The Effect of Wal-Mart on Wages and Employment Effect of Wal-Mart_Cuellar.pdfof Wal Mart’s entry are likely to diminish as one is further from that Wal Mart. Alternatively, in a

The Effect of Wal-Mart on Wages and Employment in California

Steven S. Cuellar and

Andres Estrugo*

Abstract This paper analyzes Wal Mart’s affects on wages and employment in California. We use a data set consisting of cross sectional data taken from the Current Population Survey on individuals pooled over time from 1986 to 2004 to examine the affects of Wal Mart’s entry into California in 1991. We also introduce a new measure of Wal Mart’s affects that accounts for the distance of Wal Mart to the affected workers and allows for cross regional spill over effects. Our results indicate a negative and statistically significant effect on the wages of workers in regions located within 20 miles of a Wal Mart. We fail to find a consistent effect on employment resulting from Wal Mart’s entry. We use a fixed effects model and correct for endogeneity of Wal Mart’s decision to enter into a region by using instrumental variable regression. *Steven Cuellar: Associate Professor of Economics, Sonoma State University, Rohnert Park, CA 94928. Phone 707 664-2305, Fax 707-664-4009, E-mail [email protected]. Andres Estrugo, Research Assistant, Department of Economics, Sonoma State University, Rohnert Park, CA. Phone 707 664-2305, Fax 707-664-4009, E-mail [email protected]. We would like to thank the Sonoma State University School of Business and Economics for financial support as well as the participants at the Sonoma State University Department of Economics research seminar series for their helpful comments.

I. INTRODUCTION

Although there has been a considerable amount of discussion surrounding the

effect of Wal Mart on wages and employment, there has been very little academic

research. Some of the more notable studies have been Neumark, Zhang and Ciccarella

(2006) who examined county level employment and earnings by sector and found that

Wal Mart adversely effects both employment and wages of retail sector workers, the

sector believed most adversely affected by the entry of Wal Mart. Basker (2005)

examines employment by sector and county and finds that entry of Wal Mart results in a

net gain of jobs in the retail sector but a net loss of jobs in the wholesale sector. Basker

attributes the loss of jobs in the wholesale sector to Wal Marts stream lined supply chain

management. Dube, Eidlin and Lester (2007) find that Wal Mart reduces both wages and

health benefits among retail sector workers. In a study examining Wal Mart’s affect on

poverty rates, Goetz and Swaminathan (2006) find that county level poverty rates

increase as a result of Wal Mart entry. In an early study, Ketchum and Hughes (1997)

examine the effects of Wal Mart on employment and wages in Maine but find no

statistically significant results using a difference in difference in difference methodology.

This paper adds to the research on Wal Mart and is an improvement over previous

research in several areas. First, we use individual data not aggregate data. This is

especially important when examining the effect of Wal Mart on wages where we can

examine specific sub groups of workers believed most affected by the entry of Wal Mart.

Most of the previous studies mentioned above use aggregate industry level data by

county, and construct an average wages for everyone in an industry. We believe that the

affects of Wal Mart on wages and employment will be greatest not just among retail

Contemporary Economic Policy Wal Mart, Wages and Employment

2

workers but among low skilled retail workers. Using the March rotation of the Current

Population Survey (CPS) data on individual wages and employment pooled over time, we

are able to examine the employment and wages of those believed most effected by Wal

Mart. This has the potential to more accurately discern the effects of Wal Mart on the

labor market than previous studies. Additionally, the CPS data allows us to overcome a

major shortcoming of County Business Patterns data used for example by Neumark et al

(2006) and Basker (2005) which does not contain information on individual wages. As a

result, Neumark et al must construct a “wage” variable by examining payroll per person

in a county by sector while Basker ignores wages altogether. While Dube, Eidlin and

Lester (2007) do use a similar CPS data set, the ir examination of wages using CPS is

confined to state level effects.

Second, we believe our measure of Wal Mart’s entry into a region is an

improvement over previous measures. For example, Dube et al (2007), Neumark et al

(2006), Goetz and Swaminathan (2006), Basker (2005) and Ketchum and Hughes (1997)

all assume that the effects of Wal Mart on a region are uniform once Wal Mart enters

anywhere in that region. That is, their measure of Wal Mart’s entry is a simple binary

variable indicating the presence or absence of Wal Mart. Our measure, on the other hand,

explicitly accounts for the distance of the affected groups to the nearest Wal Mart. While

other papers use distance as part of their instrument to account for the endogeneity of

store openings, ours is not an instrumental measurement but rather a different set of

criteria for examining the affects of Wal Mart on wages and employment.

For example while Neumark et al, and Dube et al, use the distance from Arkansas,

where the first Wal Mart was opened, as an instrument for Wal Mart openings, their


3

measure of the effect of Wal Mart’s opening is simply whether a Wal Mart is located in

region i at time j. The implicit assumption in this measure is that everyone in that region

is affected equally. It would seem reasonable to assume that those further away from

Wal Mart are likely to be less affected than those closer to Wal Mart. This is especially

true if the effects of Wal Mart are concentrated on low wage workers who may not be

willing to commute long distances for relatively low paying jobs.

For example, in a large region with a uniformly distributed population, the affects

of Wal Mart’s entry are likely to diminish as one is further from that Wal Mart.

