. To test
mum wage. To address this possibility, rows               the first-differenced specification we reesti-
10 and 11 present estimation results based                mated models for the change in employ-
on subsamples of stores in two narrowly                   ment including wave-1 employment as an
defined areas: towns around Newark (row                   additional explanatory variable. To over-
10) and towns around Camden (row 11). In                  come any mechanical correlation between
each case the sample area is identified by                base-period employment and the change in
the first three digits of the store's zip code.22         employment (attributable to measurement
Within both areas the change in employ-                   error) we instrumented wave-1 employment
ment is positively correlated with the GAP                with the number of cash registers in the
variable, although in neither case is the                 store in wave 1 and the number of registers
effect statistically significant. To the extent           in the store that were open at 11:OO A.M. In
that fast-food product market conditions are              all of the specifications the coefficient of
constant within local areas, these results                wave-1 employment is close to zero. For
suggest that our findings are not driven by               example, in a specification including the
unobserved demand shocks. Our analysis of                 GAP variable and ownership and chain
price changes (reported below) also sup-                  dummies, the coefficient of wave-1 employ-
ports this conclusion.                                    ment is 0.04, with a standard error of 0.24.
   A final specification check is presented in            We conclude that the first-differenced spec-
row 12 of Table 5. In this row we exclude                 ification is appropriate.
stores in New Jersey and (incorrectly) de-
fine the GAP variable for Pennsylvania                      D. Full-Time and Part-Time Substitution
stores as the proportional increase in wages
necessary to raise the wage to $5.05 per                    Our analysis so far has concentrated on
hour. In principle the size of the wage gap               full-time-equivalent employment and ig-
for stores in Pennsylvania should have no                 nored possible changes in the distribution
systematic relation with employment growth.               of full- and part-time workers. An increase
In practice, this is the case. There is no                in the minimum wage could lead to an in-
indication that the wage gap is spuriously                crease in full-time employment relative to
related to employment growth.                             part-time employment for at least two rea-
                                                          sons. First, in a conventional model one
                                                          would expect a minimum-wage increase to
                                                          induce employers to substitute skilled work-
                                                          ers and capital for minimum-wage workers.
                                                          Full-time workers in fast-food restaurants
   21~ssumingaverage employment of 20.4 in New
Jersey, the 14.92 GAP coefficient in row 1, column (ii)   are typically older and may well possess
im lies an employment elasticity of 0.73.                 higher skills than part-time workers. Thus, a
   "The "070" three-digit zip-code area (around           conventional model predicts that stores may
Newark) and the "080" three-digit zip-code area           respond to an increase in the minimum
(around Camden) have by far the largest numbers of
stores among three-digit zip-code areas in New Jersey,
                                                          wage by increasing the proportion of full-
and together they account for 36 percent of New Jersey    time workers. Nevertheless, 81 percent of
stores in our sample.                                     restaurants paid full-time and part-time
VOL. 84 NO. 4            CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT                                         785

                                                                                    Regression of change in
                                                 Mean change in outcome              outcome variable on:
                                                  NJ        PA     NJ - PA   NJ dummy     Wage gapa     Wage gapb
Outcome measure                                   (i)       (ii)     (iii)      (iv)         (v)          (vi)
Store Characteristics:

 1. Fraction full-time workersc (percentage)

 2. Number of hours open per weekday

 3. Number of cash registers

 4. Number of cash registers open
    at 11:OO A.M.

Employee Meal Programs:

 5. Low-price meal program (percentage)

 6. Free meal program (percentage)

 7. Combination of low-price and free
    meals (percentage)

Wage Profile:

 8. Time to first raise (weeks)

 9. Usual amount of first raise (cents)

10. Slope of wage profile (percent
    per week)

Notes: Entries in columns (i) and (ii) represent mean changes in the outcome variable indicated by the row heading
for stores with available data on the outcome in waves 1 and 2. Entries in columns (iv)-(vi) represent estimated
regression coefficients of indicated variable (NJ dummy or initial wage gap) in models for the change in the
outcome variable. Regression models include chain dummies and an indicator for company-owned stores.
   aThe wage gap is the proportional increase in starting wage necessary to raise the wage to the new minimum
rate. For stores in Pennsylvania, the wage gap is zero.
   b ~ o d e lin
               s column (vi) include dummies for two regions of New Jersey and two regions of eastern Pennsylvania.
   'Fraction of part-time employees in total full-time-equivalent employment.

