The Board of Regents of the University of Wisconsin System
Trucking Regulation, Unionization, and Labor Earnings: 1973-85
Author(s): Barry T. Hirsch
Reviewed work(s):
Source: The Journal of Human Resources, Vol. 23, No. 3 (Summer, 1988), pp. 296-319
Published by: University of Wisconsin Press
Stable URL: http://www.jstor.org/stable/145831 .
Accessed: 19/04/2012 18:22
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
University of Wisconsin Press and The Board of Regents of the University of Wisconsin System are
collaborating with JSTOR to digitize, preserve and extend access to The Journal of Human Resources.
http://www.jstor.org
Trucking Regulation, Unionization,
and Labor Earnings: 1973-85
Barry T. Hirsch
ABSTRACT
Current Population Survey (CPS) data for the years 1973 to 1985
are used to examine the earnings of union and nonunion truck
drivers during and after ICC regulation of the motor carrier industry. Hourly earnings for union drivers fell following deregulation, whereas wage changes among nonunion drivers closely
mirrored economy-wide changes among nonunion operatives.
Significant narrowing of the union-nonunion wage differential
occurred primarily in the previously regulated for-hire sector of
the trucking industry and in those regions with the most extensive nonunion operations. Despite some narrowing, the union
premium remained sizable following deregulation. Significantly
larger wage concessions would have been necessary to have
prevented the marked decline in union trucking and Teamster
membership that followed deregulation.
I. Introduction
Administrativederegulationduringthe late 1970s, coupled
with subsequent passage of the Motor CarrierAct of 1980, effectively
removed entry and rate restrictionsfrom an industrylong shelteredfrom
The author is a professor of economics at the Universityof North Carolinaat Greensboro. He received helpfuldiscussion and commentsfrom WilliamLinde, John Neufeld,
Nancy Rose, Frank Scott, anonymousreferees at thisjournal, and seminarparticipants
at the Universityof Kentuckyand the Universityof North Carolina-ChapelHill. The initial draft of the paper was preparedduringa faculty researchleave funded by the Universityof North Carolinaat Greensboro.CPS tapes wereprovidedthroughthe Inter-university Consortiumfor Political and Social Research (ICPSR).
[SubmittedJune 1987;accepted January1988]
THE JOURNAL
OF HUMAN
RESOURCES
* XXIII
*3
Hirsch 297
importantforms of competition. Previous studies (e.g., Moore 1978)have
found that organized labor achieved significantbargainingpower in the
trucking industry and captured a sizable share of the monopoly rents
associated with regulation.Substantialderegulationof the truckingindustry has encouragedrate competition and entry, particularlyby nonunion
firms, and significantly decreased union employment and bargaining
power. Because deregulationoccurred during a period of erratic macroeconomicperformance,however, sufficientevidence to evaluatederegulation's longer-runeffects on the earnings of drivers only recently has
become available.1
In this paper, data from the May CurrentPopulation Survey (CPS)
tapes for the years 1973through1985are used to examineunion coverage
and hourly earningsof union and nonuniontruck driversbefore and after
deregulation.The response of wages to deregulationare examined separately for truck drivers in the for-hire and private carriersectors of the
industryand across regions. In order to control for changes in economywide conditions, the wages of truck drivers are also compared with a
controlgroupof operativesoutside of trucking.The regulatoryand collective bargainingstructure of the industry is summarizedin Section II,
while the effects of deregulationon employmentand wages are analyzed
in Section III. Section IV presents the data and modeling employed in
subsequentempiricalwork. Empiricalresults and analysisare providedin
Sections V and VI, with conclusions following in Section VII.
II. Regulation and Deregulation in the Trucking
Industry
The Motor CarrierAct (MCA) of 1935provided the Interstate CommerceCommission(ICC)with regulatoryauthorityover motor
freightcarriage.Supportedby the alreadyregulatedrailroadindustryand
the newly created AmericanTruckingAssociation, the ICC severely restrained entry and price competition in the industry. Federal regulation
primarilycovered for-hirecommon carriersengagedin intercityandinterstate cartage. Private and local carriers, as well as carriers of exempt
1. Rose (1987) provides a closely related analysis. Althoughearlier drafts of our papers
differedsignificantly,the final versions are very similar.This study differsfrom Rose's in
that the wages of driversin the privatecarrieras well as the for-hiresectors of the industry
are analyzed in detail, regional differences in the effects of regulationare examined, an
occupationalcontrol group ratherthan broaderindustrycontrol groups is employed, and
sampleselection criteriaand regressionspecificationsdifferslightly.Rose providesa fuller
discussion of the National Master Freight Agreements(NMFAs) and an estimate of the
aggregateloss in rents to the Teamstersdue to deregulation.
298
The Journalof HumanResources
commodities (e.g., unprocessed agriculturalproducts), were largely exempt from directfederalregulation.Whilethe regulatedsector involved a
distinct minorityof carriers, their total operatingrevenues far exceeded
the noncovered sector.
Regulatedcommon carrierswere requiredto operateunderICCauthority that designatedpoints of originand destination,the routes over which
freight must be carried (in the case of regularroute carriersof general
freight), and types of freight (in the case of irregularroute carriers of
special commodities). Regular route carrierstypically operated through
terminals where less-than-truckloadcargoes were sorted into full-load
lots for movements to customers or other terminals.
After "grandfathering"many carriersthat existed prior to passage of
the MCA of 1935, the ICC severely constrainedissuance of new route
certificates.As interpretedby the ICC, certificatesof publicconvenience
and necessity were issued only where routes were not previously being
served and if entry would not economically damageexisting carriers.In
practice, the primarymechanismthroughwhich companies expandedor
entered into new lines of service was through purchase of route
certificatesin the secondary market.The aggregatemarketvalue of these
route certificates, which provides a rough measure of the monopoly returns to original certificate and capital owners from regulation,was reduced from 1.5 to 2 billion dollars in the 1972dollars to close to zero by
deregulation(Moore 1978, 1983).2
Because the setting of detailedfreightrates would have been extraordinarily complex, the ICC encouragedregulatedcarriersto establish rate
bureaus through which they could engage in joint ratemaking.In 1948
Congress passed the Reed-BullwinkleAct, which exempted these rate
bureausfrom the antitrustlaws. By 1980,approximatelyten majorand 55
smallerrate bureauswere in operation.Thus, individualfirmratemaking
in the regulated sectors of the marketwas severely constrainedprior to
deregulation.
Restrictions on entry and rate competition provided an opening for
union gains in the trucking industry. The longtime goal of the International Brotherhoodof Teamsters(IBT) was the signingof a nationalwage
2. Frew (1981) provides a detailed analysis of the determinantsof route certificateprices
during1971-77. Route prices exhibiteda downwardtrendover the period, ceteris paribus,
suggestingthat deregulationwas partiallyanticipated.Rose (1985)providesan events study
in which changesin share prices of 32 publiclytradedfirmsengagedin motorcarrieroperations are examinedover the 1977-81 period. She finds large (and highlyvariable)losses in
market value resulting from deregulation:an average 31 percent among general freight
carriersand an average 15 percent among specializedcommoditycarriers.She concludes
thatregulatoryrentswere realizedin partby ownersin the formof higherfirmprofits,andin
partdissipatedthroughhigherlabor and service costs.
