1,. 12‘

 

 

OI‘!‘ 303

a , iiiiiiiiiiii

This is to certify that the

dissertation entitled

PUBLIC POLICY CHANGES AND THE VIABILITY OF
LOCAL COLLECTIVE BARGAINING RELATIONSHIPS ACROSS
MANUFACTURING INDUSTRIES DURING THE 19805

presented by

MATTHEW M . BO DAH

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Socia] Sciences

W

Major professor

DateW7I’/m‘
I

MSU it an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

 

 

 

 

 

 

 

 

IF‘II

 

LIBRARY

Michigan State
Unlverslty

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES Mum on or before date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSUIsMAmmdh‘ ' '1 "r, A ',2.....'2.‘.1...

 

 

.~ W “"“3

0‘
It‘

.‘1
L‘“ V.

    
 
 
 
 
  
  
  
  

mic FQLICY CHANGES AND THE VIABILITY OF
mean BARGAINING RELATIONSHIPS ACROSS

WBING INDUSTRIES DURING THE 19808
By
Matthew M. Bodah

‘z

study "31'.
WNEWHA ad"
in nga-.: .

m flrtab ~'.

tflv

 
 
   
 
 

5E:

“mum"; "

' A W a!"
t‘ ;A DISSERTATION

. to bark}
I N,

0: “ha :5 Submitted t0

5 w‘, Michigan State University

. . .; rfglfillmant of the requirements
s..- for the degree of

tho pol!" .
:3! A DOCTOR OF PHILOSOPHY

.{Ol‘y (51.11%)
zgf Labor and InduStrial Relations
1.: - '

0991133 1' I M
5"».

   
   
  
  
   
    
  
 

  

1996
Of DI’PSQV‘Jn' . 5 '.'., 1+“

  

5,

  
   

)nCIudinQ van 1.32%; ea (stream: L319

“ions. lhp'u Swat. Mutton;

of the Augl‘ir;-- :. (that-[1.2.5 L5 '.’.- exazuue

firm

and

 

ABSTRACT
PUBLIC POLICY CHANGES AND THE VIABILITY OF
LOCAL COLLECTIVE BARGAINING RELATIONSHIPS ACROSS
MANUFACTURING INDUSTRIES DURING THE 19808
BY

Matthew M. Bodah

In this study the author examines the impact of public policy
changes under the National Labor Relations Act during the 1980s on
local collective bargaining relationships in fourteen.manufacturing
industries. The areas of policy studied relate to a model of local
collective bargaining unit disappearance, specifically: an
employer’s duty to bargain over capital decisions, union security,
and standards of behavior in decertification elections. Political
and institutional changes at the National Labor Relations Board,
which preceded the policy changes, are also examined.

A An event history analysis using a pooled time—series
regression model is applied to data across the fourteen industries
measured in four waves of observations between the mid- 19708 and
late- IOOOS. The purpose of the regression analysis is to examine
the stability of a model including variables representing firm

‘ .9Q‘|‘-‘ :Qia'.<‘
failures, plant relocations, input substitution, and

 

 

decertification. The main hypothesis of the study is that
bargaining unit disappearances will be linked to industry economic
distress early in the period, but more closely associated with
relocations, input substitutions, and decertification elections
later in the period as mid-19805 policy changes take hold. Results
nay assist in resolving the debate between those who argue that
union decline is largely the result of structural economic changes
and those who point instead to increasingly unfavorable public
policy.

Qualitative findings indicate that indeed there were sharp
reversals in most of the salient policy areas following the Reagan
administration’s ideological reordering of the National Labor
Relations Board in 1983. The quantitative findings show that firm
failure rates across industries were a strong significant predictor
of bargaining unit disappearances for the entire period, for the
period following 1980, but not for the period following 1983.
After 1983, the evidence suggests that disappearances may actually
have been more typical in relatively healthy industries, where
capital investments and changes in the cross—state distribution of
firms were occuring. The lack of nwre recent data makes it
difficult to arrive at more definite conclusions.

The study ends with a discussion of the unsettled question of
the appropriate role of the federal government in the development

of labor relations policy.

   

AC KNOWLEDG EMENTS

I would like to thank everyone who assisted me in the
completion of this dissertation. First, my thanks go to my
committee: Richard Block, Joel Cutcher—Gershenfeld, Edilberto
Montemayor, Lisa Fine, and Harry Perlstadt. But, indeed, I
thank everyone at Michigan State: the faculty and staff of
the School of Labor and Industrial Relations, friends and
acquaintances in the Departments of Sociology and History, and
particularly my colleagues in the doctoral program, especially
my office mates Pat McHugh and Victor Nichol.

I also owe a large debt of gratitude to my current
colleagues at the University of Rhode Island who have
supported me throughout this endeavor: Ted Schmidt, Mary
Pinch, Scott Molloy, Carl Gersuny, Larry Rothstein, Mark
Grossman, Laura Beauvais, Suzanne Taylor, and Richard
McIntyre, to name just a few.

Further, I must acknowledge the Support of my family: my
parents William and Teresa, my sister Paula, and my brothers
Henry and Timothy.

Rinally, I thank Theres for her kindness, patience, and

understanding.

 

 

TABLE OF CONTENTS

LIST OF TABLES

CHAPTER I. INTRODUCTION AND OVERVIEW
INTRODUCTION

THE IMPORTANCE OF LOCAL COLLECTIVE BARGAINING

RELATIONSHIPS
THE IMPORTANCE OF LAW AND PUBLIC POLICY

A FRAMEWORK FOR STUDYING THE DISAPPEARANCE OF

LOCAL RELATIONSHIPS
RESEARCH OBJECTIVE
THE CHAPTERS

CHAPTER II. LITERATURE REVIEW
INTRODUCTION .
UNION GROWTH AND DECLINE
ECONOMIC EXPLANATIONS
POLICY EXPLANATIONS
INTERNATIONAL AND COMPARATIVE
RESEARCH .
DOMESTIC RESEARCH
BRIDGING THE LITERATURE

CHAPTER III. INSTITUTIONAL CHANGES
INTRODUCTION
A FORESHADOWING OF CONFLICT
OPEN CONFLICT AT THE BOARD
IMPACT OF THE NEW BOARD
CONCLUSIONS

CHAPTER IV. DOCTRINAL CHANGES

INTRODUCTION .

THE DUTY TO BARGAIN IN GOOD FAITH
AN EMPLOYER’ S DUTY TO BARGAIN OVER
DECISIONS AFFECTING THE BARGAINING UNIT
MOVING WORK DURING THE LIFE OF AN
AGREEMENT . .

THE LOSS OF CERTIFICATION . .
QUESTIONS OF MAJORITY SUPPORT
UNION DISCIPLINE

ii

H

\Infi

12
18

20
20
20
22
28

28
3O
34

37
37
37
50
59
69

71
71
72

76

93
97
97
101

 

STANDARDS OF CONDUCT DURING

DECERTIFICATION ELECTIONS . . . . . . . . 105
CONCLUSIONS . . . . 101
CHAPTER V. INDEPENDENT VARIABLES AND HYPOTHESES . 112
INTRODUCTION . . . 112
BUSINESS DECISIONS AFFECTING THE BARGAINING
UNIT . . . . . . . . . . . . . . . . . . . . . 113
PLANT CLOSINGS . . . . . . . . . . . . . 113
PLANT RELOCATIONS . . . . . . . . . . . . 122
SUBCONTRACTING . . . . . . . . . . . . . 125
TECHNOLOGICAL CHANGE . . . . . . . . . . 129
LOSS OF CERTIFICATION . . . . . . . . . . . . 132
SUMMARY . . . . . . . . . . . . . . . . . . . 134
HYPOTHESES . . . . . . . . . . . . . . . . . . 135
CONCLUSIONS . . . . . . . . . . . . . . . . . 144
CHAPTER VI. DATA AND METHODS . . . . . . . . . . . 147
INTRODUCTION . . . . . . . . . . . . . . . . . 147
THE SAMPLE . . . . . . . 147
CALCULATION OF THE DEPENDENT VARIABLE:
PANEL ATTRITION . . . . . . . . . . . . 152
PROBLEMS WITH AGGREGATION'.
ECOLOGICAL FALLACY . . . . . . . . . . . . . . 158
THE LONGITUDINAL METHOD:
AN EVENT HISTORY DESIGN . . . 161
PROBLEMS: INTERACTION AND MULTICOLLINEARITY . 164
FORMULAE . . . . . . . . . . . . . . . . . . . 165
MODEL 1 . . . . . . . . . . . . . 165
MODELS 2 THROUGH 4 . . . . . . . . . . . 166
MODELS 5 THROUGH 7 . . . . . . . . . . . 167
DESCRIPTIVE STATISTICS . . . . . . . . . . . . 168
VARIABLES . . . . . . . . . . . . 169
DISAPPEARANCE RATE . . . . . . . . . 169
FIRM FAILURE RATE . . . 171

CHANGE IN CROSS- STATE DISTRIBUTION . 172
RATE OF GROWTH IN NEW CAPITAL

EXPENDITURES . . . . . 173
ANNUAL RATE OF CHANGE IN SMALL

ESTABLISHMENTS . . . . . . . . 174
DECERTIFICATION ELECTIONS . . . . 175

AVERAGE ANNUAL ESTABLISHMENT CHANGE 176
AVERAGE ANNUAL CHANGE IN

PRODUCTION WORKER EMPLOYMENT . . . . 178
CAPITAL INTENSITY . . . . . . . 179
UNION PENETRATION RATES . . . . . . 180
INDUSTRIES . . . . . . . . 181
SIC 20, FOOD PRODUCTS . . . . . . . 182
SIC 22, TEXTILES . . . . . . . . . . 183
SIC 24, LUMBER . . . . . . . . . . . 184
iii

 

SIC 25, FURNITURE . . . . . . . . . 185

SIC 26, PAPER . . . . . . . . . . . 186
SIC 28, CHEMICALS . . . . . . . . . 187
SIC 29, PETROLEUM . . . . . . . . . 188
SIC 30, RUBBER . . . . . . 189
SIC 32, STONE, GLASS, AND CLAY . . . 190
SIC 33, PRIMARY METALS . . . . . . . 191
SIC 34, FABRICATED METALS . . . . . 192
SIC 35, MACHINERY . . . . . . . 193
SIC 36, ELECTRICAL EQUIPMENT . . . . 194
SIC 37, TRANSPORTATION EQUIPMENT . . 195
SUMMARY . . . . . . . . . . . . . . . . . 196
CHAPTER VII. RESULTS . . . . . . . . . . . . . . . 198
INTRODUCTION . . . . . . 198

TREND IN LOCAL BARGAINING RELATIONSHIP
DISAPPEARANCE . . . . . . . . . . . . . 200
ZERO- ORDER CORRELATIONS . . . . . . . . . . . 203
REGRESSION MODELS . . . . . . . . . . . . . 208
CONSTANT TIME MODEL . . . . . . . . . . . 208
TIME VARYING MODEL . . . . . . . . . . . 211
ALTERNATIVE MODELS . . . . . . . . . . . . . . 220
NON- LAGGED MODEL . . . . . . . . . . 220
A MORE PARSIMONIOUS MODEL . . . . . . . . 223
CONCLUSIONS . . . . . . . . . . . . . . . . . 228
CHAPTER VIII. CONCLUSIONS AND DISCUSSION . . . . . 233
INTRODUCTION . . . . . . . . . . . . . . . . . 233
DISCUSSION . . . . . . . . . . . . . . . . . . 238
LIST OF REFERENCES . . . . . . . . . . . . . . . . 248

L
iv

 

LI ST OF TABLES

Table 6.1. Sample characteristics, selected
industries, 1977 . . . . . . . . . . . . . . . 152

Table 6.2. Disappearance rate, by industry,
3—year periods, 1977—1990 . . . . . . . . . . 169

Table 6.3. Firm failure rate, per 10,000 firms,
by industry, 3—year periods, 1974—1986 . . . . 172

Table 6.4. Percentage point change in distribution
of establishments across states, by industry,
3-year periods, 1974-1986 . . . . . . . . . . 173

Table 6.5. Average annual rate of change in new
capital expenditures, by industry, 3—year periods,
1974-1986 . . . . . . . . . . . . . . . . . . 174

Table 6.6. Average annual rate of growth in the
percentage of establishments with fewer than
100 employees, by industry, 3—year periods,
1974—1986 . . . . . . . . . . . . . . . . . . 175

Table 6. 7. Annual average decertification elections
per 10, 000 bargaining units, by industry, 3—year
periods, 1974- 1986 . . . . . . . 176

Table 6.8. Percentage change in the number of
establishments, by industry, 3—year periods,
1974—1986 . . . . . . . . . . . . . . . . . . 178

Table 6.9. Annual average percentage change in the
number of production workers, by industry,
3-year periods, 1974-1986 . . . . . . . . . . 179

Table 6 10. Percentage of assets invested in
capital equipment, by industry, 3—year periods,
1974- 1986 . . . . . . . . . . . . . 180

 

”

Table 6.11. Average union penetration rates,

by industry, 3—year periods, 1974-1986 . . . . 181
Table 6.12. SIC 20. Food Products . . . . . . . . 183
Table 6.13. SIC 22. Textiles . . . . . . . . . . 184
Table 6.14. SIC 24. Lumber . . . . . . . . . . . 185
Table 6.15. SIC 25. Furniture . . . . . . . . . . 186
Table 6.16. SIC 26. Paper . . . . . . . . . . . . 187
Table 6.17. SIC 28. Chemicals . . . . . . . . . . 188
Table 6.18. SIC 29. Petroleum . . . . . . . . . . 189
Table 6.19. SIC 30. Rubber . . . . . . . . . . . 190
Table 6.20. SIC 32. Stone, Glass, and Clay . . . 191
Table 6.21. SIC 33. Primary Metals . . . . . . . 192
Table 6.22. SIC 34. Fabricated Metals . . . . . . 193
Table 6.23. SIC 35. Machinery . . . . . . . . . . 194
Table 6.24. SIC 36. Electrical Equipment . . . . 195
Table 6.25. SIC 37. Transportation Equipment . . 196
Table 7.1. Hypotheses . . . . . . . . . . . . . . 199

Table 7.2. Production worker employment, selected
manufacturing industries, 1975, 1980, 1985, 1990,

in thousands 202
Table 7.3. Zero—order correlations . . . . . . . . 205
Table 7.4. Standardized regression coefficients,

lagged model, t—values in parentheses, N=56 . 209
Table 7.5. Standardized regression coefficients

with period effects, lagged model, t-values in

parentheses, N=56 . . . . . . . . 212
Table 7.6. Standardized regression coefficients

with period effects and cross—products, lagged

model, t—values in parentheses, N=56 . . 217

vi

 

Table 7.7. Standardized regression coefficients,
alternative models, t-values in parentheses,
N:56 . . . . . .

Table 7.8. Standardized regression coefficients,
lagged reduced model, t—values in parentheses,
N:56

Table 7.9. Standardized regression coefficients,
lagged reduced model with period effects,
t—values in parentheses, N:56

Table 7.10. Standardized regression coefficients,
lagged reduced model with period effects and
cross-products, t—values in parentheses, N:56

222

224

226

228

 

Chapter I. Introduction and Overview.

A. Introduction.

Perhaps the most widely—reported fact concerning American
industrial relations during the 1980s was the sharp decrease
in the number of individuals belonging to labor unions or
covered by collective bargaining agreements.1 From 1980 to
1990, the percent of workers belonging to unions decreased
from 23 to 16. In manufacturing alone, the number declined
from 36 to 22 percent. Although a downward trend had started
in the mid-19505, the rate of decline accelerated quickly
after 1980. (Kokkeneleberg and Sockell, 1985; Curme, Hirsch,
MacPherson, 1990; Hirsch and MacPherson, 1993).

The reasons for the recent decline in unionism and for
the dynamics of union growth and decline more generally have
been examined by a number of scholars (see the literature
reviews of Block and Premack, 1983; Chiason and Rose, 1991;

Fiorito, Gallagher, and Greer, 1986; and Wheeler and McLendon,

 

. #h m”

‘ Not all workers who are covered by collective
agreements choose to join the unions representing them.
During the 19803, coverage tended to exceed membership by
approximately two percentage points (Curme, Hirsch and
Macpherson, 1990)

 

2
1991). Several leading theories emerge from this literature.2
And although a number of scholars have attempted to achieve
a synthesis of thought (Kochan, Katz, and McKersie, 1986;
Block, 1990), there remains some divergence between those who
emphasize public policy explanations and those who focus
instead on more purely economic and market-driven factors.
Put in terns of an older and broader social—scientific debate,
the two groups represent, respectively, those favoring
historical and non—historical points—of—view (see Stepina and
Fiorito, 1986, for a similar c1assification)3 In recent
times, the principal focus of the former group has been the
growing opposition to unions by employers during in the past
couple of decades and the failure of public officials to
respond (Dickens, 1983; Dickens and Leonard, 1985; Freeman,

1985, 1988; Freeman and Medoff, 1984; Weiler, 1983, 1990).

 

’ This point will be developed further in Chapter 2.

’ Stepina and Fiorito use the terms “saturationists” and
I'historians." But in a footnote (1986, p. 248, note 1) they
explain:

The term ’saturationist’ (first
popularized by Bell (1953)) has probably
outlived its usefulness. It might be
more meaningful to distinguished between
’modelers’ and ’historians’ or
qualitative versus quantitative
approaches.

For the debate between historicists and non-historicists
in the social sciences, see Karl Popper (1957) for a classic
treatment and Theda Skocpol (1984) for a more recent
examination.

 

_
./

3

Indeed, some have argued that the posture of public officials
and the outcomes of public policy, particularly in the mid-
19805, reflect government complicity with employer opposition
(Bluestone and. Bluestone, 1992; Gould, 1986, 1993). The
latter group, the economic theorists, submit that mainly
market and concomitant structural—economic changes are
responsible for union decline, particularly the shift of jobs
away from heavy manufacturing, where most union members were
found, to light manufacturing and service (Jones, 1992;
LaLonde and Meltzer, 1991; Sheflin, Troy and Koeller, 1981;
Troy, 1969, 1990).

The debate between these schools is fascinating, and, at
times, heated.‘ A major problem, however, is that the two
sides often talk past each other by focusing their arguments
on different phenomena. The policy theorists focus most
frequently on the inability of unions to gain new members,
with their most powerful arguments concerning employer
~opposition to union organizing drives and initial contracts
(Cooke, 1985; Dickens, 1983; Weiler, 1983; Freeman, 1989,
1989a, 1989b), while the economic theorists find their most
persuasive case in the decline of unionization in areas of

previous strength (Jones, 1992; Troy, 1969, 1990). Perhaps

 

‘ The best examples are Troy’s (1990) sharp (at times
personal) attack on Freeman and the debate between Weiler and
LaLonde and Meltzer (1991).

 

m.

4
lack of growth and decline require two separate theoretical
conceptualizations?

In this research, I will attempt to bridge these two
bodies of knowledge by borrowing from both schools. Typical
of the recent public policy work, I will focus heavily on
doctrinal and institutional changes that were, I believe,
unfavorable to collective bargaining and damaging to labor
unions. But instead of walking the well—trod path of those
who have examined union organizing, I will focus, instead, on
threats to the viability of existing local collective
bargaining relationships.5 In doing so, I look at a
phenomenon more central the arguments of the economic
theorists.

B. The importance of local collective bargaining
relationships.

One characteristic of the union decline literature is
that, invariably, empirical studies use some measure of the
number of individuals belonging to unions as the dependent
variable. But as Block and Premack (1983, p. 45) note:

There is a weakness common to all
studies that use change in union
membership as the dependent variable.
This limitation is based on the fact that
a substantial percentage of union members
at any time made no explicit choice to
join their union; rather they simply took

 

’ Changing standards for union elections are not ignored
completely since they are relevant in the discussion of union
decertification.

 

W

5
a job in a firm that was already
unionized and had a union security
agreement.
In short, changes in the number of union members probably
tell us more about labor markets than about industrial
~re1ations activity. It is analogous to the problem, cited by
Buss and Redburn (1987), of using unemployment statistics to
make statements about the impact of plant closings on
communities. Rather, the essential building blocks of the
American industrial relations system——and one factor that
makes the American system somewhat exceptional—~are local,
plant-level, union—employer relationships (Bok and Dunlop,
1970; Kochan, 1980). In this study, I will depart from the
previous practice of studying membership levels, and use pre—
existing local collective bargaining relationships as the
unit of analysis.

The American system of industrial relations is almost
completely dependent upon these local relationships.
Employers (and hence employees) who are unionized are part of
the system, non—union employers (and hence employees) are
not. This highly decentralized system stands in rather sharp
contrast to many other industrial relations systems,
particularly those of Western Europe, which rely heavily on
centralized institutions and bargaining structures,
particularly union and employer associations (Bamber and

Lansbury, 1987; Dunlop and Galenson, 1978). In most other

   

6
nations, once deals are struck between employers and unions
at the national, industrial, or regional levels, terms of
employment are enforced on all employers, regardless of

whether employees at a particular location have selected

affiliation 'with any particular labor organization. The
terms nunion" and "non—union" employer are largely
meaningless.

In the United States, these terms are far from
meaningless and are essential to who is and who is not part
of the industrial relations system. Although there are also
centralized bargaining mechanisms in this country——though
many fewer than there used to be (Katz, 1993)-—they differ
from1 European structures in that agreements entered into
between employer and union associations cannot be enforced at
workplaces where employees have not chosen union
representation.‘ This is not to deny the spillover effects
of unionism, nor the impact of unions on public policy more
generally (Freeman and Medoff, 1984; Delaney, Master, and
Schwochau, 1990; Lepo, 1992). However, only unionized
workplaces are formal components of the industrial relations

system.

 

‘ When wages are subject to regulation by prevailing wage
laws, e.g. Davis-Bacon and Walsh—Healey, the American system
functions somewhat like the European. Wage and Hour officials
frequently rely on union labor agreements to set the
prevailing wage rate, which then must be paid by all
employers.

 

7

An additional point is that an examination of
relationships is likely to tell us more about employer
intentions than would an examination of membership levels.
If union workers are laid-off from an establishment, the size
of the labor movement is reduced. But if the employer
continues to bargain in good faith with the representative of
the remaining employees, it is difficult to suggest that the
employer is trying to "get rid of the union." On the other
hand, a complete severing of relations, while not conclusive,
lends more credence to the argument that an employer wishes
to escape the obligations of collective bargaining.

Therefore, while most of literature focuses on how many
employees are "union," the more important question is how
many employers are “union,“ as reflected in the number of
local collective bargaining relationships.

c. The importance of law and public policy.

Another, though somewhat less distinctive, aspect of the
American industrial relations system is that it occurs within
a fairly well—defined legal framework (Bok and Dunlop, 1970;
Cox, Bok, and Gorman, 1987). The outcomes of the bargaining
process still depend upon the relative strength of the
parties, but who must bargain with whom and over what are now
largely matters of law. Since 1935, the most important law
covering the private sector has been the National Labor

Relations Act (29 USC 169) (hereafter referred to as the NLRA

 

-a‘LiR‘VA

8

or the Act), which, under Section 7, grants workers the right
to organize, bargain collectively, and engage in concerted
activities—-the so-called "three basic rights." The Act also
defines rules of conduct for unions and employers designed to
protect the basic rights, and gives authority to an
administrative agency, the National Labor Relations Board
(hereafter referred to as the NLRB or the Board), to enforce
the Act.

Through its guidelines for union organizing and
representation, the law codifies the importance of local
relationships. Under Section 9 of the Act, employees must
select or reject union representation at the "bargaining
unit“ level—-a bargaining unit being a group of positions,
usually at a single establishment, that share an apparent
'community of interest," and are, therefore, an appropriate
unit for the purposes of collective bargaining (Weber, 1961).
Production and maintenance positions within manufacturing
often constitute an appropriate unit, as do clerical
positions. If a majority of workers occupying the bargaining
unit positions vote in favor of representation, the union is
certified by the NLRB as its "exclusive bargaining
representative" and the employer must confer with the union,
and no other organization, group, or individual before
altering “wages, hours, or other terms and conditions of

employment," what are referred to as the "mandatory subjects

 

7““.

9
of bargaining.“ Equally important, once a relationship
begins, it continues until some explicit action is taken to
end it; there need be no periodic reaffirmation of a union’s
right of representation.

With these two aspects of the system in mind——the
importance of local relationships and the related importance
of the law to the industrial relations system-—it seems
appropriate to bring the two together within the context of
current questions of union decline: To what extent does the
law, and indeed public policy more broadly, affect the
viability of existing local relationships?

D. A framework for studying the disappearance of local
relationships.

Much of the research concerning the effects of public
policy on union dynamics examines the establishment of union
representation. Authors have been particularly interested in
the impact of the law and its enforcement on the success or
failure of unions in. winning certification elections and
securing first contracts (see, Roomkin and Block, 1981;
Cooke, 1985; Dickens, 1983; Getman, Goldberg, and Herman,
1976; Weiler, 1983, 1990). Through this literature, which
benefits greatly from the availability and precision of union
election data, we have learned a good deal about the impact
of employer speech, anti—union discrimination, and

administrative delays on election outcomes. However, with

 

10
the exception of decertification elections, the
representation process better explains the lack of union
growth than the acceleration of union decline. It is
important, therefore, to look beyond elections and focus on
additional ways in which. public policy might affect the
viability of local relationships.

A common—sense model of union dynamics, as suggested by
Chiason and Rose (1991), Dickens and Leonard (1985), Chiason
and Dhavale (1990), Lawler (1990), and others, can easily be
refined by dropping the variables relating to union growth
and the contraction of employment to show that the total
number of disappearances of local collective bargaining
relationships are a function of the number of unionized
business closings, unionized business relocations, complete
substitutions of non—union labor for union labor, complete
substitutions of technology for union labor, and union loss
of certification. In equation form:

EIXS=K2UBC+XUPR+ZNUL+ZTUL+ZLOS) [H
Where:

DIS: The disappearance of a local collective bargaining

relationship.

UBC: The closing of a unionized business.

UPR: The relocation of a unionized business to a non-

union site. '

NUL: The complete substitution of non—union labor for

union labor.

TUL= The complete substitution of technology for union

labor.

LOS: A union’s loss of certification as collective
bargaining representative.

 

11

Upon a closer examination, we see that the model
contains two broad categories of actions. The first four
independent variables represent business decisions by
employers that can affect the bargaining unit, resulting in
either the immediate of ultimate loss of unit jobs and,
perhaps, the bargaining unit relationship itself. The final
variable is different; it concerns the right of a union to
represent employees, a right that can be challenged either by
employers or bargaining unit employees with no concurrent
impact on jobs.

If we assume the following...

1. Union leaders wish to preserve the collective
bargaining relationship with the employer and their
right to represent workers in a bargaining unit.
That is, they are motivated by self—preservation.

2. Employers, in most cases, prefer the right to make
business decisions unilaterally, without being
compelled to bargain with a union. That is, they
are motivated by a desire for control and
certainty.

3. The degree to which unionized employers will act to
escape collective bargaining relationships-—and,
therefore, gain greater control and certainty—-is
related to the amount of effort, associated costs,
and relative benefits of doing so.

4. A legal requirement that employers bargain over
business decisions that could result in a end to a
bargaining relationship increases the associated
costs and effort, and provides union leaders with
an additional opportunity to pursue self—
preservation.

;5.,_ A loss of certification is rare likely to occur
when unions have less recourse against dissident

 

\
«Kg-s

12

members, and when employers are allowed to
intervene in the decertification process.

...the most relevant policy questions arising from
Equation 1 are:
1. To what extent must an employer involve his or

her union counterpart in a business decision
that may, immediately or ultimately, result in

the disappearance of the bargaining
relationship?
2. To what extent can the union demand the

allegiance of its members?

3. To what extent can the employer intervene in
the decertification election process?

8. Research objective.

When Congress passed the National Labor Relations Act in
1935, it enunciated a position that collective bargaining was
a desirable institution and, as a matter of public policy,
should be encouraged. Beyond the material benefits to
workers, the framers of the Act also believed that it would
eliminate a major source of industrial conflict by giving
workers a statutory right to share in certain decisions
affecting their employment. Voluntary collective bargaining
would be good not just for labor, but for the economy and,
hence, the nation (29 USC 169, Section 1). This belief still
shows up in court decisions as the "neutral purposes of the
Act' or some equivalent expression of the social usefulness
of collective bargaining (see, for example, Fibreboard

Prggugtg Corporation v. NLRB, 379 US 203, (1964); First

 

13

National Maintenance Corp. v,. NLRB, 452 US 666 (1981)).

However, history teaches us that few participants in the
system have taken such a dispassionate view. An enormous
body of law has developed around the conflicts between
employers and unions as to what the Act truly requires. The
belief by some (particularly academic) observers at mid-
century that the Act would provide the foundation for a
stable and mature industrial relations system based on
acquiescence to the law and the mutual recognition of rights
proved illusory (Dunlop, 1958; Slichter, 1947). A. more
enduring observation was the one made by Jack Barbash (1948,
p.1) nearly 50 years ago: "No political democracy has offered
a more hostile environment to unionism than has the United
States." Even the employer-sponsored Taft—Hartley Amendments
of 1947 did not create the neutral arena or level playing
field that some seemed to have envisioned. In fact, the 1947
amendments may have led to more conflict by confusing the
purposes of the Act (Forbath, 1989; Gross, 1995; Klare, 1981;
Stone, 1981; Tomlins, 1985, 1985a). As former long—time
Board member and chair John Fanning (1984, pp. 61, 63)
commented:

Labor relations has always been a
field that arouses strong emotions—-
sometimes more emotion than reason...As
someone who has participated in some
25,000 decisions of the Board, I can

assure you that the one factor every case
has in common with the other is the

 

II)

I L

 

l4

presence of at least two people who see
things completely different.

It is the job of government to arbitrate these
differences of opinion. Whether the government ever was, or
even should. be, a neutral arbiter is subject to debate.
Regardless, since public policy is the product of political
processes, decisional outcomes are likely to reflect the
political balance of power. Historical evidence suggests
that outcomes are more likely to favor organized labor when
Democrats control critical positions, while business
interests are more warmly received. when Republicans hold
power (Kaufmann, 1991; Nagel, 1969). Certainly, organized
labor’s tendency to endorse Democratic candidates with much
greater frequency than Republican candidates reflects this
belief (Delaney, Masters, and Schwochau, 1990; Lepo, 1992);
and the same could be said for the favor shown to Republicans
by the U.S. Chamber of Commerce and National Association of
Manufacturers.

Although these characterizations are somewhat sweeping,
the 19803 provide a rather vivid-~perhaps extreme——example of
the effects of partisan politics on labor relations public
policy. Regardless of whether one believes that the policy
changes of that decade were necessary adjustments to a system
driven too far to the left (Irving, 1984; Miller, 1984;

O'Connell, 1986; Walther, 1985) or right—wing attempts to

 

15

destroy bargaining protections (Clark, 1988, 1989; Gould,
1993; Gross, 1994), one must acknowledge that the political
shift of 1980 began an era of fairly sharp departures from
past doctrines. Some observers have complained that the
changes of the 19805 went far beyond what was normally to be
expected when the presidency shifts parties (Brownstone,
1986; Gross, 1994). Others have gone so far as to question
the stability and even the adjudicatory integrity of the NLRB
(Clark, 1988, 1989; Kaufmann, 1991).

The immediate research objective is to examine the
connections between the political changes of the 19805,
changes in labor relations public policy——focusing on the
three areas listed earlier-—and the viability of local
collective bargaining relationships. It will be argued that
policy changes in these areas reflected the wishes of a
conservative federal administration, which took extraordinary
steps to ensure that its ideology be transferred to
decisional outcomes. The main features of the this ideology
were a strong emphasis on the rights of employers to dispose
of property unilaterally and without constraint, and anti-
collectivisnlwith regard to employee relations. Changes both
within the requisite governmental bodies and the resulting
decisional outcomes will be reviewed.

To explore the final part of the question, I will return

tawthe model put forth in Equation 1. First, a regression

 

16

analysis will be performed on a model predicting the
disappearance of local bargaining relationships across
selected industries. The model is based upon the one
suggested in Equation 1, but several important control
variables have been added to improve specification. The goal
of the regression analysis is to see which variables are the
best predictors of the variance in local bargaining
relationship disappearance rates across a sample of
manufacturing industries from 1977 until 1990 using a pooled
cross—sectional technique. Then, time period dummy and
cross-product variables will be added to the model to test
the stability of predictors. The technique, which is
essentially a covariance model (Blalock, 1979) has been used
in a couple of different forms, differing in the way in which
variables have been added to the model. Allison (1982, 1984)
demonstrated its use in his work on event history analysis,
and a nearly identical technique was recommended by Gujarathi
(1970) as an improvement upon the Chow (1960) test for the
comparison of regression equations.

The main hypothesis is that variance across industries
will be found to have been more closely linked to the
economic distress of industries, particularly as measured by
firm failure rates (Dun and Bradstreet, 1974—1990). early in
the period, but less associated with economic distress toward

the.endrof 19808. Depending on results, two alternative

 

17

explanations are possible: If the proxies for business
relocations and the substitution of inputs, and the variable
for union decertifications explained an increasing amount of
the variance as the decade progressed (i.e. if there is a
significant change in one or more of the slope coefficients),
it provides evidence that employers——in the case of
decertification, also employees——may have been reacting to
the policy changes that encouraged those behaviors. If the
progress of time explained an increasing amount of variance
(i.e. if there is a significant change in the model’s
intercept), but no change in slope coefficients is observed,
it provides evidence that some change, consistent with a
policy explanation, is possible, but cannot be linked
explicitly 11) a particular change 111 behavior. The null
hypothesis is that the original model will not change with
the introduction of time periods or cross-products (i.e.
there will be no significant changes in either the slope or
the intercept coefficients).

Of course, time period dummy variables pick up all
effects coincident to a particular period and not controlled
for in the model. That is why the qualitative findings are
just as important as the quantitative results. It will be
argued that the magnitude of the policy changes of the mid—
19803 provide the most likely explanation for any changes in

the predictive power of the independent variables.

18

Results may assist in the debate between the public
policy and economic theorists by providing a policy focus in
the area that has provided the economic theorists with their
strongest arguments. Favorable results would support an
argument that policy has affected the viability of previously
established collective relationships, not just the
probability of establishing new relationships.

I. Chapters.

In the next chapter, the main currents in the literature
on union decline are identified, with. particularly close
attention paid to economic and public policy explanations.
In Chapter 3, the changes within public institutions are
examined. The purpose of Chapter 3 is to show the unusual
lengths that the federal administration went to in diffusing
its conservative ideology. In Chapter 4 the products of this
transformation, the reversals in public policy, are examined
in light of the model of local union disappearance put forth
above. In. particular, changes in an employer’s duty to
bargain over business decisions, changes in a union's ability
to maintain internal discipline, and changes in an employerfs
ability to intervene in the decertification process are
examined. The purpose of Chapter 4 is to demonstrate how
decisional outcomes reflected more conservative notions of
property relations and individual rights of association, and

deemphasized bilateral solutions and collective rights. In

1.1:}. .

 

19
Chapter 5, I return to the independent variables mentioned in
Equation 1 and examine the phenomena of plant closings,
relocations, input substitution and decertifications. I
detail changes and speculate on possible causes the for
trends. Research hypotheses are then generated. In Chapter
6 the data and methods for the quantitative portion of the
study are described. Findings are relayed in Chapter 7, and

conclusions and final observations are found in Chapter 8.

 

Chapter II. Literature Review.

A. Introduction.

The purpose of this chapter is to highlight main currents
in the theoretical and empirical literature on the dynamics
of union growth and decline. After a very brief overview of
the varieties of this scholarship, I examine more closely two
schools of thought. I first consider the more “purely
economic" research, of which Ashenfelter and Pencavel’s 1969
article is seminal. Scholars working in this tradition tend
to deemphasize historical phenomena and stress, instead, both
cyclical and secular economic effects. Next, I examine the
work of political economists and public policy analysts. In
this research, unique historical trends embedded in the
political histories of nation—states are central. Finally,
I consider some of the problems in trying to bring these two
schools of thought together.

8. Onion growth and decline.

Over the years, scholars have proffered a number of
theories for the growth and decline of the labor movement and
the institution of collective bargaining. Some have focussed

on individual and social attitudes toward unions (Farber,

20

'1'
..n .-

 

at

21

1987; Hammer and Smith, 1978; Kochan, 1979; Lipset, 1986;
Moody, 1988; Premack and Hunter, 1988); others have examined
the performance of unions in attracting and maintaining
members (Craft and Extejt, 1983; Dickens and Leonard, 1985).
Still others have contributed to a large body of literature
concerning the economic——usua11y macroeconomic——determinants
of unionization, e.g. the business cycle, structure of
employment, and foreign competition (Adams and Krislov, 1974;
Ashenfelter and Pencavel, 1969; Bain and Elsheikh, 1976;
Doyle, 1985; Dunlop, 1948; Hirsch and Berger, 1984; Hirsch
and Addison, 1986; Jones, 1992; Manke, 1971; Neumann and
Rissman, 1984; Troy, 1969, 1990). And lately, amidst calls for
labor law reform, there is a growing literature on how public
policy-—mainly the regulation of employer and union behavior—
-affects the size and strength of the labor movement (Block,
1994; Chiason and Rose, 1990; Freeman, 1989, 1989a, 1989b;
Goldfield, 1984; Gould, 1986, 1993; Gross, 1994; Weiler, 1983,
1990).

Often, emphasis more than exclusion provides the basis
for categorizing studies. But for present purposes, I will
set aside the literature on individual and social attitudes,
since it addresses a level and unit of analysis not found
here. Rather, in dealing with industry trends over a period
of time, the economic and public policy explanations seenlmost

appropriate. These two fields borrow liberally from each

 

 

LII

s .
.‘f '— - "

22

other, but there is a fairly large intellectual divide between
those who advance historical explanations for union growth and
decline and those who suggest more time invariant explanations
(Stepina and Fiorito, 1986). The "purer economic" perspective
links growth and decline with causes that could occur anywhere
at any time, so long as certain economic antecedents, such as
a rise in prices or employment, are present. The historical
perspective, which dominates the political economy and public
policy literature, focuses more directly on peculiar patterns
of events leading up to or surrounding periods of growth and
decline--war, legislative changes, changes in the national
political mood, periods of worker unrest, etc.

1. Economic explanations.

The idea that macroeconomic factors are the basic
predictors of unionization trends dates back to at least World
War I when George Barnett (1916), an early researcher of union
growth, noted that changes in the size of the labor movement
at the turn of the century followed closely the relative
prosperity of the time. This general line of reasoning has
been revisited many times since in analyses of the effects of
prices, wages, and the structure of employment on unionization
(Ashenfelter and Pencavel, 1969; Bain and Elshiekh, 1976;
Dunlap, 1948; Hirsch and Addison, 1986; Jones, 1992; Manke,
1971; Sheflin, Troy and Koeller, 1981, Troy, 1990).

Much of today’s research is in response to a seminal

 

an“ I.’ .Iii

23

article written in 1969 by Ashenfelter and Pencavel. The
dependent variable used in the study was the annual rate of
change in the percentage of workers belonging to American
unions. The authors examined the effects of price changes,
employment and unemployment trends, union saturation, and
political climate on the growth and decline of the labor
movement between 1904 and 1960. Their strongest finding
concerned prices, with a one percent change in consumer prices
associated with a .65 percent change in the unionization rate.
Ashenfelter and Pencavel concluded, therefore, that unions are
largely defensive organizations which protect workers against
a declining standard of living. When they entered into their
equation several categorical variables representing particular
time periods, they found no major changes in the coefficients.
They write (p. 434):

Although it is difficult to find one

cohesive strand in the evidence, the most

widely accepted view ascribes the bulk of

trade union growth to what Bernstein has

called periods of "social unrest.“ This

paper seeks to show that there is no need

to resort to such an ad hoc device to

explain American union growth; a single

behavioral relationship can explain the

progress of the American labor movement

in the twentieth century.

There have been a number of responses to the Ashenfelter

and Pencavel model. Generally, scholars have experimented

with the model by updating the time series, improving

measurement, and altering the specification (Adams and

 

v
i
a

 

 

 

 

24
Krislov, 1974; Bain and Elsheikh, 1976; Jones, 1992; Mancke,
1971; Moore and Pearce, 1976; Neumann and Rissman, 1984;
Sheflin, Troy, and Koeller, 1981).

Mancke (1971) was one of the first to respond to
Ashenfelter and Pencavel. Using as his dependent variable
the level of unionization among workers in ”unionizable
employment" he respecified Ashenfelter and Pencavel’s model
by including a dummy variable to separate the periods of 1904
to 1937 and 1938 to 1960. He found that a significant shift
had occurred following 1937, and concluded that "the Wagner
Act caused a major structural change in American labor
markets (p. 193)."

Not long after Mancke, Bain and Elshiek (1975) provided
one of the most comprehensive treatments of this question.
First developed on data in Great Britain, the authors
extended their analysis to the United States. They made
several fairly minor adjustments to the Ashenfelter and
Pencavel model to improve specification, and included a
measure of wage change to reflect the "credit“ of
unionization (and to balance Ashenfelter and Pencavel’s
”threat” effect of prices). They found that the basic
formula suggested by Ashenfelter and Pencavel continued to
perform well. In a response to Moore and Pearce (1976), who
argued that the Ashenfelter and Pencavel model was sufficient

for explaining trends prior to, but not after, War World II,

 

25
Elsheikh and Bain (1978) showed that their alternative model
held constant for the entire period.

In 1981, Sheflin, Troy, and Koeller retested both the
Ashenfelter and Pencavel and Bain and Elshiekh models. The
authors not only extended the time series to 1975, but
corrected the sketchy early twentieth century membership data1
and used a maximum likelihood estimate of parameters rather
than the more typical ordinary least squares technique. They
found that, indeed, a shift had occurred, not in 1945 as
Moore and Pearce had found, but in 1937 as the earlier work
of Mancke had indicated.

After about fifteen years of reshaping and rerunning the
Ashenfelter and Pencavel model, it seemed fairly clear that
prices and employment were good predictors of union trends,
and that passage of the Wagner Act had had a significant
effect on union growth as well. The stability of this model
was once again questioned, however, after the sharp and rapid
decline in unionization in the early 19805. Scholars
returned to the question of whether a “single behavioral
relationship" could explain growth and decline over the long—
term. Leo Troy (1990), one of the co—authors of the 1981
study, argued that the basic econometric model needed to

include some measure representing the structural shift in the

 

‘ Most of the early estimates were based on union self-

reports, some of which included Canadian membership.

 

26
economy away fron1 heavy manufacturing (e.g. basic steel)
toward light manufacturing (e.g. computer components) and,
particularly, services. He also took note of the increases
in part—time, female, and public sector employment.

To test this added speculation, Jones (1992) created a
modified version of the Ashenfelter and Pencavel model.
Instead of using Ashenfelter and Pencavel’s measure of total
union saturation, she employed a measure of union penetration
within the four historically most unionized industries.
Indeed, Jones concluded that Sheflin, Troy, and Koeller and
later Troy were both correct. A change had occurred in the
explanatory power of the Ashenfelter and Pencavel model after
1937. and structural shifts in the economy away from the
traditionally more unionized sectors did provide most of the
explanation for the decline in recent years. She writes (p.
266):

Thus it would appear that the union
membership declines of the 19705 and
19805 do not require period—specific
explanations but, instead, represent a
continuation of factors affecting union
membership growth that were evident
before the current decline began to
appear.

‘The one minor concession to historicism allowed by
scholars following in the mode of Ashenfelter and Pencavel is

the single transition point of 1937, which. was, without

question, a pivotal period in American industrial relations.

 

_ u“. L

27
Otherwise, they argue that the dynamics of unionism are tied
to economic trends——not political shifts, nor changes in
public policy, nor decreasing militancy by unions, nor
increasing militancy by employers.

A number of authors have challenged this view. Farber
(1987) in an article appearing in the journal Science noted
that, first, structural economic changes do not provide a
causal explanation, but at best can "account for" the loss in
union members. Moreover, he explored how much of the recent
decline could, in fact, be accounted for by structural shifts
in the economy, and found that of the 5.4 percentage point
decline in unionization between 1977 and 1984, shifts in
employment between industrial sectors accounted for only 1.1
percentage points, or roughly 20 percent of the drop.