Alternatively, in a region where most of the population is located at one end of the region

while Wal Mart is located at the other end, the affects of Wal Mart’s presence is expected

to be muted.

The ideal solution to this problem would be to identify affected workers in a

smaller unit of geographic location (e.g., zip code, rather than county) allowing us to

examine the affects of Wal Mart at different distances. While we do have detailed

information on the location of each Wal Mart store, unfortunately, we do not have this

level of data for individual workers.

As a second best approximation to this ideal, we measure the average distance of

each Wal Mart to each zip code in a region, irrespective of which region Wal Mart is

located. Not only does this explicitly account for the distance to the nearest Wal Mart,

our measure also allows us to examine the affect of Wal Mart on wages and employment

of workers for whom the nearest Wal Mart is located in a region different from their own.

These spatial spill over effects are not accounted for in previous papers and constitute a

significant improvement in the Wal Mart literature.


4

Finally, we explicitly account for potential simultaneity of Wal Mart’s decision to

enter a region with wages in that region by us of instrumental variable regression.

II. DATA

The data on wages and employment is taken from the March rotation of the

Current Population Survey pooled over the years 1962-2006. However, since Wal Mart

entered California in 1990, most data analyzed is from 1986 to 2004. We analyze

employment and wages in constant1983-84 dollars, of workers 16 years and older,

working at least 10 hours per week, earning at least the minimum wage of $3.35 per hour

in 1983-84. The Wal Mart data was obtained from the WalMart.com web site by Thomas

J. Holmes at the University of Minnesota, and contains the date and location of 3,243

Wal Mart’s opened since the first Wal Mart opened in Rogers Arkansas on July 1, 1962

up through October 26, 2005. We concentrate on the data for California which contains

the opening date and location of 157 Wal Marts from the first Wal Mart opened in

Lancaster in 1990 up to 2005. We examine twenty regions in California identified in the

CPS by metropolitan statistical area (MSAFP). These are shown in Table 1 along with

their msafp code used in the CPS data set.

TABLE 1

III. THE MODEL

We use a fixed effect model to examine wages and employment by region over time in

California. The wage equation is specified as follows:

(1) Wijt = ijtijtj

jjtktt

tjtkjtk uRGTxQWMTQWMWM +++++++ ∑∑∑ θχλγδβββ 10


5

To analyze the employment effects of Wal Mart, we examine the employment population

ratio of each region over time. The model used to examine employment is similar to that

used to examine wages minus the vector of individual characteristics.

(2) EPRjt = jtj

jjtktt

tjtkjtk uRGTxQWMTQWMWM ++++++ ∑∑∑ λγδβββ 10

Wijt is the log wage of individual i in region j at time t.

EPRjt is the employment population ratio for region j at time t.

WMjtk is an indicator variable for each year that a new Wal Mart enters region j at

distance k.

QWMjtk is the cumulative number of Wal Marts in region j, year t and distance k.

Tt is a time period indicator representing the number of years since Wal Mar’s entry in

region j in five year increments.

TxQWMjtk is an interaction between time and the number of Wal Mart’s in region j at

distance k.

RGj is an indicator variable for region j

?ijt is a vector of personal characteristics of individual i , including indicators for Black,

and female, levels of potential experience defined as age minus schooling minus 6,

experience squared and years of schooling attained.

We measure the effect of Wal Mart on wages and employment in a region based

on both the year of entry and average distance of the nearest Wal Mart to the center of

that region. We believe that the closer the proximity of Wal Mart to the affected groups,

the greater the effect will be on wages. This is shown in Figure 1, where we estimate the

effect of distance on the annual percentage change in wages.

FIGURE 1


6

Our measure of distance is unique to the Wal Mart literature and we feel is an

improvement over previous measures which measure entry simply as the period in which

Wal Mart enters a region. Using the average distance to Wal Mart in a region takes into

account the fact that in large areas, the affects of Wal Mart may be irrelevant for a

majority of a county if Wal Mart is located in a far end of that county. To provide a more

accurate measure of the affects of Wal Mart, we measured the distance, in a given year,

of each zip code in an MSA to every Wal Mart opened that year. We then pick the Wal

Mart within a pre-specified distance to each zip code. The distances we chose to examine

are Wal Marts within an average of 5, 10, 15, 20 and 30 miles of each zip code in an

MSA.1

To calculate the average distance of each Wal Mart, we start by obtaining the

opening date and zip code of each Wal Mart in the US from the first Wal Mart opened in

Rogers Arkansas on July 1, 1962 up through October 26, 2005. We then combine this

data with the zip codes contained in each MSA. Since each zip code has a corresponding

latitude and longitude, we use the Haversine formula to calculate the distance from each

Wal Mart to each zip code in an MSA. Our measure of distance is the mean distance of

each zip code in an MSA to the nearest Wal Mart irrespective of which MSA Wal Mar is

located. This allows us to examine the affect of Wal Mart on wages and employment of

workers for whom the nearest Wal Mart is located in an MSA different from their own.

This is not accounted for in previous papers. We then examine the effect of Wal Mart on

wages and employment for all workers in an area as well as for low skilled workers and

retail sector workers.