workers exactly the same starting wage in                     workers are more productive (but equally
wave 1 of our survey.23 This suggests either                  paid), there may be a second reason for
that full-time workers have the same skills                   stores to substitute full-time workers for
as part-time workers or that equity concerns                  part-time workers; namely, a minimum-wage
lead restaurants to pay equal wages for un-                   increase enables the industry to attract more
equally productive workers. If full-time                      full-time workers, and stores would natu-
                                                              rally want to hire a greater proportion of
                                                              full-time workers if they are more produc-
                                                              tive.
   231n the other 19 percent of stores, full-time workers        Row 1 of Table 6 presents the mean
are paid more, typically 10 percent more.                     changes in the proportion of full-time work-
786                                    THE AMERICAN ECONOMIC REVIEW                       SEPTEMBER 1994

ers in New Jersey and Pennsylvania be-                        ployment costs unchanged. The main fringe
tween waves 1 and 2 of our survey, and                        benefits for fast-food employees are free
coefficient estimates from regressions of the                 and reduced-price meals. In the first wave
change in the proportion of full-time work-                   of our survey about 19 percent of fast-food
ers on the wage-gap variable, chain dum-                      restaurants offered workers free meals. 72
mies, a company-ownership dummy, and re-                      percent offered reduced-price meals, a i d 9
gion dummies [in column ( 4 1 . The results                   percent offered a combination of both free
are ambiguous. The fraction of full-time                      and reduced-price meals. Low-price meals
workers increased in New Jersey relative to                   are an obvious fringe benefit to cut if the
Pennsylvania by 7.3 percent (t ratio = 1.841,                 minimum-wage increase forces restaurants
but regressions on the wage-gap variable                      to pay higher wages.
show no significant shift in the fraction of                     Rows 5 and 6 of Table 6 present esti-
full-time workers.24                                          mates of the effect of the minimum-wage
                                                              increase on the incidence of free meals and
  E. Other Employment-Related Measures                        reduced-price meals. The proportion of res-
                                                              taurants offering reduced-price meals fell
   Rows 2-4 of Table 6 present results for                    in both New Jersey and Pennsylvania after
other outcome variables that we expect to                     the minimum wage increased, with a some-
be related to the level of restaurant employ-                 what greater decline in New Jersey. Con-
ment. In particular, we examine whether                       trary to an offset story, however, the reduc-
the rise in the minimum wage is associated                    tion in reduced-price meal programs was
with a change in the number of hours a                        accompanied by an increase in the fraction
restaurant is open on a weekday, the num-                     of stores offering free meals. Relative to
ber of cash registers in the restaurant, and                  stores in Pennsylvania, New Jersey employ-
the number of cash registers typically in                     ers actually shifted toward more generous
operation in the restaurant at 11:OO A.M.                     fringe benefits (i.e., free meals rather than
Consistent with our employment results,                       reduced-price meals). However, the relative
none of these variables shows a statistically                 shift is not statistically significant.
significant decline in New Jersey relative to                    We continue to find a statistically in-
Pennsylvania. Similarly, regressions includ-                  significant effect of the minimum-wage in-
ing the gap variable provide no evidence                      crease on the likelihood of receiving free or
that the minimum-wage increase led to a                       reduced-price meals in columns (v) and (vi),
systematic change in any of these variables                   where we report coefficient estimates of the
[see columns (v) and (vi)].                                   GAP variable from regression models for
                                                              the change in the incidence of these pro-
                IV. Nonwage Offsets                           grams. The results provide no evidence that
                                                              employers offset the minimum-wage in-
   One explanation of our finding that a rise                 crease by reducing free or reduced-price
in the minimum wage does not lower em-                        meals.
ployment is that restaurants can offset the                      Another possibility is that employers re-
effect of the minimum wage by reducing                        sponded to the increase in the minimum
nonwage compensation. For example, if                         wage by reducing on-the-job training and
workers value fringe benefits and wages                       flattening the tenure-wage profile (see
equally, employers can simply reduce the                      Jacob Mincer and Linda Leighton, 1981).
level of fringe benefits by the amount of the                 Indeed, one manager told our interviewer in
minimum-wage increase, leaving their em-                      wave 1that her workers were forgoing ordi-
                                                              nary scheduled raises because the minimum
                                                              wage was about to rise, and this would
                                                              provide a raise for all her workers. To de-
   2 4 ~ i t h i nNew Jersey, the fraction of full-time em-
                                                              termine whether this phenomenon occurred
ployees increased about as quickly at stores with higher      more generally, we analyzed store man-
and lower wages in wave 1.                                     agers' responses to questions on the amount
VOL. 84 NO. 4            CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT                                           787