Furthering Farber's point, Weiler (1989, cited in
Chiason and Rose, 1991) criticized the structural theorists
for their silent assumption that, for some reason, female and
part—time workers in emerging industries are unorganizable.
Like Farber, he argued that their models do not explain why
workers in emerging industries are not organizing. After
all, Weiler noted, the single most highly organized
occupation in the country is a female-dominated, white collar

profession--teaching.

 

28

2. Policy issues.

Those who reserve the most criticism for the structural,
economic explanations are political economists and public
policy experts who note differences, both across nations and
within nations over time, that do not seem to be linked to
particular economic trends. As substitutes for the economic
and structural explanations, these scholars prefer to focus
on differences and changes regarding the posture of employers
toward unions, the ability of unions to respond, and,
particularly, the role of the state in regulating labor—
management relations. Historical trends and political
changes are central to this line of research.

a. International and comparative research.

Observers of international and comparative industrial
relations provide some of the most compelling reasons for
policy analysis. They note that while some of the economic
studies have shown empirically impressive results, they have
not always held up well to cross—national comparisons. A
number of scholars (Visser, 1988; Freeman and Medoff, 1984;
Freeman, 1989, 1989a, 1989b) have observed that purely
economic variables cannot account for the disparity in
unionization trends across nations experiencing similar
economic conditions, and suggest that other international
differences, particularly' government policies, need. to be

taken into account (Chiason and Rose, 1990; Freeman, 1989,

 

29
1989a, 1989b; Freeman and Medoff, 1984; Freeman and
Pelletier, 1990; Murphy, 1988; Weiler, 1983).

Britain and Ireland and the United States and Canada
provide vivid examples. While the two pairs of nations faced
similar economic pressures during the 19805, their political
histories diverged significantly. Britain and the U.S. were
led by governments with strong conservative and anti—
collectivist views: Thatcher’s battle with the National
Union of Miners was similar to Reagan’s action against PATCO;
what Thatcher accomplished through the Trade Union Laws,
Reagan managed through appointments to the federal courts and
the National Labor Relations Board. Meanwhile the trade
union movements of Ireland and Canada flourished (Chiason and
Rose, 1990; Freeman, 1989, 1989a, 1989b; Freeman and Medoff,
1984; Freeman and Pelletier, 1990).

Canada, in particular, has been used as a frequent basis
for comparison with the United States (Block, 1994; Chiason
and Rose, 1990; Freeman and Medoff, 1984; Gould, 1993;
Murphy, 1988; Weiler, 1983). Freeman and Medoff (1984, p.
227), in dismissing the structural change arguments, write:

The major structural changes in the U.S.
labor market-—growth of white—collar
jobs, decline of manufacturing, increased
number of women in the labor force——have
occurred in virtually all major western
economies. If structural changes were
the chief factor behind the decline in

unionism, the proportion organized would
fall everywhere....Perhaps most telling

 

30
is the fact that the country most like
the United States, Canada, where many of
the same unions and firms operate, the
percentage unionized went from below the
U.S. percentage unionized to above it.
They continue, borrowing from Weiler (1983), (p. 242):

The principal difference between
unionization in the United States and in
Canada is that U.S. laws allow management
to conduct lengthy, well—funded election
campaigns against unions. Canadian labor
laws do not permit such activity...The
result is growing private sector
unionization in Canada.

The same theme is repeated by Chiason and Rose (1990)
and again by Freeman (1989). Block (1994) also notes
differences between the U.S and Canada, but rather than
focusing on doctrinal differences in the law suggests that
Canadian public agencies have done a better job of
maintaining the system of collective bargaining. In Canada,
labor law enforcement mechanisms work better and function
more neutrally, and the provincial legislatures have been
more attentive to the changing needs of the law than has the
U.S. Congress.

b. Domestic research.

In addition to the comparative literature, there is a
good deal of research on cross-sectional differences among
states and regions within the U.S.—-usually differences

regarding the presence or absence of state right—to—work laws

or public employee bargaining statutes (Hirsch, 1980; Moore

 

31
and Newman 1975, 1985, 1988)--and on trends in policy and
policy effectiveness over time. A :major focus of the
longitudinal research is the increase in employer opposition
to unionization, and the inability or unwillingness of the
government to respond.

The issue of employer oppositrmi has been around for
some time, and gave rise to attempts to amend the NLRA in the
late—19705 (Prosten, 1979). However, several studies in the
early-19805 provided better and even more compelling accounts
of the severity of the problem. In a detailed analysis of
NLRB election results, Roomkin and Block (1981) found what
had long been suspected: the longer a representation
campaign lasted, the greater the probability that employees
would choose against union representation. vmnler (1983)
seized on this point, and in a truly influential article in
Harvard Law Review examined employer behavior during
organizing drives. Most alarming to Weiler was the increase
in unfair labor practice charges filed against employers over
the years, compared to the decrease in the number of union
elections. Making some rough calculations--which have been
criticized In! others (Goldberg, Brett anxi Getman, 1984;
Irving, 1987; LaLonde and Meltzer, 1991)——Weiler concluded
that one—in—twenty employees involved in union organizing
campaigns have faced illegal discharge. Whether precise or

not, the figure was picked up by others and became a key

32
issue during Congressional hearings held in 1984 (U.S. House
of Representatives, 1984a).

Cooke (1985), in turn, found that employer opposition
did not end with the election campaign. In about 25 percent
of the cases he surveyed, employers and unions were unable to
arrive at a first contract settlement. These relationships
had "died on the vine." Cooke also found that a significant
predictor of the failure to settle was the extent of employer
unfair labor practice activity. He concludes (p. 178):

The evidence presented in this study
clearly indicates that the rights of many
workers to bargain collectively with
employers are being denied. Ironically,
the very agency responsible for the
protection (Hf these rights, time NLRB,
appears to be in part an unwitting
accomplice tx> employers hoping t1) deny
these rights to workers.

Not everyone believed that the complicity of the Board
was "unwitting. " Indeed, there was a growing chorus of
critics who complained that the Board, acting on behalf of
the Reagan administration, was quite deliberate in eroding
the Act’s protections.

Bluestone and Bluestone (1992, p. 76) write:

In the new administration of Ronald
Reagan, corporate lobbyists found.ea willing
partner to help them rebuild profits by
slashing federal taxes and government
regulations...

Summarily discharging all of the striking
air traffic controllers in 1981 and placing new

pro—management members on the National Labor
Relations Board effectively deregulated the

33

rules under‘ which. organized labor
management had been bound.

Heckscher (1988, p. 35) adds:

The disastrous effects of the NLRB’s
anti—union tilt during the Reagan
Administration are tine most recent
indications that the fortunes of the
labor movement are closely bound with the
favor of the state.

and

Present NLRB chair William Gould (1993, p. 78) writes of

the erosion of labor’s rights under the stewardship of "the

free market buccaneers of the Reagan administration."

the NLRB, he states (p. 22):

It is certainly true that the Board,
particularly ‘under time chairmanship <3f
Donald Dotson between 1983 and 1986, was
preoccupied with a mission to reverse a
series of decisions that might promote
the collective bargaining process. The
appointments made to the Board were one—
sided and pro—management.

And of

And from James Gross (1994, p. 48), a leading authority

on the Board and author

history:

Republican Boards, particularly the
Dotson Board, elevated. managements’
authority to manage above employers’
statutory obligation to bargain. During
the Reagan presidency in particular, the
Dotson board seriously diminished the
statutory obligation to bargain, once
considered central tx> the Act, by
excluding management decisions too
important to en1 employers’ business to
have to be negotiated with a union. Many
of these decisions have the most direct
impact on jobs.

of a :multi~volume work on its

34

Exactly what the "disastrous effects (Hecksher, 1989)"
were is not certain. There is a lot of evidence that court
decisions and Board processes and outcomes became
increasingly favorable to employers during the 19805. For
example, it has been shown that dismissals of unfair labor
practice charges against employers increased. 300 percent
between 1983 and 1985 (Bernstien and Gold, 1985), and that
having a charge of anti—union discrimination upheld against
an employer decreased by 35 percent for the period between
1983 and 1988 (Leap, Hendrix, Cantrell, and Taylor, 1990).
Less certain is what the impact of this new environment was
for industrial relations more generally and particularly for
the viability of local relationships.

C. Bridging the literature.

Even if the political events of past couple of decades
are put aside, the period was an extremely turbulent one.
The post—War expansion stalled in the early 19705, as the
domestic economy became increasingly challenged by production
abroad. The oil embargo of 1973 was one of the first signs
of this vulnerability. But, in fact, during the past couple
of decades foreign trade has increased from approximately 9
to over 21 percent of gross domestic product (Dunlop
Commission Report, 1995). This change itself is believed to
have had a drastic impact on the industrial relations system.

When the U.S. was a net exporter of goods, unions could

35
organize entire industries and "take wages out of
competition," while employers could easily pass higher prices
on to consumers who were limited in their product choices.
As products made by foreign workers, who are beyond the reach
of American laws and American unions and in many cases earn
much less, began entering the market more frequently,
domestic employers sought lower wages, productivity
improvements, or both. To the extent that unions provided an
obstacle t1) these changes, employers sought txn avoid them
(Block, 1990; Block and McClellan, 1985; Farber, 1987;
Kochan, Katz, and McKersie, 1986).
Nonetheless, Stepina and Fiorito (1986, p. 248) write:

[A] limited number of variables can

account for' much. of the ‘variation of

union growth, but to more fully explain

and appreciate the dynamics of union

growth, variables must be included which

capture unique historical events. Some

movement (already Imus occurred 1J1 this

direction; Ihowever, tflma examination (of

variables, and of historical variables in

particular, has been very limited.

It is easy to agree with Stepina and Fiorito, but not
easy to find a solution. Goldfield (1987) raises the problem
when he points out that while a leading cause of recent union
decline is "changes in public policy that tend to favor more
and more the employers (p. 231)..."

The special responsibility of the Reagan
administration is not very convincing.

It is certainly true that the policies
are more overtly anti-union than any

36

administration since Herbert Hoover.
Their effect, however, are completely
entangled with the sharp economic
recession and the resulting high
unemployment that vmnflri certainly' have
found unions on the defensive even if
Jimmy‘ Carter“ or a Inore liberal ’pro—
labor’ Democrat were in office (p. 53).

Goldfield raises a difficult issue, how to disentangle
the effects of political and economic factors. Ii have
already suggested some possibilities in Chapter I: combine
qualitative and quantitative methods and borrow from the
strong points of both political and economic research.
Further, exploring changes both across industries and over
time may offer a solution, assuming that the economic trends
across industries vary sufficiently.

In the next chapter I begin the qualitative portion of

the study by examining changes in the institutions of labor

relations public policy.

Chapter III. Institutional changes.

A. Introduction.

In this chapter I describe the changes within the public
bodies responsible for labor relations policy. The primary,
though not exclusive, focus is on the National Labor Relations
Board.

The historical record reveals that labor relations
institutions that luui become somewhat stable I»! the early
19705, started to come under enormous pressure in the mid-
19705. By the early 19805, the mutual understanding between
labor, business, and the federal administration of the
protocols of policy were scrapped, and a strong ideological
agenda was pushed by the administration, its allies in
Congress, and business groups. The changes described below

occurred mainly during the;period of 1980tfluxnxfl11983, though

 

they were foreshadowed by earlier conflicts.

 

B. A foreshadowing of conflict.

During the 19705 and 19805, labor relations public policy
experienced several challenges and underwent a number of
changes (Kochan, Katz and McKersieq 1986). Following the 1973

oil embargo and concomitant economic contraction, the

37

38

stability which (to some extent) had developed between
government, labor, and management during the years of post-
war expansion began to erode. Although, as stated earlier,
the National Labor Relations Act never became the vehicle of
accommodation that some had expected, much of the bickering
over the purposes of the law and the proper role of the
National Labor Relations Board had subsided. The period
between the passage of the Landrum—Griffin Act in 1959, which
included a few amendments to the NLRA (29 USC 401), and the
health care amendments of 1974 (88 Stat. 395) was relatively
uneventful (McCulloch and Bornstien, 1972; Moe, 1985, 1987).1
However, at about mid—decade in the 19705 several episodes
began fraying relations between organized labor and management
groups at the highest levels and creating a climate in which
conflict easily grew.

Among the more noteworthy conflicts of the 19705 were the
unsuccessful battle 13/ building trades unions tx> pass a
"common-situs picketing" bill and the failed attempt of the
labor movement to amend the National Labor Relations Act more
generally (Moe, 1985, 1987; Moody, 1988). The former measure

would have allowed construction unions to picket entire work

 

1 Gross (1995) notes that in the mid-19605 employers were
beginning to unify their opposition to the NLRB. The most
notable manifestation of this activity was the Labor Law
Reform Group, which was given a forum for its concerns during
hearings held by Senator Sam Earvin. None of the Group’s
ideas became law.

39

sites during a labor dispute, in effect stopping all work,
even among contractors or subcontractors not party to the
primary dispute. The legislation was engineered by Secretary
of Labor John Dunlop as a quid pro quo for union agreement on
a more orderly (and less inflationary) system of construction
bargaining. The measure was initially supported by President
Ford, who withdrew his support amidst pressure from the
Republican right-wing. Ford’s veto foi to Dunlop’s
resignation from the Labor Department (Mills, 1980; Kochan
and Osterman, 1994).

The latter' measure ‘was more sweeping and involved
several amendments tr>tflu2 NLRA designed primariLy to make
union organizing easier by, among other things, increasing
penalties for employer unfair labor practices during union
organizing campaigns. The battle over the Labor Law Reform
Act was a particularly frustrating one for organized labor.
The measure had enough votes for passage in both houses of
Congress, and.was supported.by President Carter, but died when
a motion to stop a Republican filibuster failed by one vote
(Prosten, 1979).

What was particularly troubling to:many labor leaders was
the active opposition of a number of employers who were
simultaneously engaged in a top-level committee known simply
as the "Labor Management Group." The Labor Management Group,

which. included, among' others, AFL—CIC) Secretary-Treasurer

4O

(later President) Lane Kirkland and the chief executive
officers of DuPont and General Electric, was formed in 1976
to discuss concerns of mutual interest (e.g. inflation) and
tc>chart a course for constructive labor—management relations.
The Group gained semi—official status during Jimmy Carter’s
presidency. Labor' leaders immi hoped tfimu: their‘ employer
counterparts in the Group would remain neutral on the labor
law reform bill. They did not (Moody, 1988).

Speaking before a group in Harriman, N.Y., a clearly
agitated Lane Kirkland (1978) stated:

And there is a filibuster-—
organized, promoted, and sustained, not
just by some primitive sections of
industry, but all of corporate America,
through its major organs of influence.
The circle is unbroken. No voice of
compassion for (fine plight of ea worker
stripped of his livelihood for having the
character to overcome fear and exercise
his legal rights has been heard in public
from any official of a major corporation.

There is a term for the kind of
campaign that American industry has
launched to kill the hopes of the most
oppressed and deserving workers in this

country.
It is class warfare. Like the South
African "Bruderbund," the unbroken

phalanx, the conspiracy of silence of the
supposedly enlightened and the hate—
ridden invective of your Helmses and
Hatches who do the corporate dirty work
make it plain.

By the late—19708, relations between labor and business
groups were at a low point. Some were worried that the rift

would lead to even more serious consequences. Senator Jacob

41

Javits, for example, warned that, "the relative peace between
labor and business that the U.S. has experienced for several
decades is now sorely strained and could break into open
conflict (Daily Labor Report, August 28, 1978, p. A—7)."2
Perhaps it was almost inevitable that the problems would
spread t1) time National Labor Relations Board” which. had
become a remarkably stable agency after a tremendously
turbulent beginning (Gross, 1981).

That the Board should get dragged into such disputes is
not surprising since it is at the heart of labor relations
public policy, and has been caught in the crossfire between
labor and management, liberals and conservatives more than
once. Almost immediately eafter if}; establishment tn! the
Wagner Act of 1935, the Board was attacked for its pro—union
and, moreover, pro-CIO leanings. Added allegations of
communist-influence and infiltration turned the Board into a
political liability for President Roosevelt and. provided

ammunition for many anti—New Deal members of Congress (Gross,

 

2 Since 1960, the Bureau of National Affairs, a private
publishing company, has published the Dailnyabor Report. The
Egg is an invaluable historical resource. It provides not
only news items, but full texts of Board memoranda, Board_and
court decisions, and speeches of Board members, labor and
management advocates, and others. In fact, several
information requests that I made to the Board were answered
with references to the _D_L_R. Since news items in the DD};
appear without by—lines, they are cited as Daily Labor Report
with the date and page number of the article. Speeches,
memoranda, etc., are cited by author's name.

42
1974; Millis and Brown, 1950; Scher, 1956).

The Board also became a favorite target for the press,
with the steady drumbeat of bad publicity leading to a rather
low opinion of the Board and the Act on the part of the
public. Several major newspapers-—some of which were facing
organizing drives-—attempted to lay blame for the late—30s
economic slowdown (the so—called "recession within the
depression") on the Act. In the 1938 election, Republicans
gained 81 seats in house, doubling their presence there, and
picked up 8 Senate seats. Not a single incumbent Republican
was defeated. Of course there were many issues in the
campaign, but Gallup polls revealed that, indeed, Americans
had. doubts about the .Act. Senator~ Wagner; the Board’s
architect (and chair of its NIRA prototype), although
victorious in his first post—Act election, received 350,000
fewer votes in 1938 than he had in 1932 (Gross, 1981 pp. 73-
74).

By 1940, amendments to curtail the Board’s power were
being' sought :hi Congress. And vwnfli Republicans finally
gaining a majority in both houses in 1946, the stage was set
for some measure to pass. In the following year, major
revisions did come with the Taft—Hartley amendments, paSsed
over President Truman’s veto (61 Stat. 136, 1947).

Taft—Hartley sought to "balance" the Act by, among other

things, providing for a set of union unfair labor practices

43

to mirror the Wagner Act’s employer unfair labor practices,
and In! instituting ea system cu? union decertification that
would allow employees a means of ousting an incumbent union.
The Act also made institutional changes at the Board by
outlawing "economic" (read socialist) research, and by making
the General Counsel, who serves as the Board’s prosecutor of
unfair labor practices, a separately appointed position.

Prior to the amendments, the General Counsel served at
time pleasure <3f tlmz Board. Therefore time investigation,
prosecution, and ultimately the judgement of unfair labor
practices were contained within a single structure of
authority. While the Board had made an effort to minimize
potential conflicts-of—interest, some problems remained and
business interests and Congressional conservatives complained
that the Board was unable to carry out all of its functions
fairly and objectively (Gross, 1981; Millis and Brown, 1950:
Scher, 1956; Stryker, 1989).

The separation between the General Counsel and Board has
proven, at times, to be itself a source of conflict. In fact
the first independent General Counsel, Robert Denham, a pro—
Taft—Hartley Republican appointed by Harry Truman, battled
continually and publicly with chairman Paul Herzog over the
former’s claim that the Board had remained, even after the
1947 amendments, softer on unions than on employers. Pro—

labor Democrats tried to get the General Counsel’s provision

44

rescinded, but were unsuccessful. Ultimately, on September
15, 1950, President Truman was persuaded to force Denham’s
resignation (Gross, 1995, pp 58—71). But although Denham was
out, the General Counsel’s independent authority remained
intact. However, following’ DenhanVs departure relations
between the Board and General Counsel calmed, at least
temporarily. (Gross, 1995; Kaufman, 1991; McCulloch and
Bornstein, 1974).

Amidst the relative calm, one truly pivotal conflict did
occur during the Eisenhower years that seemed to set the tone
for the Board selections for years to cxmma .As might be
expected, the Eisenhower Board was relatively more favorable
to business interests than the Truman or Roosevelt Boards had
been. But organized labor, at its apex in numbers, was in a
fairly good position to make sure that business groups did
not have a completely free hand in reshaping labor policy.
When a distasteful nominee, Albert Beeson, was put forward by
the President, a serious challenge was mounted (McCulloch and
Bornstein, 1974; Scher, 1956).

Labor’s fight was carried forwand by Senator John F.
Kennedy, who subjected Beeson to an extraordinarily rigorous
confirmation hearing. Kennedy argued that Board appointees
deserved particularly close scrutiny because of the quasi-
judicial nature of their work. Unlike cabinet posts, where

the president had.ea right t1) expect ideological loyalty,

45

Kennedy argued that Board appointees should be above partisan
politics and as ideologically neutral as possible. Although
Beeson was ultimately confirmed by a 45 to 42 vote, Kennedy’s
arguments seemed to provide a philosophical basis for future
Board nominations (McCulloch and Bornstien, 1974, p. 62).
Eisenhower’s next appointment was John Fanning, a Democrat
and career civil servant who would end up being reappointed
by' both. Democrats and Republicans, ultimately serving a
record 25 years on the Board.

When Kennedy himself assumed the presidency, the Board
again swung in a direction somewhat more favorable to labor,
but trilateral relations between labor, management, and the
Board were maturing and stabilizing. No major attempts were
made to reshape the Board, and for the next decade or so both
Republicans and Democrats were routinely appointed to second,
third, and even fourth terms. One of Kennedy’s appointments,
Howard Jenkins, 53 Republican, would serve fin: the next 20
years, retiring at about the same time as Fanning. Jenkins
and Fanning took the same side on many issues, and together
embodied the stability which had developed at the Board. A
couple of Congressional hearings were held during the 19608-
—first by Illinois Representative Roman Pucinski in 1961 and
then by North Carolina Senator Sam Earvin in 1969--to air
complaints about the Eisenhower and Kennedy/Johnson Boards

respectively, but neither resulted in any significant

46
Congressional action (McCulloch and Bornstien, 1974).

Even the change in White House politics in 1968 failed
tr) make 61 dramatic difference. 131 1974, McCulloch. and
Bornstien (p. 76) noted:

The Nixon NLRB, contrary to labor’s worst
fears and.management’s highest hopes, has
not engaged in any systematic program to
reverse prior Board precedents on any
wholesale basis.

Although changes in.tflu2 political affiliathmi of the
president had always meant some change in the ideological
make-up of the Board, a system of accommodation had clearly
developed” Organized ljflxn: realized timn: its role 1J1 the
selection process would be an active one during Democratic
administrations, 1mm: more advisory 111 nature during
Republican years. Business groups shared a similar
understanding of the relative balance of power, and expected
more active involvement when Republicans held. the ‘White
House. Referring to the Board in the mid—19705, Moe (1985,
p. 1096) writes:

The political controversy and passionate
disputes that surrounded the NLRB in its
early years have subsided dramatically,
and for the last two decades it has
rarely been a salient focus of partisan
politics. Indeed it has come to be
regarded by both sides of the political
fence as one of the most professional,
efficient, and successful government
agencies, processing vdxii fairness and

dispatch untold thousands of cases every
year.

47

Comments surrounding the appointment of Betty Southard
Murphy as NLRB chair by President Ford give evidence of how
the systemiworked. AFL~CIO Public Relations Director Al Zack
told the press that while Murphy was "obviously not our
nominee," the federation had no objection to her appointment.
Zack went on to describe Murphy as fair and honest, and
stated that the nomination was "well-deserved (Daily Labor
Report, January EL 1975, FL A-8)." At 1mm? confirmation
hearings, endorsements of a:1nnflxmr of unions were entered
into the record.

However, even Murphy was unable to escape the growing
tensions of the late—1970s. When Jimmy Carter became
president he abruptly replaced her, giving the Chair’s job to
John Fanning. While Board members have five year terms
independent of tine president’s” the chair serves an; the
president's pleasure. It was no surprise, therefore, that
Carter should replace Murphy, but Fanning’s appointment had
been leaked to the newspapers even before Murphy had been
told of the change in command. A terse letter to her
successor (Murphy, 1977, p. D—1) began:

Although I have heard nothing at all
from you, or from the current Secretary

of Labor, or from anyone in the White
House, I understand from my newspaper
sources that President Carter will name
you today to replace me as Chairman of
the NLRB.

In interviews, Murphy blamed her shabby treatment on

48

Secretary of Labor Ray Marshall, who she accused of wanting
to become the "labor czar" (Daily Labor Report, April 14,
1977). Her bitter feelings did not subside quickly. When
rumors began to stir that she would be asked to resume the
chair in the mid—1980s, Murphy responded, "There is no way in
God’s world that I would go back to the NLRB." Adding that
she would "rather starve" than return.tx> the Board (Daily
Labor Report, June 14, 1984, p. A4).

Carter’s appointment of Fanning to succeed Murphy was
not welcomed by the business groups, but there is no evidence
that any serious challenge to it was mounted. However, when
John Irving resigned as General Counsel in 1979 and William
Lubbers was nominated to replace him, a major battle broke
out.

Lubbers had been an attorney at the Board for many
years, serving on John Fanning’s staff for two decades. His
nomination was vigorously opposed by business groups and at
least two former Republican Board appointees. The Chamber of
Commerce’s presence at Lubbers’s confirmation hearings marked
the first time ever that a representative of that
organization had actually appeared before Congress to oppose
a Board nominee (Daily Labor Report, December 14, 1979, pp.
F-4—F-10). The objections to Lubbers’s nomination. were
centered (n1 his lack cflf private sector“ experience, but,

moreover, his close relationship with Fanning, who ihad,

49

imprudently, endorsed him publicly. Complaints were raised
that the Board would function in a pre—Taft—Hartley fashion,
with little separation between the Board and General Counsel.
Former Board chair Edward Miller likened the situation to the

"Judge’s brother running for District Attorney (Daily Labor

 

Report, December 14, 1979, p. AA—2)."

It was equally clear that business was at least as
concerned with the potential decisional outcomes as with
protecting tine integrity cu? Board processes. In separate
statements both the National Association of Manufacturers and
the U.S. Chamber of Commerce made reference to Lubbers as the
person responsible for "writing the many pro—union decisions
of John Fanning (Daily Labor Report, November 30, 1979, p. A-
11)." I find it difficult to believe that the business
groups would have been as vociferous in their opposition if
the public records of the tmm>rmai had been more to their
liking. But, in fact, both Fanning and Lubbers were career
civil servants, both were Democrats, both were liberal, and
both typically held positions favorable to labor.

The Lubbers matter dragged on for some time. He mas
given a recess appointment when Congress adjourned in 1979,
and was finally confirmed in 1980 when a Republican
filibustem‘ was overcome. Ifima fight over‘ the :nomination
Proved a mere foreshadowing of what was to come in the early—

19808. When Ronald Reagan was elected to the presidency the

50

system of mutual accommodation was scrapped completely, and
a major ideological overhaul of the Board began (Moe, 1987;
Clark, 1986; Kaufmann, 1991).

C. Open conflict at the Board.

As is by now well—known, Ronald Reagan is something of
a paradox in industrial relations history. He is the only
U.S. president (1) have held tjma presidency cu? a national
union, the Screen Actors Guild, but enjoyed little support
from labor leaders during his time in the White House. His
dismissal of striking air traffic controllers-~one of the few
unions that had actually endorsed him——early in his
presidency is still alluded to by many as a defining moment
for labor-management relations in the U.S. during the 19808
(Bluestone and Bluestone, 1992; Trumka, 1987).

I believe it is significant that Reagan himself (Reagan,
1990) writes that the pivotal events of his life were his
battles math cxmmmnist—dominated unions :U1 Hollywood, his
time spent filming a movie in Emgland during the post-war
Labour government, and his relationship with corporate
leaders while serving as a spokesman for General Electric.
Together, he relays, these experiences taught him the virtues
of industrial capitalism and the dangers of a labor movement
that is too powerful and defines its social mission too
broadly (Reagan, 1990, p. 119). He had supported the Screen

Actors Guild, he writes, simply because it helped fledgling

51
actors during the days of Hollywood's studio system (Reagan,
1990, pp. 89—90).

Throughout his years in political life, Reagan nurtured
these views of capital and labor, and by the time he ran for
president in 1976, his association with the labor movement
was well in the past (Johnson, 1991). He was, for example,
the only presidential candidate in the 1976 election to
affirm his support for Section 14(b) of the Act, which allows
states to adopt so—called "right—to-work" laws prohibiting
union security agreements. All the other candidates,

including George Wallace, expressed opposition (Daily Labor

 

Report, March 17, 1976).

With his election to the presidency in 1980, Reagan was
given the opportunity to enact his beliefs and have a direct
influence on labor policy. Urged on by the business lobby,
and buoyed by time new Republican majority II] the Senateau-
and with conservative Orrin Hatch leading the Labor
Committee——President Reagan disposed of the tradition which
had granted labor (particularly the AFL—CDO) a say in the
selection process (Moe, 1987). Instead, tflma White HOuse

pushed the candidacy of John Van De water to replace John

 

3 The Republican majority in the Senate is critical to

this story. Without it the Administration may not have been
as successful in the appointment process as it was. AFL-CIO
Building Trades leader Robert Georgine felt that the Senate
provided a larger obstacle to labor than did the White House
(Daily Labor Report, January 2, 1981).

52
Fanning as Board chair and Robert Hunter to take the seat of
member John Truesdale.

Hunter had been counsel to the minority members of the
Senate Labor' Committee during the late 1970s. In. that
capacity he worked for Orrin Hatch, and had been credited
with helping Hatch engineer the filibuster that killed the
Labor Law Reform Act. He was also co—author of a 1981
Heritage Foundation report entitled Mandate for Leadership-
—Policy Management in a Conservative Administration. In his
chapter on labor law, Hunter called for repeal of the Act’s
provisions allowing for union security agreements, a ban on
the compulsory use of union hiring halls in construction, and
a number of pro—business changes in the Act’s administration.
Although labor groups did not support the Hunter appointment,
his confirmation hearings were, nonetheless, uneventful
(Daily Labor Report, September 10, 1981).

The same could not be said for the concurrent hearings
on nomination of John Van De Water for chair. The labor
movement, having been excluded from the selection process,
did not view Van De Water as the moderately pro—business type
they had come to expect of a Republican nominee for chair.
Rather, during much of his career he had worked as a legal
consultant to businesses facing union organizing drives, and
had once been quoted in a trade publication as saying that he

had been involved in "130 union drives with the unions losing

53

125 (Daily Labor Report October 16, 1981, pp. F-l—F-2)." Van
De Water was granted a recess appointment, but when the
.Senate Labor Committee returned to his nomination, labor was
able tx> gain enough support t1) deadlock the \Kflxa. Senate
Majority Leader Howard Baker lumi considered relieving the
committee of its control over the nomination and holding a
floor vote despite the deadlock, but Van De Water had, in the
meantime, managed to anger the National Right—to—Work
Committee with a conciliatory gesture toward unions (Daily
Labor Report September 1, 1982, p. A-3). (He stated a
philosophical agreement vnlji agency sflmx) clauses). With
opposition from kxmfli sides, Baker allowed van Ike Water's
nomination to die (Daily Labor Report, November 19, 1981, p.
AA—l).

In retrospect, union leaders may wish that they had let
the Van De Water nomination go through. Undaunted by the
battle over Van De Water, the White House sent up to the Hill
the name of Donald Dotson to be the next chair. Dotson was
described as, anmxx; other things, "staunchly antiunion, a
crusader for the Reagan cause, and a protege of Jesse Helms
(Moe, 1987, p. 270)." He had once written a letter to the
American Bar Association Journal which read, in part (U.S.
House of Representatives, 1983, p. 44):

[C]ollective bargaining frequently means labor

monopoly, time destruction <3f individual
freedom, and the destruction of the marketplace

54

as the mechanism for determining the value of
labor.

Fortunately for Dotson, he had previously gained Senate
approval for a deputy secretary’s position at the Department
of Labor, which made it very difficult. for that body to
oppose his appointment to the Board. It appeared, as well,
that the Senate ‘was tiring' of the conflicts over Board
appointments. Seeing no hope of winning this time, the AFL—
CIO dropped its formal opposition to the Dotson nomination
and he was approved by a voice vote of the Senate (Daily
Labor Report, February 17, 1983 p. A—5).

Confirmed at about the same time as Dotson was Los
Angeles Attorney Patricia Diaz Dennis. Although a registered
Democrat, Dennis was a Reagan supporter, in—house counsel to
the American Broadcasting Company, and a self—professed
conservative with regard to labor relations matters (Dennis,
1986).

Dotson and Dennis joined Hunter, a soon—to-retire Howard
Jenkins, and Donald Zimmerman on the Board. Zimmerman was a
political independent and Carter appointee, who dissented on
a number of critical decisions (see following chapter).
Therefore, a solid conservative majority gained control of
the Board on May 5, 1983, the day of Dennis’s swearing-in.

Recall, however, that General Counsel Lubbers was

appointed in 1980. Since the General Counsel also has a five

55

year term, Lubbers was to remain at his post another 18
months or so. This overlap was perhaps the most contentious
in Board history since the Herzog—Denham days, and revealed
the depth of the ideological rift between the Dotson Board
and its predecessors (Gross, 1995; U.S. House of
Representatives, 1983).

The largest public dispute between Dotson and Lubbers
involved the Chairman’s plan to move control over the Board’s
Enforcement Section, which seeks federal court enforcement of
Board orders, from the General Counsel's office to the
Solicitor’s Office. It was Dotson’s view that attorneys in
the section had not been providing the Board with adequate
representation before the courts. Under his plan, attorneys
would serve at the pleasure of the Board and would be
supervised by the Board Solicitor and Deputy Solicitor, who
are not independently appointed. All pleadings before the
courts would, henceforth, contain the signature of the
Solicitor, not the General Counsel. Dotson was careful to
point out that the enforcement function had not been placed
with the General Counsel by the Taft—Hartley amendments, but
had been ceded to that office by the Board itself in 1955
(U.S. House of Representatives, 1983). The Chairman now
wanted the authority back.

The question union leaders, members of Congress, and

labor journalists were asking was, Why now? Was there some

56

emergency to warrant changing a 30 year old practice? The

ostensible issues CRHI be seen i11 the following exchange,

which took place between Chairman Dotson and Member Zimmerman

at a Board meeting on May 4, 1983. The meeting had been

closed to the public, but the minutes surfaced following a

Congressional
1983).

1301'. son :

Zimmerman:

Dotson:

Zimmerman:

DOCSOI’I:

Zimmerman:

Dotson:

Zimmerman:

DOt SOD:

subpoena (National Labor Relations Board,

What liemn concerned about here is

the way this Agency is perceived by

the public. I'd like for us not to

have the reputation and image of
being high—handed. I’d like us to

be effective iji enforcing the
Statute. I think we can do that and

I think at the same time we will
achieve more respect and credibility

with the public. ZUmi I don’t see

that attitude. I don't see that
from the General Counsel. 21 have

had complaints universally about

that and a lot of that centers
around the activities of these
people [the enforcement lawyers].

Well, our experience has apparently been
quite different.

Well, how many cases have you tried

in the Court of Appeals?

I haven’t tried any, but I have been in
touch in the 2 1/2 years I’ve been on the
Board with attorneys who do across the
country and on both sides and...

Maybe we've been in touch with
different attorneys...

Possibly so, but the statements I have
heard are quite complimentary of this
Agency generally.

I’ve heard and Mdtnessed some the
sleaziest tactics I have ever seen,

the kind of stuff I would not retain

if they [sic] were working for me.

Can you give me an example?

I am not going to give you examples

57

because I don’t want to single
people out.

Upheaval followed. Congress intervened, and a series of
very stormy hearings were held. Massachusetts Representative
Barney Frank pressed Dotson to be mpre specific about what
"sleazy tactics" he was talking about and by whom. When the
Chairman was unable or unwilling to do so, Frank stated (U.S.
House of Representative, 1983, p. 58):

II want to lma straightforwand and
tell you that I am persuaded in part by
what I have read and what I have heard
here today. But the decision to do this
was motivated more by a sense that views
of the General Counsel and those employed
in the Enforcement Section were no longer
representative of the Board majority and
of the Solicitor.

And. that the reason for this, I
think, fairly substantial shift in
authority had to do with a desire to
change the philosophy being expressed.

What Zimmerman, Frank and others were concerned about
was that Immson’s comments about time public image CHE the
Board, etc., were largely subterfuge. “flue Chairman’s real
concern was that Lubbers would represent Board decisions to
the courts in manner assuring an outcome more consistent with
General Counsel’s pre—Reagan Board views of the Act than with
the views expressed by the Board’s new majority. Dotson’s
true agenda, the dissenters felt, was to consolidate

conservative control of the Board (U.S. House of

Representatives, 1983).

58

What was particularly troubling to those testifying at
Congressional hearings was tflmfl: the person Vflu) would gain
immediate authority under the new plan was sitting Board
Solicitor, lhmfli Reilly. Reilly lmmi been 5m1 assistant to
Dotson at the Labor Department, but more important, had acted
as counsel for the plaintiff in Beck v. Communication Workers
(108 S. Ct. 2641, 1988), the now—famous case which placed
strict controls on the allowable use of union dues revenues.
Moreover, Reilly’s time while working on the Beck case was
paid for by the National Right to Work Legal Defense
Foundation, the legal action wing of the National Right to

Work Committee, the most prominent anti—union group in the

nation. The Deputy Solicitor would be Eugene Slade, an
attorney from Member Hunter’s staff (U.S. House of
Representatives, 1983) . Since Dotson had received the votes

he needed to push the plan ahead, not much could be done to
stOp it, and the General Counsel’s role was diminished.

When Lubbers’s term finally expired, the General
Counsel’s job went to Rosemary Collyerq‘ formerly of the Mine

Safety and Health Commission. Collyer was attacked. by

 

‘ For a short period, Wilford Johansen, former head of

the NLRB's Los Angeles office, served as Acting General
Counsel. Later, Johansen was given a seat on the Board along
with. Marshall Babsony a management attorney and. Reagan—
Democrat. The two were appointed in the summer of 1985 to the
seats formerly occupied by Members Zimmerman and Jenkins.
Jenkins’s seat had been vacant for over a year.

59

organized labor for having had relatively little labor law
experience and for her reputation for siding with employers
during her time on the Commission. The AFL-CIO formally and
rather ‘vigorously' opposed lmnr nomination. (AFL—CIO, .1983).
She was confirmed nonetheless, and by the summer of 1984 both
branches of the NLRB were in conservative hands.

D. Impact of the new Board.

The effects of the Board’s transformation were quick and
dramatic. Former member Peter D. Walther (1986), a supporter
of the Dotson Board, noted 29 major reversals of doctrine in
less than two years. Legal scholars, and even a few
employers, became concerned. over the unpredictability of
Board. decisions (Brownstone, 1986; Estreicher, 1986;
Kaufmann, 1991).

A number of Board practices came under immediate
scrutiny. There were the obvious complaints about outcomes,
for example the 300 percent increase (in Dotson’s first full
year) in the number of dismissals of unfair labor practice
charges filed against employers, while dismissals of union

unfair labor practice charges held constant (Bernstein and

Gold, 1985). But, additionally, critics were concerned about
a number of other Board practices: its extraordinary
disregard for precedent (Estreicher, 1986), the huge and

growing of backlog of cases, and its practice of "icing"

cases (Brownstone, 1986).

1"N
J
d

is."

a-

 

“A

5“;

 

 

60

Policy oscillations at the Board are nothing new (and

will be discussed in the next chapter). The Kennedy Board,
for example, reversed a number of Eisenhower Board
precedents, and the Nixon board overruled a few
Kennedy/Johnson Board decisions (Gross, 1995). But by the

mid-1980s legal experts, advocates, and several of the
circuit courts, were growing increasingly restless with the
tendency toward zero—based jurisprudence displayed by the
Board. The Seventh Circuit angrily denied enforcement of a
case simply because the Board had flip—flopped on the issue
so many times (Mosey Manufacturing, 701 F2d 610, 1983). But
even though the Board was never known for a slavish adherence
to precedent, the vigor with which the Dotson Board pursued
policy reversals seemed to some observers extraordinary. In
a number of leading cases, the Board-—now through the

Solicitor's Office-—asked for cases back before the courts

even had a chance to decide on enforcement or remand. New
York University Law Professor Samuel Estreicher (1986)
commented:

The National Labor Relations Board has
made the labor law professor’s job a
nightmare... These days the BNA labor
service treats us to a reversal a week by
the NLRB, sometimes coupled with notice
that the Board has asked a court of
appeals to remand a case to it for
reconsideration. How is my tennis game
going to improve under these
circumstances?

61

The Board’s backlog of cases was also a major concern
during 19803. Again, the issue was not a new one. During
the Kennedy administration a bill referred to as Plan No. 5
was introduced in Congress. The legislation was designed to
speed-up Board processes by granting more authority to lower—
level officials to dispose of unfair labor practice cases.
While the plan had a good deal of support, it was opposed by
key employer groups, and ultimately failed. Again in the
19705, Chairwoman bhummmr had mmestled Mdth.em1 increasing
backlog, particularly after the Board’s jurisdiction ‘was
extended to cover hospitals by the 1974 health care
amendments (House of Representatives, 1975, 1976). As bad as
the problem had been, it got worse in the 19805. But more
disturbing, the backlog was growing even after the number of
cases coming into the Board began declining (U.S. House of
Representatives, 1983a, 1984b).

The House Government Operations Committee met
continually during the 19805 to consider the problem of the
backlog (U.S House of Representative 1983a, 1984b, 1985,
1986, 1987, 1988). Scores of aggrieved workers testified on
how they had been left hanging while their unfair labor
practice charges crawled through the system. Although a few
employers also expressed concern——due to the effects of delay
on the compounding of potential backpay liability——the issue

was particularly a problem for unions and union activists.

62

A clear message is sent about the risks of mulitancy when
workers must wait for several years to find out whether they
will be returned to their jobs. By 1988, the final year of
Dotson’s chairmanship, the median interval between filing a
charge and judicial enforcement of an order was 1,192 days,
or over three years (LaLonde and Meltzer, 1991, p. 964).5

To compare the backlog problem during the Dotson years
with earlier administrations, consider that the average
number of "old" cases still pending at the beginning of a
fiscal year had been 595 for the five years prior to Dotson’s
chairmanship, but 1107 during the five years of his
chairmanship. This increase occurred despite a total decline
in case intake from nearly 60,000 cases in 1980 to fewer than
40,000 by 1988 (U.S. House of Representatives, 1988; National
Labor Relations Board, 1990). In short, intake dropped by
one—third while the backlog doubled.

Chairman Dotson claimed that the extraordinarily high
turnover of Board seats in the early 1980s was the leading
reason for the backlog (Dotson, 1984; U.S. House of

Representatives, 1983a, 1984b). He did acknowledge, however,

 

5 These types of delays occur when the parties can not
agree on the settlement of an unfair labor practice charge.
Most cases are settled fairly quickly when either a Regional
Director finds no merit to a charge or an employer (or union)
agrees to settle a charge informally. If decisions are
contested, however, the process can become extremely
protracted.

63

that at least part of the problem was due to "iced" cases.
Icing is the practice of holding back cases that contain
issues the Board is considering for reversal until a case
with a.rmnxe favorable fact pattern—-and, hence, easier to
argue before the courts-—emerges. Because of the unusually
large number of areas the Board wanted to change, at the end
of Dotson's first year as chair, the Board had 600 cases on
ice, or about half its total backlog. Once off the Board,
member Zimmerman complained, "It would have been preferable
in my judgement to have slated fewer cases [for icing] and to
have decided the rest according to existing precedent rather
than hold up as many as they did (Brownstone, 1986, p. 271)."

By the mid-1980s, it was difficult to find a neutral
voice when the Board was being discussed. The centerpiece
of the controversy was a volume published by the U.S. House
of Representatives in October of 1984 entitled The Failure of
American Labor Law--A Betrayal of American Workers (U.S.
House of Representatives, 1984a). The report followed a
series of Committee hearings during which an overwhelming
number of witnesses——mainly union leaders, lawyers, and
academics—-pilloried the Board and its practices. Several
labor leaders even called for a root-and—branch repeal of the
Act as a means of getting rid of the NLRB (U.S. House of
Representatives, 1983, pp. 6—27; Gould, 1993, p. 153). The

Committee concluded that the law was failing due to

u

u- a.