Hirsch 299
agreementthat would remove labor costs from competitionamongfirms.
The goal's attainmentwas facilitated by the nature of ICC ratesetting,
which based rates on operating ratios, measuringthe ratio of operating
costs to revenues. Hence, increases in labor costs, which accounted for
about 65 percent of total freightcosts, could largelybe passed throughin
rates. While individualtruckingfirms still had incentive to lower costs,
Teamstercontrol of majorterminalsand their effective use of secondary
boycotts until the 1960s allowed the IBT to organize major segments of
the industry. The Teamsters' major strengthwas, and to a lesser extent
remains, among regularroute carriersof general freightin the less-thantruckload (LTL) market. Less highly organized are irregularroute and
special commoditycarriersand privatecarriersoperatingin the truckload
(TL) market.
National bargainingby the Teamsterswas achieved in 1964in the first
National Master Freight Agreement (NMFA), which shifted bargaining
power from the regionalto nationallevel. While the NMFA was a major
victory for Jimmy Hoffa in his move to centralizebargaining,there continued to exist regional bargainingand diversity in contracts. NMFA
contracts have followed every three years since 1964, the most recent
agreementgoing into effect in April, 1985.(Detailedhistoriesof collective
bargainingin the industryare providedby Levinson 1982and Perry 1986.)
Whileit would be convenient to date deregulationof the truckingindustry with the finalpassage in June of the Motor CarrierAct of 1980,-union
labor faced increasingcompetition in the industrythroughoutthe 1970s.
This competition arose in part as a naturaleconomic response to inefficiencies and cost differentialsengenderedby past regulationand union
wage increases. In addition, there were increased legal restrictions on
Teamster activities and administrativederegulationwithin the ICC. Beginningin the mid-sixties, the Teamsterswere increasinglyconstrainedin
their use of secondary boycotts and in their ability to restrict nonunion
traffic.By 1980,nonuniontruckingoperationshad alreadyincreasedtheir
share of the market, as evidenced in particularby the rapid growth of
private carriersand of nonunionowner-operatorsin the TL market.
Beginningin 1977,ICC administrativederegulationexpandedthe areas
under which truckingoperations were exempt from federal control and
eased entry restrictions, although most awards of certificates were for
extensions in service by existing carriersand not entry by new carriers.
And in a series, of importantdecisions beginningin late 1978, the ICC
allowed companies haulingtheir own goods to apply for authorityto haul
for others, abolished regulationslimitingcontract carriersto service no
more than eight shippers, expanded airport zones which were exempt
from regulation, and announced they would consider rates a factor in
grantingoperatingrights (Moore 1983).
300 The Journalof HumanResources
The Motor CarrierAct of 1980 effectively eliminatedbarriersto entry
througha weakeningof the publicneed test, removedmost restrictionson
operations(e.g., route and commoditydesignations),and allowedindividual variationin rates until January 1, 1984, at which time collective rate
bureauswere no longer exempt from antitrustaction. Muchof the bill was
an attempt to codify (and even limit) changes that were already taking
place. The direct effects of the MCA of 1980were not easily discernible,
however, since Congressional deregulationsimply sped changes in the
industrythat had been occurringfor some time, and because of a recession that broughtabout a significantdecline in traffic.
While the magnitude of the effects from deregulationremain to be
determined,there is no question that it facilitatedexactly the qualitative
responses that theory would predict (Perry 1986). There has been increased entry of low-cost and nonunion firms, a collapse in prices for
operatinglicenses, a high rate of business failurein truckingmarketsmost
affected by entry, increased alternativesand price discountingavailable
to shippers, rates that more closely reflect marginalcosts and a lessening
of cross-subsidizationbetween lines of service, improvedefficiency and
greaterinnovationsin operations,and fuller coordinationwith alternative
freight systems (e.g., "piggybacking"freight with railroads).Deregulation reversed the previous shift of trafficfrom regulatedfor-hire motor
carriersto unregulatedprivatecarriers;between 1977and 1983the shares
of tonnage carried increased sharply among for-hiremotor carriers, decreased moderatelyamongprivatemotorcarriers,and fell sharplyamong
railroads(Boyer 1987).The marketshare of union truckingand Teamster
employmentcontinued their decline, with significantdownwardpressure
on wage rates.
III. Employment and Wages of Union and
Nonunion Drivers
As discussed above, regulation in the trucking industry
made possible supra-competitivewages for several reasons. ICC restrictions on entry and rate competition permittedcost inefficiencies to survive. Ratemakingthrough bureaus allowed regional and industry-wide
wage increases to be largely passed throughto shippers and took wage
rates out of competition. And the nature of ICC regulationin the motor
carrierindustryallowed the Teamstersto acquire significantlymore bargainingpower than could have been obtained by a union operatingin a
more competitive environment. In fact, evidence of large union wage
premiumsin the trucking industry contrast with evidence from electric
Hirsch
utilities and other regulated industries (Hendricks 1977, 1986), where
firm-level rate regulation and wage bargaining appear to limit union power
(but see Ehrenberg 1979 and Perloff and Wachter 1984 for evidence of
substantial wage premiums in the unionized New York Telephone Company and U.S. Postal Service, respectively). Because regulation of the
trucking industry allowed substantial rents to accrue to labor, deregulation is expected to have affected employment, labor earnings, and the
union-nonunion wage differential.
Industry employment is likely to be positively affected in the long run,
due to substitution and scale effects associated with lower labor costs,
and to an increased volume of freight resulting from greater rate competition.3 For given levels of freight, however, the ratio of drivers to capital
(and possibly other labor) may be lower following deregulation. Decreased labor intensity is possible owing, in part, to the reduction in
inefficiencies previously engendered by regulation or union work rules
(e.g., return routes with empty trailers and restrictions on loading at
intermediate points on intercity routes). In addition, the use of operating
ratios in setting rates, whereby labor (i.e., operating) costs but not all
capital costs could be passed through to shippers, may have led to a
higher ratio of labor to capital during the regulatory period than is likely to
exist in the absence of regulation.4 Finally, if regulatory constraints
caused trucking firms to have suboptimal capacity utilization, deregulation may lead to decreases in the use of all factors per ton mile. Kim
(1984), who examines this possibility using Canadian evidence, concludes
that labor is the largest loser from trucking deregulation.
Union bargaining power made possible by ICC regulation clearly increased union wages, whereas its effect on nonunion wage rates is less
clear. To the extent that nonunion drivers were in relatively competitive
3. If the unionandfirmsbargainedsimultaneouslyover wages andemployment,"efficient"
bargainingoutcomes along a vertical contractcurve could obtain, wherebyunion employmentwould vary with the nonunion(opportunitycost) wage but not with the unionpremium
(MaCurdyand Pencavel 1986).Efficientbargainingis unlikelyto arise in the for-hiretrucking industrysector, however, since unionwages are determinedby nationaland/orregional
bargainingagreements,but employmentis set by individualfirms.Hence, sequentialwage
and employmentdeterminationmakes it likely that outcomes lie on firms' labor demand
curves.