7

IV. ENDOGENEITY

A potential problem with the models specified in equations (1) and (2) is that the

location and timing Wal Mart’s decision to enter a region may not be exogenously

determined. For example, if Wal Mart’s decision to locate in an area is determined in

part by local wages and employment, then wages in a region and Wal Mart’s decision to

enter a region will be simultaneously determined causing the error term to be correlated

with the Wal Mart opening variable as well as the variable representing the number of

Wal Marts in a region. We tests for the presence of endogeneity, by examining both the

Hausman (1978) test for endogeneity and a test for the difference between the Ordinary

Least Squares (OLS) and Instrumental Variable (IV) estimates. The results of both are

shown in Table 2. The Hausman test indicates the presence of endogeneity in all models

except for the model measuring the effects of Wal Mart within 5 miles. In the lone

instance where endogeniety is not a problem when testing at the 5% or 1% level of

significance, it does exist at the 10% level of significance. The test for a difference in

estimators finds that the OLS and IV models produce statistically different estimators in

models 4 and 5 measuring the distance to Wal Mart at 15 and 25 miles respectively,

while finding no difference in estimates in the models measuring the average distance to

Wal Mart of 5, 10 and 30 miles. We thus conclude that endogeneity does exist.

TABLE 2

To correct for endogeneity, we instrument Wal Mart’s decision to enter a region

in each year by constructing a binary location model of entry. We then use logistic

regression to estimate the time and location of Wal Mart’s entry into a region. The model

is specified as follows:


8

(3) WMjt = jtjtjtjtjtjt TTPWHTPRTPLSPCEW ςεεεεεεεε ++++++++ 276543210

Where,

WMjt is an indicator variable for the year a Wal Mart opened in region j.

Wjt is the average wage of region j in year t.

PCEjt is the percentage change in the employment population ratio in region j year t.

PLSjt is the percent of low skilled workers, measured as the percent of the population

with a high school degree or less, workers in region j and year t.

PWHTjt is the percent of the population that is White in region j year t.

T and T2 are time and time squared respectively.

Our instrument differs from that used by others in several respects. For example,

in their study on employment and wages in the retail sector, Neumark et al (2006) use an

instrument based on the pattern of expansion of Wal Mart in concentric circles from the

original location in Rogers Arkansas in 1962. In contrast, Basker’s (2005) instrument

differs considerably from Neumark et al in that Basker accounts for endogeneity of Wal

Mart openings by instrumenting the actual number of store openings with the planned

number of store openings based on the store number assigned by Wal Mart. Finally,

Goetz and Swaminathan (2006) in their analysis of county wide poverty levels use an

instrument for Wal Mart entry consisting of a “retail pull factor,” highway access, the

number of female heads, average length of commute, purchasing power and educational

level in each county.

Although, the Neumark et al instrument is convincing in light of the pattern of

growth of Wal Mart stores nationally, it is not appropriate for our purposes of analyzing


9

store openings in California. In particular, as Neumark et al correctly note, although the

pattern of Wal Mart openings do appear to follow concentric rings centered on Rogers,

Arkansas, this pattern is only appropriate as Wal Mart initially expanded from Arkansas.

Once California is reached, the pattern of Wal Mart locations differ in that the first Wal

Marts in the state were initially located across the center of the state in rural areas and

then spread out from each location. To accommodate this pattern of growth, we use a

combination of the Neumark et al and Basker instruments. Our instrument is similar to

Basker’s in that we use a set of demographic characteristics of each region which affect

the probability of Wal Mart entering. We also include economic characteristics such as

the percentage change in the employment population ratio and the percent of workers

employed in the retail sector. To account for Wal Mart’s “filling in” or saturation of the

state from the initial central California locations, we use time and its square.

To test our instrument we compared the actual opening of each Wal Mart at each

distance analyzed with the instrumented variable for each opening at those distances.

Table 3 shows the regression results for the actual Wal Mart openings regressed against

the predicted values. As the table shows, the coefficients on the instruments are close to

one and all are statistically significant at the 1% level of confidence.

TABLE 3

V. RESULTS

Wages

As a first approximation of the affect of Wal Mart on wages, Figures 2 through 6

show real mean wages before and after Wal Mart’s entry into a region of California for


10

each of the distances examined. Wages are indexed to 100 in the year of Wal Mart’s first

entry, while years are indexed to zero in the year of first entry. For example, consider

first, Figure 2 which shows the trend mean wages of those within five miles of the

location of Wal Mart’s initial entry. The trend in wages, shown in Figure 2, illustrates

the trend before and after Wal Mart’s entry into a region denoted time zero. Figure 2

appears to indicate a decrease in wages for approximately the first eight years following

Wal Mart’t entry. For those located within ten miles of the nearest Wal Mart, Figure 3

shows no apparent affect on wages. Figures 4,5 & 6 which show the affect of Wal Mart

on wages for those within 15, 20 and 30 miles respectively, indicate a similar pattern of

wage growth where wages seem to rise following Wal Mart’s entry.

FIGURES 2-6

We begin our empirical analysis of the affect of Wal Mart on wages by estimating

Equation 1 for wage earners within 5, 10, 15, 20 and 30 miles of the nearest Wal Mart.

We assume that Wal Mart will have its largest affect on the low wage, low skill labor

market, so we examine two groups of workers: The affected group consisting of low

skilled workers defined as those with at most a high school diploma. The second group

that we examine is a control group which we expect to be relatively unaffected by Wal

Mart’s entry. The control group consists of high skilled workers defined as those with a

four year college degree. The OLS regression results are shown in Table 1A of the

appendix for reference, and are unadjusted for endogeneity thus should approximate the

pattern of wage growth shown in Figures 2-6.