..A‘

an-

 

64
increasing employer opposition, and the Board's
"administrative crisis," "pmditicization," aumi "bias (UflS.
House of Representatives, 1984a, p. 25)."

Since 1985 marked the Board’s 50th anniversary, there
were plenty of occasions at numerous conferences and in
scores of special edition publications to continue the
debate. Battle lines were clearly drawn. On one side stood
those, like Professor Frank Murphy, who stated (Murphy, in
Morris 1987, p. 77):

The year 1985 marked the 50th year
since the enactment of the National Labor
Relations Act. What should have been a
joyful golden anniversary instead has
been a year characterized by a number of
sad ironies.

It is a sad irony that the statute
which in its early years encouraged the
growth of unionism has increasingly
become a promise to the ear broken to the
hope and is today viewed in many respects
as a hollow mockery.

It is a sad irony that the statute
passed.to promote unionismland collective
bargaining ii; today' consciously' turned
against itself to the statutory
objectives by the most blatantly anti-
union Board in history.

On the other side, stood those, like labor attorney
Charles Bakaly, who in response to Murphy said (Bakaly, in
Morris, 1987, p. 94):

It is self-evident that the Board is
not responsible for foreign imports or

deregulation. Nor is ii; directly
responsible for the recent ability of
employers to operate during

strikes...[C]ases are being processed in

65
much the same way as they have in the
last 20 years. Admittedly unions are not
winning 84 percent of the cases against
employers as they were during the Carter
years...[But] [t]o argue that this Board
is not fulfilling its statutory
obligations is simply without foundation.

It is even more interesting to compare the comments of
labor officials at the time of the Board’s 50th anniversary
with those Ci :hmfi: ten years before. While attending a
celebrathmi of the Emerd’s 40th anniversary-—also, by the
way, during the time of a Republican administration-~AFL-CIO
President George Meany (1975, D—l) remarked:

[I]t is only fair to state the obvious-

-the NLRB, laboring in cmmacof the most
adversary—prone areas of government, has

a commendable record. With few
exceptions, none current, it has been led
by men and women of honor, whose

capability and expertise are well known.

It has been staffed by men and women
who are truly public servants, whose hard
work and devotion to duty is often
overlooked...

I have had the opportunity to
observe the workings of a host of state
and federal agencies and my experiences
is that the PHJUB has been consistently

far better than most. In this imperfect
world, your sustained high level of
performance is ea substantial
accomplishment.

No such kudos were in store for the Board upon its
golden anniversary; I found no evidence the AFL—CIO's
participation in any anniversary festivities. Instead,
federation representatives wrote (Bernstein and Gold, 1985,

213):

66

[I]f labor chooses to celebrate this
landmark anniversary, it will do so with

anger toward the Act’s principal
custodian, the National Labor Relations
Board.

Why is the act’s golden anniversary
so bittersweet? ..Ronald Reagan has
packed the NLRB in management’s favor to
an unprecedented degree, converting it
ideologically into a weapon against the
labor movement.

And concerning the Chairman himself, AFL—CIO General
Counsel David Silberman added (Daily Labor Report, June 5,
1985):

The short of time matter‘ is that the
principal responsibility for
administering the Wagner Act has been
entrusted to a man who is hostile to the
fundamental purposes of the Act, and to
labor organizations which seek to promote
those aims. And under Mr. Dotson’s
leadership the Labor Board has proceeded
to systematically subvert the purposes of
the law.

Whether Dotson and the Board "subverted the purposes" of
the law will be considered in the next chapter when doctrinal
changes are reviewed. However, the evidence presented above
indicates that the first half of 19803 were, indeed,
turbulent years at the Board. From 1980 until 1983,
conflicts involved mainly the appointment process as the
Reagan administration took shape. Later, between 1983' and
1986, disagreements arose over major doctrinal changes.

After 1986, things did calm a bit. First, the

Republican Party lost control of the Senate, which meant that

67

Orrin Hatch yielded the chair of the Labor Committee to
Edward Kennedy. There was an obvious impact on the
appointment process as much more moderate individuals were
put forward by the White House. In fact, conservative forces
opposed the appointments of Chairman James Stephens and
Member John Higgins (Daily Labor Report, December 21, 1987,
A-7).

But, moreover, there had been so many major reversals
between 1983 and 1986, that there were very few areas of
policy left to be disturbed. Therefore, the remaining years
of the decade were relatively peaceful, with the Board seeing
many of its decisions upheld by an increasingly conservative
judiciary (Lansing and Podhajsky, 1990), which itself had
been subject t1) an equally rigorous ideological reshaping.
Goldman writes (1989, p. 319):

The Reagan administration.was responsible
for major innovathmns in the selection
process [of federal judges] which endured
throughout the two terms. The Office of
Legal Policy was the locus of judicial
selection activity. The screening
process was systematized and for the
first time III the history CHE judicial
selection, all leading candidates for
judicial positions were hmought to
Washington for extensive interviewing by
Justice Department personnel...Arguably,
the Reagan administration was engaged in
the most systematic judicial
philos0phical screening of candidates
ever seen in the nation’s history,
surpassing Franklin Roosevelt’s
administration.

68

One manifestation of this new system was that throughout
Reagan’s entire presidency, not one Democrat gained an
appointment to the Circuit Courts of Appeal. Not since the
presidency of Warren G. Harding had the opposition party been
passed over completely for such appointments. Further, the
Circuits Courts are the first forum for the appeal of NLRB
cases.

However, because the federal court system is so many
times larger than the Board, and appointments are for life
rather than five year terms, change comes much more slowly.
It was not until timeemmi of George Bush’s presidency that
those appointed after the election of Reagan constituted a
majority on the federal bench. Also, due to court practices
and procedures, it is difficult to determine just how much of
a "Reagan court effect" there may have been on labor
relations decisions. Sometimes cases are heard by individual
judges (who may or may not have been Reagan appointees),
other times they are heard en banc, or by a committee of
judges; cases may be heard in the circuit within which the
dispute arose or e1sewhere——particularly in the District of
Columbia Circuit; some judges eagerly apply precedents from
other circuits, some do rmm” In short, there was almost
certainly a shift toward a more conservative judiciary during
the 1980s, the pmecise extent to vflfirfli it affected labor

relations public policy is more difficult to determine

69
(Smith, 1993).
E. conclusions.

During the early to mid—1980s, the NLRB became the
battleground for disputes over national labor policy to an
extent not seen in four decades. The conflict was
foreshadowed In! the chill ii1 relations Ibetween. top-level
business and labor groups during the 19705 following,
particularly, the failure of the Labor Law Reform .Act.
Things became much worse, however, when the Reagan
administration, with much encouragement from the Senate, cut
organized labor out of the selection process and embarked on
an ideological reordering of the Board. Selecting nominees
with <direct ties to some: of <organized labor's bitterest
opponents seemed openly provocative and confrontational.
Although moderation returned to the Board by the end of the
decade, its reputation among unionists, members of Congress,
academics, lawyers, enmi others was severely' damaged. A
number of national unions were even asking local affiliates
to boycott the Board.

Finally, as the House of Representatives (1984a) wrote:

[Ilt has been the decisions, more than
anything else, which have caused the
collapse of confidence in the objectivity
of the current Board. Looking at the
body of the current Board’s decisions as
a whole, two facts stand out. The first
is that virtually all the unusually high

number of policy reversals of this Board
have favored management. In addition to

70
undermining time perception.<3f fairness
and objectivity, the decisions of the

Board have altered the substance of law
in a manner contrary to the Act.

In the next chapter, I consider this final point in

light of the model of local union disappearance discussed

earlier.

Chapter IV. Doctrinal changes.

A. Introduction.

In this chapter I examine the doctrinal changes of the
mid—1980s, which followed the reshaping of the institutions
described in the previous chapter. The issues examined relate
to the model of local union disappearance presented in Chapter
I - rI‘hey are: 1) an employer’s obligation to bargain over

I3’LISiness decisions that affect the bargaining unit; 2) a

union's right to maintain internal discipline and allegiance;

3 ) an employer's right to object to the status of the union
as bargaining representative and to intervene in the
de C e rtification process . The evidence presented indicates

 

M, although the Board has always oscillated in its

Msions in these areas, reversals in doctrine between 1983
m1986 were some of the sharpest and most broad-ranging
%. All of these changes were favorable to business owners
or dissenting union members, reflected a conservative view of
property relations and rights of association, and may have

presented new difficulties for unions trying to maintain their

presence in the workplace.

71

72

B. The duty to bargain .in good faith.

When Congress first passed the Act in 1935, it included

within Section 8 the clause: ”It will be an unfair labor

Loria<:t:ice for an employer to refuse to bargain collectively

vwit:r1 the representative (n3 his employees..." Originally,

Chorigngess did not define what "to bargain collectively" meant.

TfijeEITEEfore, in 1947, Section 8(d) was added to the Act. The

sseaczt;j_on is quite long, but its essential elements are

(3C3r1tléiined in the first paragraph:

For time purposes <3f ttus section, to
bargain collectively is the performance
of the mutual obligation of the employer
and the representative of the employees
to meet at reasonable times and confer in
good faith with respect to wages, hours,
and other terms and conditions of

employment...
Section 8(d) answered a few questions, but raised

C) . . .
tllileirs: What constitutes "good faith?" and what is meant by

C>t:1‘1er terms and conditions of employment?" As the Supreme
C: . . .
CDIJJETt pOinted out in NLRB v. Insurance Agents International

W (361 US 477, 485 (1960)), "[C]ongress was generally not

(2 . . . .
C>r1Cerned vnifli the substantive terms (x1 which time parties
(:Cbtltlracted." Justice Brennan, the decision's author, went on

thD quote Senator Wagner who had stated:

When the employees have chosen their
organization, when they have selected
their representatives, all the bill
proposes to do is to escort them to the
door of their employer and say ’Here they
are the legal representatives (ME your

73
employees.’ What happens behind those
doors is not inquired into, and the bill
does not seek to inquire into it (361 US
477, 484).
While accepting the general philosophy of the Senator's
sstéat:<ement, the Court also realized that there would be

EXKCDID:LemS. Brennan wrote:

Obviously there is tension between
the principle that the parties need not
contract (N1 any' specific terms emui a
practical enforcement of (fine principle
that they are bound to deal with each
other in a serious attempt to resolve and
reach a common ground (361 US 477, 486).
3:I1 fzact, what happens "behind those doors" must be inquired
j—IWCCLC3 if the Board is to determine whether the duty to bargain
‘ljrl ggood faith 1mm; been breached. Nonetheless, the Court
C3‘:>I'1t:_inued to hold that the government should, to the extent
1E><:’Ssssible, refrain from CMctating the substantive terms of
negotiations.
As the law developed, however, questions continued to
arise. A couple of years before Insurance Agents, the
SL1IDreme Court began developing a line of reasoning that would
1‘6361Ci eventually to a three part taxonomy of issues (NLRB v.
2332S1§ter Division of Borg Warner Corp., 356 US 342, (1958)).
ISSues that have as their primary focus the employment
rGalationship are mandatory subjects of bargaining and must be

discussed if raised by either party (e.g work rules,

IDEnsions, and benefits); issues that clearly violate the Act

74

or any other statutes are illegal subjects of bargaining and
must: not be discussed at all (e.g. race—based work
assignments, closed shop provisions); all other issues are
permissive subjects of bargaining, and may be discussed by

mutual consent of the parties.
Today, there is not a great deal of controversy over
i .1 legal subjects,1 but there are volumes of cases in which the
parties have disagreed on whether an issue should be
Classified as mandatory or merely permissive.2 The
r“<31’1Ciatory/permissive distinction is an important one (Sockell

and Delaney, 1986, Sockell, 1986, 1989). If a union raises

X

1 An exception is the controversy over work preservation
CJ~6l‘Lises (Hieb and Moody, 1981; Moon, 1987; Rappa, 1986;

 

 

ROSenberg, 1984; Steffen, 1989; Zidich, 1984). In general,
C lauses which preserve work are legal, but clauses designed
to acquire work are not (National Woodwork Manufacturers
Mciation v. NLRB 386 us 203, 1964). With changing
aechnology, this distinction is not always easy to make. In

S eries of Board and Supreme Court cases, it was determined

:ha t- a provision in the Longshoremen’s Association’s contract
w:quiring that its members unload ship containers in
s i réhouses miles from shore was legal since the work done was
Lomllar to the work ILA members had always done (International
Whoremen’s Association v. NLRB (ILA I) 447 US 490 (1980);
Wnational Longshoremen’s Association v. NLRB (ILA II) 973
DJ: 61 (1985)). A further issue could arise if a work
.. eLaaervation clause appeared to be a illegal attempt at
foeatherbedding." Unions can demand that employees be paid
V unnecessary work, but not for doing no work at all (NLRB

1':
Wmle Enterprises, Inc. 345 US 117 (1953)).

3 In most cases it is the union arguing that an issue
<3 atgement sees as its sole prerogative to decide should be
Casandatory subject of bargaining. But this is not always the
mate - Sometimes employers Wish to bargain over internal union

QQnCEIs, such as the procedure by which members ratify a
tract; in fact, Borg Warner was one such case.

man

75
a permissive issue—~say, the price a company charges for its
product--the employer may simply refuse to discuss the
subject without violating the law. Just as important, there
is no obligation for either party to fulfill information
requests, nor may the parties engage in strikes or lockouts
over permissive issues. If a subject is mandatory, however,
the parties must bargain and must provide each other with
necessary information about the issue. Further, in the
absence of an agreement an employer may implement a mandatory
issue only after the parties have bargained to a bona fide
impasse. And, finally, when dealing with a mandatory
EsLllDject, the parties may reinforce their positions with
St- rikes, lockouts, or other such pressure tactics (assuming

no Contractual agreement to the contrary is in force) (Cox,

80k and Gorman, 1986) .

The mandatory/permissive distinction becomes most
(ii f ficult to apply in cases involving business decisions that
a f f ect bargaining unit jobs or the bargaining unit itself:
QJ‘Qsaing or relocating an establishment, investing in new
t:eel’inology, or subcontracting unit work. The difficulty is
QCDIT‘lDOunded when it is unclear whether a business decision is
“IQ: i\fated by a labor issue, or is made simply in response to,
Say . changing market conditions or the availability of new
t_QQ1-'1riology. The critical question is, Does a business

de - .
Qlsions automatically involve a “term or condition or

76
employment" when it has a collateral effect on the bargaining
unit and bargaining unit employment? Although this has never
been an easy question to answer, during the 19805, and
particularly between 1983 and 1986, the answer became more
favorable to employers (Barton, 1986; Dube, 1986).

1. An employer's duty to bargain over business
decisions affecting the bargaining unit.

There are differences in substance between plant
Closings and relocations, subcontracting, and technological
SUbstitutions, but the legal issues concern the same point:
leSt; an employer bargain over such decisions? As Gacek
( 1981) noted, legal reasoning in this area has developed
along three lines. At times the Board and courts have
at(jQEDt:ed a standard of "open bargaining," which offers a
S t— rorig presumption in favor of mandatory decision—bargaining.

At other times, such decisions have been considered
presumptively "management rights" with a requirement only to
bargain over the "effects" of a decision, but not the
decision itself.3 And the third approach has been a

pragmatic balancing of rights," where the benefits of

be
I gaining versus unilateral action are conSldered on a case-

/\

3 It must be noted that even when an employer has no

mil lgation to bargain over a busmess deClSion, the employer
uh . t- bargain over the effects of that decision on bargaining
relt employees. Issues might include severance packages,

hat— raining, outplacement assistance, etc. At least one author
es questioned whether there is a meaningful difference
t"Ween decision and effects bargaining (Kohler, 1983) .

77

by—case basis.

Because of the differing and wavering positions of the
Board, the Circuit Courts, and the Supreme Court on this
iesssLie over the years, it is a very difficult area of law to
:fc31_lxow (Gacek, 1981; George, 1985; Irving, 1982; Millspaugh,
1984; O’Connell, 1986). It is equally difficult to make
g;ear1&en:al statements about what the law has been at any given
CLignnee.. Frequently, different standards have been applied in
C341 ffifeerent Circuits, and, at times, Board decisions seem to
have contradicted even the Supreme Court. Nonetheless,
if(DJ—lowing a series of mid—1980s Board decisions, the
1:>EiI-Ԥ3'aining mandate reached its narrowest point in the history
(2’13 the Act (Ashford and Ayers, 1987; Barton, 1986;

COliingswortn, 1993; Dube, 1986; Hedlund, 1986).
Case law in this area began with Gerity Whitaker Company
( :3 :3 IqLRB 393 (1941)). In that case the Board determined that
t:'171633 <:ompany had breached its duty to bargain with the Metal
‘E3<:)-1—ii¢shers Union when it unilaterally terminated operations in
<:>r)”63 pdant and moved them to another. Five years later in
‘Zglé‘111125en Roller Bearing (70 NLRB 500 (1946)) the Board issued
51 EsJiInilar finding iii a subcontracting case. However, at

€11:N:)\l .

t the same time, the Board suggested that an employer
C2<:)Lll<:1, perhaps, take a unilateral action that would terminate
:3<:it353 so long as the employer was not doing so out of anti—

\111‘
J‘CDYI animus (Pepsi Cola Bottling Company of Montgomery (72

78

rJLJRB 601 (1947)).
When the Eisenhower Board was seated, it preferred the
reasoning of Pepsi Cola to that of the earlier decisions
(:seeea, for example, Bickford Shoes, 109 NLRB 1346 (1954)).
ZAIICi in its final year before yielding to a Kennedy—appointed
rnaa:ic>rity, the Board made one its most sweeping decisions in
t:r1j.ss area. In Fibreboard Products Company Immx (130 NLRB
-1—ESESE3 (1961)) the Board held that an employer had the right to
tLEEJCITIinate all bargaining unit employees in favor of a
Subcontractor without bargaining over the decision. The
E3<3>Eiitxj distinguished the present case from earlier ones in

E""]=-'7|-:3L<:l“1 the union would continue to represent remaining

earnI:>J.<oyees. In those cases, the Board argued, the union had
an interest in the impact of the employer’s decision on
:tT‘EBIWHEiining employees. But where no bargaining unit employees

ZET‘EEIT161ined, the union would have no interests.
The Kennedy Board was quick to reverse the Fibreboard
Q‘l‘eeision. And, in Town and Country Manufacturing Company
( :L-:3 ES NLRB 1022 (1962)) the Board stated unequivocally:

[E]liminathmi of unit jobs, albeit for
economic reasons, is a matter within the

statutory phrase ’other terms and
conditions of employment’ and is a
mandatory subject of collective

bargaining within the meaning of Section
8(a)(5) of the Act.

1:
I)‘ tilie same year, it echoed that logic in Adams Dairy, Inc.

(137 NLRB 815 (1962)).

79

The Circuit Courts had considered this issue several
tzirnes (see, for example, Jay Foods Inc. v. NLRB, 292 F.2d.
.3Ll'7 (1961); NLRB v. Rapid Bindery, 293 F.2d. 170 (1961)), but
until the mid—19608 the lone Supreme Court case had been
(Dirciearr<af Railroad Telegraphers \1. Chicago and.llflv. Railway
<2c>rnr>éany (362 [K3 330 (1960)). While supporting mandatory
Cieec::Lssion—bargaining, time Telegraphers case lumi been issued
under the Railway Labor Act (45 USC 151) not the NLRA.
-P1CDXA7€3\/er, in 1964 the Supreme Court granted certiorari in the
-1:Zi.k31rweboard II case mentioned above (Fibreboard Products v.

W 379 us 203 (1964)).

The issue in Fibreboard was whether an employer could
‘LlliliL-luaterally subcontract bargaining unit work upon the
EB:"(ISIM‘Lration of an agreement without first bargaining over the
<j1€33(:=j_sion with the union. The Eisenhower Board had said yes,
t:;171633 IKennedy Board had said no, and the Circuit Court had said
Iil<:) The Supreme Court agreed with the latter two decisions.
IE3 . .

€33<:56Iuse labor costs appeared to be such a prominent factor in
tllrlese . - -
company s deCiSion to subcontract, and the work to be
$3 by the subcontractor was identical to the work being
63 by bargaining unit employees (they were janitors), the
egg . - ,
II‘t: determined that the company had a duty to bargain over
it
t:'55 (decision to subcontract.

The Court was clear in stating that the facts of the

Q
EiESEE were extremely important, and that it did not wish to

80
(zireate a strong presumption in favor of mandatory decision—
kpaurgaining in all cases where business decisions eliminated
chsz. Vhiting for the Court, Chief Justice Warren urged a
cnaasee—by—case approach, and iilea concurring opinion, Potter
Stewart wrote (379 US 203, 223):
[T]here are other areas where

decisions by management may quite clearly
imperil job security, or indeed terminate

employment entirely. An enterprise may
decide to invest in labor—saving
machinery. Another may resolve to

liquidate its assets and go out of
business. Nothing the Court holds today
should be understood as imposing a duty
tt) bargain. collectively' regarding' such
managerial decisions, which lie an: the
core of entrepreneurial control.

In response to the Supreme Court, the Board developed a
Esth-Elrlcflard for subcontracting cases. In Westinghouse Electric
( :L-ES C) NLRB 1574 (1965)) the Board held that an employer would
171(2):: 13e required to bargain if the decision to subcontract was
motivated by economic reasons only; it had been the past
DraCtice of the employer to subcontract various kinds of
‘AICDZEiki; the employer’s present action did not vary to any great
CiEESQTrree from its established practice; there was no
ES ‘ . . . . . .

lSgnificant detriment to bargaining unit employees; and, the
r11On had had an opportunity to bargain over subcontracting
i. , . .
SSESJJeas during the most recent round of negotiations.
5&7 . . . .
‘EEEStilnghouse resulted in 51 strong presumption in favor of

k>-
Qrgaining (Gacek, 1981).

81

Shortly after Fibreboard II the Supreme Court returned
to the issue of mandatory bargaining under a somewhat
different set of facts. In Darlington Manufacturing v. NLRB
(380 US 263 (1965)) the Court was presented with a case in
which the employer had closed one of its facilities in
response to a successful union organizing drive. The Board
asked the Court to find the employer guilty of anti—union
animus in its decision (a violation of NLRA, Section 8(a) (3))

and of a failure to bargain over the decision to close.

The Court determined that, indeed, the company was in
Violation of the Act, resting its decision on the key fact
that. the Darlington plant was one of many owned by Deering
Mi lliken of New York City. The Court accepted the NLRB’s
View that in multi-plant situations the closing of a single
J‘E<31Cility in response to an organizing victory could put a
.. C1'1i11“ into union drives elsewhere, and was, therefore, the
CYpe of behavior Section 8(a) (3) sought to proscribe.‘ The
CQUrt largely skirted the duty to bargain issue, however, but
did add that a complete shutdown of the firm would have
Q1iatnged its decision entirely (380 US 263, 274):

\

Co ‘ Some commentators have noted the irony of the Supreme
Qlul‘t's concern about “the. potential impact on employees
DaSevvhere, but relative indifference to the direct effects on
Ehrlington workers. Clyde Summers wrote: "This is to see in
Die execution of. hostages nothing more than an intimidation
1‘1 l the liVlng; it is to make murder a crime only when the
lers purpose is to instill fear (Summers, 1965, p. 67)."

82

We hold that as far as the Labor Act is
concerned, an employer has the absolute
right to terminate his entire business
for any reason he pleases.

During this same period, several of the Circuit Courts

were relying heavily on Justice Stewart’s Fibreboard dicta to
create a presumption in favor of management’s unilateral
right to make business decisions (see, for example, m
Royal Plating and Polishing Company, 350 F.2d. 191 (1965);
Adams Diary 350 F2d 108 (1965)). The Board, however, was
heading in the opposite direction, and using Darlington as an
argument for creating a presumption in favor of mandatory
dec ision-bargaining over partial closings.

This outcome is most obvious in the case of M
W (161 NLRB 561 (1966)). Although Ozark was not
C eCElfinically a multi—plant operation, the Board considered its
CCDI”T‘IInon ownership with several other companies and the
integrated nature of its operation in concluding that Ozark
was a multi—plant operation "under the Act." Further, the
Board gave very little weight to company's assertion that its
mQVe had been made for solely economic reasons and focused,
instead, on the impact on employees. In a sweeping

philosophical statement, which deserves to be quoted at

l
eh91th, the Board wrote (161 NLRB 561 at 565):

83

With all respect to Court of Appeals for
the Third and Eighth Circuit,5 we do not
believe that the question of whether a
particular management decisions must be
bargained about should turn.cm1 whether
the decision involves the commitment of
investment capital or on whether it may
be characterized as involving a ’major’
or ’basic' change in the nature of the

employer’s Zbusiness. True i1: is that
decisions of this nature are, by
definition, of significance for the
employer. It is equally true, however,

and not tx> be lost sight cfifi that an
employer's decision tx> make 51 'nejor'
change ijl the nature cu? his business,
such. as tflma termination. of 61 portion
thereof, is also of significance for
employees whose jobs will be lost by the
termination. For just ems the employer
has invested capital in the business, so
the employee has invested years of his
working life, accumulating seniority,
accruing pension rights, and developing
skills that may or may not be salable to
another employer. And just as the
employer’s interest in the protection of
his capital investment is entitled to
consideration 111 our interpretation of

the Act, so too is the employee’s
interest in the protection of his
livelihood.

In effect, the Board was creating a job right on par
‘VVTi—t:}1 an employer’s property right (O’Connell, 1986). It is
IDITCDIDablynot surprising that the Circuit Courts generally had
‘tLITCDIihdtzwith this View (see, for example, NLRB V. Transmarine
ESEEL“EC-iézation Company 308 F.2d. 933 (1967)). Further, when the

IQ” .
len Board was seated, it retreated from such a strong

\

5

E3<2> The Third and Eighth Circuits had refused to enforce
53133 orders in Royal Plating and Adams Dairy, respectively

84

bargaining presumption in a couple of noteworthy cases. In
general Motors Corporation (191 NLRB 951 (1971)), the Board
held that the sale of a retail business constituted a major
Change in operations and was, therefore, not a mandatory
subject. And, in Summit Toolinq (195 NLRB 479 (1972)), the
Board came to a similar conclusion in a case involving a
partial closing.

But following Jimmy Carter's election, and the elevation
Of John Fanning to chair, the Board resurrected its Ozark
Trailer standards. In fact, in at least one case, the Board
af f :ered6 an opinion that relied almost exclusively on Ozark
Trailers and completely ignored General Motors and Summit
TOOling (ABC Trans—National Transport Inc. 247 NLRB No. 25
( 1980)). In other cases, the Board was more careful to

raise, but then dismiss, the General Motors and Summit
Tooling rationale (Midland—Ross Corporation, 239 NLRB 323
( l 9'78); National Family Opinion, 246 NLRB No. 84 (1979)).

In Brockway Motor Trucks v. NLRB (582 F2d 720, 1978) the
Third Circuit was given the opportunity to review the issue
of whether a partial closing was subject to mandatory
decision-bargaining. Recall that the Ozark Trailer standards

'\

16:09te Aecording to Board procedures, if a charged party wishes to
ab fOrQ a Regional Director’s decision, it is entitled to'a‘ trial

1% an Administrative Law Judge. If the ALJ’s dec151on 18
to the full Board, the Board may simply affirm it without

COInment, as was done in this case.

 

ea
furthgr

85

vveare developed in the breach: the Board had used a Supreme

(:cnart decision stating that total closings were not subject

t;c) Inandatory bargaining to create a presumption that partial

Brockway the Circuit Court refused

Cl osings were. In
eeriifcarcement on technical grounds——the Board had not presented

facts on the case——but went on to

jgt vvith sufficient
sgaee<2111ate about what the standards for mandatory bargaining

iii SSLJCh cases might be. Most notably, the Court stated that

tldeezfee should be a presumption in favor of bargaining unless

t:k1€2 eemployer is in dire economic straits and quick decisions

must; be made, some third party action has already made

.kDEiIIngéiining fruitless, or bargaining would jeopardize pending

r1e§J<Dtiations with another entity. The Court went on to state

that: (582 F2d 720, 732—733),

Saying that there is no duty to bargain
about a partial closing when the employer
is faced with the 'economic necessity’ of
moving or of otherwise restructuring the
operations of a ’failing business’ is a

is far cry from indicating that an
employer has no duty to bargain about a
partial closing when its action is

founded on ’economic considerations’

whatever they may be.
Ix year after Brockway' the Board found 61 janitorial

Es . . . . -
6313\f143es company in Violation of the Act when it refused to

k) .
éifthfiin over the dismissal of workers follow1ng the

c: . .
EarleGallation of a service contract (First National

b4j -
W. 242 NLRB 462 (1979)).

The firm in the case was

86

First National Maintenance (FNM), a supplier of janitorial

services in New York City. When Greenpark Nursing Home
cancelled its contract with FNM over a disputed fee, the firm
laid-off its Greenpark-based workers. The company refused to
discuss the matter with union representatives, claiming that
the decision was irreversible and that any negotiations would
be futile.

The Board, however, considered FNM's layoff a partial
closing since the company had contracts elsewhere. It,

therefore, found FNM in violation of the Act for refusing to
bargain over its decision to terminate the workers following
a partial closing. The Board ordered FNM to give the workers
backpay and find them jobs at other locations serviced by the
C:C>l”‘npany. The Second Circuit Court enforced the Board’s
Order, but called into question its per se standard, derived
f rom Ozark Trailers, that partial closings were a mandatory
ESkllsject of bargaining. Instead, the Court suggested a
1Telbuttable presumption in favor of bargaining that could be
Qvércome if strong enough mitigating circumstances were found
( 627 F2d 569, (1980)).

The Supreme Court, with Justices Marshall and Brennan
dissenting, overturned both the Board and the Second Circuit.
It ruled that a management decision, if its purpose was

"apart from the employment relationship" was a mandatory

a .
L113ject of bargaining only when it can be shown that the

87

benefil:s to labor—management relations émui the collective
bargairiirmz process outweigh the burden. to the employer.
Applyiiig that rationale to the case at hand, it found that a
decisix:n.to close part of a business was beyond the scope of
mandatrary bargaining (452 US 666 (1981)).

Jiistice Blackmun, writing for the majority, stated that
the dexoision of the employer (452 US 666, 677),

had a direct impact on employment, since
jobs are inexorably' eliminated 13/ the
termination, but had as its focus only
the economic profits of the contract with
Greenpark, a concern under fact wholly
apart from the employment relationship.
This decision is akin tt)ea decision of
whether to be in business at all.

Iiea then wrote (452 US 666, 679):

In view of an employer’s need for
unencumbered decision making, bargaining
over :management decisions that have (a
substantial impact on the continued
availability of employment should be
required only if the benefit, for labor—
management relations and the collective
bargaining process, outweighs the burden
placed on the conduct of the business.

Paid in applying such criteria to the FNM case, he
ccwuzltujed (452 US 666, 686):

[T]he harm likely to be done to the
employer and an employer’s need to
operate freely' in. deciding ‘whether to
shut down part of its business purely for
economic reasons outweighs the
incremental benefit that might be gained
through the union’s participation in
making the decision; and we hold that the
decision itself is not part of 8(d)’s
terms and conditions over which Congress

88
has mandated bargaining.
One cuirtious aspect of the Court’s decision is that it, at
first, :recommends a balancing test, but rather than allowing
this teest to be conducted on a case—by—case basis, concludes
\Nith ea new lper se standard against mandatory decision—
bargaquing in partial closing cases (Somson, 1984).

'True FNM decision was a Rejor reversal of doctrine and
resultxed.:u1 the immediate dismissal of scores of partial
Closing cases. However, General Counsel Lubbers, in his
menmnaarida to the NLRB regional offices, urged a very careful
reading of the FNM decision. In particular he focused on
Footrnot:e 22 (452 US 666, 686), which states:

In this opinion we of course intimate no
view as to other types of management
decisions, such as plant relocations,
sales, other kinds of subcontracting,
automation, etc. which. are t1) be
considered on their particular facts.
In Iliss Memoranda to Regional Directors, the General Counsel
wrtflle (Lubbers, 1981, p. E-l).:
[T]he Court expressly limited its holding
in First National Maintenance to
economicallyr motivated.(decisions tx> go
partially (mm; of hmsiness. The Court

stated that it intimated no view as to
other types of management decisions, such

as plant relocations, sales,
subcontracting, automation, etc.
...Thus, for example, decisions to

relocate a plant, to subcontract, to
eliminate unit jobs through automation,
or to consolidate operations are not akin
to a decision whether to be in business
at all. That is, the employer does not

89

intend to withdraw from the business
activity, either wholly or partially.
Instead the employer intends to remain in
the same business, albeit elsewhere
(relocation), or through the use of
subcontractors (subcontracting), or with
different equipment (automation) (n: at
one location rather than several
(consolidation).

.. If the factors which indicate
that time decision if; amenable tc> the
process of collective bargaining clearly
outweigh the factor that indicate that
bargaining would be ’burdensome’ to the
decision-making employer, the Region
should issue 61 complaint. This result
would be consistent to Board law prior to
First National Maintenance.

TTue Lubbers guidelines may have blunted the full effects
of Firmst:.National Maintenance for several years. The Board
contirniexi to order bargaining routinely in cases of
relocatixon, subcontracting, and technological change. But
When thee Reagan Board was in place (following Member Dennis’s
SWearirug-—in on May 5, 1983) mandatory bargaining was severely
limited :in.these other areas as well. The Board effected this
Change vvilih a reversal of its 1981 decision in Otis Elevator
(225 N1J{E3 235, 1981). The new Board asked for the case back
While tlaee company’s petition for review' and. the ZBoard’s
DEtitiOrl ‘for enforcement stood before the D.C. Circuit Court
(296 NLRB 891, 1984).

IT“? <:ompany, a division of United Technologies, sought
to (“NHSCXLidate its New Jersey operations with operations in

C01) . .
nectJ—Cut. The move would have ended a collective

9O
bargairring relationship with the United Auto Workers which
represented some of the New Jersey workers. The company
claimemi that the consolidation was a major capital decision
and, truerefore, beyond the scope of bargaining, but the Board
had dishagreed and ordered the company to bargain.

.Affl:er achieving remand, the Dotson Board reversed the
earlier‘ decision. As evidence of the difficulties in this
case, tliree separate decisions were handed down by the four
sitting) Board members (Member Jenkins’s vacated seat had yet
to be fiilled). Chairman Dotson and member Hunter formed the
pluralcit§/, and in their decision wrote that Otis’s move (296
NLRB 891. at 893),

turned upon a fundamental change in the
nature and direction of business, and for
that reason is excluded from the limited
area of bargaining described by 8(d) (my
emphasis).

They contrinued (296 NLRB 891 at 894):

Despite time evident effect on
employees, the critical factor to a
determination whether the decision is
subject to mandatory bargaining is the
essence of the decision itself, i.e.
whether it turns upon a change in the
nature or direction of the business, or
turns upon labor costs; not its effect on
employees nor the union’s ability to
offer concessions.7

An ‘ . . .
d‘1J1 Language clearly aimed at the Lubbers Guidelines (269

NLRB 891 at 894)

\—

7
COI“pare this language to Ozark Trailers above.

91

Such decisions include, inter alia,
decisions tx) sell ea business cm (a part
thereof, t1) dispose cfif its assets, to
restructure or consolidate operations, to
subcontract, tt) invest iii labor-saving
machinery, to change the methods of
finance or of sales, advertising, product
design, and all other decisions akin to
the forgoing (my emphasis).

Urnjer the plurality’s rationale, unless a business
decisixanq in an increasingly limited number of areas, "turned
upon" lxabor costs, the issue was not a mandatory subject of
bargairring regardless of its impact on the bargaining unit or
unit jcibs. Members Dennis and Zimmerman urged a less
restricfi:ive standard that would have allowed unions a role in
discmunsiiig a broader range of economic issues (although
Dennis st:ill felt that the burden remained on the union to
Show tluat: an issue should be bargained). Dotson and Hunter
were IKJC persuaded. The Ckis rationahe was adopted and
represeuiteed the narrowest view of bargaining ever advanced by
the Boarci,

FUIWllier, the CMis rationale was quickly spread by the
B . . .

Card to 51 number of other areas (see, in addition to those
di . . . .
scusseci below, Columbra City Freight Lanes 3TWL NLRB 12

(1984); gigaser shipyard 272 NLRB 496 (1984); Boetrem

D' ' -
‘£!$§ASELL_JQOR, Inc. 272 NLRB 999 (1984); Kroger Co. 273 NLRB
46 .

2 (19854)). For example, in the subcontracting case Gar

WOO .
.iiiiLLEflggggig (274 NLRB 23, 1985), a Dotson/Hunter plurality

92
decidexi that the company’s action, although not unlike many
previCNJS subcontracting cases, («n5 undertaken 'Wxi prevent
econondx: chaos," and was, therefore, not amenable to
resoltu:ion through bargaining.

Sijnilarly, the Board’s rationale with regard to
technological substitution also changed. In one of its
earlixest; cases in this area, Renton News (136 NLRB 1294 at
1297 (1A962)) the Board wrote:

[T]he impact of automation on a
specific category of employees is a
matter of grave concern to them. It may
involve not only their present but future
employment in the skills for which the
have been trained. Accordingly, the
effect of automation on employment is a
joint responsibility of employers and the
representatives of employees.

TTMe Board continued this Jaime of reasoning an: least
through. 16982. In the case of Plymouth Locomotive (261 NLRB
87 at 88 (1982)), it stated:

Where the automation of bargaining
unit work will result in the elimination
of unit jobs, it is the chnaz of the
employer to bargain with its employees’
representative over the decision to
install automated equipment, as well as
the effects of such a decision.

HoweVer, in the 1985 case of Central Broadcastinq (280

NL .
RB 56' 1985), the Board was careful to note that the

CO 0 v '
mpany violated the Act by requing to bargain over the
eff . . .
em3ts (If its decision to automate, not the deCiSion itself.

The .
‘Boalki’s post—Otis position on technological change

93
prompted Sockell to write (1993, p. 62):

By keeping technology out of the
joint control of the union and
management, those who shape public policy
have emasculated the role of the union in
the firm, a role the law has continued to
shrink over time.

2. Moving work during the life of an agreement.

The role of the union within the firm was further
diminished when the Board removed the union’s ability to
resist mid—term movements of bargaining unit work. In the
cases mentioned above, the issue in dispute had been whether
an employer needed to bargain to impasse over a decision to
eliminate bargaining unit work, usually upon expiration of an
agreement. In the case of Milwaukee Spring Division of
Illinois Coil Spring Co. v. UAW (265 NLRB 206 (1982), the
question concerned an employer’s ability to move work during
the life of an agreement in order to avoid paying UHiOH
wages.

For a number of years, the Board had held that an
employer could not move work from a union to a nonunion
f ' ' . . .
aCllltY during the life of a contract Without the consent of
th ' . . -

e union if the obVious purpose of the move was to aVOid a
r ' - . .
p OVlSlOn of the collective bargaining agreement (Bosanac,
1983; I . . '
rVing, 1983). To move work mid—term Without the

unio I ‘ . .
n S Consent had been conSidered a unilateral change in

Vic) -
latiOn of Section 8(d) .

94
Ekarlier Boards had extended the 'WK) movement without
agreenuent" standard by relying on the "recognition clause"
presern: :ni most labor contracts. The typical recognition
clausea states that the employer recognizes the union’s right
to repnresent a specified group of employees. To move work to
indivixiuals not covered by the recognition clause
constiizuted, in the Board’s view, a unilateral change. This
broad :reeding of such clauses was not given very favorable
reviemms by the courts. The Seventh Circuit, in the case of
Univeresity of Chicago v.. NLRB (514 F.2d. 942 (1975))
overturuied the Board, and stated (514 F.2d. 942, 949):
[U]nless transfers are specifically
prohibited by the bargaining agreement,
an employer is free to transfer work out
of the bargaining unit if: 1) the
employer complies with Fibreboard Paper
Products v. NLRB, by bargaining in good
faith to impasse; and 2) the employer is
not motivated by anti-union animus, [as
is] Textile Workers v. Darlington.
3T“? iNinth Circuit, however, enforced the Board’s order
in a Geese; in which it was able to prove that an employer had
m . . . .
OVQd WCXr1< a short distance to aVOid paying union wages (Los
A .
Wrine Hardware v. NLRB, 602 F.2d. 1302 (1979)).
So . .
on afilEtr, the Board had an opportunity to apply its Los

An . . ~ - -
geles Marine rationale to a case ariSing back in the

Seventh (Zircuit.

MilWaukee Spring Company was a division of Illinois Coil

Spr ' .
1ngCOrporation, which owned several plants. both union

¥

95

and non—union. In 1982 the company asked UAW Local 547,
which represented workers at the Milwaukee division, to
reopen its contract, particularly with regard to the wage
schedule. The company pointed out that Milwaukee workers
were earning almost twice as much as workers at its nonunion
facilities. The parties bargained to impasse, but the union
voted down the company’s final proposal. The company
responded by closing the Milwaukee facility and shifting work
to McHenry Spring, a nonunion subsidiary. The union filed an
unfair labor practice which the Board upheld based on its Los
Angel es Marine rationale.

When the Board’s majority changed, however, it asked the
Seventh Circuit to remand the case to it for reconsideration,
and in Milwaukee Spring II (268 NLRB 601 (1984)) overruled
the prior decision. Although the earlier Board had been
careful to distinguish this case from University of Chicago,
Whieh involved a performance issue not wages, the Dotson
Board held that the Seventh Circuit’s reasoning in University
Of Chicago not the Ninth Circuit’s reasoning in Los Angeles
Marine should be controlling.

DOtson, Hunter, and Dennis determined that in order for

an . .
L11'1fair labor practice to stand, the Board would have to

fi . .
nd a specific clause relating to the relocation of work
th .
at had been violated. They did not believe that the wage

c a , , ,
L‘lSae had been altered by the company in this case, arguing

96

that: t:11e payment of nonunion wages at McHenry did nothing to
charngte the wages at Milwaukee——even though, as Member
Zinunexrrnan pointed out in his dissent, there were no workers
left; .iii Milwaukee to be paid according to the union scale!
Efirrtlieary the Board rejected the idea that recognition clauses
deserved the broad reading earlier Boards had given them.
iiueree \Nere contracts that contained actual work preservation
clatusees, and in their absence, the Board held, they should
not be inferred.

IIri his dissent, Member Zimmerman focused (M1 the wage
provisions. His frustration was obvious (268 NLRB 601 at
606) :

It is disingenuous to argue, as do
my colleagues, that [Milwaukee Spring’s]

relocation. did run: disturb the
contractual wages and benefits of the
Milwaukee facility. If [Milwaukee

Spring] had implemented its decision,
there would be no assembly employees at
the Milwaukee facility to receive the
contractual wages and. benefits...Under
these circumstances, my colleagues’s
conclusion that [Milwaukee Spring] left
the wage and benefits provision ’intact’
at Milwaukee is illogical and without
legal significance.

T - . . . .
he Circuit Court for the District of Columbia, however,

ac . . . . .
CEeDtZed the majority’s View and the Seventh Circuit’s

StéiIICLard became Board doctrine (765 F. 2d~ 179' 1985):

97
c- The loss of certification.

1. Questions of majority support.

As the model of union disappearance in Equation 1 shows,
wi thout any of the types of business decisions outlined above
occurring, a relationship ends if the union loses its right
to represent bargaining unit workers. Typically this loss
would occur through a decertification election—~either an
employee-initiated RD election or an employer-initiated RM
e1 ection‘-—held under Section 9 of the Act. Alternatively, an
employer may simply repudiate the relationship based on a
good faith doubt that the union continues to enjoy the
support of a majority of unit members, or a union may simply
give up its right of representation (usually when facing
Certain decertification).

The basic procedures in this area of law are statutory,
and , therefore, have not changed since 1947. A
deCertification election occurs much like a certification
e1ection. First, the Board must receive a petition
indicating that at least thirty percent of bargaining unit
Inertfibers seek an election. Once this "showing of interest" is

established, an election like any other is held. The union

mus t receive majority support to continue as bargaining
epresentative. A tie vote goes against the union Since
Vldence of majority support would be lacking. Since a

QCEertification election occurs much like a certification

98
election, the standards of election conduct are germane to
both situations.