of time before a normal wage increase and                     factor cost. The average restaurant in New
the usual amount of such raises. In rows 8                    Jersey initially paid about half its workers
and 9 we report the average changes be-                       less than the new minimum wage. If wages
tween waves 1 and 2 for these two variables,                  rose by roughly 15 percent for these work-
as well as regression coefficients from mod-                  ers, and if labor's share of total costs is 30
els that include the wage-gap variable.25Al-                  percent, we would expect prices to rise by
though the average time to the first pay                      about 2.2 percent ( = 0.15 X 0.5 X 0.3) due to
raise increased by 2.5 weeks in New Jersey                    the minimum-wage rise.27
relative to Pennsylvania, the increase is not                    In each wave of our survey we asked
statistically significant. Furthermore, there                 managers for the prices of three standard
is only a trivial difference in the relative                  items: a medium soda, a small order of
change in the amount of the first pay incre-                  french fries, and a main course. The main
ment between New Jersey and Pennsylvania                      course was a basic hamburger at Burger
stores.                                                       King, Roy Rogers, and Wendy's restaurants,
   Finally, we examined a related variable:                   and two pieces of chicken at KFC stores.
the "slope" of the wage profile, which we                     We define "full meal" price as the after-tax
measure by the ratio of the typical first raise               price of a medium soda, a small order of
to the amount of time until the first raise is                french fries, and a main course.
given. As shown in row 10, the slope of the                      Table 7 presents reduced-form estimates
wage profile flattened in both New Jersey                     of the effect of the minimum-wage increase
and Pennsylvania, with no significant rela-                   on prices. The dependent variable in these
tive difference between states. The change                    models is the change in the logarithm of the
in the slope is also uncorrelated with the                    price of a full meal at each store. The key
GAP variable. In summary, we can find no                      independent variable is either a dummy in-
indication that New Jersey employers                          dicating whether the store is located in New
changed either their fringe benefits or their                 Jersey or the proportional wage increase
wage profiles to offset the rise in the mini-                 required to meet the minimum wage (the
mum wage.26                                                    GAP variable defined above).
                                                                 The estimated New Jersey dummy in col-
     V. Price Effects of the Minimum-Wage                     umn (i) shows that after-tax meal prices
                    Increase                                  rose 3.2-percent faster in New Jersey than
                                                              in Pennsylvania between February and
  A final issue we examine is the effect of                   November 1 9 9 2 . ~ T~ he effect is slightly
the minimum wage on the prices of meals at                    larger controlling for chain and company-
fast-food restaurants. A competitive model                    ownership [see column ($1. Since the
of the fast-food industry implies that an                     New Jersey sales tax rate fell by 1 percent-
increase in the minimum wage will lead to                      age point between the waves of our survey,
an increase in product prices. If we assume                   these estimates suggest that pretax prices
constant returns to scale in the industry, the                rose 4-percent faster as a result of the
increase in price should be proportional to
the share of minimum-wage labor in total