11

In general, the OLS regression results are mixed. For all low skilled workers, the

coefficient measuring the affect of Wal Mart openings on wages is positive for those

within 5 miles of nearest Wal Mart but negative for those at the remaining distances of

10, 15, 20 and 30 miles. Moreover, only the positive coefficient is significantly different

from zero. The affect of the number of Wal Mart’s in an area on wages is negative for all

five of the models examined and is statistically different from zero for the three models

measuring the affect on wages for those within 15, 20 and 30 miles of the nearest Wal

Mart. Finally, the affect of Wal Mart on wages after the first five years of entry, is

negative for those within 5 miles of the nearest Wal Mat but is not statistically different

from zero. For those further from Wal Mart, the coefficient is positive but only

significantly different than zero for one model. For the remaining time periods 6-10

years after entry, 11-15 years after entry and 16-18 years after entry, the coefficients are

mixed and show no clear pattern on the wages of low skilled workers at the various

distances from Wal Mart.

For high skilled workers, the results are surprisingly more consistent. The affect

of Wal Mart’s entry on high skilled wages is negative for all five models and statistically

different from zero for the models measuring the affect of wages for those within 10, 15,

20 and 30 miles of Wal Mart. The coefficient measuring the number of Wal Mart’s

positive and significant for those within 10-30 miles of a Wal Mart. While these two

opposing affects seem contradictory, they could be explained by spillover effects of Wal

Mart’s entry. That is, if Wal Mart acts as an “anchor” for other businesses, then although

high skill wages may initially be negatively affected by Wal Mart’s entry into an area, as

more Wal Mart’s enter an area bringing with them more economic activity, high skill


12

wages can be positively affected. This, however, appears inconsistent with the negative

affect on high skilled wages for the first five years following Wal Mart’s entry.

We also examine the affect of Wal Mart on low and high skilled wage earners in

the retail sector, which is believed to be the sector most affected by Wal Mart’s entry.

Again, the OLS results are shown in Table 2A for reference. These results indicate

mixed affects for both low skilled and high skilled workers and are generally not

statistically significant. Interestingly, however, the affect on high skilled workers in the

retail sector is similar to that of all high skilled workers shown in Table 1A. That is,

similar to all high skilled workers, the affect of Wal Mart’s entry on high skilled retail

sector wages is negative for all five models although none of the coefficients are

statistically different from zero. Also, similar to all high skilled workers, high skilled

workers in the retail sector are positively affected by the number of Wal Marts within 15,

20, and 30 miles. Again, however, the coefficients measuring the affect on high skill

wages in the retail sector over the first 10 years following Wal Mart’s entry are mixed

and none are statistically different from zero.

To correct for endogeneity problems associated with Wal Mart’s entry noted

above, we re-estimate Equation 1 instrumenting Wal Mart’s entry with Equation 3. The

instrumental variable (IV) regression results for all low skill and high skill wage earners

are shown in Table 4. The results, however, are mixed. The affect of Wal Mart’s entry

on low skilled workers is negative for those within 10, 20 and 30 miles of Wal Mart but

are only statistically significant for those within 20 and 30 miles. Similarly, low skilled

wages are negatively affected for those within 5 and 15 miles of Wal Mart but positively

affected for within 10, 20 and 30 miles. The affect on low skill wages five years after


13

Wal Mart’s entry is negative and significant for those within 5 and 15 miles but positive

and significant for those within 20 and 30 miles of Wal Mart. For all high skilled

workers, the coefficients are mixed and rarely statistically significant.

The instrumental variable retail sector analysis, shown in Table 5, and once again

the results for low skilled retail workers are mixed. The affect of Wal Mart’s entry on

low skilled retail wage earners is positive for four of the models but none are statistically

different from zero. The coefficient on the number of Wal Marts is negative for three of

the models but once again none are statistically significant. The affect on low skill retail

wages after the first five years of Wal Mart’s entry is positive for those within 10, 20 and

30 miles of Wal Mart but only significant for those within 10 miles of the nearest Wal

Mart. For high skilled retail wage earners the coefficient on Wal Mart entry is positive

for those within 5, 10 and 20 miles of Wal Mart, but negative those within 15, and 30

miles with only those within 30 miles being statistically different from zero. For the

number of Wal Mart’s in a region, the coefficient is positive for 5, 15 and 30 miles with

only those within 30 miles of Wal Mart being statistically different than zero. The time

dummies measuring the affect of Wal Mart on high skilled retail wages after Wal Mart’s

entry are mixed in sign and non are statistically significant.

Employment

To investigate the affect of Wal Mart on employment, we examine the

employment population ratio of each region in California by skill level and industry


14

sector. We define the employment population ratio as the number of those 16 years and

older working positive hours divided by the number of people 16 years and older in that

region. Consider first the groups most affect by the entry of Wal Mart, low skilled

workers. The instrumental variable results are shown in Table 6. The effects of Wal

Mart openings on the employment population ratio of low skilled workers are mixed in

that they vary in sign from model to model. The affect of Wal Mart’s presence over time

however indicate a negative and statistically significant effect on the

employment/population ratio of low skilled workers within 5, 10 and 15 miles of Wal

Mart.

