Copyright
Nancy L Rose.

Labor rent-sharing and regulation : evidence from the trucking industry online

. (page 3 of 4)
Online LibraryNancy L RoseLabor rent-sharing and regulation : evidence from the trucking industry → online text (page 3 of 4)
Font size
QR-code for this ebook


declined somewhat in 1983-84, in sharp contrast to the trucking pattern. No
pattern in the union premia in the printing industry emerges from the data. The
estimated premia vary considerably from year to year, and the post-1978 average
level is only slightly below the pre-1978 average level.

Although it is difficult to draw decisive conclusions from these comparisons,
it does not appear that the decline in the union wage premium in the trucking
industry is simply an artifact of the particular data set used. The pattern
observed in the trucking industry indicates a fail- amount of stability in the union
premium over the regulatory period, followed by a dramatic and lasting reduction
in the premium in 1979 and thereafter. This decline is consistent with the
contract evidence and aggregate wage data evaluated in section I, although the
CPS data suggest that deregulation may have reduced the union's relative wage
advantage before the effects show up in contract negotiations and aggregate
earnings indices.



28
III. Nonunion Rent-Sharing

The final possibility I investigate is nonunion rent-sharing. The analysis
thus far has measured the decline in union wages relative to nonunion wages in
the trucking industry. If nonunion workers also shared in the rents available in
the regulated trucking industry, this comparison will tend to understate both
union rents and total labor rents. An emerging literature in labor economics
suggests that nonunion rent-sharing is a distinct possibility.

A number of recent papers have investigated the existence of inter-industry
wage differentials and the role of rent-sharing in explaining their persistence.
Krueger and Summers (1986a, b) and Dickens and Katz (1986a, b) provide empirical
evidence on the substantial dispersion in wages across industries. These studies
report high correlations of industry wage differentials across occupations,
countries, and union and nonunion sectors. The authors consider a variety of
possible explanations for their findings, including efficiency wage models and
possible union and nonunion rent-sharing. Their findings are generally consistent
with noncompetitive models of wage determination, although they do not provide
clear evidence supporting a single theory.

Recent theoretical models also predict nonunion rent-sharing. Dickens (1986)
provides a model of union threat effects, in which unorganized workers are able
to capture higher wages as the firm tries to deter workers from organizing a
union. In Dickens's model, as long as nonunion wages are near union wage levels,
workers will have no incentive to bear the cost of organization. Firms avoid the
costs associated with having a unionized workforce by paying workers enough to
eliminate workers' net benefits from organizing a union An alternative model,
which does not rely on the presence of a unionized sector in the industry, is



29
proposed by Rotemberg and Saloner (1986). They construct a game-theoretic
model in which incumbent workers are more valuable than new hires, technology
is characterized by putty-clay features (so that the capital-labor ratio is fixed ex
post), and workers have some bargaining power vis-a-vis firms. In this model,
unorganized workers are able to capture a share of potential rents and quasi-rents
in their wages. The model predicts that wages will be a function of profitability
and the capital-labor ratio.

The studies suggest the desirability of investigating nonunion rent-sharing in
the trucking industry. An analysis of nonunion wage r-esponses to deregulation is
important not only for understanding the distribution of rents in the trucking
industry, but also for evaluating the nonunion rent-sharing hypotheses proposed in
these recent labor studies. To evaluate the extent of nonunion rent-sharing, we
need a benchmark against which to measure nonunion wages; I use an economy-
wide average wage for this purpose. These are estimated from separate union and
nonunion equations for all private sector employees reported in the May CPS for
each year. The equations ?ire similar to (1), with occupational, regional, and one-
digit industry fixed effects. The all-industry average wage is computed as

the wage for a blue collar worker with the characteristics described for the
representative driver in table 4, based on the average of the industry fixed
effects. Estimated union wages are included for comparison.

These wages are reported in table 6, along with the predicted wages for

n 1

union and nonunion trucking drivers from table 4.' The results suggest declines



I am grateful to Henry Farber for making his moment matrices of these
data available to me. Because of minor differences in the data construction,
the wage equation for the all-industry regressions differs slightly from (1).