4. For an explanationof the laborbias underregulation,see Moore(1978,332). Of course,
since operatingratios were not appliedat the firmlevel, individualcarriersretainedincentives (if not pressure)to minimizecosts. Daughetyand Nelson (1986)analyzeinputuse and
the productionstructurein the motor carrierindustryusing a time-series, cross-section
databaseof truckingfirmsfor 1954-58, 1968, 1978,and 1982.They find similaritiesin the
cost functionsfor the fifties and for 1982,but largedifferencesamongthe regulatoryperiod
samples.
301
302 The Journalof HumanResources
or unregulatedsectors with easy entry, nonunionwages would have been
close to competitive levels during the regulatoryperiod. On the other
hand, the presence of a strongunion may have increasednonunionwages
due to threat effects, since nonunion carrierswere willing to pay some
premiumin order to deter union organizing.5Threat effects presumably
were largeramongnonuniondriversin the regulatedfor-hiresector of the
industry than among drivers working in the largely unregulatedprivate
carrier sector, not only because of differences in the ease of entry, but
also because the formermarketwas where the Teamstershad theirgreatest bargainingstrengthand organizingability. Similarly,threateffects are
more likely to have occurred in regions of the countrywith low levels of
nonuniontruckingoperations.
Deregulationis likely to lead to a decreased wage level in the industry
as increased rate competitionand entry place cost pressures on carriers.
A decrease in union coverage and bargainingpower shouldmake possible
decreases in union and possibly nonunionwages and, in the long run, lead
to somewhat higher turnover and lower labor quality. Evidence that
nonunion wages following deregulationdo not fall, or that they move
closely with the wage rates of similarnonunionworkersoutside of trucking, would suggest that threateffects were small and that nonunionwage
rates for truckerswere close to competitive levels duringthe regulatory
period.6Indeed, the average industrywage level would decrease following deregulationthrough a shift of employmentfrom union to nonunion
drivers, even if there were no changes in union and nonunionwage rates.
While average wage levels will decrease, deregulation'simmediateeffect on the union-nonunionwage differentialis ambiguous. Decreased
union bargainingpower and a lessened ability to restrictthe expansionof
nonunion carriers are likely to reduce the spread between union and
nonunionwages in the long run. On the other hand, nonunionwages could
exhibit greatershort-runflexibilityin the face of entry and price competition. To the extent that the Teamstersmaintainbargainingpower withina
narrowsegment of the industry(i.e., the for-hireLTL long-haulmarket),
or are relatively intransigentin accepting wage concessions, the unionnonunionwage differentialis more likely to be maintainedor increased.
5. Recent studies providing evidence that nonunion wages in manufacturingand nonmanufacturingare positively and significantlyaffected by union density in workers'threedigit industryinclude Moore, Newman, and Cunningham(1985);and Hirsch and Neufeld
(1987).
6. Partiallyoffsetting the downwardpressureon wage levels are any increases in freight
trafficandlabordemandowingto ratedecreases.To the extent thatthe laborsupplyof truck
drivers is highly elastic, increased trafficshould have little long-runeffect on equilibrium
wages.
Hirsch 303
Maintenanceof a high wage differential,however, is likely to be associated with declining union employmentand long-runbargainingpower in
the industry.
IV. Data and Modeling
In order to examine union and nonunionwages of truckers,
to
and
prior
following deregulation,the May CurrentPopulationSurvey
(CPS) tapes for the years 1973-81 and 1983-85 are used (there was no
union question in the 1982 survey). The sample includes all male truck
driversin the laborforce, ages 16 to 64, who have data providedon usual
weekly earnings, usual hours worked per week, union status, and whose
hours worked are 30 or greater. Following our earlier discussion, the
period of ICC regulationis definedas 1973-78, whereas the period 197985 is consideredthe deregulationperiod.7Even thoughthe MotorCarrier
Act was not passed until June 1980, administrativederegulationin late
1978 significantly eased entry and set expectations for continued and
permanentderegulationwithin the industry.Perhapsthe most direct evidence of deregulation'sexpected effects is the averagesale price of operating licenses, which fell from $531,000 in 1977, to $370,000in 1978, to
only $55,000 in 1979 (all figures are in 1982 dollars). Both the price and
volume of traded licenses collapsed after 1979(Moore 1983).
Analyzed separatelyare truckdriversin the for-hireor commoncarrier
sector of the industry, defined as driverswho identifiedtheir industryof
employment as the trucking service industry, and drivers in the private
carriersector, definedas drivers who identifiedtheir industryof employment as something other than the trucking service industry (owner
operatorswhose designated class of worker is self-employedare not included, due to the absence of earnings informationfor these workers).
Drivers in the for-hire sector sample, which includes both general commodity and contract carriers, were affected directly by ICC regulation
and deregulation,whereas drivers in the largely unregulatedprivate carrier sector were affected only indirectlyby regulation.The wages of both
groups of drivers are subsequently compared to those of a one-in-two
7. Rose (1986)also treats 1979as the beginningof the deregulationperiod. Administrative
changesby 1978had effectively deregulatedthe TL market,whereasthe LTL marketwas
not deregulateduntil after passage of the MCA of 1980.We are unableto separateTL and
LTL drivers in the data. Year-to-yearchanges in wages and union premiumsare subsequentlysummarized,however, in orderto examinethe sensitivityof resultsto the datingof
the regulatoryperiod.
304
The Journal of Human Resources
control group sample of nontransport operatives meeting the same criteria
listed above.8
We first estimate the following log wage equations for the truck driver
samples in both the for-hire and private carrier sectors:
(1) ln(W)ij = 43kXijk + alDREGij +
L2UNij+ t3UN 'DREGij + eij,
where i indexes individuals, j indexes year, ln(W) is the natural logarithm
of hourly earnings in 1985 dollars, DREG is a dummy variable equal to 1
for the deregulation years 1979-85, UN is a dummy variable equal to 1 if
the worker is a union member, UN ? DREG is a union-deregulation interaction variable, and e is a stochastic error term with zero mean and
constant variance. The control vector X includes a constant (XO= 1) and
k variables including years of schooling completed (S); years of experience and experience squared, proxied by Age-S-5 (EXP and EXP2); and
dummy variables equal to 1 if the worker is nonwhite (NW), married with
spouse present (MARRIED), a veteran (VET), or in each of 9 Census
regions (REGION).9
The coefficient al measures for nonunion truckers the log differential in
8. The controlgroup of nontruckoperativeswas chosen because it seemed most likely to
contain"comparable"individualsnot directlyaffectedby truckingderegulation.Amongthe
occupations included in the control group are assemblers, inspectors, packers, machine
operatives, sewers and stitchers, and welders. Because of the use of revised occupational
codes in the CPS beginningin 1983, only occupationsfor which exact matches could be
made over the two subperiodswere includedin our sample. Drivers are classifiedby the
CPS as "heavy" or "light" truckdriversbeginningin 1983.The relativelysmallnumberof
"light" driverswere excluded since it is believedmost were classifiedas "deliverymenand
routemen"priorto 1983.Overallsamplesizes are smallerafter 1978becausefewer persons
were asked the earnings and unionizationquestions. The survey question on which the
union status variableis based changedseveraltimes over the 1973-85 period.The 1973-75
surveys asked, "Does . . . belongto a laborunion?"The 1976-78surveysaddedto the end,
". .. or employeeassociation."The 1979-81surveysrevisedthis to read, ".. . or employee
associationsimilarto a union." The May 1983CPS Pensionand RetirementPlanCoverage
surveyasked, "Is ... coveredby a unionor employeeassociationcontract?"(Earningsand
hours informationfor part of the 1983samplewere matchedwith the June questionnaire.)