                                                                 " ~ c c o r d i n ~to the McDonald's Corporation 1991
                                                              Annual Report. payroll and benefits are 31.3 percent of
                                                              operating costs at company-owned stores. This calcula-
    2 5 ~ nwave 1, the average time to a first wage in-       tion is only approximate because minimum-wage work-
crease was 18.9 weeks, and the average amount of the          ers make up less than half of payroll even though they
first increase was $0.21 per hour.                            are about half of workers, and because a rise in the
    2 6 ~ a t aznd Krueger (1992) report that a significant   minimum wage causes some employers to increase the
fraction of fast-food stores in Texas responded to an         pay of other higher-wage workers in order to maintain
increase in the minimum wage by raising wages for             relative pay differentials.
workers who were initially earning more than the new               he effect is attributable to a 2.0-percent increase
minimum rate. Our results on the slope of the tenure          in prices in New Jersey and a 1.0-percent decrease in
profile are consistent with their findings.                   prices in Pennsylvania.
THE AMERICAN ECONOMIC REVIEW                                          SEPTEMBER 1994

                TABLE7-REDUCED-FORMMODELSFOR CHANGE
                                                  IN THE PRICEOF A FULLMEAL
            -   -   -   -   -   -   -   -   -   -   -   -   -   -   -         -

                                                                        Dependent variable: change in the log price
                                                                                      of a full meal
           Independent variable                                         (i)        (ii)      (iii)       (iv)       (v)
           1. New Jersey dummy                                       0.033         0.037      -           -         -
                                                                    (0.014)       (0.014)
           2. Initial wage gapa                                         -           -        0.077       0.146     0.063
                                                                                            (0.075)     (0.074)   (0.089)
           3. Controls for chain andb                                   no         yes        no         yes       Yes
                ownership
           4. Controls for regionC                                      no          no        no          no       yes
           5. Standard error of regression                          0.101         0.097     0.102       0.098     0.097

           Notes: Standard errors are given in parentheses. Entries are estimated regression
           coefficients for models fit to the change in the log price of a full meal (entrCe, medium
           soda, small fries). The sample contains 315 stores with valid data on prices, wages, and
           employment for waves 1 and 2. The mean and standard deviation of the dependent
           variable are 0.0173 and 0.1017, respectively.
              aProportional increase in starting wage necessary to raise the wage to the new
           minimum-wage rate. For stores in Pennsylvania the wage gap is 0.
              bThree dummy variables for chain type and whether or not the store is company-
           owned are included.
              'Dummy variables for two regions of New Jersey and two regions of eastern
           Pennsylvania are included.

minimum-wage increase in New Jersey-                                                  One potential explanation for the latter
slightly more than the increase needed to                                          finding is that stores in New Jersey compete
pass through the cost increase caused by the                                       in the same product market. As a result,
minimum-wage hike.                                                                 restaurants that are most affected by the
   The pattern of price changes within New                                         minimum wage are unable to increase their
Jersey is less consistent with a simple                                            product prices faster than their competitors.
" pass-through" view of minimum-wage cost                                          In contrast, stores in New Jersey and Penn-
increases. In fact, meal prices rose at                                            sylvania are in separate product markets,
approximately the same rate at stores in                                           enabling prices to rise in New Jersey rela-
New Jersey with differing levels of initial                                        tive to Pennsylvania when overall costs rise
wages. Inspection of the estimated GAP                                             in New Jersey. Note that this explanation
coefficients in column (v) of Table 7 con-                                         seems to rule out the possibility that store-
firms that within regions of New Jersey, the                                       specific demand shocks can account for the
GAP variable is statistically insignificant.                                       anomalous rise in employment at stores in
   In sum, these results provide mixed evi-                                        New Jersey with lower initial wages.
dence that higher minimum wages result in
higher fast-food prices. The strongest evi-                                                          VI. Store Openings
dence emerges from a comparison of New
Jersey and Pennsylvania stores. The magni-                                          An important potential effect of higher
tude of the price increase is consistent with                                     minimum wages is to discourage the open-
predictions from a conventional model of a                                        ing of new businesses. Although our sample
competitive industry. On the other hand, we                                       design allows us to estimate the effect of the
find no evidence that prices rose faster                                          minimum wage on existing restaurants in
among stores in New Jersey that were most                                         New Jersey, we cannot address the effect of
affected by the rise in the minimum wage.                                         the higher minimum wage on potential
VOL. 84 NO. 4          CARD AND KRUEGER: MINIMUM WAGE AND EMPLOYMENT                                         789