The conclusions discussed below are not materially affected if the
table 4 trucking wage estimates are replaced by estimates from separate union
and nonunion trucking wage equations for each year This robustness is not



30
TABLE 6

COMPARISON OF TRUCKING WAGES TO AVERAGE ECONOMY WAGES







NONUNION




UNION


YEAR


ALL-INDUSTRY

AVERAGE TRUCKING


ALL- INDUSTRY
AVERAGE


TRUCKING


1973




$5.34




$4.06


$5.68




$6.26


1974




5.70




4.53


6.06




6.36


1975




6.17




4.56


6.23




7.09


1976




6.39




4.88


6.89




7.48


1977




6.72




5.45


7.49




8.38


1978




7.04




6.24


8.33




8.88


1979




7.68




7.23


7.93




9.00


1980




8.21




7.39


8.81




9.27


1981




9.50




8.64


9.59




10.62


1983




10.71




8.03


11.65




10.88


1984




10.38




8.63


13.30




12.01



Trucking wage estimates reproduced from Table 4. All-
industry average wage estimates are for a married white
male with 12 years of education and 20 years of experi-
ence, and are estimated from separate union and nonunion
wage equations including all private sector employees.
Equations control for sex, education, experience, marital
status, and race, and include occupational, regional, and
one-digit industry fixed effects. All-industry averages are
based on average of industry fixed effects, and are
evaluated for blue collar workers.



31

in both union and nonunion trucking wages relative to the all-industry averages
during the deregulation period. These patterns are clearly evident in figures 3a
and 3b, which plot trucking and all-industry wage? for nonunion and union
workers, respectively. Until 1979, nonunion trucking wages rise steadily relative
to average nonunion wages. After 1979, trucking wages fall increasingly behind
average nonunion wages. All-industry average nonunion wages increase 35 percent
between 1979 and 1984, as compared to only 19 percent for the trucking industry.
The nonunion wage gap rises from an average of S.63 per hour in 1978-79 (8
percent of industry average nonunion wages) to an average of S2.21 per hour in
1983-84 (21 percent of industry average nonunion wages) ' "• Similar patterns are

no

evident in the union wage comparisons.

These results should be interpreted with some caution The sample sizes of
nonunion trucking drivers are quite small, averaging only 35 to 45 workers in the
1980s. These suggest quite substantial standard errors on the wage estimates. In
addition, the magnitude of the trucking wage decline depends critically upon the
period over which changes are measured. Because trucking wages gain relative to
average wages during the 1970s, using 1978 or 1979 as the regulation base year
and comparing changes through 1983 or 1984 will lead to larger estimates of
nonunion rents than would be predicted using earlier years as the regulation base.



surprising, given the pattern of sample mean wages reported in table 2.

Nonunion trucking wages are roughly the same percentage of all-industry
average nonunion wages in 1983-84 that they are in 1973-74. This may somewhat
complicate the interpretation of the post-1979 increase.

TO

The rate of increase in union trucking wages is dramatically lower than
that for all-industry average union wages after 1979. There is a 67 percent
increase in all-industry average union wages between 1979 and 1984, compared to
a 33 percent increase in trucking wages over the same period. This disparity
is sufficient to reverse the historic superiority of trucking wages; by 1983,
union trucking wages have for the first time in the sample fallen below all-
industry average union wages.



FIGURE 3
COMF/yv:SOS CF TRIXKING WAGES WITH AL'^- ISDl'STRV AVERAGE WAGES

3A; NOKUNION WORKERS




ALL-INDUSTRY



TRUCKING



1 673 1S7< 1676 1»7e 1 fr7

D TRUCKING



1676 1676. 19eC 1961 1 &53 1&64

AU. INPU5TFY AVE



3B: UNION WOR};ERS




ALL-IN2L'STRY



TRUCKING



INli'J'I-'^T Av^



33

Witti thesp caveats in mind, however, the findings nevertheless appear to support
the hypothesis that nonunion, as well as union, trucking drivers shared in the
rents available under motor carrier regulation.