The 1984-85 surveys return to the 1979-81 form of the union question. In later years,
nonunionworkerswere asked if they were covered by a unioncontract,but use of a union
coverage variablein those years producedhighlysimilarresults. None of the resultsappear
to be affected measurablyby changes in the union status variable.
9. Also estimatedwere log earningsequationswith the log of hoursworkedper week on the
right-handside. LettingE be usualweekly earnings,HRS usualhoursworkedperweek, and
W the wage, ln(W) = ln(E/HRS) = ln(E) - ln(HRS). Hence, an earningsequationof the
form ln(E) = O3X+ Tln(HRS)is equivalentto the wage equationln(W) = p3Xiff T = 1. In
estimates here, the parameter r for truckers is found to be significantlyless than one,
indicatingthat earningsincreaseless thanproportionatelywith reportedhoursof work. No
importantconclusions are affected by the use of a wage ratherthan earningsequation;
hence, the more commonlyused wage specificationresults are reportedhere.
Hirsch
wages in the deregulatory relative to the regulatory period, a2 is the
union-nonunion log wage differential for truckers in the regulatory period,
and 03 measures the change in the union premium during the deregulatory
period (the union premium after 1979 equals ao2 + 0(3). Percentage differentials, D, are approximated by D = (ea - 1)100.10
In order to compare the wage performance of truck drivers with a
control group of non-transport operatives, a wage equation of the following form is estimated:
(2) ln(W)ij =
3kXijk+ 81UNl + 82DREGij + 83TRij+ 84TR DREGij
+ 8sUN *DREGjt + 86UN ' TRi + 87UN * TR *DREGj
+ e.,
where TR is a dummy variable equal to one if the worker is a truck driver
and 0 if a nontruck operative, and other notation is as above. In this
specification 81 is the union-nonunion log wage differential for nontruck
operatives in the regulation period and 81 plus 85 is the differential during
the deregulation period, 82 is the log wage change between the deregulation and regulation periods for nonunion non-truck operatives, 83 is the
differential between nonunion truck drivers and the operative control
group during the regulatory period while 84 is the change in that differential following deregulation, 86 is the difference between the operative and
truck driver union premiums during the regulatory period, and 87 is the
change in that difference during the deregulatory period.
The model above is convenient for purposes of exposition in that wage
differentials for four groups-truck drivers and nontruck operatives
classified by union status-can easily be compared for the regulatory and
deregulatory periods. A more detailed analysis would allow wage differentials among the four groups to vary for each of the twelve survey years.
Such a model was estimated and results (available on request) are discussed briefly in Section VI.
The major focus in this paper is on the earnings performance of union
and nonunion truckers prior to and following deregulation, and changes in
10. For a comparisonof more exact measuresof D, see Giles (1982).Alternatively,estimates of the differentialcould be obtained based on estimation of separate union and
nonunionequations,wherebythe log differential,d, is equal to (3u - 13U)X,
whereX is the
meanvalue (union,nonunion,or weighted)of the explanatoryvariables.For ease of reporting andin orderto lower the varianceof the estimates(since samplesizes of truckersare not
large in later years), the dummy variableapproachis used in estimatingthe differential.
Whilethis may introducebias into our estimates, bias may be less seriousthanthe increase
in variance,from, say, separateregressionsby year, union status, and occupation(trucker
or nontrucker).Surveysof alternativeestimationmethodsare containedin Lewis (1986)and
Hirschand Addison (1986).
305
306 The Journalof HumanResources
the union wage premiumover time. Pensions, nonpecuniarybenefits, and
contractualprovisions (e.g., two-tierwage schedules)cannotbe analyzed
explicitlywith these data. It is expected that wages for unionand possibly
nonuniontruckersdeclined as a result of deregulation.Expectationsas to
the behavior of the union-nonunionwage differentialare not entirely
clear, however, since nonunionas well as union wages may have declined
and nonunion wage rates may exhibit greater short-runflexibility. It
seems likely, however, that in a long-runcompetitive environment,the
union wage premiummust fall, particularlyif nonunionwages were close
to competitive levels prior to deregulation.
Also of interest are differences in the above effects between truck
drivers in the regulated for-hire and largely unregulatedprivate carrier
sectors of the truckingindustry,the timingof the above changes, and any
regionaldifferences in the effects of deregulation.Ourexpectationis that
the effects of deregulationwere felt most stronglyin the for-hiresegment
of the industry, although wages in the private carriersector can be expected to have moved in a similardirection. Administrativederegulation
in the late seventies suggests downwardpressure on truckingwages at
least by 1979, while the time lag requiredfor long-runsupply responses,
negotiationof labor contracts, and the like, suggest that the full response
to deregulationwould not be observed for several years followingderegulation. Because there were changes in business cycle conditions and demand-supplyforces that affected the labor marketfor truckersindependent of deregulation,comparisonof truckers' wages to those of similar
nontruckoperatives is necessary before drawingany finalconclusions as
to the effects of deregulation.Finally, regionaldifferencesin the strength
of union organizingsuggest that separateanalyses by region may provide
insight into the effects of truckingderegulation.
V. Empirical Results: The Trucking Samples
Before analyzing regression results, it is useful to examine
the pattern of union and nonunion wages and of union density over the
1973-85 period. Table 1 presents descriptivedata on the CPS samplefor
wages, hours, and union membershipof truck driversin the for-hireand
private carriersectors of the industry.Note that overall sample sizes are
reduced after 1978 because the earnings, hours, and union status questions were not asked of the full sample. Perhapsmost notableis the sharp
decrease in union coverage duringthe deregulationperiod. This decline is
most concentratedamongthe approximateone-thirdof the total samplein
the for-hiresector of the industry. Here union density fell from about 60
percent duringthe regulatoryperiod, to about 30 percentby 1984-85. By
Hirsch
Table 1
Mean Real Wages and Hours among Union and Nonunion Truck
Drivers, 1973-85
Driversin For-HireSector
n
P
W,
231
254
228
231
294
.623
.626
.561
.615
.605
12.89
11.87
12.25
12.22
12.87
8.24
8.56
8.49
7.90
8.30
1978 295
.569
12.57
9.02
175
94
84
127
79
111
.566
.564
.607
.504
.304
.288
11.75
11.00
11.61
11.34
11.35
9.83
8.94
8.45
9.29
7.96
8.05
8.37
1973
1974
1975
1976
1977
1979
1980
1981
1983
1984
1985
Driversin PrivateCarrierSector
n
P
Wu
Wn
HRSu
HRSn
51.7
52.8
52.5
54.3
51.8
457
515
460
473
676
.396
.392
.391
.336
.334
10.36
10.34
10.47
10.27
10.72
7.42
7.03
6.88
6.87
7.04
42.6
43.1
43.2
43.1
43.1
46.0
45.5
44.5
44.6
45.9
50.6
53.9
589
.348
10.56 7.44
43.0
45.9
48.9
48.8
46.4
46.9
46.3
47.8
48.9
49.2
45.9
48.8
51.1
47.3
338
213
176
248
145
156
.352
.423
.364
.343
.290
.288
10.57
10.03
10.30
9.94
10.15
11.53
42.7
41.5
42.2
43.3
41.7
42.5
44.6
44.3
43.8
43.7
43.1
44.2
Wn HRSU HRSn
48.4
46.9
45.7
48.5
49.6
7.46
7.22
7.61
7.27
7.47
6.97
Note: n = sample size; P = union density, W = usual weekly earnings (in 1985 dollars,
adjustedby the PersonalConsumptionExpenditurecomponentof the GNP implicitprice
deflator; Economic Report of the President 1987, 248).