Employers have always been precluded from initiating an
employee decertification petition (Pierce, 1982). However,
since it is a violation of the Act for employers to deal with
an organization that does not have the support of a majority
of bargaining unit members, there are ways for employers to
push the certification question (Flynn, 1991; Schupp, 1991).8
Employers may simply refuse to bargain or may petition the
NLRB for an RM election. To forestall the tactical use of
either of these tests, the Board and courts insist that
employers have some evidence to support a "good faith doubt"
of a union’s status (Texas Petrochemicals, 296 NLRB 139,
1989). Case law suggests that employers are on the safest
footing when they can produce several separate pieces of
evidence—-copies of letters of resignation from (former)
union members, proof of the union’s inability to maintain a
slate (HE officers (Bickerstaff Clay FToducts \J. NLRB 371
F.2d. 980 (1989)), and employees' letters indicating
dissatisfaction with union representation (AMBAC

_nternational and Engineers Local 112, 299 NLRB No. 66

 

° This point needs some clarification. An employer and

minority union may agree to a "members~only“ pact, but cannot
enter into éni agreement CHI behalf of and. bargaining unit
members. Since an employer has absolutely no legal obligation
to meet with a minority union, such agreements are very rare,
although a few exist in the needle trades (Gould, 1993).

99
(1990)). Employers may also conduct a secret ballot poll of
workers, but only as a confirmatory measure (Mingtree
Restaurant 265 NLRB 409 (1982)).

The law merely requires good faith doubt, not ironclad
proof. However, if the Board finds that a union, contrary to
an employer’s belief, maintains majority support, the
employer could be found guilty of failing to bargain and be
required to reverse or remedy any unilateral action that may
have been taken.

The question of majority status often occurs after the
transfer of business ownership or, more frequently, after a
strike when replacement workers have been used. In the
former case, if and when a former owner’s employees
constitute a majority of bargaining unit members with a
"successor employer," the successor is required to begin
dealing with the union (Fall River Dyeing and Finishing 482
US 27 (1987)). Of course, due to changes in operations that
often accompany a transfer of ownership, this standard is not
always easy to apply.

The latter issue, strike replacements, has created even
more controversy. At one time, the Board presumed that
replacement workers continued to support the union (Peoples
EELS System, Inc. 214 NLRB 944 (1974)). Later, the Board
Concluded that replacement workers supported the union in the

Same proportion as striking workers (Windham Community

100

Memorial Hospital, 230 INLRB 1070 (1977)). However, the
First, Fifth and Eighth Circuits chose to adopt the views of
Professor Robert Gorman who, in a legal treatise on the
subject, argued that it was illogical to presume that
replacements continued t1) support the infirni (National Car

Rental v. NLRB, 594 F.2d. 1203 (1979); Soule Glass & Glazing

 

Co v. NLRB 652 F.2d. 1055 (1981); NLRB v. Randel-Eastern
Ambulance Service, Inc. 584 F. 2d. 720 (1978)). Finally, in
Buckley Broadcasting (284 NLRB 1339 (1987)) the Board took a
middle road and opted for a no presumption rule.

Although Buckley was a retreat from the earlier,
stronger presumption of support, the Board did not go so far
as to adopt the Gorman presumptions. To resolve the conflict
between several of the Circuits, and between the Board and
the Circuits, the Supreme Court granted certiorari in Curtin
Matheson Scientific (110 5341:. 1542 (1990) and, in.ea 5—4
vote, upheld the Board’s no presumption rule. At least one
scholar (Leroy, 1991) has found empirical support for the new
standard.

There are consequences arising from this new standard.
Since there is no longer a presumption°ir1favor of the union,

it would seem that an employer's good faith doubt maintains

\

9 When there has been no replacement of workers, a union
Still enjoys a rebuttable presumption of support, and an
irrebuttable presumption for the 12 months following a
CErtification election (Cox, Bok, and Gorman, 1987).

101
equal standing with a union’s assertion of continued support.
In such an environment, a union’s ability to stem defections
seems especially critical. However, that ability was also
eroded in several mid-1980s cases.

2. Onion discipline.

When time Act ku5 amended lIl 1947, Congress included
within it a set of union unfair labor practices to mirror the
Wagner Act's employer unfair labor practices, Section 8(b).
The first of these, Section 8(b)(1)(A), makes it an unfair
labor practice for a union, "to restrain or coerce employees
in the exercise of the rights guaranteed in section 7." The
section continues, however:

[T]his paragraph shall not impair the right of a

labor organization to prescribe its own rules with

respect to the acquisition or retention of
membership...“

It was the preceding language that guided the Supreme

Court in its seminal union discipline case, NLRB v. Allis—

 

Chalmers Manufacturing Co. (388 US 175 (1967)). In that case,
the Court held that a union could fine a member for crossing
a.picket line and returning to work during a strike, and could
sue in state court to collect the fine. The dissenters in the
case noted that Section 8(b)(1)(A) speaks only of a union’s
right to promulgate rules "with respect to the acquisition and
retention.of membership," not picket line crossing; The Court

was not persuaded, and spoke of how majority rule was at the

102
"heart of federal labor policy" and that, "the power to fine
or expel strikebreakers is essential if the union is to be an
effective bargaining agent."

It. is extremely' important to ‘note that the Supreme
Court%3 decision jjl Allis—Chalmers applied only t1) union
members. A union may not fine bargaining unit employees who
are not union members. This issue arose in NLRB v. Textile
Workers Local 1029, Granite State Joint Board (409 US 213
(1972)). Several months into a strike, several dozen members
resigned from the Textile Workers Union and went back to work.
The union attempted to fine each member $2000 pursuant to a
resolution. passed. concurrent vulji the strike ‘vote. The
Supreme (Rmnfig however, refused Ix) uphold (flue discipline
stating that "the union has no more control over a former a
member that it does over the man in the street (409 US 213,
217)." Justice Douglas noted that "neither the contract nor
the Union's constitution or bylaws contained any provision
defining or linnting'tflue circumstance under ukfirfliea member
could resign (409 US 213, 214)." A year later the Supreme
Court held that a union’s constitutional clause prohibiting
strikebreaking was not enough to allow it to levy a fine
against a former member because, once again, the union’s
Constitution was silent on the issue of resignation (Booster
Lque No. 405 v. NLRB, 412 US 84 (1973)).

In response to both decisions, many unions began

103
promulgating rules governing resignations (Oliver, 1986).
Some unions, however, exceeded the standard of reasonableness
in, their‘ pre—conditions for‘ resignation, and. found. their
guidelines voided by the Board. A sixty-day waiting period,

for example, was thrown out by the Board in General Electric

 

Co. (197 NLRB 608 (1972)). And in Dalmo Victor (231 NLRB 719

 

(1977)), the Board found the Machinists’ Union in violation
of the Act when it tried to fine an ex—member who had crossed
a picket line because he had not resigned at least fourteen
days before the strike began in accordance with the union's
constitution. In the Board’s view, the union’s action was an
unlawful attempt to control the behavior of a former member.
It further expressed its displeasure with limits on
resignations linked specifically to times of strikes, which,
the Board argued, would be when members would most want to
resign. The Ninth Circuit disagreed with the Board, however,
and refused enforcement in Dalmo Victor. (kiremand, the Board
fashioned.a compromise, which allowed for a thirty day waiting
period on resignations, so long as it applied both during

strikes and otherwise (Dalmo Victor II, 263 NLRB 984, 1982).

 

Chairman Van De Water and Member Hunter dissented in the

opinion, questioning the rationale that went into the

selection of "thirty" as the appropriate number of days.
When the Reagan Board was seated the matter was settled

With the adoption of a very clear standard. Unions could

104

promulgate IND rules restricting' the :right cflf members to
resign. In Machinists Local 1414 (270 NLRB 1330 (1984)) the
Board relied upon the Supreme Court’s holding in Scofield v.
NLRB (394 US 423 (1969)), which stated that union discipline
could not contradict national labor policy, to find
restrictions on resignations in violation of an individual’s
Section 7 right to "refrain from" collective activity.

At about the same time, the Seventh Circuit was
considering a case involving a charge by the Board against a
Patternmakers’ rule restricting resignations during strikes
(NLRB v. Patternmakers League of North America, 724 F2d 57,
1983), and the Ninth Circuit was examining the Board’s action
in Dalmo Victor II. Tb smmtle matters, the Supreme Court
granted certiorari and, in.a 5—4 decision accepted the Board’s
view that all restrictions on resignations were unlawful
(Patternmakers League of North America v. NLRB, 473 US 95
(1985)).

In a sharp dissent, Justice Blackmun (writing also for
Justices Brennan and Marshall) stated (473 US 95, 117):

Today the Court supinely defers to
a divided-vote determination by the
National Labor 1Relations Board. that 23
union commits an unfair labor practice
when it enforces a worker’s promise to
his fellow workers not to resign from his
union and return to work during a strike,
even though the worker freely made the
decision to join the union and freely

made the promise not to resign at such a
time, and even though union members

AVA.
vo-v

 

\

Ar

 

..g,’

H-

. . . i
. .s u. .

hv ~x~ r V: ,\& hks
.. .3

.
(
.
.
9A

 

105

democratically made the decision to
strike in full awareness of that promise.

Since adoption of the standard, the Board and the
Circuit Courts have extended the proscription regarding
resignations in.aa number CHE cases (Communicatrmn Workers,

Local 9201, 275 rHJHB 1529 (1985); Newspaper Guild Ci 1km)

 

EggrkL Local 31, (1984); International Association of
Iaachinist, Local 1769, (1984)). And in NLRB v. Sheet Metal
Eggrkers Local 73, the Seventh Circuit even extended the rule
tr) cover members already facing union discipline.
Phenceforth, a member facing discipline need only resign to

enscape the union's penalty.

3. Standards of conduct during' decertification
elections.

With less recourse against dissident members, unions may
eru3CNJnter increasing' decertification activity. Mbre
information about the extent of decertification elections
Wilhl be provided in the next chapter. Here, the important
issnnres concern the behavior of employers during the election
prcuzeess. In Chapter II, it was noted that employer behavior
durirlgcertification elections is advanced by several experts
as 51 leading cause of union decline (Weiler, 1983; Freeman
and Medoff, 1984; Freeman, 1989, 1989a, 1989b). or course,
bOtfln. {marties have been known to engage in egregious conduct
dulriJflgg elections, but an employer’s actions are, for obvious

re . . . . .
aSons, geared toward aVOiding or escaping a bargaining

106

relationship and are, therefore, more salient to this

discussion. As mentioned above, since certification and

decertification elections follow similar procedures,

standards of election conduct are relevant to both

situations.
At the same time that Congress passed into law the
sections of the Act regarding decertification (Section 9), it

also passed Section 8(c) which guarantees employers free

speech rights during election proceedings so long as

employers make "no threat of reprisal or force or promise of

benefit." Section 8(c) was designed to change the Board’s

Wagner Act standard, which precluded employers’ involvement

in elections altogether. The Board’s original reasoning was

that the selection of a bargaining representative was
strictly a matter for employees, and of no legitimate concern

t<1) employers. Congress disagreed, however, and gave

employers the right to intervene in the election process.
Nonetheless, it was still the Board’s policy to maintain
"laboratory" conditions for representation elections——whether
they be certification elections or, as here, decertification
eleCtions (General Shoe, 77 NLRB 124 (1948)). However, the
Board also recognized that elections did not, in fact, "occur
in a laboratory where controlled or artificial conditions may

b . . . .
e Established." Rather, it determined that it must gauge

be . . . . . .
havlor "in light of realistic standards of human behaVior

107

(Liberal Market, 108 NLRB 1481 (1954))."
Over (flue years, tine permissible "standards cfif human

behavior" lunna wavered. th‘ example, 111 Standard—Coosa-

 

Thatcher (85 NLRB 1358 (1949)) the Board adopted a strong
standand prohibiting tine interrogation (ME employees about
their union sympathies. But the Eisenhower Board moved away
from this standard several years later in Blue Flash Express
(109 NLRB 591 (1954)) when it held that an interrogation may
be lawful if there were no threats or promises made. Soon
this standard was upended as well (Marsh Express, 140 NLRB
899 (1963); Laney Tank Lines 166 NLRB 1053 (1967)). But in
1973 and 1974, the Board decided to permit the questioning of
"known union sympathizers" (B.F. Goodrich 201 NLRB 353
(1973); Stumpf Motor Company 208 NLRB 431 (1974); Federal
Paper Board 206 NLRB 681 (1974)). But, still again, this
lenient standard was changed in ITT Automotive (231 NLRB 878
(1977). And tine Board. finally cumma full circle 111 PPG
Industries (251 NLRB 1146 (1980)) when it stated that
interrogations were inherently coercive and were to be
considered a per se violation of the Act.

The interrogation standard was in obvious trouble when
the Reagan Board began taking shape. Member Hunter and
Chairman Dotson both dissented in interrogation cases earlier
in their terms (Donnelly Manufacturing Company, 265 NLRB No.

196 (1982); Lambrecht Earth Movers, 267 NLRB No. 35 (1983).

-1.

s..

 

 

 

C.

«\C

108

And, when Member Dennis finally joined the Board, the three
were able to gain a majority, with Member Zimmerman
dissenting, iji reversing lfiX? and remurning tt>ea standard
closer to Eflue Flash (Rossmore House 269 NLRB 198 (1984).
Following Rossmore House the Board routinely found
interrogations lawful (Sears anmi Roebuck ~274 IHJHS No. 2
(1985); Michael's IMarket CM? Canterbury' 274 iNLRB It). 105
(1985); Diversified Products 272 NLRB No. 162 (1984)).

The same type of oscillation can be seen in the Board’s
standards concerning the misrepresentation of facts during an
election campaign. The Board has reversed itself three times
in determining that it would, then would not, then would,
then would not set aside an election when it received
evidence that one of the parties had made false statements
(Hollywood Ceramics 140 NLRB 221 (1962); Shopping Kart Food
Market 228 NLRB 1311; General Knit of California 239 NLRB 749
(1978); Midland National Life Insurance, 263 NLRB 127
(1982)).

The Reagan Board also loosened the standard on what
would and would not be regarded as threatening or coercive
speech by an employer. The Board held that an employer could
cite "economic realities" based on "objective fact." So, fOr
example, employers could make insinuations concerning plant
Closings, etc. if they could point to a unionized facility

which. had, indeed, closed. Similarly, the Board began

109
allowing employers to rmflue such statements as "bargaining
begins from zero (Pilliord of Mississippi, Inc. 275 NLRB No,

117 (1985))," or "bargaining begins from scratch (Atlantic

 

Forest Products, 282 NLRB No. 105 (1987))." Both are obvious
references to what could happen to one’s pay or benefits if
union. representatitnl was chosen, tun: neither (would,
henceforth, be considered threats under Section 8(c).
Although these standards were adopted in response to
certification campaigns, there is 1M3 reason vflur the same
degree (ME latitude would rm»: be afforded employers during
decertification proceedings as well.

The complete record obviously shows that there have been
significant changes over the years——even before the Reagan
years--in all cflftflue legal principles cited» However, as
Sockell and Delaney (1987. L1. 28) argue, "the Reagan NLRB
decisions involve nwre tjmni the cyclical changes 1J1 NLRB
policy that have accompanied. the agency’s shifting
partisanship." The Reagan Board undertook a massive
deregulation consistent with its broader agenda. Sockell and
Delaney conclude, "The changes documented in this paper
[concerning election conduct] represent only the tip of the
iceberg...the overall pattern reflects an ideology that
collective bargaining is not a valuable institution in

society (p. 42)."

110
D. canclusions.

The purpose of this chapter has been to explore some of
the leading reversals of labor relations policy germane to
the model of local union disappearance. Hence, it was shown
that the Otis Elevator‘ standard. whittled the bargaining
obligation over business decisions to its narrowest point in
'history by extending the Supreme Court’s rationale in First
National Maintenance to most all management actions,
including subcontracting and technological substitution.
Further, in .Milwaukee Spring' II the Board decided that
employers could move work during the life of a contract, even
if to avoid a provision of a labor agreement, so long as
there was 1x3 specific language 111 the contract (x) prevent
them from doing so. The effects of both cases may have been
to lower the costs associated with actions ending the
bargaining' relationship kn! marginalizing' the Ixflja of the
union within the firm, giving the employer more freedom to
act, and allowing the union less opportunity to pursue self—
preservation.

In addition, the Board and courts have made it more
difficult for unions to maintain their internal strength by
voiding or eroding most forms of union discipline. The
Machinists and Patternmakers cases outlawed rules limiting
resignations, while the Sheetmetal WOrkers case extended the

proscription to include members already facing discipline.

lll

Taken together, these cases leave unions with almost no power
to enforce standards of member conduct.

Maintaining union status against a threat of
decertification may also have been. made more difficult.
Buckley Broadcasting removed a more favorable presumption of
union support on the part of replacement workers, but, more
important, Midland.National Life, Rossmore House, Pilliord of
Mississippi and their many progeny have, in effect,
deregulated the election process.

Several years ago Kenneth Gagala (1983) wrote a bmck
entitled Union Organizing and Staying Organized. Roomkin and
Block (1981), Weiler (1983) and many others have demonstrated
how difficult organizing can be in the face of employer
opposition and anemic legal protections. The changes of the
mid—1980s may have made staying organized more difficult as
well. And as International Typograhpical Union President J.
Woodruff Randolph said in 1949, "The first and most important
objective of any trade union is to stay alive (Tomlins, 1985,

pp. 310-311)."

Chapter V. Independent variables and Hypotheses.

A. Introduction.

In this chapter, my focus shifts from labor relations
policy to a more general examination of the variables
contained in the model of local union disappearance—~plant
closings, relocations, input substitutions, and loss of
certification. One section if; devoted to time first three
topics, another to the decertification question. The purpose
here is twofold: I examine trends in the variables during the
period (fl? study. But 1: alas look (13 prior research to
consider the role that anti—union (or anti—bargaining)
motivations played in these trends. (Although in the case of
decertification, the answer is obvious).

Examining the latter point aids hypotheses generation.
I would expect that a "pro—business" change in public policy
would have the greater impact on phenomena more clearly linked
to deunionization in the first place. Ftu‘ example, if
employers have shown a willingness to relocate in order to
escape unionization, then making that practice less
troublesome may encourage it even more. On the other hand,
if there :hs only ea weak ljmfi: between, swan technological

improvement and union avoidance, changes in bargaining

112

113

obligations should have a less profound impact on the

association between technological change and unit
disappearance.

Based on previous research, relocation seems most
strongly tied to issues of unionization. The link is less

certain with subcontracting and, particularly, technological
change. And although it is possible, due to new standards in
bankruptcy law, that some number of complete firm failures
have been associated.with a desire to escape unionization, it
is unlikely that rmnu/ employers lumna gone (mm: of business
simply to avoid unionization. The loss of certification is
almost certainly associated with a desire to escape
unionization.

B. Business decisions affecting the bargaining unit.

1. Plant closings.

Of all the variables in the model, plant closings have
received the most public and scholarly attention. Although
the ‘problem is run: rmwm its sudden increase «during the
recession of the early 19808 produced.a(groundswell of concern
for displaced workers and devastated communities, as well as
increased frustrations over what seemed to be an inadequate
government response (Buss and Redburn, 1987). By the early
1980s, a number of states, and eventually the federal
government considered or passed legislation dealing with some

aspect of the problem (Krotsen, 1984; Rothstein, 1986).

114
(Currently federal laW' :requires employers t1) provide
notification to their employees of their intention to close
(29 USC 2101).

An early and highly influential work on the plant closing
issue is Bluestone and Harrison’s The Deindustrialization of
America (1982), though many others have followed (see, for
example, nmnographs 137 Howlamd, 1988; Perrucci, Perrucci,
Targ, and Targ, 1988; Portz, 1990; Rothstein, 1986; Wendling,
1984). The information reported in these works and by the
federal government is indeed compelling. 131a 1987 study, the
Senate Human Resources Committee (U.S. Senate, 1987) revealed
that almost 12 percent of labor force participants lost their
jobs between 1981 and 1985 due to closings, with roughly 5
million. businesses-—about 2%) percent of 6UJ. businesses——
shutting down during the period. If the number of permanently
laid-off individuals (from firms still in business) are added
to those who lost their jobs due to closings, the numbers
become even more staggering. The General Accounting Office
(1986) reported that nearly 20 percent of workers lost their
jobs due to plant closings or permanent layoffs in 1983 and
1984 alone.

When Bluestone and Harrison examined the issue in 1982,
the worst was yet to come. Returning to the theme in a recent
work, Barry Bluestone and his father Irving (1992, p. 66)

write:

115

By 1987, employment in manufacturing was
actually lower than 111 1966..;Nearly a
quarter of tflua producthmn workforce in
the auto industry in 1978 would disappear
by 1989; similarly' more than Ihalf <3f
those who build farm machinery and
construction equipment disappeared from
these industries in the course of a
single decade. In furniture, glass
products, steel, metal products,
machinery, food and beverages, textiles,
clothing, paper, chemical products, and
leather, employment fell below levels
achieved ill the 1960s enmi 1970s. The
distressing consequences of plant
closings on workers, their families, and
their communities were all too evident in
terms of unemployment, income loss, and
physical and mental trauma.

Community impact was greatest in single company or
single industry' towns (Buss eumi Redburn, 1983), but the
effects (M1 particular' groups cflf workers are 51 bit less
certain. Hamermesh (1989, p. 52) writes that, "Displacement
is more likely among men then among women; among blacks and
Hispanics‘ than among non-Hispanic whites; and especially
among manufacturing workers." However, he also notes that
contradictory findings are» common. Ftn: example, (Gordus,
Jarley, and Ferman (1981) suggest that older, white males
with high seniority but relatively less education were the
most severely affected, while Remy and Sawers (1984) claim
that the most severe effects were felt by women and
minorities. Wallace anxi Rothschild (1988) gxnju: out that
findings tend to differ depending on whether the secondary

and tertiary effects of closings are considered. Men are

116

perhaps more likely to suffer from the immediate effects of
factory closings (since men comprise a larger percentage of
the workforce in most manufacturing industries), but in the
long-term, when ripple effects spread to service and retail
establishments, women end up bearing the larger burden.
Nearly all the research following the fairly early study of
Gordus, Jarley, and Ferman concludes that non—whites have
suffered disproportionately to whites.

The impact of plant closings on the industrial relations
system is two—fold. First, many of the closings during the
1970s and 1980s occurred in the highly unionized sectors of
heavy manufacturing.1 Therefore, many union jobs and entire
bargaining units were lost (Cornfield, 1987), a fact embedded
in the arguments of Leo Troy (1990) and the other structural
change theorists. And even tjmnmfli there remains ea debate
over time relative magnitude (ME structural effects, rm) one
argues that plant closings had no impact on the system at
all. Farber (1987) and. Freeman. and. Medoff (1984), while

critical of the structural change explanation in general, do

 

1 However, the image of plant closings as mainly limited

to old factories in the rust belt is far from correct. While
it is true that communities such as Youngstown, Akron, Flint,
and Gary, which were centers of old—line heavy industry, were
devastated, the bulk of closings, particularly during early—
1980s, were among relatively small and relatively new
businesses-—which, historically, have had high failure rates-
-many of which were located in the interior South (Wallace and
Rothschild, 1988).

117
conclude that somewhere between 10 and 20 percent of the
decline in unionization is accounted for by
deindustrialization.

In addition, bargaining outcomes may have been affected
by the increase in plant closings. Some argue that the image
of closed plants created enough fear to force unions into
agreeing to historically large and perhaps misguided
concessions (Perrucci, Perrucci, Targ, and Targ, 1988;
Slaughter 1983). Kovach and Millspaugh (1987, p. 48) point
out that,

The [UAW] agreed tt>Immm3 than one
billion dollars in givebacks, in return
for both effects and decision bargaining
rights in future plant closings...

The United Food and Commercial
Workers Union agreed to production line
speed—up and cost—of—living freezes at
Armour Meats in return for management
assurances of security against
unannounced future shutdown.

Likewise, the tire industry’s most
recent pact with the United Rubber
Workers Union contains a management
guarantee against plant closings in
return for wage concessions representing
a 13 percent decrease in labor costs over
the life of the contract.

Although detractors submit that union concessions of the
early-19803 were actually small considering the state of
labor markets at the time (Hamermesh, 1989), the rhetoric of
concession—bargaining may have made organized workers uneasy
about their union status. It was not uncommon for employers

to use non—union facilities as the basis for comparison in

118
concession bargaining (see the .Milwaukee Spring' decision
above), and gflmflxxs of closed Luflflfll'plants were frequently
used. against unions during organizing campaigns (Gagala,
1983).

In short, there is much evidence that the atmosphere
created by the sharp rise in closings was turned to the
advantage of employers-—both during organizing campaigns and
in negotiations. But whether employers, to any great degree,
actually closed—down operations to escape unionization is a
more complex issue. Some plant closings were the result of
firmi failures, while others were linked tXD single
establishment closings or the (often related) relocation of
work (Bluestone and Harrison, 1983). While it would seem
implausible that ea business owner vmnflxi go out-of-business
deliberately to escape a bargaining relationship, 1978
amendments to the Bankruptcy Code may have made that option
more attractive. Debtors are now allowed to retain a certain
amount of money to be used for future investments or as
retirement income; they need not pay out everything to
creditors (Bluestone and Harrison, 1983; Barlett and Steele,
1992). Dismantling a unionized business and re-emerging with
enough cash—in—hand to start a nonunion business at a
"greenfield" site is, tummy: law, quite possible. In the
1970s there ‘were approximately 25,000 bankruptcies filed

annually, in the 1980s the number climbed to an average of

119

63,500 per year (Bartlett and Steele, 1992, p. 68).
Likewise, the number of firms that never emerged from Chapter
11 rose similarly. The concurrent increase in the number of
professional bankruptcy consultants rmn/ have made ii; even
more likely that some percentage of those failures were
undertaken for strategic, not emergency, reasons (Barlett and
Steel, 1992).

Bankruptcy reorganization, if not complete firm failure,
has certainly been used as a reason to void union contracts,
particularly after the practice was approved by the Supreme
Court in the case of NLRB v. Bildisco and Bildisco (465 US
513 (1984))(George, 1985a; Kurtz, walton énmi Wells, 1990;
Wren and Murmann, 1990). In that case, the Court considered
whether collective bargaining agreements were normal
executory contracts under Section 365(a) of Chapter 11. If
they were, then they could be rejected by debtors—in—
possession. ‘ Lower courts had a problem applying Section
365(a) rationale to labor agreements, because such agreements
have their own standing under federal law (Labor—Management
Relations Act of 1947, Section 301 (29 USC 141—197)). Some
courts had held that employers had no more right to void
collective bargaining agreements tlmni to suspend time Fair
Labor Standards Act (Keating, 1994). Certainly paying less
than minimum wage would help a firm "get back on its feet,"

but it is clearly not legal. The Supreme Court determined,

120

however, that employers in Chapter 11 should be allowed to
reject labor agreements without being required to bargain to
impasse.

At about the same time, Congress was debating an
amendment to the Bankruptcy Code following the unfavorable
constitutional review of a provision of the 1978 act.
Organized labor was able to have included in the new
legislation Section 1113, the so—called Packwood Amendments,
which require a nine—step test before an agreement can be
rejected. McClain (1992, p. 191) writes, however:

But Section JJJLL envisioned as ea pro—
labor response to Bildisco has failed to
provide armor for collective bargaining
agreements eumi has actually placed
laborers in a worse position than before
its enactment...[N]earhy every decision
since 1984 has granted the debtor's
petition for rejection of the collective
bargaining agreement.

Rejection of an agreement in itself does not end a
union-management relationship; the employer is still required
to bargain to impasse over new terms of employment. How much
of an obligation remains after rejection is a bit unclear.
George (1985a) notes that if the Supreme: Court had. not
intended to give debtors more leeway in their bargaining
obligations, then it put them in the illogical position of
having been better off before rejection, when they could make

unilateral changes, than after.

An additional point is that the local relationship may

121

be more fragile when no contract is in force. Union security
clauses are rm) longer binding sum} the so—called "contract
bar," which prevents the filing of a decertification petition
during most of the life of an agreement is also lifted.
Without such protections, and during a time of near certain
uneasiness among employees, a union may find itself under
severe pressure-—particularly if an employer views the union
an as obstacle to reorganization (Pendergast, 1993). Lawler
(1990, p. 179) writes:

Although not terminating union

representation, time trustee cu? debtor—

in—possession could act unilaterally to

change personnel policies without first

negotiating such changes with the union.

Besides undermining the union's

bargaining position, organized labor saw

this as a means cf Lfltimately ushering

out the union.

In this research, I use Dun and Bradstreet data as
reported in the Firm Failure Record.2 The measure is of firm
failures kn/jyear and industry, per 10,000 firms. .A firm
failure :us defined as ea finn going cnmeof—business owing

money. Businesses which closed for other reasons, such as

retirement or illness, are not included in the figure.

 

’ Descriptive statistics are found in the next chapter.

122

2. Plant relocations.

The link between firm failure or bankruptcy
reorganization and deunionization is a bit speculative.
However, there is Immfii mare evidence that ‘relocations—-
sometimes of entire firms, sometimes of establishments, and
sometimes of jobs——have been undertaken to escape
unionization (Gordon, 1977; Jaffee, 1986, 1988).

Throughout the 20th century there has been a southward
movement of industry, and more recently a movement "offshore"
no other countries (Bluestone anmi Bluestone, 1982). The
textile industry, which was located primarily in the
Northeast 111 the lfhfli century, is rmwv largely ea southern
industry; the center of the U.S. tire industry has moved from
Ohio to Oklahoma; and apparel manufacturers, who once
employed thousands of mmmkers in rmmtheastern cities have
shifted enormous amounts of production overseas. The
movement of rmnur other industries t1) southern aumi western
"sunbelt" regions. accelerated (during (flue 1970s and 119805
(Bluestone and Bluestone, 1992; Storper and Walker, 1984).

There are many reasons why firms relocate, particularly
Us the South» Cheaper energy costs, especialhy in areas
served by the Tennessee Valley Authority, laxer environmental
regulations, and tax inducements are among the more
frequently mentioned factors (Wallace and Rothschild, 1988).

But labor costs and (the related) union presence are factors

123

as well. Bluestone and Harrison (1982, p. 165) write:
"During the 1950s and 1960s, the practice of running from
unions greW'sM) much that inf the 1970s (flue northern—based
industrial unions had been severely weakened." They cite a

Wall Street Journal article, in which a plant location

consultant stated, "Labor costs are the big thing far and
away. Nine out of ten times you can hang it on labor costs
and unionization." And an economic development

representative from Greenville, Mississippi admitted that his
community actively discourages the location there of firms
that are likely txa be unionized, so as run: to damage the
reputation of the area as a union—free haven. Caldwell
(1981), in an article concerning the legality of communities’
activities in defense of their non—union status, writes (p.
679):

[T]he fortuitously dropped union
authorization card in a plant restroom,
or a report that a guest at a local motel
used 51 credit card issued t1) a labor
union, will send ripples of alarm
throughout the community. Elaborate
’Christmas tree’ communication networks
will spring into action like a community
DEW line, passing the dread word rapidly
from merchant to manager, from minister
to mayor, that 'the union is in town.’
Most communities will greet this
announcement with the enthusiasm of
Atlantans awaiting General Sherman.

Caldwell cites many instances :hi which individuals—-

clergy, newspaper editors, and (ordinary townspeople--even

124
those with no formal connection to the targeted employer,
organized against unionization in their community. So long
as they are not acting as "agents" of the employer, such
allies are largely beyond the scope of the Act.

It is easy to find many anecdotes similar to the ones
just mentioned. But, as well, in a quantitative study, David
Jaffee (1986, 1988) found that the strongest negative
predictor of where a firm relocated within the continental
United States between 1970 and 1980 was a state’s rate of
unionization. He concludes emphatically (p. 313):

The results from this analysis provide no
support for the conventional industrial
location perspective...TTm2most important
factor overall was the general level of
labor organization within a state.

Further evidence is provided by Verma and Kochan (1985),
who studied the creation of a nonunion sector within a firm.
Zhi 1957 the observed firm vans 80 percent unionized. But
during the next 25 years, the company developed a strategy of
vigorous opposition to unions at new locations, and favored
those locations with the placement of work. The union sector
continued to shrink and by 1982, less than half of the firm’s
employees were union members—~a reduction of nearly 50
percent in 25 years.

One exception to this trend occurred in the auto

assembly industry where a "southern strategy" of operating

non—union was abandoned in the face of union resistance in

125

the 1970s. As auto manufacturers agreed to neutrality and
even bargaining unit accretion at southern plants all
establishments became unionized (Katz and MacDuffie, 1994).

In sum, the weight of both the quantitative and
qualitative evidence suggests that unionization may be one
factor in relocation decisions. Unfortunately, there are no
systematically collected data that track actual relocations.
As 51 proxy, scholars typically 'have looked an: shifts in
employment or establishments from region—to-region. I will
continue that practice here and use as a proxy for relocation
the cumulative percent point change in the distribution of
establishments across states, by industry, for each period
under study. Statistics were derived from the Census
Bureau’s annual series County Business Patterns. This is
essentially the procedure used by Jaffee (1986, 1988). Based
on the previous research, I believe that there exists some
link between relocation and the desire to escape
unionization.

3. subcontracting.

Muoh less clear':u3 the extent (x) which unionization
plays a role in input substitution. Based on the volume of
Board and court cases involving subcontracting, and the
widespread presence of subcontracting language in contracts,
it appears that union officials are quite concerned with the

practice of substituting nonunion labor for union labor.

126
There is also some scholarly conjecture that subcontracting
has been used to escape unionization (Bluestone and Harrison,
1982; Holmes, 1986; Pfeffer emmi Baron, 1988), Inn: little
research evidence to support this speculation.

While the Europeans (Holmes, 1986) and Japanese
(Watanabe, 1971) have studied the issue quite thoroughly,
empirical vmndc in the IldS. is underdeveloped. Currently,
there is ea growing literature 1J1 response rm) "contingent
worker“ issues, but rmxfl: evidence (xxmas frle individual
industry and even firm—level case studies (Dunlop Commission
Report, 1995). Both definitional and, moreover, data
problems are stalling this research. Bureau of Labor
Statistics information shows a huge increase in employment in
the personnel services industry during the past decade, which
suggests that many employers increased their use of contract
workers and leased employees. Lhfikmtunately, there is rm)

precise information on where these workers were being

employed.3
Neither is there an accepted definition of
subcontracting (Holmes, 1986). The hiring of contractors to

 

3 Leased employees were counted on the payroll of the

leasing company. If a manufacturing firm leased all of its
production workers, it would have been counted as having zero
workers in that category. Needless to say, with the growth
in leasing arrangements the usefulness of establishment-level
employment information is questionable (Personal inquiry to
the Rhode Island Department of Employment Training’s Labor
Market Information Unit, August 9, 1995.)

127

provide clerical, accounting, and maintenance services, the
leasing of production workers, and the outsourcing of work to
job shops have all been termed "subcontracting." Although
the increase in leased production workers will likely have an
impact on future research, most of the work done to date on
manufacturing has focused on outsourcing (Kelley and
Harrison, 1990).

In either case, subcontracting is a difficult phenomenon
to measure. There is neither systematically collected data,
nor truly adequate proxies. The hatter studies have used
original surveys, which have included direct questions to
employers about their outsourcing' behavior. 2\ possible
though admittedly risky proxy is the relative change in the
number (ME small establishments within EH1 industry (Imrie,
1986; Piore and Sabel, 1984). Since large firms typically
outsource to smaller firms that specialize in the production
of components, an increase in outsourcing may be reflected in
a change in the relative number of small establishments
within an industry. Piore and Sabel (1984) note this
relationship between "craft" and "mass" producers in their
book the Second Industrial Divide.

One of the more thorough and recent treatments. of
subcontracting in the United States is the work of Kelley and
Harrison (1990). Several years ago they conducted a survey

of over 1000 production managers in metalworking and asked

128
them about the extent of subcontracting by their firms and
their motivations for subcontracting. Kelley and Harrison
made specific inquiries about industrial relations practices.
They found (p. 1283):

The results for the labor relations
variables enie somewhat surprising. By
itself, independent of the cost of labor,
we find no statistically significant
evidence that subcontracting practices of
U.S. manufacturers are part of a union
avoidance strategy. As for joint labor—
management problem solving, we find that
such collaboration is only associated
with ea significant commitment ti) avoid
subcontracting when there is also a

union. In non-union plants, such
collaborative arrangements actually lead
to an increased likelihood of
subcontracting.

Of course there~ are shortcomings to time Kelley' and
Harrison study: it covers only one industry and relies on
the responses of production managers, who may or may not have
been completely candid about their motivations.
Nevertheless, it is the best American research available on
the issue that I have uncovered, and indicates that the
conjecture concerning the connection between subcontracting
and union avoidance may be somewhat misguided.

I use as a measure of outsourcing, the annual rate of
change in the percentage of establishments employing fewer
than 100 workers, by industry. This figure is derived from
County Business Patterns. While I believe that there may be

a link between outsourcing and a desire to escape

129
unionization, the weight of the research evidence tempers my
expectations.
4. Technological change.

I am also skeptical of the link between technological

change eumi union avoidance. Haddad (1984, I3. 150) writes
that, "One of most serious problems that unions face around
the issue of high technology is membership erosion." Indeed
in a number of cases that has proven true. ,mMi for that
reason, ii: has become axiomatic that unions resist
technological change (Hieb and Moody, 1981). However, again,

the evidence is not very supportive (Bamber, 1986; Cyert and
Mowry, 1987; Keefe and Kphl, 1993). In fact, Weikle and
Wheeler (1984) report that their survey" of local unions
leaders revealed "a rather mild form of encouragement" of new
technology.

Keefe and Kohl (1993) submit that unionists in
manufacturing believe that the competitiveness of their firms
is linked tx> technological development. Several (M3 the
larger industrial unions (the Steelworkers and Mineworkers,
for example) have actively promoted technology, and even, at
times, have been critical of employers for failing to make
capital improvements. It would seem that many union leaders
accept time argument timn: improvements iii compensation and
quality' of worklife are tied to technological advances.

Helen Rainbird (1988, p. 174) writes:

130

Trade unions recognize that companies
invest in new technology to remain
competitive. Their policy is therefore
not to resist technological change, but
to attempt to control its introduction in
seeking agreements on new technology,
disclosure of information and job
security...Unions recognize timm; there
are positive and negative effects of new
technology on jobs, and invariably stress
the need to seek the positive benefits in
negotiations.

The issue becomes more contentious when entire
bargaining units or even entire crafts are threatened. For
example, in 1962 the International Typographical Union
carried out a 114 day strike against New York City newspapers
to prevent the introduction of "cold type" processes. The
union won a short-term victory, but eventually dropped its
opposition to the new technology in 1974. By the mid~1980s,
the ITU, the nation’s oldest union, was out—of—business,
forced to merge with the Communication Workers due to
shrinking' membership (Keefe and Kohl, 1993). A. similar
battle was fought on the railroads when diesel engines
eliminated. the need for firemen (Troy, 1990). And the
introduction. of ;prefabricated.:materials has created. some
problems in construction, although Allen (1986) reports that,
even there, the resistance to new technology has been
exaggerated.

Since most manufacturing establishments are rm) longer

organized along craft lines, the likelihood that

131

technological change would wipe out an entire bargaining unit
has probably lmmai diminished. rThe weight CHE the evidence
does suggest that there can be negative employment impacts of
new technology, but these vary considerably across
occupations (Cyert and Mowery, 1987). Because manufacturing
bargaining units often include a number of occupations, job
loss would have to be extreme before the bargaining
relationship itself is threatened.

The study of technological effects suffers from some of
the same shortcomings as the study of subcontracting. Cyert
and Mowery complain (1987, p. 29):

If we could measure the rates of
invention, innovation, and diffusion in
the U.S. economy, we could simplify
greatly the analysis of technology’s
impact on employment. Such measurements,
however, are far from simple, the United
States and most other industrial nations
do rmm: collect systematic time series
data on the rates of diffusion of
specific new technologies.

Without such data, proxies abound. Total factor
productivity, new capital investment, capital vintage,
computer—intensity, emmi the employment share cu? scientists
and engineers have all been used (Howell and Wolff, 1992).
In this study, I use the average annual rate of change in new

capital investment, 1J1 constant dollars, in! industry, .as

reported 111 the Annual Ehnnma/ of Manufactures. I presume

 

that there may be a link between technological change and a

132
desire to escape unionization. But, once again, based on the
research evidence, the link is likely not a strong one.
C. Loss of certification.

Here there is really no question of motivation. Nearly
the only reason why an employer or unit members would
challenge the right of a union to represent workers is out of
a desire to escape the obligations of unionization and
collective bargaining,‘ whether such obligations include the
payment of dues or agency fees on the part of employees, or
the necessity of bargaining on the part of employers.

Without a good idea of the number of bargaining
relationships in the economy, it is very difficult to
interpret time trend 1J1 decertification activity. The raw
figures indicate: a that RD (employee—iniaited) elections
surged. in mid-1970s, when the annual average went from
approximately 400 per year for the decade before 1975 to over
800 per year for the decade following. The highest year was
1980 with 902 elections. By the end of the 1980s, however,
the figure had gone back down to around 600 per year. The
number of RM (employer-iniated) elections has decreased from

about 400 per year in the early 1970s to only about 200 per

 

‘ One might argue that employers act out of a concern

with violating the law by dealing with a non-majoritarian
representative. But, considering the passive enforcement of
the Act, I cannot imagine that that concern is an overwhelming
one.

133
year in the 19803. In 1990, there were only 87 RM elections
(NLRB, Annual Reports, various years).

Even though these trends may seem favorable to unions
today, it; is important ti) note that time union "win" rate
seems to have been declining as well. When the RD election
trend increased in time 1970s, unions were successful in
maintaining representation in approximately thirty percent of
elections; by the end of the 1980s, that figure had slipped
to just over twenty percent. The statistics for RM elections
are even more dramatic. In the early—1970s, when the number
of electhmms was high, unions were consistently retaining
representation about forty percent of the time; now that
figure is also closer to twenty percent.

Perhaps a larger problem with the decertification data
is that we have no way of accurately figuring how many
elections per bargaining unit are occurring. My research
shows, énmi other research implies, that there 1mm; been a
fairly sharp decline in the number of bargaining
relationships in the economy. Based on the sample used here,
the decline :hi the number (if pre—existing relationships
between 1980 and 1990 was 24.4 percent across fourteen

manufacturing industries.5 Since we also know that the rate

 

5 This figure is simply the percentage difference between the
number of relationships contained in the panel in 1980 and in 1990.
That is: ((1032—1365)/1365)*100=—24.4.

134

of new organizing in manufacturing was quite low during the
decade, we are fairly safe in assuming that there are fewer
bargaining relationships today than there were fifteen or
twenty years ago. Therefore, even though the average annual
number of decertification elections has declined, the rate of
elections ,per LmUJ; very' likely’ has increased. Further,
because of the declining union success rate, the magnitude of
the impact (fl? decertification elections (ii the industrial
relations system may be continuing to rise (Levine, 1989).

The measure used here is the number of decertification
elections per 10,000 units, by industry. The numerator comes
from. NLRB Annual Reports, and is simply the number of
decertification elections per industry, per year. The
denominatcm‘ was figured kn! first calculating total union
membership per industry, and then dividing that figure by the
average size of a bargaining unit in the sample of cases used
in this study, resulting in an approximation of the number of
units per industry. The number of decertification elections
per bargaining unit were figured, and the result was
multiplied by 10,000 to mirror the way the firm failure data
are reported.
D. summary.

The preceding discussion was intended to aid hypotheses
generation. The first conclusion is that firm failures were

more likely to have been associated with economic distress

135

than union avoidance, though it is possible--due to
amendments to the law——that some failures were strategically
undertaken with the goal of escaping collective bargaining.
Second, there is likely to have been a mmch stronger link
between plant relocation and a desire to escape bargaining
obligations. There is quite a bit of both qualitative and
quantitative evidence to suggest that employers moved
operations away from union sites to non—union sites in other
states or other nations. Third, the connection between both
types of input substitution——labor for labor and technology
for labor——and union avoidance appears rather weak. Fourth,
there appears to have been a rising trend in decertification
activity, particularly vmmNi the decrease ill the number of
bargaining units is considered. These conclusions lead to
the following research hypotheses.