The instrumental variable regression results for low skilled workers in the retail

sector, shown in Table 7, are once aga in mixed. The coefficients on Wal Mart’s entry

and the number of Wal Mart’s in a region vary in sign, magnitude and significance.

However, similar to the affects on all low skilled workers, the affect on low skilled retail

sector workers over time indicate a negative and statistically significant effect on the

employment/population ratio of those within 5, 10 and 15 miles of Wal Mart.

For the control group of high skilled workers presumed not affected by Wal

Mart’s entry into a region, the sign, magnitude and significance of the coefficients on

Wal Mart’s entry and the number of Wal Mart’s in a region are mixed. The coefficients

on the variables measuring Wal Mart’s presence over time are similarly ambiguous with

the coefficients mixed in sign, magnitude and significance. The OLS results for the

employment population ratio are shown in Tables 3A and 4A for reference.

VI. CONCLUSION


15

The debate over Wal Mart’s affect on the labor market continues to be argued in

academia as well as in the popular press. Opponents of Wal Mart argue that the retail

giant destroys smaller mom and pop retailers and acts as a monopsonist in the low skill

labor market resulting in lower employment and wages. Others argue that Wal Mart acts

as an “anchor” for other retailers by providing a signal to enter a region, thereby

increasing employment and wages. This article provides an empirical investigation of the

labor market effects of Wal Mart’s entry into California and is a significant improvement

over previous studies. To begin with, our data set allows us to examine the effects of

Wal Mart on those believed to be most affected by Wal Mart’s entry, namely low skilled

workers and workers in the retail sector. Additionally, we are able to further identify the

affected group of workers by creating a unique measure of entry based on the average

distance of Wal Mart to the affected groups. That is, we assume that those nearer to a

Wal Mart will be affected more than those further away from a Wal Mart. This is a

significant improvement over previous measures which simply measure Wal Mart’s entry

into a region without regard to distance, thus implicitly assuming that everyone in a

county, MSA or other similarly defined region is equally affected. Finally, as with

previous studies, we were able to correct for endogeneity of Wal Mart’s decision to enter

a region by constructing an instrument for Wal Mart’s entry into a region based on

demographic characteristics of the region as well as the saturation of Wal Mart stores

over time.

Given these improvements, our analysis fails to find any consistent affect on

wages resulting from Wal Mart’s entry into a region on either all low skilled workers or

low skilled workers in the retail sector. We do, however, find negative and statistically


16

significant employment effects on all low skilled workers as well as low skilled retail

workers within 15 miles of Wal Mart. These negative affects do not appear in the retail

sector analysis. We attribute this to the small number of observations in a region, within

the retail sector. While the results of this study neither commend nor condemn Wal

Mart’s entry into a region, we do however prove that as with most polemics, there is a

little bit of truth to both sides.


17

REFERENCES

Basker, E. “Job Creation or Destruction? Labor Market Effects of Wal-Mart Expansion.”

The Review of Economics and Statistics, 87(1), 2005, 174-183.

Ciccarella, S., David Neumark, and Junfu Zhang. “The Effects of Wal-Mart on Local

Labor Markets.” National Bureau of Economic Research Working Paper No

11782, 2005.

Dube, Lester and Eidline, “Firm Entry and Wages: Impact of Wal-Mart Growth on

Earnings Throughout the Retail Sector.” Institute for Research on Labor and

Employment Working Paper Series, 2007. University of California, Berkeley.

Goetz, S. and Hema Swaminathan. “Wal-Mart and County-Wide Poverty.” Social

Science Quarterly, 87(2), 2006, 211-226.

Hausman, J. (1978), “Specification Tests in Econometrics.” Econometrica

46(6), 1251-71.

Hughes, J. and Brian A. Ketchum. “Wal-Mart and Maine: The Effect on Employment

and Wages.” Maine Business Indicators, 42(3), 1997, 6-8. Ory, D and Mokhtarian, Patricia, “The Impact of Telecommuting on Commute Time,

Distance and Speed of State of California Workers.” Institute of Transportation

Studies, University of California Davis, December 2005.

1 According to a report by the University of California’s Institute of Transportation Studies, the mean commute distance in California is approximately 20 miles.

TABLE 1 California Metropolitan Statistical Areas Analyzed

Region Area Encompassed MSAFP 1 Modesto 5170 2 Bakersfield 680 3 Stockton 8120 4 Vallejo-Fairfield-Napa 8720 5 Fresno 2840 6 Santa Rosa-Petaluma 7500 7 Visalia-Tulare-Porterville 8780 8 Riverside-San Bernardino 6780 & 7280 9 Oxnard-Ventura 6000 & 8738 10 Yuba City 9340 11 Chico 1620 12 Anaheim-Santa Ana 360 & 5945 13 San Diego 7320 14 Salinas-Seaside-Monterey 7120 15 San Jose 7400 16 Sacramento 6920 17 Los Angeles-Long Beach 4480 18 Oakland 5775 19 Santa Barbara-Santa Maria-Lompoc 7480 20 San Francisco 7360

TABLE 2 Tests for Endogeneity

Model 1 Model 2 Model 3 Model 4 Model 5 Hausman Test Wages Wages Wages Wages Wages Error Term 0.815 -0.229 -0.193 -0.159 -0.056 (1.26) (5.03)** (3.47)** (4.35)** (3.81)** Test for Difference in Coefficients t-test (1.138) (0.377) (3.432)** (4.029)** (0.210) Absolute value of t-statistics in parentheses * significant at 5% level; ** significant at 1% level

TABLE 3

Regression of Actual Versus Predicted Wal Mart Openings Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 IV 0.983 0.912 1.035 0.963 0.969 0.986 (228.78)** (331.76)** (211.42)** (305.27)** (340.66)** (427.13)** Const. -0.001 0.001 0 0.008 0.009 0.007 (1.77) (8.06)** (0.34) (10.17)** (10.60)** (7.17)** N 346400 346400 346400 346400 346400 346400

R2adj 0.13 0.24 0.11 0.21 0.25 0.34

Absolute value of t-statistics in parentheses

* significant at 5% level; ** significant at 1% level

TABLE 4 Instrumental Variable Regression of Wages

Model 1

Within 5 Miles Model 2




Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry 0.277 -0.024 -0.181 -0.128 0.377 0.034 -2.613 4.754 -0.328 -0.455

[0.00]** [0.88] [0.09] [0.72] [0.00]** [0.71] [0.01]** [0.37] [0.00]** [0.00]**

Number of Wal Marts -0.008 -0.012 0.021 0.063 -0.06 0.014 0.257 -0.453 0.018 0.046 [0.70] [0.78] [0.27] [0.43] [0.00]** [0.29] [0.01]** [0.38] [0.00]** [0.00]**

5 Years After Entry -0.228 0.046 0.017 -0.037 -0.036 -0.025 0.262 -0.503 0.02 0.009 [0.01]* [0.80] [0.20] [0.44] [0.00]** [0.02]* [0.01]** [0.34] [0.00]** [0.41]

6-10 Years After Entry -0.029 0.018 0.053 0.025 -0.052 -0.01 0.364 -0.693 0.046 0.054 [0.42] [0.77] [0.02]* [0.65] [0.00]** [0.52] [0.01]** [0.36] [0.00]** [0.01]**

11-15 Years After Entry 0.062 -0.075 0.151 0.121 -0.444 -0.033 2.68 -5.056 0.263 0.416 [0.01]** [0.09] [0.14] [0.70] [0.00]** [0.76] [0.01]** [0.37] [0.00]** [0.00]**

16-18 Years After Entry 0.083 -0.061 0.023 -0.029 0.01 -0.007 0.192 -0.162 -0.018 -0.009 [0.00]** [0.19] [0.03]* [0.37] [0.10] [0.43] [0.01]** [0.35] [0.00]** [0.34]

Constant 4.777 4.556 4.697 4.406 4.721 4.428 5.954 3.24 5.225 5.127 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.02]* [0.00]** [0.00]**

Observations 43828 25674 43828 25674 43828 25674 43828 25674 43828 25674 Adjusted R-squared 0.21 0.17 0.21 0.17 0.18 0.17 0.18 0.12 Absolute value of t-statistics in brackets * significant at 5% level; ** significant at 1% level

TABLE 5 Instrumental Variable Regression of Wages-Retail Sector

Model 1





Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry 0.18 2.065 0.296 4.147 0.277 -0.092 -1.554 1.818 -0.2 -1.629 [0.28] [0.46] [0.17] [0.22] [0.26] [0.83] [0.36] [0.66] [0.12] [0.04]*

Number of Wal Marts -0.025 0.001 -0.025 -0.874 -0.033 0.067 0.163 -0.164 0.016 0.168 [0.61] [1.00] [0.50] [0.25] [0.30] [0.31] [0.35] [0.72] [0.15] [0.01]**

5 Years After Entry -0.125 0 0.073 0.443 -0.02 0.021 0.152 -0.183 0.005 0.002 [0.50] [.] [0.01]* [0.22] [0.46] [0.64] [0.37] [0.57] [0.76] [0.97]

6-10 Years After Entry 0 -0.147 -0.034 -0.468 -0.028 0.026 0.212 -0.26 0.029 0.253 [1.00] [0.81] [0.46] [0.30] [0.44] [0.71] [0.35] [0.64] [0.18] [0.06]

11-15 Years After Entry 0.077 0.004 -0.309 0 -0.358 -0.569 1.56 -2.586 0.13 1.212 [0.16] [0.99] [0.13] [.] [0.17] [0.31] [0.37] [0.56] [0.30] [0.13]

16-18 Years After Entry 0.086 0.514 0.047 0.446 0.024 -0.018 0.123 -0.027 0.007 -0.068 [0.19] [0.24] [0.05]* [0.28] [0.09] [0.71] [0.28] [0.55] [0.62] [0.19]

Constant 4.882 5.413 4.862 6.9 4.841 5.31 5.276 5.222 5.137 8.303 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**

Observations 7633 1630 7633 1630 7633 1630 7633 1630 7633 1630 Adjusted R-squared 0.19 0.09 0.19 0.17 0.12 0.18

Absolute value of t-statistics in brackets * significant at 5% level; ** significant at 1% level

TABLE 6 Instrumental Variable Regression of the Employment Population Ratio

Model 1





Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry 0.141 0.044 -0.308 -0.924 0.303 0.161 -0.71 -3.316 -0.143 -0.035

[0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.45] [0.00]** [0.00]**

Number of Wal Marts -0.008 0.029 0.036 0.195 -0.045 -0.032 0.07 0.326 0.009 -0.003 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.46] [0.00]** [0.00]**