contrast, the fall in union density among private carrier truckers was less
marked, from about 35-40 percent during the regulatory period to about
30 percent during 1984-85.
The small sample sizes for the later years makes us cautious in attaching undue weight to the mean in any given year. The sharp decline in
union density is consistent, however, with Teamster membership losses
reported elsewhere. Perry (1986, 110) estimates that the number of Teamsters covered by the NMFA fell by between 18 and 34 percent between
the 1982 and 1985 contracts. Using union financial reports, Troy and
Sheflin (1985; personal correspondence) find that the Teamsters' U.S.
membership, which peaked in 1974 at 1.95 million, had declined to 1.52
million in 1983 and 1.54 million in 1984.1
The pattern of hourly earnings for union and nonunion truckers indicates declining real wages among union truck drivers in the for-hire sector
11. The Troy and Sheflinmembershipdataare dues-weightedannualaverages,whereasthe
CPS figuresare for May. Hence, membershiplosses after, say, May 1983,wouldshow up in
the 1983Troy-Sheflinfigures,but not until 1984in the CPS. Data on Teamstermembership
of course reflect a large numberof workerswho are not truckdrivers.
307
308
The Journalof HumanResources
following deregulation.12 Union real wages, which were relatively stable
and averaged (unweighted)$12.45 duringthe 1973-78 period, averaged
"only" $11.15 duringthe 1979-85 deregulationperiod. No clear-cutpattern is evident amongthe other three groupsof drivers.Nor are there any
systematic changes in average hours worked between the regulatoryand
deregulatoryperiods.13While the descriptiveevidence is informative,inferences about deregulation's effects on wages and the union premium
should not be drawn until we estimate a wage function accounting for
worker and labor market characteristics,as well as comparingtrucking
wage rates with those prevailingelsewhere in the economy.
Table 2 provides regression results for Equation (1), estimated separately for drivers employed in the for-hire and private carrier sectors.
Brief mention of the control variable coefficients can be made. As expected, the coefficients on schooling are low, reflectingnot just a lack of
importanceof schooling for truck drivers, but also the low marginalreturns to individuals within any narrowly defined occupation (i.e., such
samples are self-selected and schooling coefficients fail to measure returns from occupationalmobility). Otherresults indicatethat driversexhibit typical wage-experience profiles, nonwhites have wages approximately 10 percent below whites, veterans exhibit a 3 percent premium,
and marriedmales earn a 4-5 percent premium.While the coefficientson
the regionaldummiesare not shown in Table 2, they indicatethat drivers
in the Pacificand, to a lesser extent, the East North Centraland Mountain
regions, realize significantlyhigher wages than in other regions (these
results are available on request).
12. Real wages are calculatedas usual weekly earnings,in 1985dollars, dividedby usual
hours workedper week. The PersonalConsumptionExpenditures(PCE)componentof the
GNP deflatoris used to convert currentin 1985dollars(EconomicReportof the President,
248). Alternative adjustmentby the CPI, which overstates the "true" rate of inflation,
makes it appear that real wages for truckers and nontruckersalike declined duringthe
eighties. No inferencesregardingderegulationandthe unionpremiumare affectedby use of
the PCE, althoughthe sensitivityof real wages to the choice of a price index reinforcesthe
importanceof employingan economy-widecontrolgroup.The 1985union sampleof truckers appears unrepresentativeto some degree-those in the for-hiresector are atypically
young and have low wages (but see footnote 19), while those in the private sector are
atypicallyexperiencedwith higherwages.
13. Departmentof Transportation(DOT) regulationsplace limits on allowable hours by
drivers.Whilerathercomplex, the gist of the regulationsis to restricthoursto no morethan
ten hours without an eight-hourbreak, and 60 hours over seven consecutive days. It is
generallybelieved that these hours restrictionsare frequentlyviolated, althoughtrucking
firmrecordsdo not report such violations. If truckdriversaccuratelyreporttheirhours in
the CPS survey, such data would allow measurementof adherenceto the regulations.Over
the 1973-85 period, only 6.9 percent of the for-hiresector samplereportover 70 hours of
work the previous week, while 3.5 percent of the private carriersample report over 70
hours.
Hirsch 309
Table 2
Regression Results-Wage EquationEstimatesfor Truck
Drivers, 1973-85
Variable
Constant
Schooling
Experience
Exp2/100
Nonwhite
Veteran
Married
Region
DREG
UN
UN
DREG
R2
n
For-Hire Sector
Private
CarrierSector
1.569 (25.26)
.021 (4.92)
.015 (5.34)
1.415 (42.74)
.020 (8.39)
.020 (13.02)
-.025
-.105
(4.44)
(3.74)
-.033 (11.34)
-.082 (5.51)
.034 (2.06)
.036 (1.56)
included
.026 (1.05)
.392 (19.24)
.025 (2.27)
.050 (3.89)
included
.001 (0.11)
.300 (23.18)
-.128
(3.78)
.279
2,203
-.022
(0.96)
.330
4,446
Note: Itlin parentheses.
Estimates presented in Table 2 do not directly compare the earnings
differentialbetween drivers in the for-hire and private sectors. Such a
comparison,however, indicates a premiumjust under 10 percent among
drivers in the for-hire sector, relative to drivers with similar measured
characteristicsin the privatecarriersector. This may reflecta compensating differentialfor more difficultworkingconditionsamongdriversin the
for-hire sector, unmeasured quality differences between drivers in the
two sectors, or stronger union power in the for-hire sector that raises
wages for both union and nonunion drivers. We return to the issue of
threat effects on nonunionwages subsequently.