D. Hypotheses.

H1. The disappearance rate of local collective bargaining
relationships across manufacturing industries is
significantly and positively related to the number of
firm failures per 10,000 firms.

For the entire pooled sample, it is expected that the
disappearance rate of local collective bargaining
relationships will be higher in industries that experienced
more firm failures. Support for this hypothesis would

indicate that time disappearance of reiationships is, mmat

likely, associated with the economic distress of industries.

136

The hypothesis will be tested by an examination of means,

simple correlations, and partial slope coefficients in

multiple regression analysis.

H2. The disappearance rate of local collective bargaining
relationships across manufacturing industries is
significantly and positively related to the percentage
point change in the distribution of establishments
across states by industry, controlling' for' new
establishment growth and capital intensity.

For the entire pooled sample, it is expected that the
disappearance rate of local collective bargaining
relationships will be higher in industries that have
displayed 61 greater degree (if change 1J1 interstate
distribution. Support for this hypothesis would be
consistent with an argument that the disappearance of local
relationships is associated with firm relocations to escape
unionization.

However, the cross—state distribution of firms may
change if an industry grows more rapidly in one region than
in another. Therefore, the rate of change in establishment
growth by industry is controlled so that interstate shifts do
not simply mask industry expansion.

Further, relocation may be a function of the ability of
a firm to move—~which, in turn, may be a function of the
amount resources invested in plant and equipment. Therefore,

capital intensity is controlled.

This hypothesis Vail. be tested hw'eMi examination cm

137

means, simple correlations, and partial slope coefficients in

multiple regression analysis.

H3. The disappearance rate of local collective bargaining
relationships across manufacturing industries is
significantly enmi positively related. to time rate <ef
change ll] the percentage CHE establishments with fewer
than 100 employees, controlling for employment growth by
industry.

Expectations for this hypothesis are extremely guarded.
First, previous research. arrives at rm) strong' conclusion
regarding the association between subcontracting and union
avoidance. Moreover, the proxy being used, while inferred
from prior studies has neither been widely tested, nor is
widely accepted.

Since an increase in the relative number of small
establishments is expected anytime employment within an
industry contracts, employment growth is controlled.

This hypothesis will be tested by an examination of
means, simple correlations, and partial slope coefficients in
multiple regression analysis.

H4. The disappearance rate of local collective bargaining
relationships across manufacturing industries is
significantly and positively related to the rate of
change in new capital investment, controlling for
employment growth by industry.

Expectations for this hypothesis are guarded as well.
Previous research ckmms not indicate 61 strong relationship

between investment in rmwv capital and a1 union. avoidance

motive.

138

A significant result on this variable ‘would. not ‘be
meaningful when investment in new capital rises with
increased employment in the industry. In order to focus on
employment-displacing technological change, growth in
employment by industry will be controlled.

This hypothesis will be tested by an examination of
means, simple correlations, and partial slope coefficients in
multiple regression analysis.

H5. The disappearance rate of local collective bargaining
relationships across manufacturing industries is
significantly and positively related to the number of
decertification elections per 10,000 bargaining units.
The surge in decertification activity beginning in the

mid—19708 suggests that the loss of certification may explain

a significant amount of variance in the disappearance rate of

collective bargaining relationships. The expectation here is

fairly strong. This hypothesis will be tested by an
examination of means, simple correlations, and partial slope
coefficients in multiple regression analysis.

H6. There will be a positive and significant effect of the
time period on disappearances recorded after 1980,
controlling for the variables mentioned above.

With this hypothesis, I begin to study the effects of
historical periods on the model of local unit disappearance.
The period following 1980 is expected to be a positive and

significant predictor of unit disappearance since it includes

all of the events following the election of Ronald Reagan and

139

the reordering of the National Labor Relations Board.

However, since it is likely that some there is a lag in the

recording disappearances, the period following 1980 may be

mitigated tux time effects <3f time late 1970s. Therefore,
expectations for this hypothesis will not be as strong as for
those immediately following.

This hypothesis will be tested through an examination of
simple correlations and partial slope coefficients in
multiple regression analysis.

H7. There will be a positive and significant effect of the
time period on disappearances recorded after 1983,
controlling for the variables mentioned above.

The period following 1983 is also expected to be a
positive significant predictor of unit disappearance.
Recalling the lag mentioned above, this period would include
the impact of the early 1980s through the end of the time
period. The null categony would represent disappearances
logged before 1983. Since the period dummy would include the
effects of the early—1980s recession plus the policy effects
of the Reagan adminstration, the expectations are quite
strong for the this hypothesis.

This hypothesis will be tested through an examination of
simple correlations and partial slope coefficients' in
multiple regression analysis.

H8. There will be a positive and significant effect of the

time period on disappearances recorded after 1986,
controlling for the variables mentioned above.

140

The period following 1986 is expected to be a positive
significant predictor of unit disappearance since it includes
lagged effects beginning in 1983. The null category in this
case would include disappearances recorded before 1986.

Expectations for this hypothesis are somewhat guarded
since it is expected that a fairly large rise in
disappearances may have occurred in the mid—19805, which is
now part of the null category.

This hypothesis will be tested through an examination of
simple correlations and partial slope coefficients in

multiple regression analysis.

H9. The variable firm failure will lose significance
with the introduction of each period variable into
the model.

In order for the time period categorical variables to be
truly meaningful, they should correspond to some change in
significance in the other variables. Otherwise, time period
significance may simply signal a parallel shift in the effects
of the other variables. Zhi this case, I expect the firm
failure variable to lose its significance following the
introduction of each time period variable. This result would
suggest that economic distress within industries becomes a
less salient predictor of disappearance as time progresses.
The strongest expectations are for the final time period,

since the impact of the early 1980s recession. would. be

141
included within the null category.
This hypothesis will be tested through an examination of
partial slope coefficients in multiple regression analysis.

H10. All other variables will retain their significance as
time period variables are added.

Since time other continuous variables enme expected to
increase in magnitude after the reviewed policy changes, I
expect no loss of significance following the introduction of
period variables in the variables representing relocations,
input substitutions, or decertifications.

This hypothesis will be tested through an examination of
partial slope coefficients in multiple regression analysis.
H11. The disappearance rate of local collective bargaining

relationships across :manufacturirm; is jpositively .and
significantly related to the cross—product of each time
period variable and the variable representing
establishment relocation, controlling for establishment
growth and capital intensity.

Due to the policy changes reviewed earlier, it is
expected that employers will have taken greater advantage of
the relocation option as the decade progressed. Therefore,
the significance of the relocation proxy will increase with
each time period. This trend will be identified through a
significant result on a variable accounting for the
interaction of time period and the cross—state distribution
variable. The strongest expectation is for the final period,

since pre-Dotson Board.disappearances would be included within

the null category.

142

This hypothesis will be tested through an examination of
partial slope coefficients in multiple regression analysis.
H12. The disappearance rate of local collective bargaining

relationships across manufacturing is positively and
significantly related to the cross—product of each time
period variableémmithe variable relatingtmethe increase
in. establishments ‘with 100 cm? fewer‘ employees,
controlling for employment growth.

Duetmetlmepolicy changes reviewed earlier:n:is expected
that employers 'will ihave taken. greater' advantage (if the
outsourcing option as the decade progressed. Therefore, the
significance of the outsourcing proxy will increase with each
time period. This trend will be identified through a
significant result on a variable accounting for the
interaction of time period and an increase in establishments
with fewer than 100 employees. In light of previous research,
the expectation here is not strong. Although, again, the
strongest expectations are for the final period, since pre—
Dotson Board disappearances would be included within the null
category.

This hypothesis will be tested through an examination of
partial slope coefficients in multiple regression analysis.
H13. The disappearance rate (n5 local collective bargaining

relationships across manufacturing is positively and
significantly related to the cross-product of each time
period variableanmithe variable relatingtmethe increase
in new capital investment, controlling for employment
growth.

Due to the policy changes reviewed earlier it is expected

that employers will have taken. greater advantage: of the

143

technology for labor option as the decade progressed.
Therefore, time significance (ME the rmnv capital investment
variable will increase with each time period. This trend.will
be identified through 51 significant result (ii a variable
accounting for the interaction of time period and an increase
new capital investment. .Again, the expectations for this
hypothesis are extremely guarded, and the strongest
expectations are for the final period.

This hypothesis will be tested through an examination of
partial slope coefficients in multiple regression analysis.
H14. The disappearance rate of local collective bargaining

relationships across manufacturing is positively and

significantly related to the cross—product of each time
period variable and the variable relating to
decertification elections.

Due to the policy changes reviewed earlier it is expected
that employers and employees will have taken greater advantage
of the decertification option as the decade _progressed.
Therefore, the significance of the decertification variable
will increase with each time period. ‘Nihs trend will be
identified through a significant result on a variable
accounting for the interaction of time period and
decertification election per 10,000 bargaining units.
Strongest expectations are for the final period.

This hypothesis will be tested through an examination of

partial slope coefficients in multiple regression analysis.

144
H15. The disappearance rate of local collective bargaining
relationships across .manufacturing :Me negatively' and
significantly related to the cross-product of each time
period variable and the variable relating to firm
failures.

Due to the policy changes reviewed earlier it is expected
that the economic distress of industries will become a less
salient explanation for the disappearance of local collective
bargaining relationships across industries. Therefore, the
negative significance of the firm failure variable ‘will
increase with each time period” This trend.will be identified
through a significant result on a variable accounting for the
interaction of time period and firm failure. Strongest
expectations are also for the final period.

This hypothesis will be tested through an examination of
partial slope coefficients in multiple regression analysis.
a. conclusions.

In this chapter, I have discussed the variables in the
model of local union disappearance. Based. on. previous
research, I have indicated that firm failures, while
increasing dramatically throughout the period are difficult
to tie to anti-union behavior. Therefore, I use the firm
failure variable as ea measure (ME economic distress within
industries and a means of testing the economic theory of
relationship disappearance.

In contrast, there is a great deal of evidence that plant

relocations are associated with a desire to escape

145
unionization. When controlling for new establishment growth,
changes iii the cross—state distribution (n3 establishments
become one test of the policy theory of relationship
disappearance. Employers may have taken advantage of more
lenient standards in mandatory bargaining to move facilities
to greenfield sites.

The expectations with regard to the input substitution
variables are not strong, since previous research reveals at
best a weak connection between either capital substitutions
or outsourcing and anti—union behavior. Nonetheless,
significant results with regard to variables relating to new
capital investment and small firm growth could be consistent
with the theory that employers took greater advantage of such
substitutions when the duty to bargain was weakened.

Although there is little doubt that decertification
elections are undertaken to escape unionization, the extent
tt>which this phenomenon explains the variance in relationship
disappearance across industries during time pernmd will be
tested. A significant result would be consistent with the
proposition that looser standards of employer election conduct
and tighter standards of union discipline and union security
encouraged more union decertification activity.

The first time period variable is associated with
disappearances logged after 1980; the second with those logged

after 1983, and the third with those recorded after 1986.

146

In the next chapter, the data and methods used to test

the preceding hypotheses are discussed.

Chapter VI. Data and Methods.

A. Introduction

In this chapter I describe the data and methods used to
test the hypotheses presented at the end of the Chapter V.
First, I discuss the sample and sampling technique used to
calculate the dependent variable. I then discuss the methods
for analysis of the data, and problems likely to be
encountered. I conclude with a description of formulae and
descriptive statistics relating to the dependent and
independent variables.

B. The sample.

There is virtually no data measuring the characteristics
of local collective bargaining relationships. 1x:is, in fact,
impossible tmetanswer precisely how mmmu/ such relationships
exist. Neither the economic censuses nor the Bureau of Labor
Statistics’ establishment survey contain information on local
bargaining relationships. Hence, most of what we know about
union growth and decline comes by way of the Census Bureau’s
Current Population Survey; which” as 61 household. survey,
provides informatitml about individuals not firms <or

establishments. The demographic information on union members

147

148
is, therefore, relatively good, but we can only guess at the
number and characteristics of bargaining units and unionized
employers.

One potentially good source of information is the Federal
Mediation and Conciliation Service’s (FMCS) records of F—7
filers. The FMCS requires a one month notice of the
expiration of a collective bargaining agreement. Notice is
given by filing an F—7 form, which contains some interesting
information on both the employer and local union. There are
gaps involving extremely small bargaining units, the public
sector, and those situations in which a mid-term extension of
contract terms was agreed to. Otherwise, an approximately
three year cycle of F—7 information should yield nearly a
complete census of private sector collective bargaining
relationships. Unfortunately, the data are available only
from the late—1980s forward and, therefore, would have been
of limited use in this study.

Instead, the sample was constructed from the 1978 key
files of the U.S. Department of Labor’s collective bargaining
agreement archive. While the archive is not a random sample,
and relies on voluntary contributions from unions and
employers, its coverage in the 1970s was quite broad. The key
file used for this study listed over 5700 agreements covering

relationships of 100 workers or more across the entire private

149

economy.1 The total number' of (workers covered In! these
agreements was nearly 10 million, or approximately 55 percent
of the private sector unionized labor force at the time. The
key file itself contains a number of interesting variables,
including employer name, location, and four—digit Standard
Industrial Classification; union name and local number; size
of the bargaining unit and type of worker covered; and the
structure of the bargaining relationships, e.g. single
employer—single union, multiemployer, multiunion, etc.

Not all relationships found in the key file were used for
this study; rather several selection criteria were applied.
First, the final sample contains only relationships found in
manufacturing. Manufacturing was selected in order to have
relationships at risk on all of the variables contained in the
model of local relationship disappearance. It.would.have:made
no sense to examine plant relocations, for example, with firms
bound to a local market, such as hotels, retail stores, or
construction contractors. Second, only single-
employer/single—location/single—union relationships were
considered. This was a practical necessity; it is virtually
impossible to check for the disappearance of individual
relationships within multi—employer, multi—plant, ‘ or

associational bargaining structures. .A similar situation

 

1 There is a separate file for public sector agreements.

150

exists with. multiplant agreements at a single employer.
Attempting to gain accurate information for more than a few
such situations would have been an enormous undertaking and,
very likely, produced rather unsatisfactory information.
Fortunatelyy in most:manufacturing industries, single location
agreements are quite typical. Sometimes these agreements are
the sole contracts covering workers at a particular plant; at
other times, as with the Big Three Automakers, local
agreements address matters not covered in higher—level
agreements.

Third, due to the way in which I accounted for
disappearances, it was necessary for the union to be
identified by its parent organization and local union number,
e.g. UAW Local 652. Finally, to equalize the time periods
being studied, the relationship had to have been in existence
since at least the beginning of 1977. It was easy to verify
this requirement by examining the active dates of the
agreement listed in the key file.

I remained concerned that the non—randomness of the
sampling frame and the selection criteria applied would yield
an achieved sample that would be an unrealistic representation
of the actual population (i.e. a large number of relationships
in an industry known to have low union penetration and.a small
number of relationships in an industry' with high union

penetration). Since there is no information on the actual

151

population of relationships, a direct comparison of sample to
population Characteristics wasrmm:possible. Instead, thelma/
file data on bargaining unit size was used to estimate the
number of covered employees per industry in the sample.
Estimates were then created for the population with Current
Population Survey data for the same period. Both sets of data
were ranked by the number of union members per industry, and
a non—parametric Spearman’s rank order correlation test
(Siegel, 1956) was carried out to reject the hypothesis of no
rank order correlation between the sample and.populationu The
null hypothesis was comfortably rejected.with less than a .008
probability of committing a Type I error.

However, in time process Ci? conducting time test, two
industries in the sample were found to digress substantially
from the population: Apparel (SIC 23), and Printing (SIC 27).
The reason for these digressions is very clear: both
industries rely heavily on associational agreements.
Therefore, both were excluded from the analysis.

A further decision was made tie exclude several other
industries which.produced uncomfortably small within—industry
sample sizes, or where there was a lack of data on more than
one or two of the independent variables for the entire time
period. Industries excluded for these reasons were, Tobacco
(SIC 21), Leather Goods (SIC 31), Instruments (SIC 38) and the

residual category of Miscellaneous Manufacturing (SIC 39).

152
The final sample used in the analysis contains 1411
relationships across fourteen two-digit industries.
Altogether the relationships cover approximately 856,000
workers, or nearly 18 percent of the organized labor force in

those industries in 1977.

Table 6.1. Sample characteristics, selected industries, 1977.

 

 

SIC/Industry N Average unit size (S.D.) Median unit size
20 Food 134 654 (952) 425
22 Textiles 43 522 (441) 400
24 Lumber 42 358 (308) 250
25 Furniture 31 500 (482) 300
26 Paper 149 450 (489) 250
28 Chemicals 138 543 (614) 400
29 Petroleum 44 275 (215) 200
30 Rubber 44 669 (603) 475
32 Stone, Glass, Clay 85 370 (309) 300
33 Primary Metals 160 602 (530) 500
34 Fabricated Metals 98 495 (623) 250
35 Machinery 189 692 (695) 500
36 Electrical Equip. 132 975 (1286) 600
37 Transport. Equip. 122 782 (1197) 400
Total 1411 606 (780) 350

 

Source: U.S. Department of Labor, Key File, Private Sector Agreement Archive, 1978

C. Calculation of the dependent variable: .panel attrition.
The 1411 cases formed a panel of relationships known to
exist in 1977. Each relationship was then followed forward

until 1990, using' the ZDepartment of Labor’s Register* of

 

Reporting LabortOrganizations, and<disappearances were logged.
For most panel studies, attrition of subjects is a problem;
for this study it.:he the variable of interest. The 1959
Labor—Management Reporting and Disclosure Act, more commonly
known as the Landrum—Griffin Act (29 USC 401—531), requires

that all labor organizations representing workers in the

153

private eumi federal sectors file emmumd. reports vnlii the
Department of Labor's Bureau of Labor—Management Standards.
Registers of active filers were published approximately every
three years between passage of the Landrum-Griffin Act in 1959
and 1990.2

If a labor organization ceases operation, it must file a
"terminal report" and is taken off the active filer list. The
terminal reports contain information on the reasons why an
organization ceases operation, the value of assets at time of
termination, etc., and.would, therefore, have been of enormous
value to this study. Unfortunately, the Department of Labor
does not retain such reports for more than one year. Nor is
it clear what happens to the reports once they leave the
Department. (Officials there, including time chief of the
section, chmi not know). They apparently gm) into federal
storage, and are retrievable only if one knows the box number
in which a particular report is contained. Inquiries at the
National Archives were rm) more helpful. Officials there
stated that they deal only in executive branch records that
are at least 25 years old. Apparently, it is almost
impossible to gain access to records more than one year and
less than 25 Years old!

Without the terminal reports themselves, the Register had

 

’ For budgetary reasons, the Department has not produced
a Register since 1990.

154

to suffice. If a local organization (as identified. by
location, national union affiliation, and local union number,
e.g. UAW Local 652, Lansing, MI) continued to be listed as an
active filer, it was assumed that the relationship was on—
going, and. the case ‘would receive: a score <of 'WV' on. a
categorical ‘variable lammflitml DISAPPEARANCE. If ea local
organization was not listed, it was assumed that the
relationship had ceased and the case would receive a score of
“1'I for DISAPPEARANCE. The baseline year was 1977, and
Registers published in 1980, 1983, 1986 and 1990 were
consulted. Apparent disappearances were always re—checked in
the following period (except, obviously, for those
disappearing in the final period) to see if an organization
had simply failed to file during a publication year; none were
found to have done so. In addition, compliance with the Act
was checked by cross—referencing those organizations listed
in the key file as having had active agreements in 1977 with
those listed in the 1977 Register. Approximately 40
organizations appeared to tmecmu: of compliance, and were,
therefore, excluded from the analysis. With such an
apparently high rate of compliance, the fear of false
positives (i.e. a non—reporting relationship being coded 1 on
DISAPPEARANCE) is small.

But obviously, the technique described above is not

flawless. First, it should be obvious that the term

155

"disappearance of a relationship" is not completely accurate,
since there is no way of telling whether one party to a
relationship-~the employer—-changed. Ei)ltmm;as a local union
appears to have continued representing employees at a
particular location, the relationship was coded as having no
disappearance, even if there) had. been. a (change in firm
ownership. On the other hand, if there had been a complete
change in the parent organization and number of a local union
(as may occur when one union "raids" another) the relationship
would have been counted as disappearing——an obvious source of
error. But due ti) "no raid" agreements among all mmmmer
unions of the AFL-CIO, the phenomenon is quite rare. Also
rare is the re—emergence of a new union at location that had
formerly been "deunionized".

It more (mimmni phenomenon is LmUtm1 mergers (Chiason,
1986), which did change the national union designation for
some cases during the period of study and created.a hazard for
identificatitmr Fortunately, the research revealed that local
union identification numbers are typically retained after a
merger, with, perhaps, a slight modifier added. For example,
when the Clothing Workers and Textiles Workers merged into the
Amalgamated Clothing and.Texti1eeWorkers (ACTWU), local unions
retained. their local identificatitmi numbers, but (added (a
letter to indicate their former affiliation, For example,

Textile Workers Local 110 would become ACTWU Local 110T.

156

Similarly, following the merger of the Cement Workers with
the Boilermakers, local organizations were simply listed in
the Register as BBF Local 110, Cement Workers.

Another problem. with this system. of coding is the
assumption that the end of a bargaining relationship and the
termination of a local organization is relatively coincident.
However, labor organizations may remain in existence for some
periodcoftjnmeafter collective bargaining ceases. 115a plant
closes, for example, the local union may remain active—-
albeit often under the trusteeship of its parent organization-
-to resolve any outstanding claims against the company, settle
its bills, and liquidate its own assets. Therefore, it is
possible that eni organization filing vniii the KKK, in, say
1983, but not filing in 1986, may actually have been a "lame
duck" local for somejyears before 1983. Although these errors
may occur, as long as they are not systematically related to
any particular industry or time period the analysis should not
be too seriously flawed. However, the results may be
conservative with regard to the number of total
disappearances, since :mxme disestablished locals rmn/ have
continued to file annual reports.

Once all cases were coded, they were aggregated to the
industry level (2-digit Standard Industrial Classification)
and time period, with the time periods being weighted to

account for their slightly differing durations” The resulting

157

dependent variable is time disappearance rate tin: fourteen
industries over four approximately three—year time periods
between 1977 and 1990. Taken.as pooled cross-sections (Sayrs,
1989) 56 cases are observed. To account for the fact that
there will always be some time lapse between the end of
relationship and the filing of a terminal report, independent
variables in the model are lagged by one time period. For
example, firm failures between 1983 and 1986 are associated
with the disappearances recorded for the period of 1986
through 1990. (Consideration. will be given. in the next
chapter, however, to alternative, non—lagged, models.

Although an individual level analysis may have been
preferable, theztask of gathering establishment—level data for
over 1400 locations, and for a period of over a decade, would
not tmnme been feasible. Archival sources (e.g. [hm1 and
Bradstreet’s Middle Market and Million Dollar directories),
would.provide only some data, typically at the firm-level (not
the establishment level), and only tin: a relatively small
portion of the firms in the sample. A survey would, most
likely, also have proven unsatisfactory. First, the burden
on the respondent to provide information for such a long
period would have nearly ensured a low response rate and/or
shoddy information“ Second, some number of the disappearances
are the result of firms going completely out—of-business,

leaving no one to respond.

158

A third option would have been to carry out the analysis
using individual cases as the dependent variable, but
employing mostly industry level data for the independent
variables. However, since there would have been no variance
across cases within industries——i.e. the industry mean would
be used to represent the firmtmean-—the result would have been
a de facto industry study with an artificially inflated sample
size.

The most reasonable and honest approach was to aggregate
cases to the 2-digit SIC level and conduct an industry-level
analysis. It is important to make the distinction between the
unit of analysis, which is still local collective bargaining
relationships, and level of analysis, which is the 2-digit
industry.

D. Problems with aggregation: ecological fallacy.

Whenever the dependent variable is aggregated, a strong
caution regarding ecological fallacy must be issued. At one
time social scientists quite freely made inferences about
individual behavior from group—level data. That began to
change :hi 1950 'when tNilliamt Robinson. demonstrated, ‘using
evidence on literacy’ and. national origin, that both. the
strength and the direction of a correlation changed When
individual—level rather than state-level data. were used.
Employing 1930 Census data, Robinson showed that there was a

positive correlation between being a native—born American and

159
being literate, Ibut 51 fairly strong' negative (morrelation
between the relative size of a state’s native—born population
and its rate of literacy (Langbein and Lichtman, 1978).
There 1mmne been ea number Ci? suggested approaches to
cross—level analysis, which are appropriate under certain
circumstances. Some'approaches,samiias ecological regression

(Langbein and Lichtman, 1978), result in better estimates, but

require a large sample of groups. No approach solves the
problem completely. Yet aggregate—level research continues
for a variety of reasons. In some cases——as here--there are

simply too many obstacles to gathering individual level data.
At other times, it is the aggregated comparisons which are of

greater theoretical interest.
As Langbein and Lichtman (1978) note:

Public policy analysts have examined
the determinants of both fiscal and
nonfiscal policy adoptions in such units
as counties, cities, states, and nations,
employing aggregate level data in the
seemingly perennial search ti) determine

whether socioeconomic or political
variables are more powerful predictors of
policy outputs. Economists employ

aggregate data almost exclusively.

Perhaps the most notable examples of
economic studies are the time series
econometric:models<oftimeAmerican.economy
and. the cross-sectional studies of
expenditure~ patterns iii capitalist and
socialist nations.

Aggregate—level research is also frequently found in the

sociological literature (Hannan and Freeman, 1977, 1988). In

160

a stream of research that is methodologically, if not
substantively, very close to this study, there is a rather
substantial literature on the "ecological" causes of marital
divorce. Typically variables such as female labor force
participation, urbanization, and economic activity are
examined. across regions and/or‘ over time ti) gauge their
effects on divorce rates. Fukari and Alston (1992) recently
found that sex ratio, female labor force participation,
female in—migration patterns, population increase, and net
household income “were all significant predictors of the
divorce rate across Japanese prefectures. And yet we can see
where the possibilities for error exist. Although the female
labor force participation rate within a region may be a
strong predictor of the divorce rate, the authors present no
evidence that couples receiving divorces actually included a
working woman.

The same short-coming exists in this research. The
connection. between industry and. organizational level
characteristics it; speculative. A tmtfli fimn failure rate
within an industry may be associated with a high bargaining
relationship disappearance rate; however, I cannot pinpoint
the precise cause of any individual disappearance. This

limitation must be kept in mind.

161
E. The longitudinal method: an event history design.

The basis for the method employed here is an event
history design (Allison, 1982, 1984; Tuma, Hannan, and
Groenwald, 1979; Yamaguchi, 1991). Marriages, divorces,
child birth, prison recidivism, etc., are examples of
phenomena which have been examined through event history
techniques. Industrial relations examples include the study
done by Ichniowski (1988), in which he looked at the impact
of changes in the law on the spread of public sector
bargaining; Dobbin, Meyer and Scott’s (1993) examination of
internal labor markets following the passage of Title VII;
and Roberts and Zonia’s (1994) work on the impact of workers’
compensation amendments. Although event history designs are
used heavily for individual level survey research, Hannan and

Tuma (1979, p. 306) write:

Perhaps the most widespread
application of the event~history design is
in archival research. For example, C.

Tilly“s pioneering study (ME trends in
collective violence iii small French
political units records the dates of all
events.of collective violence greater than
some minimal scope. The fact that Tilly
typically aggregates over units (to the
nation) and over time (to the year) in his
analysis should not obscure the fact that
the design itself records event histories
to a population of small areal units.

Therefore, simply' because time dependent 'variable 1J1 this
study has been aggregated, does not change the fact that an

event history analysis is being carried cum; Further, by

162
avoiding time binomial situation, logistics regression can
also be avoided and the more straightforward ordinary least
squares model used.

A design for discrete—time event history analysis is
suggested by Allison (1984, 1982) and is simply a model whose
dependent variable is 1 if an event occurs during a
particularly time period, and 0 if it does not. Independent
variables are included for each case/time period observation,
some of which may vary for each time and individual case and
some of which may be constant over a number of observations.
Because of aggregation, the dependent variable here is the
disappearance rate in an industry for a time period and can
range from 0 to 1.

A. model is estimated with the assumption that the
probability of an event occurring is constant over time,
predicted by a matrix of exogenous variables, as is typical
:hi all regression—based research. Further mpdels include
dummy variables representing time periods. Here, each of the
periods before and after 1980, 1983, and 1986 are coded
either 1 or 0. Hence, a regression can be run with a vector
of exogenous variables, and then rerun with. a variable
separating effects before and after a critical year to see if
there is a need for a separate specification depending upon
the time period. That is, model fit and slope significance

are compared to the original model to see if the

163
disappearance rate is also a function of time. The interest
may be in time (i.e. duration) itself——as with the time
between a divorce and second marriage or the time period
coincident to some historical event. In this case, it is the
latter of these which is of interest.

By placing a period dummy in the model, the question of
whether the time period is significant and improves the fit
of the model can be answered. However, cross-products
involving the time dummies and other independent variables
can also be emamined to see if time effects of particular
variables change over time. This is a classic analysis of
covariance model (Blalock, 1979). If the introduction of
cross—products does not change the fit of the model, perhaps
there was simply a parallel shift in slopes over time.

With Allison’s approach, several discrete time periods
are added to the model at once, and then separate equations
are calculated containing the cross—products of each of the
time periods with one cm time other independent variables.
Another approach is offered by Gujarathi (1970) as an
improvement over the Chow test for comparing regression
equations. In this model a single dummy variable
representing the change after one point in time is added with
cross—products representing time same time change in
independent variables. It is Gujarathi’s technique that is

used here.

164

F. Problems: interaction and multicollinearity.

The cross—products indicate whether a variable is
interacting with a time period. For example, did relocation
interact with the time period between 1986 and 1990 to become
a more salient predictor of disappearance in that period?
However, if the magnitude (not simply the effect) of a
variable (also increases with. time, 51 problem) with
multicollinearity* may arise (Blalock, 1979). When
multicollinearity’ is present standard errors tend. to be
exaggerated enmi slope coefficients difficult ti) interpret
(Jaccard, Turrissi, and Wan, 1990). A classic symptom of
multicollinearity is a relatively high coefficient of
determination (i.e. R“, but few significant slope
coefficients and/or slope coefficients with perverse signs.
There are several solutions which require rather complex
transformations and deviate from the ordinary least squares
approach—-for example, principal components regression and
ridge regression (Chatterjee and Price, 1991).

There are also 51 coupbe of somewhat less complicated
Options for dealing with the problem as well. Iverson (1991)
and Blalock (1979) suggest adding the variables in the
desired order and studying the change in R"I rather than
looking at the slope coefficients. In addition, one can
include less than a compete set of variables in the model,

dropping those most responsible for collinearity. Some

165
computer programs drop variables automatically, although this
may create problems for model specification.
G. Formulae.

The first model examined here includes independent
variables for the entire period. Since it is believed that
there is some length of time between the ending of a
bargaining relationship and the filing of a terminal report,
explanatory variables are lagged one time period. Therefore,
all continuous independent variables are calculated for T—1.
The purpose of this model is to test hypotheses H1 through
H5.

In the second model I add categorical variables, or time
dummy variables, to account for shifts in the model's
intercept coincident with the periods associated with
disappearances logged as occurring after 1980, 1983 and 1986.
The purpose of this model is to test hypotheses H6 through
H10.

The third model contains all variables--lagged.continuous
variables, timme period chmmuea, emmi cross—products. The
purpose of this model is to test hypotheses H11 through H15.

1. model 1.

131 Model 1, time variables relating ti) Hypothesis 1
through Hypothesis 5 are tested. In addition to the
variables <)f pmiumuar theoretical interest——that ism those

mentioned in Equation 1-— and the control variables mention

166

in Chapter 5, an additional variable for union density (or

penetration) by industry is added to the model. Union

penetration may signal the inevitability of unionism and be

indication of the relative power of unions within an

industry.
may spread the

employers,

Perhaps more important, a high penetration rate

"burden" of unionization mpre evenly among

causing fewer to pursue escape strategies.

Therefore:

Y: a+b1(xfft-1)+b2(xdiet-1)+b3(xsft-1)+b4(cht—1)+b5(xdct-1)+

b6(xestt-1) +k)7()(empt-l) +b8(Xcapt-1) +b9(xpent-1) [2]

Where:

‘1: disappearance rate across 14 industries and 4 time periods.

1““: firm failure rate, by industry, for T-l.

1“,“: the percent point change in the distribution of establishments across
states, by industry, for T—l.

x,,,,,= average annual rate of change in the relative percentage of
establishments with fewer than 100 employees, by industry, for T-l.

1”“: average annual rate of change in investment in new capital, by industry,

16cm: decertification rate, by industry, T—l.

1“,“: annual percentage change in the number of establishments, by industry,
for T-l.

x_,._,= annual percentage change in the number of production workers, by
industry, for T-1.

Xmm— capital intensity, by industry, for T-1.

Int“: union penetration rate, by industry, for T-l.

2. Mbdels 2 through 4.

In Models 2 through 4 time period dummy variables are

added to detect
periods,

the significance of slopes.

changes in Model 1’s intercept across

and to examine the effect of intercept changes on

The purpose of the models is to

167
test hypotheses H6 through H10. A significant result in the
dummy variable would indicate that some effect coincident to
a time period is relevant to predicting the behavior of the
dependent variable. However, if individual slope
coefficients become (n: remain significant, time effects of
those variables must also be considered along with any impact
of the time period. If the time period variable is
significant, with no change in slope values, it could signal
a parallel shift in effects over time.
Therefore:

Y: a+bl(xfft-l) +b2(xdist-1) +1:>3(Xsft:—l)+b4(xnct-1)‘i—bS(xdct-1)+bG(xestt—1) +
b7(Xempt-1) +b8(xcapt—1)+b9(xpent-1)+b10(xtd) [3]

Where the additional variable is:

1(a): categorical variable representing periods after 1980, 1983, and 1986.

It should be noted that Models 2, 3, 4 are each run once
to test changes associated with disappearances recorded after
1980, 1983 and 1986. Results of the three runs will be
placed in the same table so that they may be compared across
time periods.

3. models 5 through 7.

In Models 5 through 7, cross—products of the time period
dummies éumi variables (if primary theoretical interest are
included in the model. The purpose of this model is to test
hypotheses H11 through H15. While the dummy variable tests

for changes in the intercept across time periods, the cross—

168

products test for significant changem; in. slope ‘variables
across time periods. 'Was there 61 significant change away
from firm failure and towards the other variables as the
periods passed?

Therefore:

Y: a+b1(xfft-1)+b2(xdist-1)+b3(xsft-1)+b4(xnct-1)+b5(xdct-1)+
b6(xtd)+b7(xfft-1) *(Xtd)+b8(xdist-1) 1k(Xtd)+b9(xsft-l)1'r
(Xtd)+b10(xnct—1)* (Xtd)+b11(xdctt-1)*(Xtd)+b11(xestt-1)+
b12(X +b13(X +b14(xpent-l) [4]

erupt-1) capt-l)

Where the additional variables are:

(x,,.,,)*(x,,)= firm failure rate, by industry, for T-l times period effect.

(1.1..-,)*(Xu)= the percent point change in the distribution of establishments
across states, by industry, for T-l times period effect.

(x,,,,,)*(xm)= average annual rate of change in the relative percentage of
establishments with fewer than 100 employees, by industry, for
T-l times current period effect.

(lub,)*(xm)= average annual rate of change in investment in new capital, by
industry, for T-l times period effect.

(L¢,_,)*(Xtd)= decertification rate, by industry, for T-l times period effect.

H. Descriptive statistics.

Before turning ti) the research. results in time next
chapter, I examine the attributes and patterns of the
variables through descriptive statistics. First I present
cross—industry comparisons on single variables, then cross—

variable comparisons for single industries.

169

1. variables.

a. Disappearance rate.

 

 

Table 6.2. Disappearance rate, by industry, 3-year periods, 1977-1990.
SIC/Industry Period 1 Period 2 Period 3 Period 4 Mean (S.D.)
20 Food .0448 .1641 .0583 .0647 .0830 (.0547)
22 Textiles .0698 .0750 .1719 .1490 .1164 (.0517)
24 Lumber .0238 .0976 .1903 .0455 .0893 (.0741)
25 Furniture .0323 .0000 .0684 .0000 .0252 (.0326)
26 Paper .0268 .0414 .0303 .0538 .0381 (.0122)
28 Chemicals .0072 .0365 .0562 .0898 .0474 (.0347)
29 Petroleum .0000 .0000 .0964 .0660 .0406 (.0485)
30 Rubber .0682 .0500 .0544 .1100 .0706 (.0274)
32 Stone, Glass, C1ay.0353 .0488 .0939 .1301 .0770 (.0434)
33 Primary Metals .0125 .0506 .0565 .1348 .0636 (.0513)
34 Fabricated Metals .0202 .0417 .0807 .0932 .0590 (.0338)
35 Machinery .0370 .0604 .0868 .1058 .0725 (.0301)
36 Electrical Equip. .0379 .0551 .0526 .0579 .0509 (.0089)
37 Transport. Equip. .0328 .0847 .1060 .0953 .0797 (.0325)

Mean .0321 .0576 .0859 .0854

(S.D.) (.0321) (.0409) (.0455) (.0403)

 

Note: Calculated from the key files of the Department of Labor's private sector
collective bargaining agreements archive and the Register of Reporting Labor
Organizations, 1980, 1983, 1986, 1990.

In Table 6.2, several characteristics of relationship
disappearance can be observed. It should be noted that the
disappearance of relationships and the disappearance of firms
is not the same thing. Firm failures, which will be
considered below, Hey kxa one cause of reflationship
disappearance, Inn; a relationship rmn/ cease for tine other
reasons implied by Equathmi l as well. First, we can see
that the weighted mean disappearance rate jumped 100 percent
from 1980 to 1983, another 50 percent between 1983 and 1986,
but remain about the same to the end of the decade. What is
interesting, however, is that the standard deviation of the

disappearance rate narrowed slightly during the final period.

l70

This would suggest that the phenomenon of relationship
disappearance became more generalized and less confined to a
few strongly impacted industries. Indeed, we see that in
Period 3, the jump in the overall mean can be accounted for,
almost. exclusively, in! increases :hi tlme means cfif several
industries of 9 enui lO percentage points (i.e. Lumber,
Textiles, and Petroleum). In the final period there are no
such rapid increases. However, nearly half of the industries
under study (6 of 14) show disappearance rates of at least 10
percent. This narrowing of standard deviation is consistent
with the argument that disappearances became more diffuse
later 111 the decade. It will lxe important t1) see whether
there was a concurrent diffusion of economic distress.

Examining the disappearance trend across industries
reveals that most display rising trends throughout the
period, Inn: several industries sflxnvzmore complex patterns.
Lumber, for example, rises sharply from Period 1.tx> 2 and
flmmn Period 2 t1) 3, but falls rather dramatically between
Periods 3 and 4. This pattern may reflect first the slowdown
in construction caused by high interest rates in the late—
1970s and the economic recession of the early 19808, and then
the building boom of the mid—to-late 19805. Similarly, the
trend in Petroleum seems to reflect the fortunes of that
industry as well. As oil prices fell during the mid—1980s,

the industry, for the first time, begins registering

171
disappearances. The Imma: puzzling tnmaxi occurs If] Food
Products. Periods 1, 3, and 4 all show disappearance rates
of approximately 5 or 6 percent, but Period 2 climbs to 16
percent. When statistics for all variables are considered by
industry, the reason for this unusual pattern may become more
apparent.
b. Firm failure rate.

Observing the trend in lagged firm failures (Table 6.3),
we do see some evidence that supports the argument that
economic distress did, indeed, become greater and more
generalized as the decade progressed. The change from Period
L1 to Period 2_1 does not correspond to the increase in
disappearance rate seen above, but the later trend becomes
much more consistent. The large jump in firm failure rate of
140 percent from Period 21 to Period 34, and the somewhat
smaller jump from Period 3_1 to Period 4_,, track fairly
closely the trend in disappearances. So too does the
narrowing of the standard deviation in firm failures. The
more widespread incidence of firm failures between 1983 and
1986 (Period 44) may explain the more widespread incidence of

relationship disappearances recorded between 1986 and 1990,

Period 4.

172

 

 

 

Table 6.3. Firm failure rate, per 10,000 firms, by industry, 3-year periods, 1974-
1986.

SIC/Industry Period 1,1 Period 2,,Period 3,, Period 4,l Mean (S.D.)
20 Food 32.33 29.00 70.00 122.33 63.42 (43.46)
22 Textiles 71.33 65.33 104.33 118.67 89.92 (25.72)
24 Lumber 33.00 29.33 94.67 133.00 72.50 (50.25)
25 Furniture 82.67 77.33 184.67 156.00 125 17 (53.50)
26 Paper 25.67 28.33 58.33 90.33 50.67 (30.31)
28 Chemicals 33.67 32.00 68.33 113.67 61.92 (38.35)
29 Petroleum NA NA NA 187.00 187.00 (00.00)
30 Rubber 40.00 41.00 82.00 119.33 70.58 (37.94)
32 Stone,Glass,C1ay 26.00 25.00 70.33 96.67 54.50 (35.17)
33 Primary Metals 33.00 34.33 120.67 165.67 88.42 (65.84)
34 Fabricated Metals 32.00 24.67 76.33 132.00 66.25 (49.42)
35 Machinery 31.67 27.33 98.00 111.33 67.08 (43.77)
36 Electrical Equip. 73.00 45.67 102.00 148.67 92.33 (44.04)
37 Transport Equip. 78.33 69.00 138.67 179.67 116.42 (52.26)

Mean 45.59 40.64 97.56 133.88

(S.D.) (21.77) (18.24) (34.73) (29.63)
Source: Dun and Bradstreet, Firm Failure Record 1974-1986.

 

c. Change in cross-state distribution.
In Table 6.4 we can observe the proxy variable for plant

relocation. The variable represents the overall percentage

point change in the distribution of establishments, by

industry, across states per three year period. We can seen

that over the period of 1974 tx> 1986, average change was

remarkably constant, and actually decreased 51 bit between

Period L1 and Period 44. Across industries, however, there

was some variation and even a few counter trends.

Transportation Equipment, for example, showed a fairly strong

increase :hi Period. 44- II: is interesting tub speculate
whether this represents a policy effect since both major

:relocation. decisions discussed. earlier—-Otis Elevator" and

173
Milwaukee Spring——involved this industry and its major union,
the United Auto Workers. Interesting as well is the spike in
Period 21 for the Food Products industry, which may help
explain the Period 2 increase in the recorded disappearance

rate observed earlier.

Table 6.4. Percentage point change in distribution of establishments across states,
by industry, 3-year periods, 1974-1986.

 

 

SIC/Industry Period idPeriod 24Period 34Period 4.1 Mean(s.D.)
20 Food 3.00 5.35 4.06 3.18 3.90(1.07)
22 Textiles 7.29 4.38 6.26 3.61 5.39(1.69)
24 Lumber 3.99 4.83 3.80 3.90 4.13(0.47)
25 Furniture 7.72 6.26 5.55 6.92 6.61(0.92)
26 Paper 6.93 5.72 4.09 5.67 5.60(1.16)
28 Chemicals 3.83 4.83 5.28 4.00 4.48(0.69)
29 Petroleum 5.67 8.56 7.69 6.57 7.12(1.27)
30 Rubber 6.30 5.99 4.63 6.17 5.77(0.77)
32 Stone, Glass, Clay 7.13 3.84 4.74 4.83 5.14(1.40)
33 Primary Metals 4.94 4.82 5.32 5.60 5.06(0.50)
34 Fabricated Metals 5.61 3.83 4.88 3.43 4.44(0.99)
3S Machinery 5.48 4.79 4.78 3.47 4.63(0.84)
36 Electrical Equip. 7.84 7.41 7.89 5.13 7.07(1.31)
37 Transport. Equip. 5.10 5.29 5.05 8.41 5.96(1.63)

Mean 5.74 5.42 5.28 5.06

(S.D.) (1.54) (1.32) (1.24) (1.57)

 

Source: Bureau of the Census, County Business Patterns, 1974, 1977, 1980, 1983, 1986.

d. Rate of growth in new capital expenditures.