5 Years After Entry -0.15 0.097 -0.028 -0.11 -0.032 -0.015 0.069 0.334 0.009 0.002 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.45] [0.00]** [0.00]**

6-10 Years After Entry -0.048 -0.015 0.062 0.146 -0.04 -0.015 0.1 0.48 0.022 0.006 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.44] [0.00]** [0.00]**

11-15 Years After Entry 0.003 -0.021 0.232 0.851 -0.303 -0.128 0.761 3.555 0.129 0.059 [0.02]* [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.45] [0.00]** [0.00]**

16-18 Years After Entry -0.023 0.032 -0.012 -0.066 0.01 0.012 0.056 0.106 0.004 0.004 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.42] [0.00]** [0.00]**

Constant 0.617 0.792 0.522 0.498 0.624 0.86 0.52 0.357 0.576 0.838 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.59] [0.00]** [0.00]**

Observations 43931 25707 43931 25707 43931 25707 43931 25707 43931 25707 Adjusted R-squared 0.46 0.27 0.25


TABLE 7 Instrumental Variable Regression of the Employment Population Ratio-Retail Sector

Model 1





Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry -0.078 0.079 0.106 -0.103 -0.217 -0.035 0.696 -0.504 0.074 0.141 [0.00]** [0.67] [0.00]** [0.53] [0.00]** [0.21] [0.03]* [0.54] [0.00]** [0.00]**

Number of Wal Marts 0.031 -0.01 -0.002 0.03 0.034 0.012 -0.068 0.062 -0.002 -0.007 [0.00]** [0.71] [0.54] [0.41] [0.00]** [0.01]** [0.05]* [0.51] [0.01]** [0.07]

5 Years After Entry 0.033 0 -0.004 -0.008 0.016 -0.007 -0.072 0.029 -0.008 -0.016 [0.02]* [.] [0.21] [0.63] [0.00]** [0.02]* [0.03]* [0.66] [0.00]** [0.00]**

6-10 Years After Entry -0.003 -0.005 -0.024 0.011 0.031 0 -0.092 0.064 -0.009 -0.033 [0.60] [0.90] [0.00]** [0.62] [0.00]** [0.94] [0.03]* [0.57] [0.00]** [0.00]**

11-15 Years After Entry 0.007 -0.003 -0.114 0 0.2 0.038 -0.746 0.537 -0.078 -0.135 [0.09] [0.88] [0.00]** [.] [0.00]** [0.30] [0.03]* [0.55] [0.00]** [0.00]**

16-18 Years After Entry -0.038 0.014 -0.008 -0.006 0.003 -0.008 -0.044 -0.005 0.002 0.002 [0.00]** [0.64] [0.00]** [0.77] [0.12] [0.01]** [0.04]* [0.55] [0.05] [0.66]

Constant 0.91 1.003 0.953 0.961 0.919 0.985 0.969 0.916 0.936 1.004 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**

Observations 7633 1630 7633 1630 7633 1630 7633 1630 7633 1630 Adjusted R-squared 0.1 0.07 0.05


TABLE 1A Ordinary Least Squares Regression of Wages

Model 1





Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry 0.083 -0.051 -0.017 -0.041 -0.004 -0.028 -0.003 -0.021 -0.009 -0.023 [0.02]* [0.46] [0.13] [0.01]** [0.66] [0.02]* [0.68] [0.07] [0.23] [0.04]*

Number of Wal Marts -0.019 -0.002 -0.01 0.044 -0.011 0.023 -0.009 0.018 -0.007 0.014 [0.36] [0.96] [0.14] [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**

5 Years After Entry -0.075 0.07 0.034 -0.027 0.003 -0.021 0.003 -0.026 0.001 -0.02 [0.23] [0.65] [0.00]** [0.04]* [0.58] [0.01]** [0.57] [0.00]** [0.88] [0.00]**

6-10 Years After Entry 0.032 0.018 0.022 0.013 0.002 -0.002 -0.005 -0.005 -0.006 -0.01 [0.25] [0.72] [0.05] [0.45] [0.78] [0.81] [0.34] [0.57] [0.18] [0.14]

11-15 Years After Entry 0.066 -0.079 0.003 0.053 -0.055 0.032 -0.057 0.033 -0.04 -0.016 [0.01]** [0.08] [0.95] [0.62] [0.04]* [0.57] [0.02]* [0.52] [0.03]* [0.64]

16-18 Years After Entry 0.086 -0.064 0.033 -0.023 0.012 -0.01 0.001 -0.005 -0.006 0.004 [0.00]** [0.17] [0.00]** [0.07] [0.05] [0.24] [0.93] [0.52] [0.30] [0.61]

Constant 4.773 4.543 4.763 4.426 4.844 4.427 4.842 4.43 4.864 4.484 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**

Observations 44653 26128 44653 26128 44653 26128 44653 26128 44653 26128 Adjusted R-squared 0.21 0.17 0.21 0.17 0.21 0.17 0.21 0.17 0.21 0.17


TABLE 2A Ordinary Least Squares Regression of Wages -Retail Sector

Model 1





Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry 0.216 -0.14 0.047 -0.052 -0.004 -0.026 0.016 -0.035 0.024 -0.054 [0.01]** [0.67] [0.09] [0.51] [0.83] [0.64] [0.38] [0.49] [0.18] [0.28]

Number of Wal Marts -0.036 0.155 0.008 0.056 0.001 0.053 -0.001 0.038 -0.003 0.036 [0.44] [0.41] [0.61] [0.18] [0.87] [0.01]** [0.81] [0.02]* [0.54] [0.00]**

5 Years After Entry -0.16 0 0.056 0.015 0.007 0.018 -0.002 -0.037 -0.01 -0.05 [0.26] [.] [0.02]* [0.81] [0.63] [0.61] [0.88] [0.24] [0.43] [0.14]

6-10 Years After Entry -0.005 0.252 0.012 0.038 0.012 0.015 0.004 -0.013 -0.004 -0.01 [0.94] [0.29] [0.64] [0.63] [0.41] [0.73] [0.73] [0.71] [0.69] [0.75]

11-15 Years After Entry 0.08 -0.167 -0.107 0 -0.07 -0.655 -0.074 -0.606 -0.079 -0.362 [0.14] [0.43] [0.29] [.] [0.24] [0.03]* [0.19] [0.03]* [0.06] [0.02]*

16-18 Years After Entry 0.087 0.215 0.038 -0.053 0.027 -0.014 0.021 -0.027 0.016 -0.009 [0.18] [0.37] [0.07] [0.35] [0.05] [0.70] [0.11] [0.41] [0.19] [0.80]

Constant 4.959 5.642 4.846 5.648 4.934 5.479 4.913 5.48 4.921 5.677 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**



TABLE 3A

Ordinary Least Squares Regression of the Employment Population Ratio

Model 1





Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry 0.038 0.036 0.004 -0.014 0 -0.008 -0.001 -0.008 -0.011 -0.006 [0.00]** [0.00]** [0.00]** [0.00]** [0.77] [0.00]** [0.01]* [0.00]** [0.00]** [0.00]**

Number of Wal Marts -0.007 0.025 -0.015 -0.01 -0.006 -0.008 -0.005 -0.007 -0.003 -0.005 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**

5 Years After Entry 0.109 0.009 -0.001 0.002 0 -0.065 [0.00]** 0 [0.00]** -0.001 [0.02]* -0.001 [0.00]** 0.001 [0.50]

6-10 Years After Entry [0.00]** -0.012 [0.41] 0.018 [0.02]* 0.009 [0.00]** 0.009 [0.00]** 0.002 -0.018 [0.00]** 0.006 [0.00]** 0.002 [0.00]** 0.001 [0.00]** 0.001 [0.00]**

11-15 Years After Entry [0.00]** -0.02 [0.00]** 0.089 [0.00]** 0.047 [0.00]** 0.038 [0.00]** 0.029 0.005 [0.00]** -0.036 [0.00]** 0.01 [0.00]** 0.017 [0.00]** 0.001 [0.00]**

16-18 Years After Entry [0.00]** 0.033 [0.00]** 0.013 [0.00]** 0.006 [0.00]** 0.007 [0.25] 0.005 -0.022 [0.00]** 0.005 [0.00]** 0.009 [0.00]** 0.008 [0.00]** 0.01 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**

Constant 0.605 0.787 0.612 0.824 0.598 0.831 0.599 0.835 0.602 0.834 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**



TABLE 4A Ordinary Least Squares Regression of the Employment Population Ratio-Retail Sector

Model 1





Within 30 Miles

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Low Skill

High Skill

Wal Mart Entry -0.037 -0.001 -0.01 -0.025 -0.003 -0.002 -0.007 0.007 0.002 0.003 [0.00]** [0.97] [0.00]** [0.00]** [0.09] [0.53] [0.00]** [0.04]* [0.11] [0.33]

Number of Wal Marts 0.027 0 0.015 0.012 0.006 0.007 0.005 0.004 0.003 0.005 [0.00]** [1.00] [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**

5 Years After Entry 0.003 . -0.012 0 -0.003 -0.009 -0.002 -0.011 -0.003 -0.011 [0.77] [.] [0.00]** [0.99] [0.01]* [0.00]** [0.09] [0.00]** [0.01]* [0.00]**

6-10 Years After Entry -0.014 0.012 -0.005 -0.001 0.002 -0.004 0.001 -0.004 0.002 -0.009 [0.01]** [0.45] [0.02]* [0.91] [0.18] [0.10] [0.18] [0.09] [0.00]** [0.00]**

11-15 Years After Entry 0.009 -0.009 -0.017 0 -0.022 0.002 -0.013 -0.009 -0.009 0.009 [0.05] [0.54] [0.03]* [.] [0.00]** [0.94] [0.00]** [0.64] [0.01]** [0.36]

16-18 Years After Entry -0.038 0.004 -0.014 0.004 0.003 -0.006 0.001 -0.005 -0.001 -0.004 [0.00]** [0.82] [0.00]** [0.31] [0.01]* [0.02]* [0.34] [0.03]* [0.56] [0.08]

Constant 0.913 1 0.919 0.983 0.925 0.988 0.923 0.987 0.921 0.982 [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]** [0.00]**



FIGURE 1

FIGURE 2

FIGURE 3

FIGURE 4

FIGURE 5

FIGRUE 6

The Effect of Wal-Mart on Wages and Employment Effect of Wal-Mart_Cuellar.pdfof Wal Mart’s entry are likely to diminish as one is further from that Wal Mart. Alternatively, in a

Documents