Of principalinterest in Table 2 are the coefficients on DREG, UN, and
UN - DREG. Priorto deregulation,we estimate a 48 percent union wage
premium in the regulated for-hire sector, as compared to a 35 percent
premiumin the unregulatedprivate carrier sector. Following deregulation, nonunionwages were little changedin either sector, suggestingthat
threat effects were not importantduring(or have not changed since) the
regulatory period. Most interesting is the evidence on changes in the
union premiumwith the onset of deregulation.The premiumfell significantly in the for-hire sector, from 48 to 30 percent (for equivalent evi-
310
The Journalof HumanResources
dence, see Rose 1987). In marked contrast, the union premiumin the
private carriersector fell only slightlyto about 32 percent. Whereasprior
to 1979, the regulated sector of the trucking industry exhibited higher
union coverage and a largerunion premiumthan the unregulatedsector,
similarunion wage premiumsand union density are observed following
deregulationin both the private carrier and for-hire sectors of the industry.
Despite the substantial deteriorationin Teamster bargainingstrength
and the union's modest contract "gains" in its 1979, 1982, and 1985
contracts, the differentialbetween union and nonunionwages remained
large in both sectors, however. It appearsthat there would have to have
been more substantialconcessions in wages in order to have slowed the
inroads made into the industryby nonunioncompanies and drivers.
VI. Empirical Results: Truckers and Nontruck
Operatives
The analysis presentedabove comparesthe wages of union
and nonunion drivers before and after deregulation. It cannot be concluded, however, that all changes between the two periodsresultedfrom
deregulation. Coinciding with deregulation were substantial increases
(and, subsequently, decreases) in gasoline prices, a sharp recession followed by a recovery, a structuralshift of employmentaway from production and manufacturingjobs toward those in the service sector, and
changes in the economic environmentthat weakened union bargaining
power and decreased membershipeconomy-wide. Hence, it is useful to
compare the wages of union and nonuniontruckersduringthis period to
those of otherwise similarworkers outside of trucking.We have chosen
as a control group a sample of nontransportoperatives.
To examine the issues discussed above, Equation(2) is estimated. The
partialresults presented in Table 3 allow comparisonof wages for four
groups during the regulatory and deregulatoryperiods: nonunion nontruck operatives, union nontruck operatives, nonunion truckers, and
union truckers (i.e., the seven included variables account for all eight
possibilities; the excluded group is nonunionnontruckoperatives during
the pre-1979period). Separateregressionsare estimatedfor driversin the
for-hireand privatecarriersectors (both equationsincludethe same operative control group).14
14. When we calculate an F test comparinga "restricted" regression with all truckers
(includinga for-hireinterceptdummy)with separateregressionsfor driversin the for-hire
and private sectors of the industry, we obtain F(17, 6613) = 2.961. Using the standard
Hirsch
The coefficients on TR show that truck drivers in the for-hire sector
earn about 4 percent more, and drivers in the private sector about 4
percent less, than the control group of operatives. The DREG coefficient,
82, indicates that the real wages of nonunion nontruck operatives grew by
only 1 percent between the 1973-78 and 1979-85 periods. The coefficients
84 on TR ' DREG are close to zero and insignificant, indicating no change
in the relationship between the wages of nonunion truck drivers and the
control group of nonunion operatives during the deregulation period. The
evidence, therefore, provides no support for the hypothesis of significant
industry-specific threat effects, which would have been evidenced by
decreasing nonunion wages in response to lower union density in the
trucking industry. Of course, increased demand for nonunion drivers
might have offset declining threat effects were the supply of nonunion
drivers relatively inelastic, but it is more likely that supply was highly
elastic. Thus, the wages of nonunion drivers appear to have been competitively determined, with changes over time mirroring wage changes of
similar workers outside of trucking.
Equation (2) also allows comparison of the union-nonunion differentials
among truckers and nontruck operatives. The differential among the control group of operatives averaged about 29 percent from 1973 through
1978, and exhibited a small and insignificant decrease during the eighties
(as seen in the coefficients on UN *DREG).15 Relative to nontruck operatives, truckers exhibited a significantly higher union premium during the
seventies (coefficients 86)-about 13 percentage points higher among
drivers in the regulated sector and 6 percentage points higher among those
in the unregulated sector. (The calculated union-nonunion log differentials for truck drivers, alnW/IUNTR, are also presented in Table 3.) Deregulation brought about a significant narrowing of the union premium in the
for-hire sector, however, decreasing the differential by about 10 percentage points (coefficient 87). By contrast, there was virtually no change in
the union premium in the private carrier sector. Following deregulation,
union premiums were similar in magnitude among nontruck operatives
criterion, we reject (at the .05 significance level) the null hypothesis of equal slope
coefficients in the two sectors. However, using the weaker mean squarederror criterion
proposed by Wallace (1972), or a test proposed by Learer (1978, 114) that adjusts for
samplesize, the null hypothesiscannotbe rejected(criticalvalues at the .05 level are 3.081
for the Wallacetest and 9.005 for the Leamertest).
15. Separate estimates on a yearly basis indicate higher premiumsduring 1976-78 than
during1973-75 and variabilityin the union premiumduringlater years. Such results have
been evidentin previousCPS studies. See, for example,Freeman(1986)and, for a surveyof
studieswith datathrough1979,Lewis (1986).Both authorsdiscuss differencesbetweenCPS
estimates and those obtainedfrom other data sources.
311
Table 3
Partial Wage Equation Results Comparing Truck Drivers and Non-Transport Operatives
Region, 1973-1985
Private Carrie
For-HireSector Comparison
Variable
Nation
Northeast
North
Central
South
West
Nation
Northeast
UN
.254
(31.62)
.183
(10.24)
.241
(17.81)
.288
(18.81)
.293
(14.51)
.253
(32.28)
.179
(10.34)
DREG
TR
TR ?DREG
UN DREG
.012
.033
.010
(1.15)
(1.37)
(.50)
-.008
(.46)
.025
.012
.034
(1.03)
(1.17)
(1.46)
.040
.115
.008
.031
.054
(2.73)
.008
(.33)
-.018
(1.19)
(2.85)
-.148
(2.30)
-.042
(1.27)
(.28)
-.023
(.50)
-.022
(.85)
(1.28)
.075
(1.79)
.005
(.17)
(1.60)
.046
(.83)
-.009
(.23)
-.037
(4.05)
-.013
(.77)
-.017
(1.17)
.020
(.88)
-.098
(2.31)
-.041
(1.28)
UN
TR
.124
.074
.164
.142
.046
1973-78
(6.45)
-.108
(3.10)
.319
12,750
.378
(21.32)
(1.54)
.009
(.11)
.251
2,650
.258
(5.72)
(4.98)
-.050
(.85)
.295
4,107
.405
(13.29)
(3.96)
-.167
(2.56)
.306
3,839
.430
(13.18)
(1.00)
-.182
(2.21)
.344
2,154
.339
(8.22)
(3.84)
-.007
(.25)
.313
14,993
.308
(24.75)
(1.09)
.091
(1.47)
.236
3,068
.214
(7.83)
IdnW/UNTR
1979-85
PTR: 1973-78
PTR: 1979-85
.252
(9.69)
.599
.482
.225
(3.66)
.720
.563
.333
(7.56)
.678
.537
.267
(5.46)
.460
.418
.149
(2.43)
.566
.420
.284
(14.58)
.364
.349
.264
(5.70)
.482
.428
UN TR DREG
R2
n
alnW/OUNTR
.055
.035
Note: Itlin parentheses.All regressionsinclude a constant;years of schooling, experience, and experience squ
white, veteran, and maritalstatus. The nationalregressionsincludeeight subregionaldummies,while the large
region dummies. The differentials alnW/dUNTR are calculated as 61 + 85REG + 86TR + 87TR *DREG (see Eq
mean union density of truck driversfor the groupand perioddesignated.The controlgroupsample of nontruc
the for-hireand privatesector comparisonregressions.The omittedreferencecategoryis nontrucknonunionop
314
The Journalof HumanResources
and drivers in both the previously regulatedand unregulatedsectors of
the truckingindustry.16
Furtherinsight into the effects of truckingderegulationis gained from
an analysis of inter-areadifferences in wage determination.Table 3 presents estimates of Equation (2) estimated separatelyfor each of the four
major Census regions.17The bottom two rows of Table 3 also present
average union density among truck drivers in each region during the
regulatory and deregulatory periods. Evident from the coefficients on
UN . TR .DREG is that the significantreductionin the union premium
previously reportedfor the for-hire sector occurredmost stronglyin the
South and West, where levels of nonunionoperationswere highest. For
example, the union wage premiumfor truckersin the South is estimated
to have fallen from 54 percent duringthe regulatoryperiod to 31 percent
following deregulation.A similarpattern is found in the West and, to a
lesser extent, the North Central. By contrast, the union-nonunionwage
differentialamong drivers in the highly unionized Northeast was 25-29
percent duringboth periods. These results support the view that in the
absence of regulatoryprotection, union truckingwas most able to maintain its regulatory-periodwage differential in regions with high union
density, and least able to maintainthese premiumsin regions with extensive nonunion operations. The decrease in union density in all regions,
however, suggests weakened union bargainingpower and continuedpressure to narrowunion-nonunionwage differentials.