Table 6.5 reveals that, in general, new capital
expenditures folloW'lgeneral economic trends. During' the
recession of the early 19805, only Petroleum showed an
average annual increase. But as the oil glut of the mid—
19808 depressed petroleum prices, capital expenditures began
to fall in that industry as well. Below, we will compare new
capital expenditures to employment trends to see if capital

expenditures may have been labor displacing. It will be

particularly interesting to see how the two variables compare
in Rubber,

all of which showed fairly large increases

expenditures in Period 44.

Fabricated Metals,

174

and Transportation Equipment

in

new capital

 

 

 

Table 6.5. Average annual rate of change in new capital expenditures, by industry,
3-year periods, 1974-1986.
SIC/Industry Period ldPeriod 24Period 34Period 4.1 Mean (S.D.)
20 Food 4.33 1.26 -6.25 3.02 .59 (4.73)
22 Textiles 2.43 -3.01 -4.20 -1.45 -1.56 (2.89)
24 Lumber -1.55 -0.57 -21.02 9.15 —3.50 (12.64)
25 Furniture -0.25 4.33 -4.87 5.84 1.26 (4.84)
26 Paper 6.19 5.02 -6.86 3.79 -2.04 (6.01)
28 Chemicals 9.15 -7.26 -12.36 0.58 -2.47 (9.40)
29 Petroleum 1.03 6.89 5.27 -19.83 -1.66 (12.36)
30 Rubber -0.94 -0.27 -10.80 18.73 1.68 (12.35)
32 Stone,Glass,C1ay -0.87 7.33 ~22.13 8.60 -1.77 (14.21)
33 Primary Metals —1.22 -2.84 -12.41 -11.11 -6.90 (5.68)
34 Fabricated Metals 2.58 3.93 -1S.18 14.33 1.42 (12.24)
35 Machinery 2.92 10.18 -11.62 -0.36 0.28 (9.08)
36 Electrical Equip. NA NA NA NA NA
37 Transport. Equip. 10.50 8.40 -15.67 27.95 7.79 (17.93)
Mean 2.64 2.56 -10.62 4.56
S.D. (3.96) (5.17) (7.38) (12.25)
Source: Bureau of the Census, Annual Survey of Manufactures, 1974—1986

 

e. Annual rate of change in small establishments.
The trend in small establishments may tell us something

about outsourcing behavior in industries. Every industry has

shown some overall increase in the relative number of smaller

establishments, with Primary Metals showing the largest

overall increase. Again, it anJ_ibe necessary t1) compare

employment patterns with the trend in small establishments

before any stronger conclusion. can. be <drawn. At first

glance, it appears as though relative changes in the number

ZXS

of small establishment tracks general economic trends.

175
the economy contracts (Period l_1 and Period 34) the relative
number of small establishments grows. But as these
establishments add workers, some jump into the category of
larger establishments. Particular attention should be paid
to coincident growth in both employment and the relative
percentage of small firms. With a parallel trend, a stronger
argument could be made that small firms are being relied on

more heavily for industry output, perhaps due to outsourcing.

Table 6.6. Average annual rate of growth in the percentage of establishments with
fewer than 100 employees, by industry, 3-year periods, 1974-1986.

 

SIC/Industry Period 14Period 24Period 34Period 4.1 Mean (S.D.)

 

.72 -0.25
.39 -0.36
.41 -0.22
.05 -0.32

20 Food

22 Textiles
24 Lumber

25 Furniture

.90 -1.18
.67 -0.28
.19 -0.22
.67 -0.55

.26 (0.89)
.60 (1.28)
.04 (0.31)
.21 (0.77)

0 0 0
0 2 0
0 0 0
0 1 0
26 Paper 0.96 -0.42 1.49 -0.77 0.31 (1.08)
28 Chemicals 0.06 -0.26 0.52 0 10 0.10 (0.32)
29 Petroleum 0.29 -0.54 0.75 0 59 0.27 (0.58)
30 .Rubber 0.76 -0.34 1.20 -0.65 0.24 (0.88)
32 Stone, Glass, Clay* 0.50 -0.47 0.72 -0.17 0.14 (0.56)
33 Primary Metals 1.13 -1.03 3.02 -0.80 0.58 (1.89)
34 Fabricated Metals 1.34 -0.42 1.11 -0.25 0.44 (0.91)
35 Machinery 0.38 -0.50 0.79 -0.14 0.13 (0.57)
36 Electrical Equip. 1.10 -1.59 1.88 -0.02 0.34 (1.51)
37 Transport. Equip. i-0.04 -0.51 1.47 0 51 0.10 (0.94)
Mean 0.64 -0.60 1.25 -0.27
(S.D.) 0.42 0 40 0.75 0 36

 

Source: Bureau of the Census, County Business Patterns, 1974-1986.

 

f. Decertification elections.
Unlike:most decertification data, Table 6.7 indicates how
many elections were held relative to the approximate number
of bargaining units within each industry. So, for example,

even though relatively few decertification elections occurred

each year in Furniture——since it is not a highly unionized

industry--the number

possible venue was quite high.

follows fairly

of

closely

176

the

trends

decertification

in

elections

firm

disappearance rates, but unlike those variables,

deviation increases as well. The increase in standard
deviation appears to be caused by the strong rise
magnitude of decertification activity in Chemicals
Machinery, relative to Paper and Transportation Equipment,

which remain fairly low throughout the period.

The decertification trend
failure

the standard

 

 

 

Table 6.7. Annual average decertification elections per 10,000 bargaining units,
by industry, 3-year periods, 1974-1986.

SIC/Industry Period 14Period 24Period 34Period 4,l Mean (S.D.)
20 Food 64.81 106.86 114.22 132.70 104.65 (28.70)
22 Textiles 43.24 75.25 64.05 126.72 77.32 (35.51)
24 Lumber 47.08 87.61 106.50 124.17 91.34 (33.07)
25 Furniture 51.94 93.58 167.15 133.89 111.64 (49.89)
26 Paper 21.09 31.77 25.95 35.02 28.46 (6.18)
28 Chemicals 72.20 113.39 158.80 241.42 146.45 (72.52)
29 Petroleum 52.16 93.23 102.28 111.58 89.81 (26.20)
30 Rubber 82.27 135.38 108.01 119.73 111.35 (22.40)
32 Stone,Glass,Clay 27.08 46.99 73.96 111.05 64.77 (36.34)
33 Primary Metals 26.44 36.80 58.73 75.27 49.31 (21.93)
34 Fabricated Metals 45.28 66.56 111.03 103.75 81.66 (31.10)
35 Machinery 65.72 81.29 162.31 217.27 131.65 (71.07)
36 Electrical Equip. 64.91 104.81 123.16 111.57 101.11 (25.30)
37 Transport. Equip. 21.65 47.98 51.23 58.15 44.75 (15.98)

Mean 48.99 80.11 101.96 121.59

(S.D.) (19.64) (30.94) (43.24) (54.15)
Source: National Labor Relations Board, Annual Reports, 1974-1986.

 

g. Average annual establishment change.

Table 6.8 presents the first control variable, the

eaverage annual rate of change in the number of establishments

1mg industry. This is an important control with regard to the

177

cross-state change in the distribution of establishments. It
is surprising that the rate of change has been, on average,
positive. We know by examining Table 6.9 and Table 7.2 that
employment in manufacturing has declined quite dramatically.
Obviously, newer establishments have relatively fewer
employees than the older establishments.

The increase observed in Period 34, seems particularly
surprising since that period includes the recession of the
early—1980s. However, when I examined the underlying data,
I discovered that a sharp rise in the number of
establishments began in 1983, negating decreases in 1981 and
l982. In absolute terms, however, few industries grew to
their pre—recession levels. Further, following the decade-
long contraction, small absolute increases caused relatively

large percentage increases.

178

Table 6.8. Percentage change in the number of establishments, by industry, 3—year
periods, 1974-1986.

 

 

SIC/Industry Period 1_,Period 2_,Period 3,,Period 4.1 Mean (S.D.)
20 Food —1.46 -4.69 1.24 —0.94 -1.46 (2.45)
22 Textiles -1.18 -1.85 0.32 -1.10 -0.95 (0.92)
24 Lumber 1.06 -2.67 2.54 -0.58 0.09 (2.24)
25 Furniture 1.72 -0.89 4.23 2.77 1.96 (2.16)
26 Paper 1.57 —0.20 0.70 0.28 0.59 (0.75)
28 Chemicals 1.08 -0.73 2.25 0.13 0.68 (1.28)
29 Petroleum 0.59 1.20 1.41 1.39 1.15 (0.38)
30 Rubber 4.11 3.52 3.35 2.47 3.36 (0.67)
32 Stone, Glass, Clay 1.56 -1.99 1.10 -0.33 0.08 (1.60)
33 Primary Metals 0.86 -0.57 0.38 -1.64 —0.24 (1.10)
34 Fabricated Metals 2.62 0.14 2.96 0.92 1.66 (1.35)
35 Machinery 2.89 1.32 3.28 -0.50 1.75 (1.72)
36 Electrical Equip. 3.29 0.47 5.37 2.62 2.94 (2.02)
37 Transport. Equip. 2.82 -2.91 2.94 1.25 1.03 (2.74)

Mean 1.54 -0.70 2.29 0.48

(S.D.) (1.57) (2.07) (1.51) (1.45)

 

Source: Bureau of the Census, Annual Survey of Manufacturers, 1974-1986

h. Average annual change in production worker
employment.

The data in Table 6.9 represent percentage changes in
production worker employment by industry. This variable is
particularly salient as a control for the possible effects of
the proxies for outsourcing and technological change on the
disappearance rate. The cwerall trend reweals ea general
decrease, particularly during the early—19805 recession. It
is interesting to note that the turnaround in establishment
decline in 1983 is not mirrored in the employment trend.

The sharpest declines occurred in Primary Metals, which
also experienced an overall decline in establishment growth
(but a relatively substantial increase in the number of small

establishments, perhaps mini—mills). The most robust

179

increase during the mid-19808 recovery occurred in
Transportation Equipment. Since this 2-digit SIC would
include many industries involved in the production of
military craft, I would guess that the turnaround in

employment in this industry is closely linked to the defense

build—up of the mid-19805.

 

 

 

Table 6.9. Average annual percentage change in the number of production workers,
by industry, 3-year periods, 1974-1986
SIC/Industry Period queriod 24Period 34Period 4 Mean (S.D.)
20 Food -0.04 0.40 -1.76 0.52 -0.22 (1.05)
22 Textiles -1.77 -2.38 -4.55 -1.58 -2.57 (1.36)
24 Lumber 1.15 -1.87 -1.31 3.01 0.24 (2.26)
25 Furniture -1.09 -0.37 -1.64 3.85 0.19 (2.50)
26 Paper -1.31 0.31 -1.81 1.11 -0.43 (1.37)
28 Chemicals -0.26 0.53 -2.56 -0.61 -0.72 (1.31)
29 Petroleum 1.39 -1.56 -1.55 -3.53 -1.31 (2.03)
30 Rubber 1.54 0.28 -0.44 3.83 1.30 (1.87)
32 Stone, Glass, Clay -1.89 —1.12 -5.28 1.23 -1.77 (2.70)
33 Primary Metals -3.31 -1.39 -10.60 -2.85 -4.54 (4.12)
34 Fabricated Metals -1.20 0.09 -5.85 2.02 -1.23 (3.35)
35 Machinery -1.14 4.15 -9.29 1.01 -1.32 (5.74)
36 Electrical Equip. -1.60 2.63 -2.74 0.46 —0.31 (2.37)
37 Transport. Equip. 0.80 -1.09 -3.67 5.15 0.30 (3.72)
Mean -.62 -.10 -3.79 .97
(S.D.) (1.43) (1.77) (3.05) (2.54)
Source: Bureau of Labor Statistics, Employment, Hours andggarnings, 1994.

capital intensity of industries.
important
.intensity may affect the ability of a firm to relocate, since
large investments in buildings and equipment would make the

Inoving process

The following table

i.

control

in

 

several

Capital intensity.

(Table 6.10)

presents data on the

regards.

relatively expensive.

This variable provides an

First,

Second,

capital

employers

180

within capital intensive industries Hey lxe less concerned

with the impacts of unionization, since labor represents a

relatively smaller portion of total costs. As the table

indicates, Paper is the most capital intensive of the

industries under study while Electrical Products is the least

capital intensive.

 

 

 

Table 6. 10. Percentage of assets invested in capital equipment, by industry, 3-
year periods, 1974-1986.
SIC/Industry Period 1,,Period 2_,Period 3,,Period 4.1 Mean (S.D.)
20 Food 34.00 34.67 35.33 34.33 34.58 (0.57)
22 Textiles 31.67 33.00 33.33 32.67 32.67 (0.72)
24 Lumber NA NA NA NA
25 Furni tu re NA NA NA NA
26 Paper 49.33 50.00 56.00 61.00 54.08 (5.50)
28 Chemicals 39.33 39.33 41.33 37.67 39.42 (1.50)
29 Petroleum 46.67 48.67 55.33 57.33 52.00 (5.14)
30 Rubber 33.33 33.67 35.67 37.00 34.92 (1.73)
32 Stone, Glass, Clay 45.33 46.67 48.33 45.33 46.42 (1.42)
33 Primary Metals 46.67 45.00 48.33 45.33 46.33 (1.52)
34 Fabricated Metals 27.67 28.67 31.33 30.00 29.42 (1.60)
35 Machinery 27.33 29.00 30.67 27.33 28.58 (1.60)
36 Electrical Equip. 23.67 23.67 26.33 26.00 24.92 (1.45)
37 Transport. Equip. 25.67 26.67 29.33 26.67 27.08 (1.57)
Mean 37.53 36.58 39.28 38.39
(S.D.) (9.91) (9.13) (10.33) (11.68)
Source: Bureau of the Census, Quarterly Financial Reports, 1974-1986.

 

j. Onion penetration rates
The final control variable is union penetration rates
(Table 6.11). A high union presence may affect the balance
of power within an industry, and may also distribute the
effects of unionization more evenly across employers. (Union

penetration ranges from 10 percent in Textiles in Period 4_1

tt) 62 percent in the Primary Metals industry in Period 14.

All of the

181

industries sflxwv large declines throughout the

 

 

 

period.
Table 6.11. Average union penetration rates, by industry, 3—year periods,
1986.
SIC/Industry Period 14Period 24Period 34Period 4.1 Mean (S.D.)
20 Food 44.23 43.10 37.00 32.07 39.10 (5.66)
22 Textiles 14.87 14.60 12.80 10.20 13.12 (2.15)
24 Lumber 26.87 24.67 19.87 14.97 21.59 (5.30)
25 Furniture 26.60 27.67 21.73 16.27 23.07 (5.22)
26 Paper 55.83 55.20 51.63 46.43 52.28 (4.31)
28 Chemicals 31.00 30.40 23.57 16.47 25.36 (6.82)
29 Petroleum 34.53 38.83 34.87 32.97 35.30 (2.50)
30 Rubber 38.73 32.67 29.10 23.97 31.12 (6.21)
32 Stone, Glass, Clay 48.33 47.97 39.97 29.87 41.53 (8.68)
33 Primary Metals 62.07 59.46 54.87 47.67 56.02 (6.31)
34 Fabricated Metals 40.77 40.30 34.00 28.50 35.89 (5.81)
35 Machinery 33.83 31.80 23.30 17.43 26.59 (7.62)
36 Electrical Equip. 33.73 30.03 25.03 19.43 27.06 (6.21)
37 Transport. Equip. 57.03 55.17 46.80 42.97 50.49 (6.71)
Mean 39.17 37.99 32.47 27.09
(S.D.) (13.24) (13.02) (12.56) (12.22)
Source: Rokkenelenberg and Sockell (1985), Curme, Hirsch, and MacPherson (1990).

2. Industries.
In this section, I explore the same variables presented

above, but on an industry—by-industry basis. This allows for
easier cross-comparisons to identify possible causal effects.
In each case several questions will be consider: Was there
a substantial rise from one period to the next in firm
failures or decertifications? Was there an increase in the
(ihange 1J1 cross—state distribution coupled with IN) or low
increases in the number of establishments? Was there a rise
in new capital expenditures coupled with no or low increases
:hi small establishments

in employment? Was there a rise

182

coupled with no or low decreases in employment?
a. SIC 20, Food Products.

Food products displays a somewhat unusual trend. While
the typical logged disappearance rate appears to be between
approximately five éHKi six percent, time rate jumps
substantially if} Period 2 (x) more trawl 16 percent. This
pattern does not appear to be explained by a rise in firm
failures, which are quite low in Period 2.3 The
decertification rate does rise substantially in Period 2, but
it stays high in Period 3 and Period 4 when the disappearance
rate drops. With tjue exception (ME Period 1” new capital
expenditures and employment rise and fall together, so there
is no strong argument for capital for labor substitution.
Similarlyy small establishment growth anmi employment are
negatively correlated, as I would expected.

However, the fairly sharp rise in the change in cross-
state distribution, coupled with 51 decrease in the
establishment growth for Period 2 does signal that there may
have been substantial plant movement during the period,
perhaps owing to trends in population distribution favoring

the sunbelt.

 

3 Keep in mind, all independent variables are lagged one

period.

183

Table 6.12. SIC 20. Food Products.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0448 .1641 .0583 .0674
Firm Failures/10,000 firms 32.33 29.00 70.00 122.33
Cross-State Distribution (% change) 3.00 5.35 4.06 3.18
Decertifications/10,000 units 64.81 106.86 114.22 132.70
Small Firm Growth (ann. % change) 0.90 -1.18 0.72 -0.25
New Capital Expenditures (ann. % change) 4.34 1.26 -6.25 3.02
Establishment Growth (ann. % change) -1.46 -4.69 1.24 -0.94
union Penetration (% workers) 44.23 43.10 37.00 32.07
Capital Intensity (% assets) 34.00 34.67 35.33 34.33
Employment Growth (ann. % change) -0.04 0.40 -1.76 0.52

 

Sources: Key files of the Department of Labor's private sector collective bargaining
agreements archive and the Eggister of Reporting Labor Organizations, 1980, 1983,
1986; Dun and Bradstreet, Firm Failure Record 1974-1986; Bureau of the Census,
County Business Patterns, 1974-1986; Bureau of the Census, Annual Survey of
Manufactures, 1974-1986; National Labor Relations Board, Annual Reports, 1974-1986;
Bureau of Labor Statistics, EmploymentLgfiours andggarniggs, 1994; Bureau of the
Census, Qggrterly Financial Reports, 1974-1986; Rokkenelenberg and Sockell (1985),
Curme, Hirsch, and MacPherson (1990).

b. SIC 22, Textiles.

The rate (ME unit disappearances :hi textiles :hs high.
Between 1983 and 1986, nearly one—in-five units were recorded
as having disappeared. Examining the variables for Period 3,
we ck) notice 61 large increase in time finn failure rate.
Further, there appears to be a fairly large shift in
establishments across states combined with nearly flat
establishment growth-—a signal that the industry experienced
a geographical shift. Neither capital substitutions nor a
heavier reliance (x1 small establishments appear associated
with this period. If capital for labor substitution did
occur, they took place in Period 1, much earlier than our
hypotheses would predict. However, the Period 4 escalation

in decertification ratee.may' be consistent. with: a policy

184

explanation.

Table 6.13. SIC 22. Textiles.

 

 

Period 1 Period 2 Period 3 Period 4
Disappearance Rate .0698 .0750 .1719 .1490
Firm Failures/10,000 firms 71.33 65.33 104.33 118.66
Cross-State Distribution (% change) 7.29 4.38 6.26 3.61
Decertifications/10,000 units 43.24 75.25 64.05 126.72
Small Firm Growth (ann. % change) 0.69 -0.28 2.39 -0.36
New Capital Expenditures (ann. % change) 2.43 -3.01 -4.20 -1.45
Establishment Growth (ann. % change) -1.18 -1.85 0.32 -1.10
Union Penetration (% workers) 14.87 14.60 12.80 10.20
Capital Intensity (% assets) 31.67 33.00 33.33 32.67
Employment Growth (ann. % change) -1.77 -2.38 -4.55 -1.58

 

Sources: see table 6.12.

c. SIC 24, Lumber.

Lumber (Table 6.14) CUsplays eni enormous increase in
disappearance rate ill Period EL explained, ii: appears, by
increases in both the firm failure and decertification rates.
Otherwise there appear to be no trends relating to capital
substitution, or a reliance on small firms. In Period 2 and
4 there do appear to be some changes in interstate
distribution that do not correlate with an overall increase
in firms. The rise the decertification rate and changes in

cross—state distribution in Period 4 could be consistent with

185

the policy hypotheses.

Table 6.14. SIC 24. Lumber.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0238 .0976 .1903 .0455
Firm Failures/10,000 firms 33.00 29.33 94.67 133.00
Cross-State Distribution (% change) 3.99 4.83 3.80 3.90
Decertifications/10,000 units 47.08 87.61 106.50 124.17
Small Firm Growth (ann. % change) 0.19 -0.22 0.41 -0.22
New Capital Expenditures (ann. % change)-1.55 -0.57 -21.02 9.15
Establishment Growth (ann. % change) 1.06 -2.67 2.54 -0.58
Union Penetration (% workers) 26.87 24.67 19.87 14.97
Capital Intensity (% assets) NA NA NA NA

Employment Growth (ann. % change) 1.15 -1.87 —1.31 3.01

 

Source: see Table 6.12.

d. SIC 25, FUrniture.

Furniture, ems well, records ix}; highest disappearance
rate in Period 3 (Table 6.15). Again, firm failures and
decertification in Period 3 appear to have the closest
association with this trend. In Period 2, both changes in
cross-state distribution and in possible capital for labor
substitutions were quite substantial, However, the sample
revealed no disappearances for that period. In this

industry, the trends in later periods are clearly not

186

consistent with the hypotheses advanced.

Table 6.15. SIC 25. Furniture.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0323 .0000 .0684 .0000
Firm Failures/10,000 firms 82.67 77.33 184.67 156.00
Cross-State Distribution (% change) 7.72 6.26 5.55 6.92
Decertifications/10,000 units 51.94 93.58 167.15 133.89
Small Firm Growth (ann. % change) 0.67 -0.55 1.05 -0.32
New Capital Expenditures (ann. % change)-0.25 4.33 -4.87 5.84
Establishment Growth (ann. % change) 1.72 -0.89 4.23 2.77
Union Penetration (% workers) 26.60 27.67 21.73 16.27
Capital Intensity (% assets) NA NA NA NA

Employment Growth (ann. % change) -1.09 -.37 -1.64 3.85

 

Source: see Table 6.12.

e. SIC 26, Paper.

Paper displays fairly stable patterns (Table 6.16). The
disappearance rate remains at about three to five percent
throughout the period. Although firm failures increase in
Period 3 and Period 4, the decertication rate remains about
the sane. There was ea trend consistent with capital for
labor substitution in Period 1, but not since. It is
noteworthy that changes in cross-state distribution. have
remained relatively robust, even though there has been little
(or rm) overall establishment growth. since time late—19708,

perhaps signalling relocation.

187

Table 6.16. SIC 26. Paper.

 

Period 1 Period 2 Period 3Period 4

 

Disappearance Rate .0268 .0414 .0303 .0538
Firm Failures/10,000 firms 25.67 28.33 58.33 90.33
Cross-State Distribution (% change) 6.93 5.72 4.09 5.67
Decertifications/10,000 units 21.09 31.77 25.95 35.02
Small Firm Growth (ann. % change) 0.96 —0.42 1.49 -0.77
New Capital Expenditures (ann. % change) 6.19 5.02 —6.85 3.79
Establishment Growth (ann. % change) 1.57 -0.20 0.70 0.28
Union Penetration (% workers) 55.83 55.20 51.63 46.43
Capital Intensity (% assets) 49.33 50.00 56.00 61.00
Employment Growth (ann. % change) -1.31 .31 —1.81 1.11

 

Source: see Table 6.12.

f. SIC 28, Chemicals.

The chemical industry (Table 6.17) displays an increase
in the disappearance rate nearly consistent with increases in
firm failures and decertifications. There has also been very
little small establishment growth. Possihflte capital for
labor substitutions have occurred, but mainly in Period 1.
The large number of decertification elections in Period 4,
along with possible relocation and capital for labor

substitution activity provide moderate support for the

188

hypotheses.

Table 6.17. SIC 28. Chemicals.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0072 .0365 .0562 .0898
Firm Failures/10,000 firms 33.67 32.00 68.33 113.67
Cross-State Distribution (% change) 3.83 4.83 5.28 4.00
Decertifications/10,000 units 72.20 113.39 158.80 241.42
Small Firm Growth (ann. % change) 0.06 -0.26 0.52 0.10
New Capital Expenditures (ann % change) 9.15 -7.26 -12.36 0.58
Establishment Growth (ann. % change) 1.08 -.73 2.25 0.13
Union Penetration (% workers) 31.00 30.40 23.57 16.47
Capital Intensity (% assets) 39.33 39.33 41.33 37.67
Employment Growth (ann. % change) -0.26 0.53 -2.56 -0.61

 

Source: .0. Tale 5 . 12 .

g. SIC 29, Petroleum.

In petroleum (Table 6.18), the rise in disappearances in
Period 3 seems most clearly foreshadowed by capital
expenditures at the expense of labor. For both Period 2 and
Period 3 new capital expenditures grew along with a rise in
capital intensity, while employment fell. The effects of the
mid—19808 oil glut are also seen. Prior to the mid—19808,
Inu1 and. Bradstreet did run: even report firnl failures in
Petroleum, and yet between 1983 and 1986 the failure rate for
the industry was one of the highest in manufacturing. In the

final period employment continued to fall, and, for the first

189

time,

Table 6.18. SIC 29. Petroleum.

new capital expenditures began to decline as well.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0000 .0000 .0964 .0660
Firm Failures/10,000 firms NA NA NA 187.00
Cross-State Distribution (% change) 5.67 8.56 7.69 6.57
Decertifications/10,000 units 52.16 93.23 102.28 111.58
Small Firm Growth (ann. % change) 0.29 -0.54 0.75 0.59
New Capital Expenditures (ann. % change) 1.03 6.89 5.27 -19.83
Establishment Growth (ann. % change) 0.59 1.20 1.41 1.39
Union Penetration (% workers) 34.53 38.83 34.87 32.97
Capital Intensity (% assets) 46.67 48.67 55.33 57.33
Employment Growth (ann. % change) 1.39 -1.56 -1.55 -3.53

 

Source: see Table 6.12.

h. SIC 30, Rubber

Rubber displays a large rise in the disappearance rate in

Period 4 (Table 6.19). Firm failure is associated with this

increase, but so too is a fairly large increase in the change
in the cross—state distribution of establishments and a large
increase ill new capital expenditures relative tx> gains in
employment. These observations may reflect changes that were
occurring in the tire industry, a former union stronghold,

which was particularly hard hit by plant closings and

relocations.

190

Table 6.19. SIC 30. Rubber.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0682 .0500 .0544 .1100
Firm Failures/10,000 firms 40.00 41.00 82.00 119.33
Cross-State Distribution (% change) 6.30 5.99 4.63 6.17
Decertifications/10,000 units 82.27 135.38 108.01 119.72
Small Firm Growth (ann. % change) 0.76 -0.34 1.20 —0.65
New Capital Expenditures (ann. % change)-0.94 -0.27 —10.80 18.73
Establishment Growth (ann. % change) 4.11 3.52 3.35 2.46
Union Penetration (% workers) 38.73 32.67 29.10 23.97
Capital Intensity (% assets) 33.33 33.67 35.67 37.00
Employment Growth (ann. % change) 1.54 0.28 -0.44 3.83

 

Source: see Table 6.12.

i. SIC 32, Stone, Glass, and Clay.

Like the rubber industry, Stone, Glass, and Clay shows a
increase in disappearance rate in the final period, although
the trend is steadier throughout the decade (Table 6.20).
Both firm failures and decertifications rise at about the
same rate as disappearances. In Period 2, there is evidence
that capital investments came at the expense of labor, and in
Period 4 cross—state changes are not justified by an increase

in the number of establishments, a trend consistent with

191

hypotheses.

Table 6.20. SIC 32. Stone, Glass and Clay.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0353 .0488 .0939 .1301
Firm Failures/10,000 firms 26.00 25.00 70.33 96.67
Cross-State Distribution (% change) 7.13 3.84 4.74 4.83
Decertifications/10,000 units 27.08 46.99 73.96 111.05
Small Firm Growth (ann. % change) 0.50 -.47 .72 -0.17
New Capital Expenditures (ann. % change)~0.87 7.33 ~22.13 8.60
Establishment Growth (ann. % change) 1.56 -1.99 1.10 -0.33
Union Penetration (% workers) 48.33 47.97 39.97 29.87
Capital Intensity (% assets) 45.33 46.67 48.33 45.33
Employment Growth (ann. % change) -1.89 -1.12 -5.28 1.23

 

Source: see Table 6.12.

3'. SIC 33, Primary Metals.

A rise in the firm failure rate and changes in location
during Period 3 and Period 4 appear to be closely related
with the extreme rise in disappearances (Table 6.21). Period
4 displays the most precipitous increase in failures and also
a stark divergence between geographic location and
establishment growth. Although there is also an increase in
the decertification rate entering the final period, in
comparison to other industries the rate is actually quite
low. There has been relatively high small firm growth,
coupled vnifll decreases ij1 employment. Capital bar labor
substitutions are not apparent at the industry—level, since

the rate of new capital expenditures decreases in every

192
period.

Table 6.21. SIC 33. Primary Metals.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0125 .0506 .0565 .1348
Firm Failures/10,000 firms 33.00 34.33 120.67 165.67
Cross-State Distribution (% change) 4.94 4.82 5.32 5.60
Decertifications/10,000 units 26.44 36.80 58.73 75.27
Small Firm Growth (ann. % change) 1.13 -1.03 3.02 -0.80
New Capital Expenditures (ann. % change) -1.22 -2.84 -12.41 -11.11
Establishment Growth (ann. % change) 0.86 -0.57 0.38 -1.64
Union Penetration (% workers) 62.07 59.67 54.87 47.67
Capital Intensity (% assets) 46.67 45.00 48.33 45.33
Employment Growth (ann. % change) —3.31 -1.39 -10.60 -2.85

 

Source: see Table 6.12.

k. SIC 34, Fabricated.Metals.

The steady increase in<disappearance in Fabricated Metals
reflects closely an increase in firm failure rates, while the
decertification rate actually declines in the final period
(Table 6.22). There may also have been some relocation
activity 11] Period 2 enui Period.I3, although tine change 1J1
cross—state distribution for this industry is relatively low.
The only period in which capital expenditures were correlated

negatively with employment was in Period 1.

193

Table 6.22. SIC 34. Fabricated Metals.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0204 .0417 .0807 .0932
Firm Failures/10,000 firms 32.00 24.67 76.33 132.00
Cross-State Distribution (% change) 5.61 3.83 4.88 3.43
Decertifications/10,000 units 45.28 66.56 111.03 103.75
Small Firm Growth (ann. % change) 1.34 -0.42 1.11 -0.25
New Capital Expenditures (ann. % change) 2.58 3.93 -15.18 14.33
Establishment Growth (ann. % change) 2.62 0.14 2.96 0.92
Union Penetration (% workers) 40.77 40.30 34.00 28.50
Capital Intensity (% assets) 27.67 28.67 31.33 30.00
Employment Growth (ann. % change) -1.20 0.09 -5.85 2.02

 

Source: see Table 6.12.

l. SIC 35, Machinery.

The disappearance rate in Machinery seems particularly
affected by a significant rise in the decertification rate
(Table 6.23). In fact, decertification elections in
Machineny are surpassed cmly in! decertification levels in
Chemicals. Firm failures also rise steadily throughout the
period, and generally reflect the pattern of disappearances.

Again capital for labor substitution appear a possibility in

194

Period 1 and relocation in Period 4.

Table 6.23. SIC 35. Machinery.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0370 .0604 .0868 .1058
Firm Failures/10,000 firms 31.67 27.33 98.00 111.33
Cross-State Distribution (% change) 5.48 4.79 4.78 3.47
Decertifications/10,000 units 65.72 81.29 162.31 217.27
Small Firm Growth (ann. % change) 0.38 -0.50 .79 -0.14
New Capital Expenditures (ann. % change) 2.92 10.18 -11.62 -0.36
Establishment Growth (ann. % change) 2.90 1.32 3.28 -0.50
Union Penetration (% workers) 33.83 31.80 23.30 17.43
Capital Intensity (% assets) 27.33 29.00 30.67 27.33
Employment Growth (ann. % change) -1.14 4.15 -9.29 1.01

 

Source: see Table 6.12.

m. SIC 36, Electrical Eguipment.
In Electrical Equipment, Period 2 may have been the
greatest period. for infirx1 escape ibehavior. During' that
period decertifications jumped sharply and a high change in

cross—state distribution vex; not ndrrored kn! establishment

growth. Further, the firm failure rate was low during that
period. For Period 3 and Period 4, the disappearance rate
remained at about five percent. Also, establishment growth

began to increase, while the decertification rate fell in

 

Ta;

Pi}
Cr<
De<

Ne:
Bet
Uni
Ca;

Sou

"(:1
D)

 

195

Period 4.

Table 6.24. SIC 36. Electrical Equipment.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0379 .0551 .0526 .0579
Firm Failures/10,000 firms 73.00 45.67 102.00 148.67
Cross-State Distribution (% change) 7.84 7.41 7.89 5.13
Decertifications/10,000 units 64.91 104.82 123.16 111.57
Small Firm Growth (ann. % change) 1.10 -1.59 1.88 -0.02
New Capital Expenditures (ann. % change) NA NA NA NA

Establishment Growth (ann. % change) 3.29 0.47 5.37 2.62
Union Penetration (% workers) 33.73 30.03 25.03 19.43
Capital Intensity (% of assets) 23.67 23.67 26.33 26.00
Employment Growth (ann. % change) -1.60 2.63 -2.74 0.46

 

Source: see Table 6.12.

n. SIC 37, Transportation Equipment.

Firm failures grew in Period 3 and Period 4 in
Transportation Equipment (Table 6.25), a trend at least
partially mirrored in the rise in disappearances. There is
a sharp rise in decertifications from Period 1 to Period 2,
but the rate of <decertifications is quite lOW’ for this
industry. A fairly sharp increase is also seen in the change
in cross—state distribution for time industry 1J1 Period 4.
And although the overall number of establishments also
increases for the period, the rate of change is quite small
relative to the change in interstate distribution. If
capital for labor substitution did occur, again it appears to

have occurred in Period 1.

196

Table 6.25. SIC 37. Transportation Equipment.

 

Period 1 Period 2 Period 3 Period 4

 

Disappearance Rate .0328 .0847 .1060 .0953
Firm Failures/10,000 firms 78.33 69.00 138.67 179.67
Cross-State Distribution (% change) 5.10 5.29 5.05 8.41
Decertifications/10,000 units 21.65 47.98 51.23 58.15
Small Firm Growth (ann. % change) -0.04 -0.51 1.47 -0.51
New Capital Expenditures (ann. % change) 10.50 8.40 -15.67 27.95
Establishment Growth (ann. % change) 2.82 —2.91 2.94 1.25
Union Penetration (% workers) 57.03 55.17 46.80 42.97
Capital Intensity (% of assets) 25.67 26.67 29.33 26.67
Employment Growth (ann. % change) 0.80 -1.09 -3.67 5.15

 

Source: see Table 6.12.

3. Summary.

In preceding discussion of variable trends across
industries, several patterns emerge. First, firm failure and
decertification elections tend txn vary together, and tmmh
track the disappearance rate fairly closely. Second, capital
for labor substitutions seemed most likely to have occurred
the mid-19705 (Period 1), long before the policy reversals of
a decade later. Third, while the average change in cross-
state distribution remained largely the same in all periods,
in Period 4L establishment growth stalled txb near zero,
suggesting timu: plant relocation rmn/ have increased,
consistent with hypotheses. The trends in small firm growth
show no definite patterns consistent with an outsourcing
explanation. Instead, with a few exceptions, the relative
number of small firms rises as employment declines.

In time next chapter, results (M3 the correlation and

197

regression analysis are presented.

Chapter VII. Results.

A. Introduction.

In this chapter I discuss the results of the statistical
analyses described in the previous chapter. First, I present
the trend in the dependent variable, with a brief discussion
of what it may and may not reveal. Second, I focus on the
simple correlations among the dependent and independent
variables, and discuss the possible significance of patterns
that emerge. linixL I present regression results from.several
models both with and without time period effects and cross—
products. Fourth, II consider several alternative models,
including one without lagged variables and one with a
combination of lagged and non—lagged variables. I conclude
with 51 general discussion CHE findings, shortcomings, and
reflections on the hypotheses.

Before presenting results, it is helpful to recall the

hypotheses generated in Chapter V.

198

Table

7.1. Hypotheses.

 

31.

H2.

H3.

H4.

35.

36.

H7.

89.

310.

311.

812.

813.

314.

The disappearance rate of local collective bargaining relationships across
manufacturing industries is significantly and positively related to the number
of firm failures per 10,000 firms.

The disappearance rate of local collective bargaining relationships across
manufacturing industries is significantly and positively related to the
percentage point change in the distribution of establishments across states
by industry, controlling for new establishment growth and capital intensity.

The disappearance rate of local collective bargaining relationships across
manufacturing industries is significantly and positively related to the rate
of change in the percentage of establishments with fewer than 100 employees,
controlling for establishment growth by industry.

The disappearance rate of local collective bargaining relationships across
manufacturing industries is significantly and positively related to the rate
of change in new capital investment, controlling for employment growth by
industry.

The disappearance rate of local collective bargaining relationships across
manufacturing industries is significantly and positively related to the number
of decertification elections per 10,000 bargaining units.

There will be a positive and significant effect of the time period on
disappearances recorded after 1980, controlling for the variables mentioned
above.

There will be a positive and significant effect of the time period on
disappearances recorded after 1983, controlling for the variables mentioned
above.

There will be a positive and significant effect of the time period on
disappearances recorded after 1986, controlling for the variables mentioned
above.

The variable firm failure will lose significance with the introduction of each
period variable into the model.

All the other variables will retain their significance as time period variables
are added.

The disappearance rate of local collective bargaining relationships across
manufacturing is positively and significantly related to the cross-product of
each time period variable and the variable representing establishment
relocation, controlling for establishment growth and capital intensity.

The disappearance rate of local collective bargaining relationships across
manufacturing is positively and significantly related to the cross-products
of each time period variable and the variable relating to the increase in
establishments with 100 or fewer employees, controlling for employment growth.

The disappearance rate of local collective bargaining relationships across
manufacturing is positively and significantly related to the cross-product of
each time period variable and the variable relating to the increase in new
capital investment, controlling for employment growth.

The disappearance rate of local collective bargaining relationships across
manufacturing is positively and significantly related to the cross-product of
each time period variable and the variable relating to decertification
elections.

200

Table 7.1, continued.

815. The disappearance rate of local collective bargaining relationships across
manufacturing is negatively and significantly related to the cross-product of
each time period variable and the variable relating to firm failures.

 

B. Trend in local bargaining relationship disappearance.

As mentioned in the previous chapter, the panel under
observation contains 1411 local bargaining relationships known
to exist in 1977. By the 1980, 42 or approximately three
percent of those relationships had disappeared, leaving 1369.
Of those 1369, 85 or roughly six percent disappeared by 1983.
By 1986, 91 of the remaining 1284, or seven percent had
disappeared, leaving 1193. During the final period, 1986 to
1990, 161 of 1193 relationships, or 13.5 percent disappeared.

In unadjusted figures, the trend in disappearances
appears quite compelling. The disappear rate in the final
period is nearly 400 percent greater than in the first period.
Unfortunately, the Register or Reportinquabor Organizations
was published at irregular intervals during the 1980s.
Therefore, a better model is one that weights the time periods
to reflect their slightly differing durations--particularly
of the final period which was more than one year longer than
the earlier periods. VHtllthose adjustments the disappearance
:rates for the four periods are, respectively: 3.2 percent, 5.8
percent, 8.5 percent, and 8.5 percent. In making this
adjustment, an assumption was made that the disappearances

vwere evenly distributed.within each the period” If relatively

201
more (of tine 161 disappearances ii1 Period.‘4 luui actually
occurred early in that period, then the figure of 8.5 percent
is conservative.

The substantial jump in the disappearance rate from the
first to the second period would seem to belie a policy
argument since ii: occurred during time late-1970s under a
relatively friendly federal administration. zuKithe continued
increase 1J1 disappearance IINXB into Period I3 ney' simply
reflect the impact of the early-19803 recession. However, in
the final period, which would almost certainly exclude
disappearances recorded prior to 1983, the disappearance rate
remains high, even though a period of economic expansion began
in 1983. A counter argument could be made that most
industries ‘within Inanufacturing’ did. not ‘recover' front the
recession, as economic expansion occurred primarily in the
service-producing sector.

A look at employment patterns, a coincident indicator of
economic activity (Sommers, 1993), supports this latter
argument. As Table 6.9 in the previous chapter indicated,
employment declined fairly consistently across manufacturing.
Table 7.2, which displays absolute levels of production.worker
employment by industry, shows that employment increased in 11
of the 14 industries between 1975 and 1980, but decreased in
10 of 14 by the end of the 19808. Only food, lumber,

furniture, and paper showed employment increases during the

202
decade, and paper did not surpass its 1975 level. Overall
employment intflu314 manufacturing industries declined.sharply
between 1980 EUKi 1990, with a tcmal lrxxscof more than one

million jobs.

Table 7.2. Production worker employment, selected manufacturing industries, 1975,
1980, 1985, 1990, in thousands.

 

 

 

SIC/Industry 1975 1980 1985 1990
20 Food 1120 1175 1181 1196
22 Textiles 752 737 607 593
24 Lumber 517 578 597 610
25 Furniture 337 374 394 419
26 Paper 544 523 512 524
28 Chemicals 580 626 578 605
29 Petroleum 123 125 109 103
30 Rubber 465 559 607 687
32 Stone, Glass, Clay 499 513 450 433
33 Primary Metals 887 877 611 573
34 Fab. Metals 1090 1195 1084 1097
35 Machinery 1339 1602 1307 1259
36 Elec. Equip. 1091 1328 1300 1055
37 Trans. Equip. 1154 1233 1257 1218
Total 10498 11445 10594 10372
Source: Bureau of Labor Statistics, Employment, Hours, and Earnings, 1994

 

Hence, the employment evidence suggests that the rise in unit
disappearance rate throughout the period may simply have been
the product of the precipitous drop in manufacturing activity
during the early 1980s recession, coupled with the inability
of most industries in the sector to recover.

Although the correlation between economic vitality and
unit disappearance seems fairly clear overall, in the present
analysis I attempt to probe more deeply into the issue by

(disaggregating both industry and time effects. A trend toward

203

a weaker association between economic distress and unit
disappearance across industries could signal a reaction to
policy changes as strategic behaviors in fairly healthy
industries increase relative to behaviors motivated by more
purely economic factors in depressed industries.

C. Zero-order correlations.

The first step in answering the more complex questions is
an examination of the simple correlations among variables.
Focusing first on the association between the dependent
variable and the variables of substantive theoretical
interest--i.e., those contained in Equation 1--a positive and
significant relationship between firm failure rate and
disappearance rate is found. Coupled with the observations
concerning employment made above, this result further
validates the argument that disappearances were caused by
economic hardship within industries.

But there irs also ea significant correlation. between
decertification rate and disappearance rate. This result may
signal ea response tx> more lenient standards cu: employer
conduct and stricter standards of union security. However,
the rate had been climbing fairly steadily since the early
19708, before most of these policy reversals were made;
therefore, a more careful analysis is needed before further
conclusions can be drawn.