entrants.29 To assess the likely size of such            tive effect on either the net number of
an effect, we used national restaurant direc-            restaurants or the rate of new openings. To
tories for the McDonald's restaurant chain               the contrary, all the estimates show positice
to compare the numbers of operating                      effects of higher minimum wages on the
restaurants and the numbers of newly                     number of operating or newly opened stores,
opened restaurants in different states over              although many of the point estimates are
the 1986-1991 period. Many states raised                 insignificantly different from zero. While this
their minimum wages during this period. In               evidence is limited, we conclude that the
addition, the federal minimum wage in-                   effects of minimum wages on fast-food store
creased in the early 1990's from $3.35 to                opening rates are probably small.
$4.25, with differing effects in different states
depending on the level of wages in the                        VII. Broader Evidence on Employment
state. These policies create an opportunity                          Changes in New Jersey
to measure the impact of minimum-wage
laws on store opening rates across states.                  Our establishment-level analysis suggests
   The results of our analysis are presented             that the rise in the minimum wage in New
in Table 8. We regressed the growth rate in              Jersey may have increased employment in
the number of McDonald's stores in each                  the fast-food industry. Is this just an anomaly
state on two alternative measures of the                 associated with our particular sample, or a
minimum wage in the state and a set of                   phenomenon unique to the fast-food indus-
other control variables (population growth               try? Data from the monthly Current Popu-
and the change in the state unemployment                 lation Survey (CPS) allow us to compare
rate). The first minimum-wage measure is                 state-wide employment trends in New Jer-
the fraction of workers in the state's retail            sey and the surrounding states, providing a
trade industry in 1986 whose wages fell be-              check on the interpretation of our findings.
tween the existing federal minimum wage in               Using monthly CPS files for 1991 and 1992,
1986 ($3.35 per hour) and the effective min-             we computed employment-population rates
imum wage in the state in April 1990 (the                for teenagers and adults (age 25 and older)
maximum of the federal minimum wage and                  for New Jersey, Pennsylvania, New York,
the state minimum wages as of April 1990)."              and the entire United States. Since the New
The second is the ratio of the state's effec-            Jersey minimum wage rose on April 1, 1992,
tive minimum wage in 1990 to the average                 we computed the employment rates for
hourly wage of retail trade workers in the               April-December of both 1991 and 1992.
state in 1986. Both of these measures are                The relative changes in employment in New
designed to gauge the degree of upward                   Jersey and the surrounding states then give
wage pressure exerted by state or federal                an indication of the effect of the new law.
minimum-wage changes between 1986 and                       A comparison of changes in adult em-
1990.                                                    ployment rates show that the New Jersey
   The results provide no evidence that                  labor market fared slightly worse over the
higher minimum-wage rates (relative to the               1991-1992 period than either the U.S. labor
retail-trade wages in a state) exert a nega-             market as a whole or labor markets in
                                                         Pennsylvania or New York (see Card and
                                                         Krueger, 1993 table 9l3' Among teenagers,
   29
                                                         however, the situation was reversed. In New
      Direct inquiries to the chains in our sample re-   Jersey, teenage employment rates fell by 0.7
vealed that Wendy's opened two stores in New Jersey
in 1992 and one store in Pennsylvania. The other         percent from 1991 to 1992. In New York,
chains were unwilling to provide information on new
openings.
   30
      We used the 1986 Current Population Survey
                                                            31
(merged monthly file) to construct the minimum-wage            The employment rate of individuals age 25 and
variables. State minimum-wage rates in 1990 were ob-     older fell by 2.6 percent in New Jersey between 1991
tained from the Bureau of National Affairs Labor         and 1992, while it rose by 0.3 percent in Pennsylvania,
Relations Reporter Wages and Hours Manual (undated).     and fell by 0.2 percent in the United States as a whole.
You can also read