Closely relatedto the above is the findingof evidence suggestingthreat
effects (or nonunion rent sharing)among nonuniondrivers in the highly
unionized Northeast. Whereas nonunion drivers in the Northeast received a 12 percent premiumrelative to the control group of operatives
duringthe regulatoryperiod, this premiumwas eliminatedfollowing deregulationand the accompanyingdecrease in union density from 72 to 56
percent. No evidence of threateffects is found in the other three regions.
Results from the largely unregulatedprivate carriersector indicate that
deregulationgenerally had similar qualitative, but smaller quantitative,
effects in this sector. Coefficientestimates at the regionallevel have large
standarderrors.
In a separateanalysis not shown (but availableon request), a variantof
16. The logarithmicunion-nonunionwage differentialis calculatedby IdnW/IUN= 81 +
85DREG+ 86TR + 87TR?REG for truckdriversand nontruckoperatives,beforeand after
deregulation.
17. F tests comparingthe "restricted"nationalregressionswith the "unrestricted"regional
regressionsare F(39, 12689) = 3.523 in the for-hiresector and F(39, 14932)= 4.929 in the
privatesector. The null hypothesisof equal slope coefficientsacross regionsis rejectedfor
both sectors using standardcriteriaor the Wallacetest, but cannotbe rejectedin eithercase
using Leamer'stest (see footnote 14).
Hirsch 315
Equation (2) was estimated in which individual year dummies are interactedwith UN, TR, and UN *TR. This provides separateestimatesfor
the four groups (union/nonuniontruckers/nontruckers)
for each of the 12
sampleyears. Whilethere exists year-to-yearvariabilityin results, particularlyfor the later years with small sample sizes, the overall picture that
emerges is largely identical to that presented here. One result evident
from the yearly analysis, however, is that there was a significantdecline
in the wages of both nonunion truckers and nontruckoperatives in the
years 1983-85, as comparedto 1979-81. The suggestionis that the 198182 recession, in conjunctionwith structuralchangesin the economic environment,broughtabout a significantand long-liveddecline in real wages
among male operatives. Annual estimates of the union premiumamong
for-hire sector drivers indicate that it fell during 1979-81, reboundedin
1983-84, and dropped sharply in 1985.18
While it is evident that Teamster bargainingpower and the unionnonunion differential in the for-hire sector declined relative to what it
would have been in the absence of deregulation,it remainedlarge. The
ability of the IBT to maintain sizable wage differentials, albeit for a
significantlyreducedpartof the industry,may have stemmedin partfrom
the ability of some large carriersto increase labor productivityby taking
advantageof increases in size and weight limits, by adoptingnew equipment technology, and because of union work rule concessions.
Nevertheless, the union premiumsrealized by truckersand nontruck
operatives during the eighties remained substantial;significantlyhigher
than, say, the historic premium found for production workers in the
manufacturingsector (Lewis 1986, Section 7-5). In the face of continued
sizable wage differencesbetween union and nonuniondrivers,the shift of
freightto nonunionfirmsand drivershas been large. Whilelimitedadjustments had begun prior to deregulation,administrativeactions in the late
1970s and the MCA of 1980 clearly facilitated such shifting on a large
scale. For there not to have been such an increasein nonuniontrucking,a
18. Annualcross-section log wage regressionsfor the for-hiresector (containingthe same
controlvariablesas in Equation1) producedthe followingestimatesof the union-nonunion
wage differential(Itlin parentheses):
1973:.470 (7.88)
1983:.324 (5.26)
1976:.373 (7.40)
1979:.244 (3.87)
1974:.337 (6.27)
1977:.431 (10.22)
1980:.156 (1.64)
1984:.360 (4.02)
1975:.355 (5.69)
1978:.353 (6.94)
1981:.170 (2.02)
1985:.106 (1.10)
National MasterFreightAgreements(NMFAs) went into effect in April 1973, 1976, 1979,
1982,and 1985. Counterto the above evidence on union premiums,contractprovisionsin
the 1979NMFA generallyhave been regardedas generousto drivers,whereasprovisionsin
the 1982NMFA are consideredless favorable(for a summaryof contractprovisions, see
Perry 1986and Rose 1987).
316
The Journalof HumanResources
significantnarrowingof the union-nonuniondifferentialwould have had
to occur immediatelyfollowing deregulation.Despite some concessions
and the moderatenature of the 1982and 1985NMFA contracts, at least
by historical standards,wage differentialsbetween union and nonunion
truck drivers remainedlarge following deregulation.
VII. Conclusions
The regulatoryexperience in the truckingindustryprovides
an ideal area in which to study the creation and dissolutionof economic
rents resulting from entry, price, and operating restrictions. Previous
studies (Moore 1978, Frew 1981, Rose 1985) have provided compelling
evidence that a portion of the regulatoryrents were capturedby trucking
firm owners and the initial owners of route certificates.Additionalrents
arising from deregulationwere reflected not in higher profits or in the
market value of operatingrights but, rather, in the form of higherlabor
costs. The evolution of the powerfulTeamstersunion duringthe periodof
ICC regulation facilitated the capture of sizable wage gains for union
members and, to a much lesser extent, nonuniondrivers. The deregulation experience provides evidence as to the magnitudeof the regulatory
rents that previously accrued to labor and the speed of adjustmentto
competitive wage levels.
Regression analysis on the wages of union and nonuniontruck drivers
from 1973-1985indicates a significantdecrease in wages for union drivers
in the for-hire sector of the trucking industry following deregulation.
Comparisonof truck drivers' wages with those of a nontruckoperative
control group reveals that changes in the real wages of nonuniondrivers
were very similar to those of operatives outside of trucking. While the
union-nonunionwage differentialin truckingremainedhigh throughout
the period, the union premiumfor truckersin the previouslyregulatedforhire sector of the industrynarrowedsignificantlyfollowing deregulation,
particularlyin the less highly organized South and West regions. Evidence from the unregulatedprivatecarriersector of the industryindicates
little narrowingof the union premium.
The evidence is consistent with the view that a largecompetitivesector
existed in both sectors of the trucking industry prior to deregulation.
Nonunion truckdriversexhibitedreal wage rates close to those of similar
nontransportoperativesboth before and afterderegulation.Whilederegulation had significanteffects on entry, shippingrates, traffic,firmoperating ratios, union wage rates, and the like, the wages of nonuniondrivers
reflectedthe opportunitycost of labor, as measuredby wage rates outside
Hirsch
of trucking. By contrast, deregulation sharply limited the long-run ability
of firms to continue maintenance of costly unionized trucking operations.
The significant cost pressures brought to bear upon union wages have
produced a narrowing in what was a very large union-nonunion wage
differential.
Perhaps the clearest evidence of deregulation's effect on the trucking
industry is the sharply declining union density and Teamster organizing
strength in recent years. One possible interpretation of this evidence,
given the maintenance of a large union wage premium (at least through
1984), is that the Teamsters' concessions were much too little and late to
stem the large shift away from union drivers. Indeed, one could argue that
any radical change in the economic environment, such as that brought
about by trucking deregulation, places strains on a union. Even if union
leaders anticipated and fully understood the implications of deregulation,
there existed a lag before the rank-and-file shared this realization. Hence,
union leaders may not have been able politically to engineer and gain
approval of the types of contractual arrangements that would have maintained union organizational strength.19 An alternative, but not mutually
exclusive, interpretation is that union rank-and-file fully understood the
implications of trucking deregulation, but that senior members rationally
chose to maintain a high-wage strategy in spite of the dire consequences
on membership.20
References
Blair, Douglas H., and David L. Crawford.1984. "Labor Union Objectivesand
Collective Bargaining."QuarterlyJournal of Economics 99(3):547-66.
Boyer, Kenneth D. 1987. "The Costs of Price Regulation:Lessons from
RailroadDeregulation."Rand Journal of Economics 18(3):408-16.
Daughety, Andrew F., and Forrest D. Nelson. 1986. "An Econometric
Analysis of Changesin the Cost and ProductionStructureof the Trucking
Industry, 1953-1982." Economics WorkingPaperNo. 86-41, University of
Iowa, December.
19. Perhapsthe strongest evidence of resistance by rank-and-fileto Teamsterleadership
came in the September1983defeat of a nationalriderto the NMFA thatwouldhave created
a two-tierwage andbenefitstructure.The defeatwas by a marginof morethanseven to one
(Perry 1986, 107). Wage discountingfor new hires was introducedinto many of the 1985
contractagreements.
20. Senior membersmightbe expected to preferlargewage cuts for new hires, ratherthan
moderatecuts for all workers, unless they expect a two-tier system to lead to subsequent
across-the-boardconcessions. For a formalanalysisof votingwithina collective bargaining
framework,see, among others, Blair and Crawford(1984).
317
318
The Journal of Human Resources
Ehrenberg, Ronald G. 1979. The Regulatory Process and Labor Earnings. New
York: Academic Press.
Freeman,RichardB. 1986. "In Search of Union Wage Concessions in
Standard Data Sets." Industrial Relations 25(2):131-45.
Frew, James R. 1981. "The Existence of Monopoly Profitsin the Motor
Carrier Industry." Journal of Law and Economics 24(2):289-315.
Giles, David E. A. 1982. "The Interpretationof Dummy Variablesin
SemilogarithmicEquations:Unbiased Estimation."Economics Letters
10:77-79.
Hendricks, Wallace. 1977. "Regulationand Labor Earnings."Bell Journal of
Economics 8(2):483-96.
-
. 1986. "Collective Bargainingin RegulatedIndustries."In Advances in
Industrial and Labor Relations, Vol. 3, ed. David B. Lipsky and David
Lewin, Greenwich, CT: JAI Press: 21-42.
Hirsch, BarryT., and John T. Addison. 1986. TheEconomicAnalysis of
Unions: New Approaches and Evidence. Boston: Allen & Unwin.
Hirsch, BarryT., and John L. Neufeld. 1987. "Nominal and Real Union Wage
Differentialsand the Effects of Industryand SMSA Density: 1973-1983."
Journal of Human Resources 22(1):138-48.
Kim, Moshe. 1984. "The Beneficiariesof TruckingRegulation,Revisited."
Journal of Law and Economics 27(1):227-41.
Leamer, Edward E. 1978. Specification Searches: Ad Hoc Inference with
Nonexperimental Data. New York: John Wiley.
Levinson, Harold M. 1982. "Trucking."In CollectiveBargaiining:
Contemporary American Experience, ed. Gerald G. Somers. Madison, Wis.:
IndustrialRelations Research Association.
Lewis, H. Gregg. 1986. Union Relative Wage Effects: A Survey. Chicago:
University of Chicago Press.
MaCurdy,Thomas E., and John H. Pencavel. "Testing Between Competing
Models of Wage and EmploymentDeterminationin Unionized Markets."
Journal of Political Economy 94(3, Pt. 2):S3-S39.
Moore, Thomas Gale. 1978. "The Beneficiariesof TruckingRegulation."
Journal of Law and Economics 21(2):327-43.
. 1983. "Rail and Truck Reform-The Record So Far." Regulation
(November/December):33-41.
Moore, WilliamJ., Robert J. Newman, and James Cunningham.1985. "The
Effect of the Extent of Unionism on Union and Nonunion Wages." Journal
of Labor Research 6(1):21-44.
Perloff, Jeffrey M., and Michael L. Wachter. 1984. "Wage Comparabilityin
the U.S. Postal Service." Industrial and Labor Relations Review
38(1):26-35.
Perry, Charles R. 1986. Deregulation and the Decline of the Unionized
TruckingIndustry.Philadelphia:WhartonSchool's IndustrialResearch Unit.
Rose, Nancy L. 1985. "The Incidence of RegulatoryRents in the Motor
Carrier Industry." Rand Journal of Economics 16(3):299-318.
---
. 1987. "Labor Rent-Sharingand Regulation:Evidence from the
Trucking Industry." Journal of Political Economy 95(6):1146-78.
Hirsch
Troy, Leo, and Neil Sheflin. 1985. Union Sourcebook: Membership, Structure,
Finance, Directory. West Orange,N.J.: IndustrialRelationsData and
InformationServices.
U.S. Council of Economic Advisors. 1987. Economic Report of the President.
Washington:GPO.
Wallace, T. D. 1972. "WeakerCriteriaand Tests for LinearRestrictionsin
Regression." Econometrica40(4):689-98.
319