None of the other variables in Equation 1 are

204
significant—-the (outsourcing gnrnqn change ii1 cross—state
distribution, or new capital investment. In fact, all are
signed in a direction contrary to the expectations stated in
the hypotheses, with the coefficients for new capital
investment and change in (cross—state distribution nearly

achieving significance.

205

 

 

Ho.v.m¢

mo.v.fiss
Ho.v.Qese
escom.ssemm. seam.| NH.I seam. ma.l mo. seam. ssam.nesemv. seeow. eehN. mmmd KNBhd .ma
secmm. eemm.l smN. eemN.I eH.I ma. mH.I sehN. «seam. eeeoh. sscmv. mama MNBhd .NH
NN.I mH.| bH.| mH.I OH. mo.| smN.Issemv. Issue. eesmv. coma mmahd .HH
va.l mo. mo.eeshm. ¢H.I mo. eeeww.nessmm.l eemm.n ”94¢ ZOHZD .OH
mH.| sewn. om.| mo.n esmv. mo. om. MN.I ERROKO .Bmm .m
om. e¢N.|eseHh. eeehv.l ma.) #0.: NN.I .BmflbZH AdBHmdo .m
mo. «0. ho. mH.I ma. mH.| .BMHQ NB‘BmImmOMU .b
HN.I mo. avN.u mo. pH.I MmZNBZH AGBHmdu .0
ecemm.n mo. No.1 ¢H.I mhkomO BZNSHOAQSN .m
hH.I mo. mo.n ”Gadmo IMHN AddZm .v
eeeNv. econ. MB‘N .BMNUNQ .m
seemm. mmmDaHdfl ZMHN .N
”Edd mdmmm‘mHQ .H

NH Ha OH 0 o h m m C n N H

 

.sso«usaeuuoo unencuouou .n.b sands

206

Of the control variables, only the union penetration rate
shows a significant correlation with the disappearance rate.
The negative correlation indicates that, as expected,
bargaining relationships in more highly organized industries
were relatively more likely to have survived the decade than
those in less organized industries.

The trend in the dependent variable across time periods
holds t1) our expectations. There aura strong correlations
between the disappearance rates logged in the periods
following 1980 and 1983, and a somewhat weaker though still
significant correlation for the period following 1986. The
weaker correlation in the latter period reflects the
levelling—off of recorded disappearance rates following 1983.
None of the other correlations between independent variables
and the dependent variable are significant at the .10 level.

There are, however, some interesting intercorrelations
among independent variables. For example, the association
between firm failures and decertification elections indicates
that economic distress may have made it more difficult for
unions txa:maintain representation rights. Bankruptcy may,
indeed, have been the slippery slope to deunionization that
Lawler (1990) predicted.

Also noteworthy are the significant correlations between

establishment growth and the relative frequency of small

207

firms, and between establishment growth and cross—state
change. This pattern suggests that new establishments tended
to be smaller and to have arisen in industries that displayed
greater change in geographic distribution. A likely
explanation is that older, larger establishments were being
replaced 13! smaller establishments en: greenfiebd sites ljl
other regions.

The strong correlation between new capital investment and
employment growth may confirnlthe earlier speculation that new
technology-~at the industry level——doe8 not lead.t0iwidespread
labor substitution. However, it must be remembered that the
variable, as measured here, represents new spending on all
forms of capital not just new technology.

Table 7.3 also shows a strong association between union
penetration rates and several of the independent variables.
For example, unionization is negatively associated with firm
failures and decertification elections. However, these
associations rmn/ be spurious :U1 light tn? the correlation
between unionizationenxicapital intensity. Capital intensive
firms typically experience less market competitiveness and.are
less affected by labor costs, which represent a relatively
lower proportion of total costs.

Timzpositive, though non-significant, association between
establishment growth and firm failure rate should also be

noted. Common sense would suggest that the two should be

208

negatively associated. However, time positive correlation
indicates that establishment foundings occurred. more
frequently in industries experiencing simultaneous economic
difficulty. Perhaps firm failures in one region were being
offset by (smaller) organization births in other regions.

There aura also 51 number (if significant correlations
between time periods and the independent variables. II am
particularly concerned about the strong correlations between
firm failure and the final two time period variables, which
may make it difficult to distinguished between economic and
other period effects ii1 the regression nedel. Otherwise,
noteworthy patterns occur on several variables that display
curvilinear trends as time passes. Firm failure rate,
decertification.rate, small firnxchange, establishment growth,
and the chxn) in union penetratmmn rate all correlate most
strongly with the period following 1983. Employment growth
and capital investment, on tine other hand, correlate nest
strongly with the final period.
.D. Regression models.

1. constant time model.

The first regression model contains only the continuous

\nariables and.predicts unit disappearances for the entire time

period.

 

 

7.4. Standardized regression coefficients, lagged model, t-values in
parentheses, N:56.
Model 1
Establishment Growth —.4360**
(-2.25)
Firm Failure .3415**
(2.55)
Capital Intensity -.2857*
(—1.73)
New Capital Investment —.3284*
(—1.45)
Decertification Rate .0739
(0.35)
Union Penetration Rate -.0323
(-0.16)
Small Firm Growth .0148
(0.07)
Cross-State Change .0113
(0.07)
Employment Growth .0009
(0.04)
R’ .36**
Adj R2 .19
F value 2.12
D.W. 2.10

*** p.< .01
** p.< .05
* p.< .10

 

210

The first regression model provides support only for
hypothesis H1. Indeed, firm failure rates appear to be a
significant predictor of local bargaining unit disappearance.
The decertification rate, vflfirfl1lrmi a significant positive
correlation with disappearance rate loses its significance
due, I suspect, to its correlation with capital intensity.
Cross-state change in the relative distribution of firms, for
which there was a strong expectation of significance, is not
a significant predictor of disappearance in this model. The
proxy for outsourcing, small establishment growth, also shows
no significance in the model, while the variable intended to
approximate capital for labor substitution, new capital
spending, is significant, but signed in a direction contrary
to the hypothesis.

Two of the control variables perform quite well in the
model. Establishment growth and capital intensity are both
fairly strong predictors of disappearance rate. Neither were
significant as zero-order correlations, but apparently

represent important attributes when partial effects are

observed. Union penetration rate loses the significance it
displayed as a simple correlation. Like decertification
rates, I suspect that the intercorrelation of union

penetration and capital intensity causes the former to be
overwhelmed when partial effects are observed.

In sum, the results of Model 1 (Table 7.4) indicate that

211

the<disappearance rate‘was highest in economically distressed,
labor intensive, shrinking industries—~hardly surprising, and
consistent mdth.em1 economic theory (ME decline. The were
interesting question, which will be considered next, is what
happens to these variables when the larger time period is
divided in several different ways to model effects before and
after critical years.

2. Time-varying model.

In tine follcmdrm; model, chumm/ variables representing

several time junctures are introduced.

Table 7.5.

212

t-values in parentheses, N:56

Standardized regression coefficients with period effects, lagged model,

 

 

Model 2 Model 3 Model 4
Establishment Growth -.3902** -.5877*** -.4331**
(-2.03) (-3.61) (-2.22)
Firm Failures .2364 -.3939* .1672
(1.39) (-1.81) (.64)
Capital Intensity -.3097*** -.5576*** —.3468**
(~1.91) (-3.73) (-1.92)
New Capital Investment -.2836 -.3620* -.4032
(—1.27) (-1.96) (—1.65)
Decertification Rate -.0435 -.1641 .0491
(—.20) (—.90) (.23)
Small Firm Growth .1449 -.1647 .0572
(.50) (-.87) (.25)
Union Penetration Rate -.0686 -.0077 -.0147
{-.34) (-.05) (-.07)
Cross-State Change .0180 .2999‘Ir .0766
(.11) (1.97) (.42)
Employment Growth .0534 .0390 -.0191
(.21) ( .19) (-.08)
After 1980 .2954
(1.59)
After 1983 1.0745***
(4.23)
After 1986 .2477
(.84)
R’ .40** .58*** .37*
Adj R2 .22 .46 .18
F value 2.25 4.65 1.96
1.99 1.94 2.11

*** p.< .01
.* p.‘ .05
* p.< .10

 

213

The inclusioncoftjmmaperiod variables demonstrates clear
support for hypothesis H7 but not hypotheses H6 or H8. That
is, allowing the intercept to vary for the post-1983 does have
a significant effect on model fit, but the same cannot be said
for the post—1980 and.post-1986 models. The critical juncture
is obviously 1983. In Model 3 (Table 7.5), the change in the
coefficient of determination (R5 is enormous, as the percent
of variance explained by the model nearly doubles.

It is also interesting to note what happens to several of
the slope variables when the intercept is allowed to vary.
Largely consistent with the intent of hypotheses H9 and H10,
controlling for time period after 1983 forces the industry
firm failure rate to become a significant negative predictor
of unit disappearance——an even more robust result than the
hypothesis suggested. Similarly, while hypothesis H10 stated
that changes i11 cross—state distribution. would remain :3
significant predictor of unit disappearance, this result was
made impossible by the poor performance of this variable in
Model 1. (Table '7w4). But lxnxa cross-state «distribution
becomes a significant predictor, consistent with the intent
of the hypothesis.

Significance of the dummy variable and the changes in
the firm failure and cross—state distribution variables
provide fairly strong support for the overall proposition of

this study. Changes in policy occurring after 1983 may have

214
had an impact on the environment within which labor—
management relations occurred. Although the warnings stated
:hi the previous chapter Hmst 1x2 heeded——particularly math
regard tx> ecological fallacy——trm2 evidence: indicates that
employers, even. absent the» strong* economic ;pressures, as
predicted by the firm failure rate, may have been more likely
to take advantage of the option of "escaping" (Walton,
Cutcher—Gershenfeld,énmlMcKersien 1994) collectiveebargaining
relationships after 1983. Without controlling for the time
juncture of 1983, the disappearance of relationships can
simply be explained by its association with firm failures.

The non—significant result for the post—1986 dummy
variable is contrary to hypothesis H8, but is not surprising
since it was seen that the change in the dependent variable
levels-off after 1983. As well, the dummy variable for the
post—1980 period approaches but does not reach significance.
This result is also not surprising since the post—1980
variable would include the period of relatively low
disappearance 1J1 the early—19808. Again, tjuazmost obvious
change occurs after 1983.

In Table 7.6, the analysis is taken one step further by
examining more closely the shifts in slope coefficients aCross
time periods. However, the only significant cross—product is
one which represents the product of the firm failure rate and

the post-1983 period. In the previous chapter, it was seen

215

that there was large increase in the magnitude of firm
failures after~ 1983, which, no «doubt, accounts for this
result. But while the positive coefficient for the post—1983
cross—product is largely cancelled out by the negative
coefficient (H? the independent effect (M3 the firm failure
variable. If the firm failure variable controlling for time
period would have been positive and significant, and the
cross—product had been positive and significant, an argument
could have been made for the increasing saliency of the
variable over time. But with oppositely signed coefficients
this does not appear to be the case.

Such a strong cancellation does not occur, however, with
the variable for cross—state change which remains significant
as a positive predictor of unit disappearance, and whose
cross—product, while negatively signed, is not significant
(nor even close to significant). In fact, in Model 6 the
inclusion of the cross—products causes the cross—state
distribution variable itself to become even. more highly
significant (Table 7.6).

All. of these coefficients. are somewhat difficult to
interpret because, an; Blalock (1961, g). 125) noted: "The
magnitude of rxy will depend not only in the size of the slope
but also (M1 the change 1J1 X relative t1) that iJ11nuisance
variables such as U, V, and W." In Model 3 (Table 7.5), the

time period dummy variable is clearly significant. However,

216

in Model 6 (Table 7.6), the time period dummy variable is
(just barely) non—significant. This does suggest that the
chumm/ variable, (M: least partially, nesks tine interaction
effects of the time period and (most likely) firm failure
rates, so that some of the impact of time period itself may
be explained by the increases in firm failures for the period.

For the rmxfi: part, establishment growth anxi capital
intensity remained significant. Regardless of enn/ of the
other effects, disappearances were most likely to occur in
labor intensive industries that were reducing the number of

establishments.

217

 

 

TABLE 7.6. Standardized regression coefficients with period effects and cross-
products, lagged model, t—values in parentheses, N:56.
Model 5 Model 6 Model 7
Establishment Growth -.3797** -.6834*** -.5767***
(-1.82) (-3.84) (-2.88)
Firm Failures -.2523 -1.705** .1532
(1.38) (-2.33) (.52)
Capital Intensity -.2988 -.6043*** -.4503**
(-1.67) (-3.82) (-2.36)
New Capital Investment -.2984 -.1377 -.7232**
(-.32) (-.32) (—2.47)
Decertification Rate .1226 -.1811 .1825
( .16) (—.51) ( .66)
Union Penetration Rate -.0379 -.1020 .0810
(-.17) (-.56) (.39)
Small Firm Growth .0560 -.0137 .1122
( .07) (-.05) (.44)
Cross-State Change .1813 .6636** .1712
(.60) (2.22) (.85)
Employment Growth .0484 -.0415 .0523
(.18) ( .19) (.20)
After 1980 .7681
(.92)
After 1983 .9414
(1.54)
After 1986 1.4231
(1.31)
Firm Failure x Period Effect NA 2.0446* -1.1545
(1.76) (~1.21)
Cross State Change x Period Effect -.4441 -.5925 .3058
(-.69) (-.76) (.43)
Small Firm Growth x Period Effect .0566 -.1919 -.0180
(.07) (-.69) (-.08)
New Capital Investment X Period Effect .0435 -.2297 .2602
( .05) (-.56) (1.10)
Decertification x Period Effect -.2038 -.0766 -.4131
(-.21) (-.15) (-.78)
R’ .42 .64*** .50*
Adj R' .13 .45 .23
F value 1.47 3.34 1.88
D.M. 2.06 1.93 2.20

***p.<.001, **p.<.05. *p.<.10

 

To appreciate better the story being told by the
regression results, it is helpful to examine changes in R2
particularly adjusted R3 across the models. Rzindicates the
amount of variance explained by a regression model and can
always be inflated by adding more variables to the model.
However, an inefficient inflation of R3 comes at the expense
of significance, since degrees of freedom are sacrificed for
little improvement in model fit. But when additional
variables add very little to the amount of variance explained
by a model, adjusted R3<declines.

In examining adjusted RE mezsee that its value is .19 in
Model 1, increases slightly to .22 in Model 2 and
substantially t1) .46 ill Model 3, 1mm: drops irl Model 4”
Therefore allowing the model’s intercept to vary across time
periods is somewhat useful when considering differences before
and after 1980; extremely useful when examining the periods
before and after 1983; but not useful in distinguishing pre-
and post-1986. Save for emu! significant differences that
might occur in slope coefficients, this result indicates that
the original model is as good a predictor of variance in the
post—1980 and post-1986 periods as the time—varying models.
In Model 5, which includes both the dummy variable and cross—
products for the period following 1980, adjusted Rzldrops to

.13--a nine point decrease from the original Model 2, which

219

contained only the time period dummy for post—1980.
Unfortunately, not much can be made of this result since the
cross—product involving time period effects and firm failures
could not be included in the model due to high collinearity-
—virtually all firm failures occur in the later period” ‘Model
6, which contains the only significant cross—product
coefficient, also displays a drop in Adjusted R3, suggesting
that the inclusion <3f time cross-products does little to
improve variance explanation over the model with only the
dummy variable. However, the adjusted R? in Model 7 reveals
a greater relative change in the model caused by the
introduction of the cross—products. So while the intercept
for model did not change much, the slopes may have.
Unfortunately, the change in Rh while approaching
significance, is not statistically significant at an
acceptable level.

Nonetheless this latter result does call our attention to
these cross—products. Noticeable are (fine near—significant
negative effects of firm failure on disappearance rate during
this period and the near significant positive effects of new
capital expenditures. Further, since there is Ix) strong
positive significant coefficient for firm failure in MOdel 7
(Table 7.6), there is not the same cancelling that occurred
in Model 6. Significant results would have been consistent

with hypotheses H13 and H15: that disappearances in the late

220

19808 became less a function of firm failures and more a
function of other employer behaviors. This result would also
allow a stronger conclusion than relying on the Here non-
significance of the firm failure variable itself when
controlling for the time period.
E. Alternative models.

1. Non-lagged model.

The preceding statistical analysis was conducted assuming
a one period lag in the independent variables. Hence, for
example, firm failures between 1980 and 1983 were associated
with disappearances occurring between 1983 and 1986. 'Hue
rationale for the lag was that some period of time would
elapse between a disappearance occurring and the filing of a
terminal report with the Department of Irdxmu To examine
whether a lagged model is indeed more appropriate, I present
a couple of alternative models below. First, I CUSplay a
model with no lag, and then a model with several lagged and
several non—lagged variables.

It can be seen that the direction of most coefficients

remains the same in the non—lagged model (Table 7.7).

However, two variables, cross-state distribution and
decertification, do change signs and become negative
predictors of disappearance rate when not lagged. Neither

variable, however, is significant, and in both cases the sign

reversals result 1J1 unlikeLy outcomes. Further, tine non—

221
lagged model shows non—significant results for establishment
growth, capital intensity anxi new capital investment, all
variables which were significant in the lagged model.
Overall, the lower Rzindicates that the non-lagged model does
not fit the data as well as the lagged model.

Although not displayedfi the trend with the time varying
models is similar——in general, R2.is less robust in the non-
lagged.model. A notable exception to less significance in the
non—lagged.model is that the cross—product of firm failure and
the final period dummy does achieve negative significance--
the .10 level——providing even more evidence for the earlier
assertion that economic distress may’ have ‘becone: a less
relevant factor in unit disappearance after about 1986.

A third alternative is to test a model which contains some
lagged predictors and some non—lagged predictors. I run a
model retaining the lagged variables that worked well earlier,
but do not lag variables that showed no significance in other

models, i.e. union penetration and employment growth.

 

1 Time-varying results are not displayed since it was decided
to exclude non-lagged models from the analysis due to their poor
performance. Busplaying results would. have required several
additional tables that would not have been further used in the
analysis.

222

Table 7.7. Standardized regression coefficients, alternative models, t-values in
parentheses N:56.

 

 

Lagged Model Non-Legged Combined
Establishment Growth -.4360** -.2216 -.3624***
(-2.25) (-1.15) (-2.27)
Firm Failure .3415“r .3565** .3321*
(2.55) (2.25) (1.95)
Capital Intensity -.2857* -.1711 -.1117
(-1.73) (-1.03) (—0.73)
New Capital Investment -.3284* -.0450 .0972
{-1.45) (-0.20) (0.46)
Decertification Rate .0739 -.2587 -.1350
(0.35) (-1.34) (0.74)
Union Penetration Rate -.0323 -.3075 -.2496
(-0.16) (-1.52) (-1.29)
Small Firm Growth .0148 .1491 .1008
(0.07) (0.63) ( 0.51)
Cross-State Change .0113 -.1971 —.2507
(0.07) (-1.17) (-1.60)
Employment Growth .0009 .2365 .1549
(0.04) (0.87) (0.65)
R’ .36** .30 .36**
Adj R2 .19 .13 .19
F value 2.12 1.72 2.14
D.W. 2.10 2.08 2.12

*** p0< .01
** p.< .05
* p.< .10

223

The combined model explains total variance to the same
degree as the lagged model-—however, fewer individual slope
coefficients show significance. The major difference between
the combined model and original lagged model is the trade—
off ii1 significance between infirni penetration anxi capital
intensityx Capital intensity loses significance in the
combined model as the non-lagged union penetration rate gains
significance. Further, cross—state change shows near—
significance as a negative predictor of disappearance. While
it is not possible to determine which of the three models most
closely represents reality, the best fitting and most logical
model appears to be the original lagged model.

2. A more parsimonious model.

A number of the variables included in the original model
added virtually nothing to variance explanation. It ndght
possible, therefore, to present a leaner model and recover
unnecessarily lost degrees of freedom.

Table 7.8 shows that the elimination of decertification
rate, small firm change, and employment growth has no effect
at all on the amount of variance explained——R2remeins at .35.
Further, adjusted Rzand.the level of significance increase as
degrees of freedom are recovered. It is likely that Cross-
state change could also be eliminated from this model, but is

retained based on its performance in the time varying model.

224

Table 7.8. Standardized regression coefficients, lagged reduced model, t-values
in parentheses, N:56.

 

 

Model P1

Establishment Growth -.4278**
(-2.82)

Firm Failure .3844**
(2.87)

Capital Intensity -.3157*
(-2.21)

New Capital Investment -.3404**
(—2.32)
Cross-State Change -.0077
(-0.52)

R' .35**
Adj R2 .27
F value 4.15
D.W. 2.12

*** p.< .01
** p.< .05
* p.< .10

 

225

Table 7.9 presents period effects in the more efficient
model. Again, the results are similar to those discovered
earlier. However, individual slope coefficients tend to lose
less significance vflumi period effects sure included. The
greatest impact is seen on firm failures in Model P2 and new
capital investment in Model P2 and Model P4. In those cases,
the ‘variables do run: fall belcmz an acceptable level of

significance as they did in earlier models.

226

 

 

Table 7.9. Standardized regression coefficients, lagged reduced model with.period
effects, t-values in parentheses, N:56
Model P2 Model P3 Model P4
Establishment Growth -.3585** -.6269*** —.4120**
(-2.32) (~4.64) (—2.70)
Firm Failures .2635* -.3497* .2005
(1.75) (-1.68) (0.64)
Capital Intensity -.3222** -.5010*** -.3601**
(-2.31) (-3.95) (-2.40)
New Capital Investment -.3007** -.2596** -.4436**
(-2.06) (-2.09) (-2.69)
Cross-State Change .0161 .2844** .0656
(0.11) (2.01) (0.39)
After 1980 .2500
(1.63)
After 1983 .9373***
(4.19)
After 1986 .2363
(0.96)
R' .40*** .57**" .37*
Adj R2 .29 .49 .27
F value 4.06 7.90 3.62
D.W. 2.02 1.79 2.14

*** p.< .01
** p.< .05
* p.< .10

 

227

The final model includes both time dummy variables and
cross—products. Here, time R2 associated Mutii Model FEB is
significant, while R2 in Model 5 was rune—apparenthy as a
result of firm failures and capital intensity retaining their
significance‘whentflmeperiod.dummy‘wariableenmicross-products

are added.

228

TABLE 7.10.

Standardized regression coefficients with period

effects and cross

products, lagged reduced model, t-values in parentheses, N:56.

 

 

Model P5 Model P6 Model P7
Establishment Growth -.3437**- .5254*** -.5767***
(-2.17) (-3.52) (-2.88)
Firm Failures .2723* -1.543** .2249
(1.77) (-2.42) (0.94)
Capital Intensity -.3061** -.5453*** -.4105***
(—2.11) (-4.28) (-2.73)
New Capital Investment -.3722 —.1549 -.7356**
(-.53) (-.45) (-3.29)
Cross-State Change .1633 .6321** .1498
(0.63) (2.34) (0.85)
After 1980 .6368
(1.10)
After 1983 .7691
(1.63)
After 1986 .9547
(1.60)
Firm Failure x Period Effect NA 1.920* -1.202
(1.95) (-1.52)
Cross State Change X Period Effect -.4196 —.6173 .4412
(-0.74) (—0.96) (0.71)
New Capital Investment x Period Effect .0993 -.1316 .3062
(0.15) (-0.39) (1.51)
R' .41*** .61*** .49***
Adj R' .27 .51 .35
F value 3.02 5.92 3.59
D.W. 2.05 1.96 2.22

*** p.< .01
*. p.< .05
* p.< .10

 

229

F. Conclusions.

In Models 3 (Table 7.5) and P3 (Table 7.9), the impact of
time period variable is dramatic, and nearly doubles variance
explanation. There was most certainly a structural change in
the model for the period following 1983. The dummy variable
for the post—1983 time period is significant, and the
coefficients for firm failure and changes in the cross—state
distribution of firms change substantially and in accordance
with expectations.

The interpretation of cross—products is a bit more
difficult. Although there is a significant positive change
in the firm failure variable after 1983, the broader impact
of this variable is apparently cancelled by the negative
coefficient observed. for time variable's pre-l983 effect.
However, the cross-product for firm failure and the post—1986
period.nearly becomes a significant negativeepredictor'of'unit
disappearance, without such a strong cancelling effect.
Together these findings suggest the declining importance of
this variable over the period of observation.

Although. more analysis is needed, the findings are
consistent with the overarching hypothesis of the study.
After the mid—19808, when public policy changed, the
disappearance of local bargaining relationships was less a

function of the financial distress within industries, and more

230
a function (of time movement of establishments, absent an
obvious economic motivation.

Returning again to the hypotheses stated above, support
is found for hypothesis H1: Firm failures are both positively
and significantly associated with unit disappearances for the
entire time period. Hypothesis H2, concerning relocation, is
not supported in general. However when the effects of the
time period after 1983 are partialed out of the nmdel, a
change in cross—state distribution does become significant.
Hypothesis IEL relating rm) the proxy 1km: outsourcing, and
hypothesis H5 concerning decertification are never supported
by any of the models. Hypothesis H4, concerning the impact
of neW' capital expenditures is, for the :most part, not
supported, but to the contrary new capital expenditures are
a significant negative predictor (HE unit disappearance in
several models.

There is strong support for Hypotheses H7. When the
model’s intercept is allowed to vary for the periods
coincident to disappearances logged. after 1983, variance
explanation is significantly increasemL However, an intercept
shift does not seem to occur following the earlier and latter
time periods.

Hypotheses H9 is supported. When period dummy variables
are added, the significance of firm failures diminishes. And,

when the effects of the second period dummy are included in

231
the model, firm failures become a significant negative
predictor of unit disappearance.

Hypotheses H10 states that all <other ‘variables ‘will
remain significant when period effects are included. There
is symme support for tints hypothesis. Most interesting,
however, is tfmm: the (cross-state: change ‘variable .becomes
significant in Models 6 and P6.

Hypotheses H11 through H15 relate to the cross—products,
and are only moderately supported” The cross—product for firm
failures and the time period related to disappearances
recorded. after' 1983 is significant and. positive, but is
cancelled by its independent effect. Nearly significant, in
accordance with hypotheses, are the cross—products for firm
failures and new capital spending in the final model.

Overall the models indicate that unit disappearances were
most likely t1) occur ir1 relatively Immxe labor intensive
industries that were closing establishments, and that there
was ea profound structural change in.tjne model after 1983.
Further, the change in the model may have been caused by the
shifting effects of firm failures.

The results CHE the regression analyses suggest that
while firm failure is a good predictor of disappearance for
the entire period, and for the period following 1980, it
becomes a less salient and maybe even a negative predictor of

variance after the mid—19808. Hence, we can conclude that the

232
trends arising in the mid—19808 were consistent with a public
policy explanation, although more conclusive results will have

to await the availability of more recent data.

VIII. Conclusions and Discussion.

A. Introduction.

There were a number of issues addressed in this research.
At the level of theory, an attempt was made to bridge
historical and nonhistorical modes of social scientific
thinking (Chapter II). In the immediate context, the former
emphasizes differences across and developments within nation—
states linked t1) specifit: historical events, jparticularly
policy changes. The latter focusses upon economic outcomes
linked lesstmispecific eventstflmnito normal fluctuations and
trends in the marketplace.

Second, policy issues were explored (Chapter III and IV).
The main focus was on the institutions, processes, and
outcomes associated with U.S. labor relations public policy
during time 19808. Ptditical influence over administrative
mechanisne and the products of that influence on several areas
of doctrine-—the duty to tergain over bmsiness decisions,
union rights to maintain allegiance and discipline, and an
employer’s right to intervene in the certification process-
—were addressed.

Third, the immmmm; of policy changes (n1 one extremely

233

234

important aspect of industrial relations practice-—the
viability of local collective bargaining relationships—-was
examined. This examination was aided by a model of the
disappearance of local relationships which included variables
linked to plant closings and relocations, the substitution of
inputs, and the loss of certification, as well as several
control variables.

The findings suggest the following: Economic arguments
relating to the financial distress within industries provide
a good overall explanation of bargaining unit disappearance.
However, there is a good deal of instability in the model when
time period effects are examined. This is particularly true
when differences before and after the mid—19808 are examined.
The inclusion of a variable representing the post—1983 time
period. improves rmxmfl_ specification significantly. This
categorical variable also forces the coefficient for the firm
failure variable, a positive and significant predictor in the
constant time model, to reach negative significance» The time
period variable further causes changes in the variable
representing change in the cross—state distribution of
establishments. While this latter variable was not a
Significant predictor of unit disappearances in a constant
time model, it is a significant and positive predictor in a
time varying model.

The results here clearly show that “something happened"

23S

around 1983. It is not easy (33 tme precise about ‘what
happened, but the evidence presented earlier suggests that
the magnitude of the policy changes of that period may have
caused ea substantial. change Lhi time context vniflnil which
industrial relations occurred.

For a number of reason conclusions must be treated with
caution. First, the within—industry sample size for several
of the cases may have made the measure of disappearances for
some of the industries questionable. Second, since direct
measures for several components of Equation 1 are not
available, proxies were required. It is not certain whether
the proxies for relocation, subcontracting, and technological
change were truly representing the intended underlying
variables. Third, Imilticollinearityn particularly7 between
firm failure and the period variables made the analysis more
difficult. Fourth, the lack of more recent Department of
Labor data makes it difficult to analyze a more complete time
series. Therefore, while it is clear that a structural change
occurred in the model in the mid-19808—-a8 was predicted——it
is difficult tm) discount all cflflmm: arguments énmi conclude
firmly the political changes and subsequent policy outcomes
were responsible.

Nonetheless, the first connection seems fairly certain.
Following the 1980 election, there was a sharp departure from

past protocol. From the 19508 through the 19708 a system of

236

mutual accommodation had developed concerning appointments to
the National Labor Relations Board. When Democrats were in
power, organized labor lmmi a direct hand ill the selection
process. When Republicans held office business groups had a
similar prerogative. However, each group held de facto veto
powers over the other’s selections (Moe, 1985, 1987). This
system began to erode with the conflicts of the 19708, and.was
partially ignored.with the elevation.of‘WillianlLubbers to the
General Counsel’s position over the objections of business
groups. But although Lubbers was distasteful to business, he
was a career civil servant and had spent only minimal time as
a union advocate (a six month stint with the American
Federation of State, County, and Municipal Employees in the
early—19508.)

After the 19804election, however, the administration went
outside the Board and put forward nominees who were career
management advocates. Some had worked against unions during
certification elections, and several were allied with
organizations cmmaflgr hostile tm> the LUUtMi movement. The
administration was generally successful in seating its
nominees due to Republican control of the Senate.

Following the appointment of a solid conservative
majority—~Dotson4 Hunter, and. Dennis-~by Dmn/ of 1983, an
unprecedented number of doctrinal reversals were issued. One

roadblock to the total overhaul remained in William Lubbers,

237

who was to hold the General Counsel’s position until 1984.
Chairman Dotson, therefore, sought to circumvent Lubbers——to
the extent statutorily possible—~by rescinding the General
Counsel’s 30 year control over the enforcement process, and
placing that responsibility with the Board Solicitor. At the
time, the Solicitor was Hugh Reilly, who had once worked for
the legal arm of the National Right-to-Work Committee.
Following the expiration of Lubbers’s tern» the General
Counsel’s job went to Rosemary Collyer, also a former
management advocate——though more recently'a memberciftimeMine
Safety’ Commission—~over the vigorous opposition. of labor
leaders.

Between 1983 and 1986, the Board reversed several dozen
precedents. For current purposes the most important cases
were Otis Elevator II, Milwaukee Spring II,.Machinists Local
1414, Rossmore House, and several progeny of the 1982 Midland
National Life decision (for example, Pilliord of Mississippi
and Atlantic Forest Products). Through these reversals, the
Board narrowed an employers duty tm> bargain over business
decisions affecting the bargaining unit, made it more
difficult for unions to promulgate standards of membership,
and further deregulated conduct during certification
elections.

These changes may have put local collective bargaining

relationships at greater risk by lowering the costs to the

238

employer associated with escaping unionization, by making it
more difficult for unions to act against dissident members,
and by allowing employers to have greater influence over the
outcome of the decertification process. This conjecture is
further substantiated by the quantitative results which
indicate changes iii the explanation (n5 unit disappearance
after 1983.

B. Discussion.

This research raises a number of issues that transcend
the specific research hypotheses and time period studied: the
proper role of the state in the industrial relations system,
the stability of labor relations doctrine, and the extent to
which the stability of doctrine affects the stability of the
industrial relations system.

F m mostcflfits history, industrial relations scholarship
had been concerned primarily with the relationship between
employers and the representatives of employees (Kaufman,
1993). In “weak state" nations snmfliems the United States
(Skocpol, 1985), it is not clear what the role of the state
has been, is, or should be in the industrial relations system.
Older conceptualizations indicate an early divergence between
the government and the system’s principals. Commons (1918,
;p. 9) writes:

By vetoing the laws which labour in its

political struggles has been able to
secure, the courts, joined to divergent

239

state policies, have excluded or delayed
labor from legislative influence.
Consequently, the energies of
organization are turned to the economic
field...In this way has our Federal and
judicial system of government added its
pressure on labour and forced it to
acquire kn/ trade infirm) action vflmm; in
other countries has been granted by
legislation.

Perlman (1928, p. 170) picking up on the same point,

adds:
Obviously, then, American governments are
inherently inadequate as instruments of
economic reform...

It is to this situation, more than
anything else that the stubborn economism
of the American Federation of Labor must
be traced.

Both Commons and Perlman argue that reliance on the
state was largely unavailable to labor movement at the
earliest stages of its development because of the patchwork
of local laws and an unreceptive judiciary. Therefore, after
having been (an; off from time state, industrial relations
developed as a "voluntary" system relying, instead, on
collective bargaining between employers and local unions.

Recently, critical legal scholars (Klare, 1981; Forbath,
1989) have taken a somewhat different view. They argue that
late into the 19th century the AFL and the Knights of Labor
continued to press for state action. All over the country,

Knights’ candidates ran for local office, and within the AFL

a fairly large Lassallean wing formed local labor parties.

240
Sometimes these political efforts were successful, but on a
larger scale the relations between labor and the state were
hostile as the use of the labor injunction became widespread.
Hence, the state was still very much a pdayer, but nearly
always to the benefit of employers. The system was not as
"voluntaristic" as Commons and Perlman assert.

After passage of the Clayton, Norris—Laguardia, Railway
Labor, Wagner, and Taft—Hartley Acts, it was impossible to
deny the active participation of the state in the system. In
John Dunlop’s classic Industrial Relations Systems (1958)
government was counted as an "actor" in the system equal in
status to labor and management. Two years later, Kerr,
Dunlop, Harbison, and Myers (1960, p. 25) were to write:

Governments have 61 significant role iii
determining the substantive rules of the
work. community <or 1J1 establishing' the
procedures and responsibilities of those
with tints power. In time highly
industrialized society, enterprise
managers, workers, and the government
tend to share in the establishment and
administration of rules. The industrial
relations system (if the industrial
society is genuinely tripartite.

Although it is no longer denied that government has a
role in the system, the nature of that role is still debated.
The most important subject of that debate itw Was it the
intent of legislators at the time of Act’s passage to grant

substantive rights to unions, or merely to guarantee

procedural rights ti) employees? Those holding time former

241

view argue that the intent of the Wagner Act was to
"encourage collective bargaining" by nurturing the process of
union organizing and increasing the role of unions within the
firm, hence, balancing economic interests (Stone, 1981;
Tomlins, 1985). Those subscribing to the latter view believe
that the purposes of the Act, particularly following the
Taft—Hartley amendments, were merely to protect the right of
workers to join or not join labor organizations and to make
sure that neither party walked away from the bargaining table
when issues central to the employment relationship were being
discussed——but iii no vmwr'was the jhmv intended ti) provide
substantive rights to unions, or to give unions any role in
the direction of business enterprise.

Stone (1981) argues that this latter view was encouraged
by a false belief in pluralism promoted by social scientists,
vim) downplayed. inherent econonnt: power" imbalances. They
believed instead that pie—Act relations between labor and
management were sufficiently mature to allow the government
to play only minor role. Government would merely provide
stability to the existing system by granting it legitimacy
and providing some moderate institutional support for the
adjudication énmi resolution <1f conflicts. Otherwi8e kmmii
labor and management could handle their own affairs. There
was a direct intellectual lineage from the earlier views of

Commons and Perlman to this mid—century idea. The link

242
became most direct when William Leiserson was appointed by
Franklin Roosevelt to "clean up the Board" in 1939. Tomlins
(1985a, p. 31) writes:
Leiserson argued that the Wagner Act had
projected public authority into a field
already occupied by nonpublic rule—making
bodies: unions and employers. For
government policy to succeed, the public
agency created ti) administer it: should
learn from the institutions already
established in the field and seek harmony
with their environment, not ’impose new
rules based on its notions of
reasonableness.’
When Harry Millis was appointed to the Board, Leiserson

wrote to Commons (Tomlins, 1985a, p. 33):

Now if the President only appoints Ed
Witte next when Smith’s appointment
expires, then we would have a real Board.
You would have all three of the Board
members your boys——and would be sure that
the administration of time LmN was both
proper and intelligent.

Not only did this point—of-view influence Board
decisions after 1939, kin: it provided time philosophical
underpinnings for the Taft—Hartley amendments. The sometimes
conflicting' language <)f time Wagner eumi Taft—Hartley'.Acts
allows the debate over the role of the state to continue
(Sockell, 1986). Edward Miller, the Board’s chair during
Nixon’s administration, describes the differences he has with

his immediate predecessor, Frank McCulloch (Gross, 1985, p.

15):

243

Frank says the purpose if the Act is
to encourage collective bargaining. I
say, ’But Frank, you didn’t read. the
revision of the preface to the Act at the
time of Taft-Hartley, which now clearly
sets forth two objectives. The first one
is the freedom of choice of employees in
deciding whether or not they want to join
a union anmi be represented karts union,
and, if so, by which union.’ And that
freedont of choice is ti) nw' mind the
keystone of the Act.

The broad and conflicting language of the Act’s purposes
is the main reason why interpretations are so sensitive to
political influence, and why the appointment process is so
critical. It is symptomatic, I believe, of a deeper
uncertainty of the appropriate role of the state in the
industrial relations system. Ironically, tithe most basic
question is still asked while the body of law itself has
become vast and highly—developed.

Political influence through the appointment process has
led to policy oscillation, and, in turn, the lack of a truly
stable national labor relations policy. As Estreicher (1986)
points out, because Congress left so much to "decisional
law," the Board has functioned in a "quasi—legislative," as
much as in a quasi—judicial fashion. That being so, there
has always kmmit at least :iime partisan influence (it the
construction of the law. There has also been some tendency

for Republican-dominated Boards to undo at least some of the

law created by Emmecratic—dominated Boards and vice—versa.

244
But never was this more obvious than between 1983 and 1986.

The lack of statutory clarity has also led to uddely
different interpretations of the Act by the Circuit Courts.
Although the Board has been criticized for its record before
the courts, in fact it has the task of crafting a national
labor policy in a country where the federal courts are often
in sharp disagreement over the basic purposes (if the law.
The Afllwaukee Spring decision, for example, was 51 contest
between the thimii and Seventh Circuit’s views (Mfzmid-term
alterations. 131 addition, the Supreme Court’s official
policy of deferring to the expertise of the Board (Ford Motor
Co. v NLRB 441 US 488 (1979); NLRB v. Erie Resistor Corp. 373
US 221 (1963); NLRB v. Steelworkers 357 US 357 (1958)) may
further encourage the jBoard ti) act independently' of the
judiciary.

A number of options have been discussed to halt this
wavering: time creation (if labor courts, au1 organizational
reconfiguration of time Board, and, jparticularly, a Inore
extensive use of rule—making under the Administrative
Procedures Act (Aaron, 1985; Block, 1994; Estreicher, 1986).
Others argue——ignoring, II believe, time damage done ti) the
credibility public institutions when policy -is so
unpredictable——that frequent reversals are ea necessary by—
product of flexibility, arm) that attempts ti) steady' the

pendulum may ossify the Act (Winter, 1968).

245
:n: is difficult to imagine that 51 level of consensus
necessary to clarify the state’s role in industrial relations
will ever be reached. There is simply too much inherent

disagreement if] the nation over time benefits unionism and

collective bargaining in the first place. Some people see
both institutions as social goods, others as economic
nuisances (Block, 1995). Courts exhibit this tension when

simultaneously' extolling time social ‘virtues (n5 free

collective bargaining and limiting the areas in which

bargaining need occur. And while decision—makers seem
particularly concerned with management’s need for
"flexibility" or "quick decision—making (Sockell, 1986)," few

seem to accept the argument made by Hedlund (1986, p. 961)
that,
The bargaining requirement ensures

that the ‘union ihas EH1 opportunity to
correct unsperceptions held tn! the

employer concerning the union’s
willingness to concede. In addition, it
provides more credible information to
both the employer and the union. A8 a
result, it promotes mutually beneficial
exchange and increases economic
efficiency.

The notion that bilateral solutions could increase
economic efficiency seemed wholly absent from most Board and
court decisions made chnfiim; the 19808. (M) the contrary,
these tribunals exhibited.éa fear that 51 strong and stable

union with a meaningful presence within firm would result in

246
economic dysfunction. The nearly unlimited right (ME both
employers and employees to escape the obligations of
mandatory bargaining and unionization were deemed essential
to maximize the conditions necessary for economic growth and
the free flow of commerce.

Alas, the question remains: Does the government’s
wavering over the basic questions of unionism and collective
bargaining really matter? Or ck) employers eumi unions go
about their business largely unaffected by the public policy?
The evidence of this study suggest the following answer:

From the mid-1970 until 1990, there was a large decline
in the number of collective bargaining relationships. If the
estimates presented earlier are correct, nearly 25 percent of
the bargaining relationships found in manufacturing in the
mid-19708 had vanished by 1990. In general, these
disappearances occurred in labor intensive industries that
were both contracting with regard to the number of
establishments, and suffering high levels of financial
distress as measured by firm failure rates. Hence, in the
broader span of time, not accounting for any changes in
policy regimes, the fortunes of the institutions of
collective bargaining seem most clearly linked to economic
trends.

However, while most collective bargaining relationships

survived the period, the number that did not increased after

247

1983, when policy regimes changed. Further, the variables
associated vomit disappearance Edfx) change. While capital
intensity and industry contraction continued to be important
predictors, a shift occurred in the saliency of firm failure
and changes in cross-state distribution. The former became
a negative predictor of unit disappearance, while latter a
strong positive predictor. Since this shift is consistent
with what might be expected when policy changes in a way that
is favorable to employers who wish to shift operations from
union to 51 non—union facilities, support. is lent ti) the
proposition that a change in policy affected the viability of
local relationships.

This conclusion, while important, should not be
oversold. .Along with.time earlier stated shortcomings, it
must be stressed that 75 percent of relationships made it
through the time period, and that the average annual rate of
disappearance never surpassed three percent. However, in a
small but significant number of cases employers may have been
responding to changes in public policy in a way ostensibly
favorable to their interests, but unfavorable to the
institutions of industrial relations. If these institutions
are, as the framers of the Wagner Act believed, important to
our society, then this study provides a reason for a re-
examination of federal labor policy, both with regard to the

substance of the law and the mechanisms of enforcement.

LIST OF REFERENCES

248

LIST OF REFERENCES

Aaron, Benjamin. 1985. The NLRB, labor courts and industrial
tribunals: a selective comparison. Industrial and Labor
Relations Review. 39: 35—45.

Adams, Arvil V. and Joseph Krislov. 1974. New union
organizing: a test of the Ashenfelter and Pencavel model
of trade union growth. Quarterly Journal of Economics.
88: 304—311.

AFL—CIO. 1983. NLRB and the Collyer nomination. The People's
Lobby: AFL-CIO Legislative Department Report on the 97th
Congress. Washington, D.C.: AFL-CIO.

Allen, Steven. 1986. Union work rules and efficiency in the
building trades. Journal of Labor Economics. 4: 212—242

Allison, Paul D. 1982. Discrete time methods for the analysis
of event histories. in S. Leinhardt, ed., Sociological
Methodology. San Francisco: Jossey—Bass. 61—98.