34
IV. Aggregate Union Losses

The magnitude of the rents implied by these results is substantial. To put
the findings into context, I calculate a rough measure of the total losses resulting
from deregulation. I restrict this calculation to rents captured by the Teamsters
Union, because of the difficulty quantifying nonunion rents. The calculation is
summarized in table 7. A number of assumptions are required to compute
aggregate union rents. First, I assume that deregulation reduced the 1973-78
union premium of 50 percent to the 1983-84 average premium of 39 percent. The
sensitivity of the estimate to this assumption is checked by recalculating losses
assuming a decline to the 1979-84 average premium of 30 percent. Second, I
assume that nonunion wages are unaffected by deregulation. This will understate
both union and nonunion rents to the extent that nonunion wages were bid up
under regulation. My calculation attributes the entire change in the union
premium to a reduction in rents, but omits any change due to a decline in the
overall trucking wage level. Finally, I exclude potential losses from a reduction
in union employment. There is substantial anecdotal evidence that Teamster
unemployment rates in the trucking industry have increased substantially since
deregulation, and that nonunion carriers have captured an increasing share of the
market. My computation of rent reductions excludes the union's loss from jobs

that are no longer held by union employees.



^"^ The Wall Street Journal (October 9, 1984, p.1) reports that the
Teamsters Union estimates 100,000 jobs have been lost since 1980. Lieb (1984)
and Hendricks (forthcoming) discuss the effects of increased nonunion competi-
tion on union employment. Although it is difficult to disentangle the role of
regulatory reforms from that of the 1981-82 recession, the persistence of such
high union unemployment rates appears to be considered abnormal by industry
observers.



35

TABLE 7
Aggregate Losses and Division of Rents



1. Aggregate losses for union

A. Assumptions;

■ union premia declines from 50% to 39%

sensitivity check; union premium dpclines from 50% to 30%

■ nonunion wages are unaffected by deregulation

■ unionization rate is constant at 60%



B.



Result for Class I intercity motor freight carriers in 1983;
886 firms (revenues greater than $5 million)



Aggregate employee compensation;
Implied union compensation;
Implied union loss;
Implied \inion loss at 30%;



$12.24 biUion

8.27 billion

657 million

1.24 biUion



Division of rents between firms and union; general freight carriers

A. Assumptions:

■ above assumptions l.A. for union rent share

■ present discounted value (pdv) of firm rents equal to 8.8% of
1978 aggregate revenue for general freight carriers

(from Rose, 1985a)

■ 10 percent discount rate, 50 percent corporate income tax
rate

■ rents are perpetuity

B. Result for Class I intercity general freight common carriers in
1978, 345 firms (revenues greater than $3 million)



Aggregate revenues:
Implied pdv of firms' rents:
Implied annual pre-tax firm rents:

Aggregate employee compensation;
Implied union compensation;
Implied annual union rents at 39%;
Implied annual union rents at 30%;

C. Comparison of rent shares, 1984 dollars



$17.50 billion
1.54 biUion
308 million

8.82 biUion
6.11 bUlion
448 mlUion
814 mlUion



Annual pre-tax firm rents;
Annual pre-tax union rents:
Union share:
Total rents/Total revenues:



$490 miUion
$713 miUion - $1.3 bUlion
59 - 73 percent
4.3 - 6.4 percent



36

The loss to an individual union driver is estimated in the results above at 8
to 15 percent of his current compensation, which is the decline relative to what
his compensation would have been had the 50 percent union differential been
maintained. The total union loss is estimated by aggregating over all Class I

motor carriers, which are regulated firms with more than $5 million in annual
gross revenues. Although this group excludes a high proportion of trucking firms
(which are quite small), it comprises the very largest companies in the industry,
accounting for half or mote of total revenues. Applying the assumptions
described above to 1983 total employee compensation for Class I motor carriers
implies union compensation of $8.27 billion, and an estimated annual reduction of
$657 million in rents for unionized employees of these firms. If the average

deregulation union premium of 30 percent is used in the calculations, the
estimated union loss rises to $1.24 billion.

Finally, the reduction in union rents can be compared to the losses incurred
by owners of trucking firms, to estimate the relative shares of rents for
unionized labor and capital. Rose (1985a) estimates the deregulation- related loss
in market value for a sample of general freight trucking firms at 8.8 percent of
1S78 revenues. Applying this estimate to revenues for all 1978 Class 1 general
freight common carriers yields an estimated after-tax loss of $1.54 billion in
present discounted value. This corresponds to roughly $308 million in annual pre-
tax rents in 1978, or $490 million in 1984 dollars. Applying the calculation



The loss. A, for a 39% premium is given by: A - (1.39 - 1.50) ■

nonunion wage - -.11 nonunion wage. The percent loss is A/union earnings =
-.11/1.39 = -.08.

Total employee compensation of $12.24 billion for the 886 Class 1 motor
carriers reporting to the ICC is taken from the U.S. Interstate Commerce Commis-
sion Bureau of Accounts, Transport Statistics in the United States. Motor
Carriers. Part 2 , 1983.



37

described above to employee compensation for these firms yields an estimate of
$448 to $814 million in pre-tax union rents in 1978, or $713 million to $1.3 billion
in 1984 dollars. These calculations, while crude, suggest that the Teamsters
Union may have been the dominant beneficiar-y trucking regulation capturing G0%



or more of the total rents in the industry.



37



Total rents are measured as union rents plus firm rents. This excludes
rents captured by nonunion labor and other factors of production. This finding
is consistent with the estimated union share obtained in a Tobin's q type of
model of trucking firms; see Rose (1985b). The union share is strikingly
similar to Salinger's (1984) estimate of a 77 percent average union rent share
in his cross-industry studying employing Tobin s q.



38
Conclusion

This study examines industry wage responses to motor carrier deregulation.
Union contract evidence and aggregate data on industry average earnings suggest
substantial changes in historical wage determination patterns after motor carrier
deregulation. Microdata estimates of union premia over nonunion wages in the
trucking industry indicate declines of 30 to 40 percent in the size of the union
differential beginning in 1979. Annual earnings for a "representative" union
driver in 1983-1984 were at least $2200, or 8 percent, less than they would have
been had the regulatory union wage differential been maintained. These results
suggest considerable rent-sharing by union workers. The annual loss in union
rents is estimated at $657 million to $1.24 billion for- employees of the 886 largest
regulated trucking firms in 1983. The results also indicate a decline in nonunion
trucking wages relative to economy-wide wage levels. Although the paper does
not precisely quantify this effect, the result provides some support for models of
nonunion rent-sharing. Further research in this area seems desirable.

Comparison of the results in this paper with estimates of rents accruing to
owners of capital indicate that the Teamsters Union may have been the primary
beneficiary of regulatory rents in the trucking industry, capturing as much as 60
to 70 percent of total rents. The findings suggest that the assumption of
competitive factor prices can be quite misleading, and provide strong support for
labor rent-sharing hypotheses.



to a>

fSl fM

in m



M o



39





^-^




^^


r-


fSI


a>


in


n


»-i


o


O


i—t


rg


fO


*-H


o


o


o


O


O


O


o


o



O 00

o o



T ro


CM \0


O \D


vO .-(




Tf -1


r- o


(Nj a>


^ -1


0^ 00 CT^


O -1


f-< f-i


-1 o


O


CO CO


^ ro


O O


r-l in


CO o


-1 T


in


vO


in r^




O ro


en CO


•-" CM


(N r-H


o o


ro O


CM (N


vX) CM


00


o


CO r-l


,-t a^ a^


'* CO


rH O


o o


O O


o o


O rH


O r-l


rH rH


o


rH


O -1


Cr> r-l rt



LU
_J
CD
<



X

Q

Z

UJ
Q.
C
<



c
o

a

LU

u

CD



D
C
C
<



>-^




^-^




, K




, — »


CM


CM




^-^




^-^




,. — .




^r^




,r^










CO 'S-


CM


in


CO


^


CM


00


O


O


r~


CM


ro


rH


iO


CO


CO


1^


r^


o








(


vO O


rH


in


o


rH


rH


o


o


o


c


r-


in


CO


•&


r~


en


vO


ro


CO


o


vO


»*•


in


vO CM


CM


o


o


o


O


o


o


o


CM


o


f—t


O


f-i


o


o


O


rH


o


rH


CM


CO




r-^














1




1




1




1




1












rH


4-1

c


> —




- —




^^




^^


in


CM




> V




^"^




^^




,-^




^-v








4)


O CO


^


iD


r—i


CM


CO


r-


o


C


CM


o


CTi


in


f—i


CM


CO


^


in


00








U


CO vD


^


^


rH


rH


CM


o


o


O


CO


lO


cn


r-


•^


lO


^


in


CM


in


'-~l


o


ro


(0


CM rH


ro


O


o


O


O


o


o


O


O


o


r-i


o


o


o


o


o


O


o


en

CM


CO


o


a


rH














1








1




1








t








ro


c

•H

in







^— V




^^




^-v




1 3

Online LibraryNancy L RoseLabor rent-sharing and regulation : evidence from the trucking industry → online text (page 3 of 4)