Allison, Paul D. 1984. Event History Analysis. Newbury Park,
Ca.: Sage Publishing.

Anderson, John C., CVReilly, Charles IL. III and Gloria B.
Busman. 1980.tkntm1decertification.intimeUnited.States,
1974—1977. Industrial Relations. 19: 100-107.

Ashenfelter, Orley and John H. Pencavel. 1969. American trade
union growth. Quarterly Journal of Economics. 73: 434—
448.

Ashford, Nicholas A. and Christine Ayers. 1987. Changes and
opportunities iii the environment for technology
bargaining. Notre Law Review. 62: 810-858.

Bain, George S. and Farouk Elsheikh. 1976. Union Growth and
the Business Cycle. Oxford: Basil Blackwell.

249

Bakaly, Charles G. 1987. A response to Murphy's law. Charles
J. Morris, ed., American Labor Policy. Symposium on the
Labor Board athid—Century. Washington, D.C: BNA. pp. 78—
95.

Bamber, Greg. 1986. New technology—~the challenge to unions:
a comparative view. IRRA. Spring Meeting. Labor Law
Journal. August: 502-510.

Bamber, Greg J. and Russel D, Lansbury, eds. 1987.
International and Comparative Industrial Relations.
London: Unwin and Hyman.

Barbash, Jack. 1948. Labor Unions in Action. New York: Harper
and Brothers.

Barlett, Donald L. and James B. Steele. 1992. The lucrative
business of bankruptcy. America: What Went Wrong? Kansas
City: Andrews and McMeel, pp. 66-87.

 

Barnett, George E. 1916. Growth of labor organizations in the
United States, 1897—1914. Quarterly Journal.of.Economics.
30: 780—795.

 

Barton, David. 1986. Milwaukee Spring Division of Illinois
Coil Spring Co.: the open road for employer relocations
of bargaining unit work during the terntof the collective
bargaining agreement. Toledo Law Review. 17: 973—1005.

Bernstein, Jules and Laurence E Gold. 1985. Mid-life crisis:
The NLRB at 50. Dissent. Spring: 213—218.

Blalock, Hubert M. 1979. Social Statistics: Revised Second
Edition. New York: McGraw-Hill.

 

Blalock, Hubert M. 1961. Causal Inferences in Nonexperimental
Research. New York: W.W. Norton and Company.

 

Block, Richard N. 1995. Labor law, economics, and industrial
democracy: a reconciliation. Industrial Relations. 34:
402—416.

 

Block, Richard N. 1994. Reforming U.S. labor law and
collective bargaining: some proposals based on the
Canadian system. S. Freedman, R. Hurd , R. Oswald and R.
Seeber, eds., Restoring the Promise of American Labor
Law. Ithaca, NY: ILR Press. pp. 250-259.

 

250

Block, Richard. N. 1990. American industrial relations in the
19808: transformation tn? evolution? LI. Dworkin, EKL,
Reflections on the 'Transformation in Industrial
Relations. Metuchen, NJ: IMLR Press. pp. 19—52.

Block, Richard, N. and Kenneth McLellan. 1985. Structural
economic change and industrial relations in the United
States’ manufacturing and transportation sectors since
1973. H. Juris, M. Thompson and. W. Daniels, eds.,
Industrial Relations irtta Decade (n5 Economic Change.
Madison, WI: IRRA Series. pp. 337-382.

Block, Richard N. and Steven L. Premack. 1983. The
unionization process: 63 review Ci time literature. D.
Lipsky and J. Douglas, eds., Advances in Industrial and
Labor Relations. 1: 31—70.

Bluestone, Barry and Irving Bluestone. 1992. Negotiating the
Future. New York: Basic Books.

Bluestone, Barry and Bennett Harrison. 1983. The incidence and
regulation of plant closings. Larry Sawers and William
K. Tabb, eds., Sunbelt/Snowbelt. New York: Oxford
University Press.

Bluestone, Barry and Bennett Harrison. 1982. The
Deindustrialization of America. New York: Basic Books.

Bok, Derek and Qkiui Dunlop. 1970. Labor enmi the American
Community. New York: Simon and Schuster.

Bosanac, Paul. 1983. Concession bargaining, work transfers,
and midcontract modificatitmu Los Angeles Marine Harware
Company. Labor Law Journal. February: 72—79.

 

Britt, Clifford P. 1986. NLRB narrows duty to bargain over
management decisions: Otis Elevator II and its progeny.
Wake Forest Law Review. 21: 725—755.

 

Brown Thomas Eh ZDJ and Rebecoa A. Campbell. 1991. Current
developments in labor—management relations. Employee
Relations Law Journal. 16: 529—532. '

 

Brownstone, Robert D. 1986. The National Labor Relations Board
at 50: politicization creates crisis. Brooklyn Law
Review. 52: 229-280.

251

Bureau <)f time Census. 1972, 1977, 1982, 1987. Census <)f
Manufactures. Washington, D.C: author.

 

Bureau CHE the Census. 1973-1990. Annual Survery of
Manufactures. Washington, D.C: author.

Bureau of time Census. 1973—1990. County Business Patterns.
Washington, D.C.: author.

Bureau of the Census. 1973-1990. Quarterly Financial Reports.
Washington, D.C.: author.

Bureau of Labor Statistics. 1994. Employment, Hours,
Earinings, United.States, 1909~94. Washington, D.C.: GPO.

Buss, Terrence F. and F. Stevens Redburn. 1987. Plant
closings: impacts and responses." Economic Development
Quarterly. 1: 170—177.

 

Buss, Terrence F. and F. Stevens Redburn. 1983. Mass
unemployment:_plant closings and community'mental health.
Beverly Hills, CA: Sage Publishing.

Caldwell, Michael. A. 1981. Community defense of union—free
status. Mercer Law Review. 32: 679—741.

Chatterjee, Samprit and Bertram Price. 1991. Regression
Analysis by Example. New York: John Wiley & Sons.

Chiason, Gary. 1980. Attmiie on union merger trends, 1900-
1978. Industrial and Labor Relations Review. 34: 114—
120.

Chiason, Gary. 1986. When Unions Merge. Lexington, MA: EC
Heath.

Chaison, Gary and Dileep Dhavale. 1990. A note on the severity
of the decline in union organizing activity. Industrial
and Labor Relations Review. 43: 366—373.

 

Chiason, Gary and Joseph B. Rose. 1990. Continental divide:
The direction and fate of North American unions. D.
Lewin, I). Lipsky' and I). Sockell, eds., Advances in
Industrial and Labor Relations. Greenwich, CT:.JAI Press.
8: 169—205.

 

 

252

Chiason, Gary and Joseph B. Rose. 1991. The macrodeterminants
of union growth and decline. G. Strauss, D. Gallagher
and J. Fiorito, eds., The State of the Unions. Madison,
WI: IRRA. pp. 3—46.

Chow, Gregory. 1960. Tests of equality between sets of
coefficients in two linear regressions. Econometrica.
28:591-625.

 

Clark, Gordon L. 1988. A question of integrity: the National
Labor Relations Board, collective bargaining and the
relocatitmlof work. Political Geography Quarterly. 7:209-
227.

Clark, Gordon L. 1989. Unions and Communities Under Seige.
Cambridge: Cambridge University Press.

Collingsworth, T. 1993. Resurrecting the National Labor
Relations Act——plant closings and runaway shops in a
global economy. Berkeley Journal of Employment and Labor
Law 14: 72—112.

Commons, John I3. 1918. History (ME Labour in time United
States, Vol. I. New York: MacMillan.

Cooke, Williant N. 1985. The failure to negotiate first
contracts: determinants and policy implications.
Industrial and Labor Relations Review 38: 163-178.

Cornfield, Daniel. 1987. Plant shutdowns and union decline:
the United Furniture Workers of America, 1963-1981. Work
and Occupations. l4: 434~451.

 

Cox, Archibald; Bok, Derek and Robert A. Gorman. 1987. Labor
Law. Mineola, N.Y: Foundation Press.

Craft, James A. and Marion Extejt. 1983. New strategies in
union organizing. Journal of Labor Research. 4: 19—32.

 

Curme, Michael A., Hirsch Barry T. and David A. MacPherson.
1990. [kmrii membership enmi contract, coverage 1J1 the
United States, 1983—1988. Industrial and Labor Relations
Review. 44:5—33.

Cyert, Richard M. and David C. Mowery. 1987. Technology and
Employment. Washington, D.C: National Academy Press.

 

 

253

Daily Labor Report. December 21, 1987. National Right-to—Work
Committee opposes NLRB promotions for Stephens and
Higgins. p. A—7.

Daily Labor Report. June 5, 1985. Remarks on NLRB by David M.
Silberman, AFL—CIO Associate General Counsel. pp, F-l-
F—3.

Dailijabor Report. June 14, 1984. NLRB chairman denies press
reports that White House requested his resignation. p.
A—4.

 

Daily Labor Report. February 17, 1983. Senate confirms
nomination of Donald Dotson to head NLRB. p. A—5.

 

Daily Labor Report. September 1, 1982. Right—to—Work.Committee
asks President to withdraw \kn1lie Water nomination to
NLRB. p. A-3.

Daily Labor Report. November 19, 1981. Committee refuses to
confirm Van De Water, block bid to send nomination to
Senate floor. p. AA—l.

Daily“ Labor“ Report. October" 16, 1981. Exchange <)f
correspondence between Sen. tmmritemmi NLRB Chairman—
designate Van De Water, and letter to Hatch for AFL—CIO
opposing Van De Water. pp. F-l—F—2.

Daily Labor Report. September 10, 1981. Senate Labor Committee
gives quick okay to former staff member Hunter for NLRB
post. p. A-6.

Daily Labor Report. January 2, 1981. Georgine eyes Senate,
not Reagan, as main threat to building trades. p. A—3.

Daily' Labor iReport. December 1L4, 1979. Statements Before
Senate Labor and Human Resources Committee on nomination
of William Lubbers as NLRB General Counsel. pp. F—l—F-
10.

 

Daily Labor Report. November 30, 1979. Management groups plan
to fight Lubbers appointment, but await‘ Nmrphy
renomination. p. A—11.

 

Daily Labor Report. August 28, 1978. Javits sees possibility
of confrontation between labor and business in wake of
labor reform fight. p. A—7.

 

254

Daily Labor Report. April 14, 1977. President Carter names
Fanning to serve as chairman of NLRB. p. A—6.

Daily Labor Report. March 17, 1976,. Reagan listed as only
backer of right—to—work principle. p. A-4.

Daily Labor Report. January 9, 1975. AFL—CIO to raise "no
objection" to Murphy’s nomination to NLRB. p. A—8.

Delaney, John T., Marick F. Masters and Susan Schwochau. 1990.
Infirmlmembership and\Kiij;f m COPE endorsed.candidates.
Industrial and Labor Relations Review. 43: 621-635.

Dennis, Patricia D. 1985. Principles that guide my decision—
making. William [mdson enmi Kent Lollis, eds., Second
Annual Labor and Employment Law Institute. Littleton CO:
Fred B. Rothman.

Dickens, William T. 1983. The effect of company campaigns on
certification elections. Industrial and Labor Relations
Review. 36: 560—575.

Dickens, William T. and Jonathan S. Leonard. 1985. Accounting
for the decline in union membership, 1950—1980.
Industrial and Labor Relations Review. 38: 323-335.

Dobbin, Frank, Sutton, John R., Meyer, John W. and W. Richard
Scott. 1993. Equal opportunity law and the construction

of internal labor markets. American Journal of Sociology.
99: 396—427.

Doyle, Philip. M. 1985. Area wage surveys shed light on
declines in unionization. Monthly Labor Review. 108: 13-
20.

 

Dotson, Donald. 1984. Processing cases at the NLRB. Labor Law
Journal January: 3-9.

 

Dube, Lawrence. 1986. New rules for collective bargaining: an
improved cliamte for management negotiators. Labor Law
Journal. July: 432—436.

 

Ihni and Bradstreet. 1974—1990. Firm Failure Record. lmiv
York: author.

255

Dunlop Commission Report. 1995. Final Report of the Commission
on Worker—Management Relations. Washington, D.C.: GPO.

 

Dunlop, John T. and Walter Galenson. 1978. Labor in the
Twentieth Century. New York: Academic Press.

Dunlop, John T. 1958. Industrial Relations Systems. New York:
Henry Holt.

Dunlop, John T. 1948. The development of labor organization.
R.A. Lester and J. Shister, eds., Insights into Labor
Issues. New York: MacMillan.

Elshiehk, Farouk and (minxme S. Bain, 1978. American trade
tHUtitgrowth:au1alternative1miiel.Industrial.Relations.
17: 75-79.

Estreicher, Samuel. 1985. Policy oscillation an: the labor
board: a plea forulemaking. Administrative Law Review.
15:163—181.

Fanning, John. 1984. The National Labor Relations Act: its
past and its future. William Dolson and Kent Lollis,
eds., First Annual Labor and Employment Law Institute.
Littleton CO: Fred B. Rothman. pp. 59—70.

Farber, Imin3/ S. 1985. fflme extent (n5 unionization iil the
United States. Thomas Kochan, ed., Challenges andtdhoices
FacinggAmerican Labor. Cambridge, MA: MIT Press. pp. 15—
44.

Farber, Henry. S. 1987. The recent decline of unionization in
the United States. Science. 238: 915-938.

Fiorito, Jack. 1982. Models of union growth: a test of the
Bain Elsheikh model for the U.S. Industrial Relations.
21: 123-127.

 

Fiorito, Jack; Gallagher, Daniel G; and Charles R. Greer.
1986. Determinants of unionism: a review of the
literature. K. Rowland and G. Ferris, eds., Research in
Personnel and Human Resource Management. GreenWich, CT:
JAI Press. 4: 269—306.

 

Fiorito, Jack.enmi Charles It. Greer. 1982. Determinants of
unionisnn past research anmi future needs. Industrial
Relations. 21: 1-32.

 

256

Fiorito, Jack and Cheryl L. Maranto. 1987. The contemporary
decline of union strength. Contemporary Policy Issues.
5: 12-27.

Flynn, Joan. 1991. A triple standard at the NLRB: employer
challenges to an incumbent union. Wisconsin Law Review.
653—705.

Forbath, William E. 1989. The shaping of the American labor
movement. Harvard Law Review. 102: 1111—1256.

Freeman, Richard EL 1989. Contraction eumi expansion: the
divergence of of private sector and public sector
unionism in time United States." Journal of Etonomic

 

Perspectives 2: 63-88.

Freeman, Richard B. 1989a. The changing status of unionism
around the world. Wei Chiao Wang, ed., Organized Labor
at the Crossroads. Kalamazoo, MI: Upjohn Institute. pp.
111—138.

Freeman, Richard B. 1989b. What does the future hold for U.S.
unionism. Relations Industrielles. 44: 25-43.

Freeman, Richard B. 1985. Why are unions fairing poorly in
NLRB representation elections? Thomas Kochan, ed.,
Challenges and Choices Facing American Labor. Cambridge,
MA: MIT Press. 45—64.

Freeman, Richard B. and James Medoff. 1984. What Do Unions Do?
New York: Basic Books.

Freeman, Richard B. and Jeffrey Pelletier. 1990. The impact
of industrial relations legislation (i1 British: union
density. British Journal of Industrial Relations. 28:
141-164.

 

Frier, Susan. 1988. Expansion of the successorship doctrine—
—Fall River Dyeing and Finishing Company v. NLRB. Wake
Forest Law Review. 23: 549-574.

 

Fukari, Hiroshi and (kit P. Alston. 1992. Ecological
determinants of divorce: a structural approach to the
explanation.of.Japanese divorce. Social Biology. 39: 257—
277.

257

Gacek, Stanley A. 1981. The employer’s duty to bargain on the
termination of unit work. Labor Law Journal. October:
659—678; November: 699—724.

Gagala, Ken. 1983. Union Organizing and Staying Organized.
Reston, VA: Reston Publishing.

George, Glenn B. 1985. To bargain or not to bargain: A new
chapter in work relocations decisions. Minnesota Law
Review. 69:667—717.

George Glenn B. 1985a. Collective bargaining in Chapter 11 and
beyond. Yale Law Journal. 95: 300—346.

Getman, Julius; Goldberg, Stephen and Jeanne Herman. 1976.
Lhfirxt Representatitml Elections: Law enmi Realityu New
York: Russell Sage Foundation.

Goldberg, Stephen; Brett, Jeanne; and Julius Getman. 1984.
The relationship between free choice eumi Labor Board
doctrine: differing empirical approaches." Northwest
Law Review 79: 721-733.

 

Goldfield, Michael. 1987. The Decline of Organized Labor in
the United States. Chicago: University of Chicago Press.

Goldfield, Michael. 1984. The causes of U.S. trade union
decline and their future prospects. Research.in Political
Economy 7: 81—159.

Goldman, Sheldon. 1989. Reagan’s judicial legacy: completing
the puzzle and summing up. Judicature 72: 318—330.

 

Gordon, IkndtL 1977. Class struggles énmi stages (if urban
development. Alfred Watkins and Daivd Perry, eds., The
Rise of Sunbelt Cities. Beverly Hills: Sage, pp. 55-82.

 

Gordus, Jeanne Prial; Jarley, Paul and Louis Ferman. 1981.
Plant Closings and Economic Dislocation. Kalamazoo, MI:
Upjohn Institute.

 

Gould, William B. IV. 1993. Agenda for Reform. Cambridge, MA:
MIT Press.

 

 

258

Gould, William B. IV. 1986. Some reflections on fifty years
of the National Labor Relations Act: the need for labor
board and labor law reform. Stanford Law Review. 38: 937—
944.

Gross, James. 1995. Broken Promise. Philadelphia: Temple
University Press.

Gross, James. 1994. The demise of national labor policy: a
question of social justice. S. Freedman, R. Hurd , R.
Oswald and R. Seeber, eds., Restoring the Promise of
American Labor Law. Ithaca, NY: ILR Press. pp. 45—60.

Gross, James A. 1985. Conflicting statutory purposes: another
look at fifty years of NLRB law making. Industrial and
Labor Relations Review. 39: 7—18.

Gross, James. 1981. The Reshaping of time National Labor
Relations Board. Albany: SUNY Press.

Gross, James. 1974. The Making of the National Labor
Relations Board. Albany: SUNY Press.

Gujarathi, Damodar. 1970. Use of dummy variables in testing
equalityr between. sets CHE coefficients in twm) linear
regressions: a note. The American Statistician.
February: 50—52.

Haddad, Carol. 1984. Technological change and
reindustrialization: implications for organized labor.
Donald Kennedy, ed., Labor and Reindustrialization. State
College, PA: Pennsylvania State University.

 

Hamermesh, Daniel. 1989. What do we know about worker
displacment in the U.S.? Industrial Relations. 28: 51-
58.

Hammer, W. Clay and Frank J. Smith. 1978. Work attitudes as
predictors of union activity. Journal of Applied
Psychology. 66: 415-421.

 

 

Hannan, Michael and.tJohn Freeman. 1988. "The~ ecOlogy' of
organizational mortality: American labor unions, 1839-
1985." American Journal of Sociology. 94: 25-43.

 

259

 

Hannan, Michael and John Freeman. 1977. "Population ecology
of organizations.“ American Journal of Sociology. 82:
929—964.

Hannan, Michael and Nancy B. Tuma. 1979. Methods for temporal
analysis. Annual Review of Sociology. 5: 303-28.

 

Heckscher, Charles. 1988. The New Unionism. New York: Basic
Books.

 

Hedlund, J. 1986. An economic case for mandatory bargaining
over partial terminantion.and plant relocation decisions.
Yale Law Journal. 95: 949—968.

Hieb, Allen E. Jr. and George R. Moody. 1981. Work
preservation: the union struggle against technological
innovation. Mercer Law Review. 32: 833—855.

Hirsch, Barry T. 1980. The determinants of unionization: an
analysis of interarea differences. Industrial and Labor
Relations Review. 33: 147—161.

Hirsch, Barry T. and John Addison. 1986. An Economic Analysis
of Unions. Boston: Allen and Unwin.

Hirsch, Barry T. and Mark C. Berger. 1984. Union membership
determination and industry characteristics. Southern
Economic Journal. 50: 665—679.

Hirsch, Barry T. and David A. MacPherson. 1993. Union
membership and coverage files for the Current Population
Surveys: note. Industrial and.Labor Relations Review; 46:
574—578.

 

Holmes, John. 1986. The organization and locational structure
of production subcontracting. Allen Scott and NMchael
Storper, eds., Production, Work, and Territory. Boston:
Allen and Unwin. pp. 80-103.

Howell, David R. and Edward N. Wolff. 1992. Technical change
and the demand for skills by U.S. industries. Cambridge
Journal of Economics. 16: 127—146. '

Howland, Marie. 1988. Plant Closings and Worker Displacement:
The Regional Issues. Kalamazoo, MI: Upjohn Institute.

 

260

Ichniowski, Casey. 1988. Public sector union growth and
bargaining laws: a proportional hazards approach with
time—varying treatments. C. Ichniowski and R. Freeman,
eds., Vfluai Public Ekaflxn: Workers Unionize. Chicago:
University of Chicago Press. 19—40.

Imrie, R.F. 1986. Work decentralisation from large to small
firms: ea preliminary analysis caf subcontracting.
Environment and Planning A. 18: 949-965.

 

Irving, John. 1987. TTmz Board’s representation. process:
another view. Charles (1.1Morris, ed., American Labor
Policy. Symposium on the Labor Board at Mid-Century.
Washington, D.C: BNA. pp. 108-114.

Irving, John. 1984. Failure of labor laws: an election year
issue. SpeeCh to Midwest Labor Law Conference. Daily
Labor Report, October 25, pp. E1—E6.

Irving, John. 1983. Plant relocation and transfers of work:
the NLRB’s ’inherently destructive’ approach." Labor Law
Journal. September: 549—562.

Irving, John. 1982. Closing and sales of businesses: a settled
area? Labor Law Journal. April: 219—229.

Iverson, Gudmund. 1991. Contextual Analysis. Beverly Hills,
CA: Sage Publishing.

Jaccard, James; Robert Turrisi and Choi K. Wan. 1990.
Interaction Effects :hi Multiple: Regression” Beverly
Hills, CA: Sage Publishing.

Jaffee, David. 1986. The political economy of job loss in the
United States, 1970-1980. Social Problems 33: 297-315.

 

Jaffee, David. 1988. The political-economic environment and
tfluegeographic restructingcfifmanufacturing: theoretical
perspectives and a state level analysis. Research in
Politics and Society. 3: 85—108.

Johnson, Haynes. 1991. Sleepwalking Through History: America
in the Reagan Years. New York: Norton.

Jones, Ethel B. 1992. Private sector union decline and
structural employment change, 1970—1988..JournalcflfLabor
Research. 13: 257—272.

 

 

261

Katz, Harry. 1993. The decentralization of collective
bargaining: a literature review and comparative analysis.
Industrial and Labor Relations Review. 47: 3—21.

Katz, Harry and John Paul MacDuffie. 1994. Collective
Bargaining in the U.S. Auto Assembly Sector. P. Voos, ed.
Contemporary Collective Bargaining intfluaPTivate Sector.
Madison, WI: IRRA. pp. 181-224.

Kaufman, Bruce. 1993. The Origins and Evolution of the Field
of Industrial Relations in the United States. Ithaca, NY:
ILR Press.

Kaufmann, E2. Scott. 1991. Ptditicization of DHJUB doctrine:
costs imposed by institutional alternatives in the

administration of the NLRA. Journal of Law and Politics.
20: 123—163.

Keating, Daniel. 1994. The continuing puzzle of collective
bargaining agreements in bankruptcy. Willianland Mary Law
Review. 35: 503—549.

Keefe, Jeffrey and George Kohl. 1993. Technological change and
labor’s interests. Workplace Topics. 3: 1—32.

Kelley, Maryellen R. and Bennett Harrison. 1990. The
subcontracting behavior of single vs. multiplant
enterprises in US manufacturing: implication for
economic development. World Development. 18: 1273—1294.

Kerr, Clark; Dunlop, Lkflui T.; Harbison, Frederick PL. and
Charles A.D®mnxx 1960. Industrialisn1and Industrial Man.
New York: Oxford University Press.

 

Kirkland, Lane. 1978. Speech before Work in America
Institute, 13mm, Harriman, DD? May’ 24. Daily' Labor

Report. May 30, p. D-l.

 

Klare. Karl E. 1981. Labor law as ideology: toward a new
historiography of collective bargaining law. Industrial
Relations Law Journal. 4: 452.

 

Kochan, Thomas A. 1980. Collective Bargaining and Industrial
Relations. Homewood, IL: Richard D. Irwin.

 

Kochan, Thomas, 1979. How American workers view labor unions.
Monthly Labor Review. 102: 23-31.

 

262

Kochan, Thomas A., Katz, Harry and Robert B. McKersie. 1986.
The Transformation of American Industrial Relations. New
York: Basic Books.

Kochan, Thomas A. and Paul Osterman. 1994. The Mutual Gains
Enterprise. Cambridge: Harvard Business Press.

Kohler, 'Phomas <3. 1983. Distinctitnl without <differences:
effects bargaining in light of First National
Maintenance. Industrial Relations Law Journal. 5: 402—
425.

Kokkenelenberg, Edward C. and Donna R. Sockell. 1985. Union
membership in the United States, 1973—1981. Industrial
and Labor Relations Review. 38: 497—543.

Kovach, Kenneth .A. aumi Peter E. Inillspaugh. 1987. Plant
closings: is the American industrial relations system
failing? Business Horizons. 30: 44—49.

Krotsen, Richard Vanl4.1984. Judicial and arbitral resolution
of contractual plant closing issues. Labor Law Journal.
July: 393-406.

Kurtz, Jannell M.; Walter, Robert J. and.Wayne R. Wells. 1990.
Rejection of collective bargaining agreements in
bankruptcy: a review, update, and guide for debtors.
Commercial Law Journal. 96: 31—44.

LaLonde, Robert J. and Benard Meltzer. 1991. Hard times for

 

unions: another look at the significance of employer
illegalities. University of Chicage Law Review. 58: 953-
1014.

Langbein, Laura and Allan Lichtman. 1978. Ecological

 

Inference. Beverly Hills, CA: Sage Publishing.

 

Lansing, Paul and Deb Podhasky. 1990. The conservative shift
in U.S. Supreme Court labor law decisions during the
1988-1989 term. Labor Law Journal. February: 67-79.

 

Lawler, John, 1990. Unionization and iDeunionization:
Strategies, Tactics and Outcomes. Columbia, S.C:
University of South Carolina Press.

 

263

Lawler, John. and. Gregory Hundley. 1983. Determinants of
certification and decertification activity. Industrial
Relations. 22: 335—348.

Leap, Terry L., Hendrix, Williaan., Cantrell, R. Stephen, and
G. Stephen Taylor. 1990. Inscrimination against pro—
union job applicants. Industrial Relations. 29: 469-478.

Lepo, Christine L. 1992. The Impact of Labor Leader
Pronouncements on the Voting Behavior and Issue Opinions
of Union Members. Unpublished doctoral dissertation,
School of Labor and Industrial Relations, Michigan State

University.
Leroy, Michael H. 1991. Strike crossovers and striker
replacement: an empirical test of the NLRB’s no

presumption policy. Arizona Law Review. 33: 291-335.

Levine, Marvin. 1989. Double-digit decertification election
activity: union organization weakness in the 1980s. Labor
Law Journal. May: 311-315.

Lipset, Seymour M. 1986. Labor unions in the public mind. S.M.
Lipset, ed., Unions in Transition. San Francisco:
Institute for Contemporary Studies.

Lubbers, William A. 1981. Office of the General Counsel
Memorandun181—38, July 14, 1981. Daily Labor Report. July
27: F-l-2

Lubbers, WilliaHIZX. 1981a. Office of time General Counsel
Memorandum 81—57, November 30, 1981. Daily Labor Report.
January 15: E—1—2.

Mancke, R.B. 1971. American trade union growth, 1900—1960: a
comment. Quarterly Journal of Economics 85: 187-93.

McClain, Anne J. 1991. Bankruptcy Code Section 1113 and the
simple rejecthmi of collective bargaining agreements:

labor loses again. The Georgetown Law Journal. 80: 191-
210.

 

McCulloch, Frank and Tim Bornstein. 1974. The National Labor
Relations Board. New York: Praeger.

264

Meany, George. 1975. Address by AFL—CIO President Meany at
NLRB banquet held on October 30 in Washington observing
40th anniversary of National Labor Relations Act. Daily
Labor Report. October 31: D—l—D—3.

Miller, Edward. 1984. Statements before the U.S. House of
Representatives Committee on Government Operations.
Oversight Hearings on the Subject "Has Labor LaW'Failed?"
Part I. Washington, D.C. U.S. House of Representatives.
p. 190.

Millis, Harry A. and Emily Clark Brown. 1950. From the Wagner
Act to 'Laft—Hartley. Chicago: Universityr of Chicago
Press.

Mills, D. Quinn. 1980. Construction. Gerald G. Somers, ed.,
Collective Bargaining: Contemporary American Experience.
Madison, WI: IRRA.

Millspaugh, Peter E. 1984. Midterm plant relocations: The
NLRB puts Humpty back together again. Labor Law Journal
May: 289—300.

Moe, Terry. 1985. Control and feedback in economic
regulation: the case of the NLRB. The American Political
Science Review. 79: 1094—1116.

Moe, Terry. 1987. Interests, institutions and.positive theory:
the politics of time NLRB. Annual Review CHE Political
Science. 2: 236—299.

Moody, Kim. 1988. An Injury to All: The Decline of American
Unionism. London: Verso.

 

Moon, Christie A. 1987. Technology, robotics, and the work
preservation doctrine: future considerations for labor
and management. Pepperdine Law Review. 14: 403-420.

Moore, William J. and Robert J. Newman. 1988. A cross—section
analysis of the postwar decline in American trade union
membership. Journal of Labor Research. 9: 111—125.

 

Moore, William J. and Robert J. Newman. 1985. The effects of
right-to—work laws: 53 review CHE the literature.
Industrial and Labor Relations Review. 38: 571—585.

 

265

Moore, WillianlJ. and Robert J. Newman. 1975. On the prospects

 

for American trade union growth: a cross-section
analysis. Review of Economics and Statistics. 57: 435-
445.

Moore, William J. and D.K. Pearce. 1976. Union growth: A test
of the Ashenfelter-Pencavel model. Industrial Relations.
15: 244-247.

 

Murphy; Betty :3. 1977. Letter' from PHJUB chairman.IBetty
Southard Murphy to member John Fanning summing up
administrative developments during her term as chairman.
Daily Labor Report. April 14: D—l

Murphy, William. 1987. Establishment and disestablishment of
union representation. Charles J. Morris, ed., American
Labor Policy; Symposiun1 on the Labor" Board. at lMid-
Century. Washington, D.C: BNA. pp. 61-78.

Murphy, Sheila. 1988. A comparison of the selection of
bargaining representatives in the ‘United States and
Canada: Linden Lumber, Gissel and the right to challenge
majority status. Comparative Labor Law Journal. 10: 65-
97.

Nagel, Stuart EL 1969. Regulatory commissioners and party
politics. The Legal Process fronl a Behavioral
Perspective. Homewood, IL: Dorsey Press. pp. 276-284.

National Labor Relations Board. 1983. Transcript of May 4,
1983 meeting of the Board. Daily Labor Report. June 15:
F—1—F-15.

 

National Ixfixnr Relations jBoard. 1973—1990. Annual. Reports
Washington, D.C.: GPO.

Neumann, George R. and Ellen R. Rissmam. 1984. Where have all
the union members gone? Journal of Labor Economics. 2:
175—192.

 

O'Connell, Francis A. Jr. 1986. Plant Closings: Workers
Rights, Management Rights and the Lawn New BrunSwick, NJ:
Transaction Books.

 

Oliver, Marianne. 1986. NLRB Crackdown on unions: union fines
and sympathy strikes." Duguesne Law Review. 24: 895—
922.

266

Pendergast, John J. 1993. Buying insolvent businesses and
labor law—-caveat emptor. Unpublished paper presented at
the Sixth Annual Rhode Island Labor Law Conference, West
Greenwich, R.I.

Perlman, Selig. 1928. A filmmnwrcof the Labor Movement. New
York: Augustus M. Kelley.

Perucci, Carolyn C.; Perucci, Robert; Targ, Dena and Harry R.
Targ. 1988. Plant Closings: International Context and
Social Costs. New York: Aldine De Gruyer.

Pfeffer, Jeffrey and James N. Baron. 1988. Taking the workers

 

back out: recent trends in the structuring of
employment. Researtflin<Organizational Behavior; 10: 257-
303.

Pierce, Ellen R. 1982. Employer participation in the
decertification process: how big a helping hand? Buffalo
Law Review. 31: 737-795.

Piore, Michael and Charles E. Sabel. 1984. The Second
Industrial Divide. New York: Basic Books.

Popper, Karl. 1957. The Poverty of Historicism. Boston:
Beacon.

 

Portz, John. 1990. The Politics of Plant Closingg. Lawrence,
KS: University of Kansas.

Premack, Stephen L. and John Hunter. 1988. Individual
unionizaticwldecisions. Psychological Bulletin.ZU13:223-
234.

 

Prosten, Richard. 1979. The longest season: union organizing
in the last decade, aka how come one team has to play
with its shoelaces tied? Proceedings of the Thirty—first
Annual Meeting. Madison, WI: IRRA.

 

Rainbird, Helen. 1988. New technology, training, and union
strategies. Richard.Hyman and Wolfgang Streeck, eds., Ngw
Technology and Industrial Relations. LondOn: Basil
Blackwell. pp. 174—188.

 

267

Rappa, Nancy M. 1986. Clarifying the work preservation/work
acquisition dichotomy under sections 8(b)(4)(B) and 8(e)
of the National Labor Relations Act: The National Labor

 

Relations Board v. International Longshoremen’s
Association. Catholic University Law Review. 35: 1061-
1097.

Reagan, Ronald. 1990. An American Life. New York: Simon and
Schuster.

Remy, Dorothy and Larry Sawers. 1984. Urban industrial decline
and the dynamics of sexual and racial oppression. Larry
Sawers anxi Willian1 Tabbn eds., Sunbelt/Snowbeltl New
York: Oxford University Press. pp. 128—151.

Roberts, Karen and Susan Zonia. 1994. Workers compensation
costs and health care providers income maintenance.
Journal of Risk and Insurance. 61: 117-131.

Roomkin, Myron and Richard A. Block. 1981. Case processing
time and the outcome of representation elections: some
empirical evidence. University of Illinois Law Review.
1: 75-97.

Rosenberg, Alicia (3. 1984. Automation and tine work
preservation doctrine: accommodating productivity and job
security interests. UCLA Law Review. 32: 135—176.

Rothstein, Lawrence. 1986. Plant Closings: Power, Politics and
Workers. Dover, MA: Auburn House.

Sayrs, Lois. 1989. Pooled Time Series Analysis. Beverly Hills,
CA: Sage Publishing

Scher, Seymour. 1956. The National Labor Relations Board and
Congress-~A Study of Legislative Control of Regulatory
Activity. ‘Unpublished doctoral dissertation, Department
of Political Science, University of Chicago.

 

Schupp, Robert W. 1991. When is a union not a union? Good-
faith doubt by an employer . Labor Law Journal . June:
357—365. '

Siegel, Sidney. 1956. Non—parametric statistics. New York:
McGraw-Hill.

268

Sheflin, Neil; Troy, Leo and Timothy Koeller. 1981. Structural
stability 111 models (ME American trade LHULHI growth.
QuarterlyfiJournal of Economics. 96: 77—88.

Skocpol, Theda. 1985. Bringing the State Back In. Cambridge:
Cambridge University Press.

 

Skocpol, Theda. 1984. Vision and Method in Historical
Sociology. Cambridge: Cambridge University Press.

Slaugther, Jane. 1983. Concessions: How'to Beat Thenu Detroit:
Labor Education and Research Project.

Slichter, Sumner. 1947. Trade 'Unions in ea Free Society.
Cambridge, MA: Harvard University Press.

Smith, Christopher. 1993. CourtsL Politics and the Judicial
Process. Chicago: Nelson Hall.

Sockell, Donna. 1993. iflua duty tx> bargain enmi technology.
Workplace Topics. 3(1): 55—66.

Sockell, Donna. 1989. The future of labor law: a mismatch
between statutory' interpretation. and :reality? Boston
College Law Review. 50: 987—1026.

Sockell, Donna. 1986. The scope of mandatory bargaining: a
critique auui proposal. Industrial auui Labor Relations
Review. 40: 19—34.

Sockell, Donna and John Thomas Delaney. 1986. Who wins if
fewer items and mandatory subjects of bargaining? Labor
Studies Journal. 11: 101—116.

 

Sockell, Donna and John Thomas Delaney. 1987. Union organizing
and the Reagan NLRB. Contemporary Policy Issues. 5: 28—
45.

Sommers, Albert T. 1993. The'U.S. Economy Demystified, 3rd ed.
New York: Lexington Books.

Somson, Barbara. 1984. Duty to bargaing over plant closings
relocations: a union. view. ‘Willianl Dolson and Kent
Lollis, eds., First Annual Labor and Employment Law
Institute. Littleton, CO: Fred B. Rothman and Co. pp.
155-180.

269

Steffen, John F. 1989. The work preservation exception to
sections 8(b)(4)(B) éUKi 8(e) ‘as a: means tx) control
technology caused worker displacement: a doctrine whose
utility has come and gone. Detroit College of LaW'Review.
3: 1133—1174.

 

Stepina, Lee and Jack Fiorito. 1986. Toward a comprehensive
theory<3finfirxigrowth.and«decline. Industrial Relations.
25: 248—264.

Stone, Katherine Van Wezel. 1981. The post—war paradigm in
American labor law. The Yale Law Journal. 90: 1509—1580.

Storper; Michael and. Richard lMalker. 1984. The spatial
division of labor: labor and the location of industries.
Larry Sawers and William K. Tabb, eds. Sunbelt/Snowbelt
New York: Oxford University Press.

Stryker, Robin. 1989. Limits on technocratization of the law:
the elimination of the National Labor Relations Board
Division (fl? Economic Research. American Sociologial
Review 54: 341—358.

Summers, Clyde. 1965. Labor law in the Supreme Court. Yale Law
Journal. 75: 59—88.

Tomlins, Christopher 1;. 1985. The State enui the 'Unions.
Cambridge: Cambridge: Cambridge University Press.

Tomlins, Christopher 1;. 1985a. The ruwv deal, collective
bargaining, and time triumph CHE industrial pfluralism.
Industrial and Labor Relations Review. 39:19-34.

 

Troy, Leo. 1990. Is the U.S. unique in the decline of private
sector unionism?. Journal of Labor Research. 9: 111—143.

 

Troy, Leo. 1969. Trade union growth and the changing economy.
Monthly Labor Review 92: 3—7.

 

Trumka, Richard L. 1987. Why labor law has failed. West
Virginia Law Review. 89: 871—882.

Tuma, Nancy B., Hannan, Michael T. and Lyle P. Groeneveld.
1979. Dynamic analysis of event histories. American
Journal of Sociology. 84: 820-854.

270

General Accounting Office. 1986. U.S. Business Closures
and Permanent Layoffs During 1983 and 1984. Washington,
D.C: GPO.

House (ME Representatives. 1988. Hearings before the
Committeeeon Goverment Operations. 1988 Oversight of the
National Labor Relation Board. July 13. Washington D.C:
GPO.

House CHE Representatives. 1987. Hearings before the
Committee on‘Goverment Operations. 1987 Oversight of the
National Labor Relation Board. July 21. Washington D.C:
GPO.

Iknmxa of Representatives. 1986. Hearings before the
Committee (n1 Goverment Operations. Oversight (Hi the
National Labor Relation BoardL Part 2. July 17.
Washington D.C: GPO.

House CNS Representatives. 1985. Hearings before the
Committee (M1 Goverment cmmxations. Oversight (Hi the
National Labor Relation Board. June 18. Washington D.C:
GPO.

. House of.Representatives. 1984. Joint hearings before the
Committees on Education and Labor and Government
Operations. Oversight hearings on the Subject "Has Labor
Law Failed?" June 21, 25 and 26. Washington, D.C: GPO.

House of Representatives. 1984a. The Failure of Labor
Law: A Betrayal of American Workers. Washington, D.C.:
GPO.

House of Representatives. 1984b. Hearings before the
Committee on Goverment Operations. National Labor
Relation Board Case Backlog, Part 2. May 9. Washington
D.C: GPO.

. House of Representatives. 1983. Joint hearings before the
Committees on Education and Labor and Government
Operations. CNersight CH1 the Rational Imfixnr Relation
Board. June 15 and 29. Washington, D.C: GPO.

House of Representatives. 1983a. Hearings before the
Committee on Goverment Operations. National Labor
Relation Board Case Backlog. November 2. Washington D.C:
GPO.

 

271

U.S. House of Representatives. 1976. Hearings before the
Committee on Education and Labor. Oversight on the
National Labor Relation Act. February 17, 19; March 15,
16, 22, 29, 30, 31; April 6, 7; May 3, 4, and 5.
Washington, D.C: GPO.

U.S. House of Representatives. 1975. Hearings before the
Committee on Education and Labor. Oversight on the
National Labor Relation Board. ihflqr 15; October, 23;
November 13enuil8; Novemeber 216NK122. Washington, D.C:

 

GPO.

U.S. Senate. 1987. Committee on Labor and Human Resources.
Hearings on the Economic Dislocation and Worker
Assistance Act of 1987. mashington, D.C.: Government

 

Printing Office.

Walther, Peter D. 1985. The NLRB today. Labor Law Journal. 36:
803—816.

 

Wren, John and Kent Murrman. 1990. Chapter 11 and collective
bargaining agreements. Employee Relations Law Journal.
16(1): 17—27.

Verma, Anil and Thomas A. Kochan. 1985. The growth and nature
of the nonunion sector within a firm. T. Kochan, ed.,
Challenges and Choices Facing American Labor. Cambridge,
MA: MIT Press. pp. 89—118.

Visser, Jelle. 1988. Trade unionism in Europe: present
situation and prospects. Labor and Society. 13: 125-182.

Wallace, Michael and Joyce Rothschild. 1988. Plant closings,
capital flight, and worker dislocation: the long shadow
of deindustrialization. Researtl1in Politics and.Society.
3: 1-35.

 

Walton, Richard, Cutcher-Gershenfeld, Joel and Robert
McKersie. 1994. Strategic Negotiations. Boston: Harvard
Business Press.

 

Watanabe, S. 1971. Subcontracting, industrialisation and
employment creation. International Labour Review. 105:
425-449.

 

Weber, Arnold. 1961. The Structure of Collective Bargaining.
Glencoe, IL: Free Press.

 

272

Wendling, Wayne. 1984. The Plant Closure Policy Dilemma.
Kalamazoo, MI: Upjohn Institute.

Weikle, R.D. and H. Wheeler. 1984, Unions and technological
change: attitudes and local union leaders. Proceedings
of the Thirty—Sixth Annual Meeting. Madison, WI: IRRA.

Wieler, Paul. 1990. Governing the Workplace. Cambridge, MA:
Harvard University Press.

Wieler, Paul. 1983. Promises to keep: securing workers’ rights
to self—organization under the NLRA. Harvard Law Review.
96: 418—476.

Winter, Ralph. 1968. Judicial review of agency decisions: the
labor board and the court. Supreme Court Review 53.

273

Wheeler, Hoyt and Lkflni A, McClendon. 1991. The individual
decision to unionize. G. Strauss, D. Gallagher and J.
Fiorito, eds., fflua State of time Unions. Madison, WI:
IRRA. pp. 47-84.

 

Yamaguchi, Kazuo. 1991. Event History Analysis. Newbury Park,
Ca.: Sage Publishing.

Zidich, Debra J. 1984. Robotics in the workplace: the
employers duty to bargain over its implementation and
effect on the worker. Santa Clara Law Review. 24: 917—
945.

 

MIC

$111111: