Why do unionized workers have more nonfatal occupational injuries?

Why do unionized workers have more nonfatal
occupational injuries?
Alejandro Donado1
University of Würzburg
Anecdotal evidence and labor unions’ activities suggest that unionized workers
should have less occupational injuries than non-unionized workers. A quantitative and qualitative analysis of the empirical literature reveals that unions seem to
be e¤ective in reducing fatal injuries, but that, contrary to expectations, unions
are clearly associated with more nonfatal injuries. Panel estimates at the individual level for the …rst time con…rm and reinforce the paradoxical union impact
on nonfatal injuries. The three most prominent explanations suggested in the
literature for this paradoxical result can be ruled out using econometric methods.
However, supporting empirical evidence is o¤ered for a novel explanation based on
the theory of behavioral adaptation. According to this, unions make workplaces
safer, but workers respond to this enhanced safety by changing their behavior
and reducing their own self-prevention activities. The net e¤ect is an increase in
nonfatal injuries.
JEL codes: J 51, J 28, J 81, C33
Keywords: labor unions, occupational health and safety, behavioral adaptation
1
Introduction
"There remain two puzzling results of the estimation of our model of coal mining injuries.
The …rst of these is the fact that unionized mines have higher non-fatal accident rates
than would be expected for non-union mines with the same characteristics. [. . . ]”(Boden
1977, p. 139)
"The absence of any evidence of a signi…cant union reduction of hazards runs counter to
the conclusion one might draw on the basis of one’s observation of actual union actions."
(Viscusi 1979, p. 231)
Labor unions have traditionally played an important role in improving occupational health
and safety. They have not only shaped workplaces’safety through collective bargaining and
other activities, they have also used their political in‡uence to develop and support the passage
1
Alejandro Donado: University of Würzburg, Department of Economics, 97070 Würzburg, Germany,
[email protected]. Phone: +49.931-31-82952, fax: +49.931-31-87025.
1
of government legislation like the Occupational Safety and Health Act in 1970 (Schurman et al.
1998, pp. 134-6). Historical anecdotal evidence also contains several prominent examples of
unions’safety-enhancing activities that include gaining recognition for occupational diseases
caused by exposure to coal dust (Smith 1987), cotton dust (Botsch 1993), asbestos (Rosner and
Markovitz 1991), radium (Clark 1997), and dibromochloropropane (Robinson 1991). Many
other historical examples could be mentioned.
In more general terms, labor unions have in‡uenced occupational health and safety outcomes in several important ways. These include the provision of job hazard information, the
protection of workers who refuse to accept hazardous assignments, and the assistance and
representation of workers in accident compensation claims. Moreover, apart from in‡uencing
the regulatory process and its enforcement, unions bargain for the provision of protective
equipment, for compensatory wages, and for the establishment of joint union-management
health and safety committees.2
All these examples based on anecdotes and unions’activities suggest that strongly unionized workplaces should be safer and that because of this, union workers should have less
occupational injuries and illnesses than nonunion workers.3 However, in spite of these examples, the empirical literature draws a totally di¤erent picture. Most of the empirical research
has failed to provide su¢ cient "hard" evidence showing that labor unions are e¤ective in
reducing occupational injuries. In fact, most of the studies associate labor unions with more,
not less, occupational injuries. This is a result that has puzzled many researchers, as the
quotes at the beginning of this introduction illustrate.4
The result is the more puzzling since there is empirical evidence showing that unionized
workers earn higher wages (Lewis 1986) and receive more and better fringe bene…ts such as
pensions, vacation pay, and life, accident and, health insurance (Freeman and Medo¤ 1980, ch.
4). Moreover, unionized workers are more likely to receive unemployment insurance bene…ts
(Budd and McCall 1997, 2004), to receive premium pay for overtime hours (Trejo 1991, 1993),
and are less likely to work overtime (Trejo 1993). In general, based on the empirical evidence,
it appears as if unions were doing everything just right for their members except for reducing
occupational injuries.
In section 2 of this paper, a qualitative and quantitative analysis of the empirical literature
based on 25 studies is presented. This analysis reveals that the impact of unions on occupational injuries seems to be di¤erent if fatal or nonfatal unions are considered. The analysis
suggests that, according to expectations, unions tend to be e¤ective in reducing fatal injuries.
However, when only nonfatal injuries are considered, it looks as if unions were causing more
2
See Robinson (1991, p. 40), Beaumont (1983, p. 2), Viscusi (1979, pp. 230-1), Dorman (1996, pp. 131-4)
and Schurman et al. (1998).
3
Henceforth, the term "injuries" will be used to refer to both "injuries and illnesses".
4
Very similar sentences can be found in Chelius (1974, p. 727), Boden (1985, p. 500), Fishback (1986,
p. 290), Fairris (1992, p. 205), Reardon (1996, p. 239), Smitha et al. (2001, p. 1007), and Robinson and
Smallman (2006, p. 101).
2
nonfatal injuries. Moreover, the paradoxical positive union impact on nonfatal injuries seems
to be robust in the sense that it holds for di¤erent aggregation levels, di¤erent data sets,
di¤erent countries and di¤erent estimation techniques.
Section 3 extends the empirical literature by providing the …rst estimates of the union
impact on nonfatal injuries using panel data at the individual level. The data comes from the
National Longitudinal Survey of Youth 1979 (NLSY79). This data set does not have many
of the limitations of the data sets previously used in the literature. In particular, apart from
its panel nature and its extensive set of questions on personal and job characteristics, the
NLSY79 also contains detailed information on both nonfatal injuries and union status of the
respondents. All previous studies were obliged to match injury rates at the industry level
from another data source to each individual for which they had information on their union
status and other characteristics. The NLSY79, however, allows us to calculate the probability
of having an injury based on each individual’s own experience and not on an average of the
industry where they work. According to these new panel estimates, the probability of having
a nonfatal injury is around 30% higher for union workers. Thus, the paradoxical result of
unions on nonfatal injuries from the literature is con…rmed and reinforced.
After having documented the paradoxical results from the literature, the question this
paper addresses is: Why if we expect, based on anecdotal evidence and unions’ activities,
unions to reduce nonfatal injuries is the empirical literature failing to provide "hard" evidence
of this reduction? Why are labor unions constantly been associated with more nonfatal
injuries? Several potential explanations for this paradoxical result have been suggested, but
the empirical evidence supporting or rejecting them is very limited. Section 4 studies the
three most prominent explanations from the literature.
The …rst explanation considered is called "reporting". According to this explanation,
unions are believed to reduce the number of actual nonfatal injuries but also to increase
the number of injuries that are reported. Since most data sources are not based on actual
but on reported injuries, unions appear to be associated with more injuries in most of the
cases. However, the strongest piece of evidence that can be o¤ered against this explanation is
that the new panel estimates presented in this paper are based not on reported but on actual
injuries. These estimates clearly show that unionized workers have more actual injuries. This
rules out this explanation.
The second explanation is called "causality". Proponents of this explanation argue that it
is not labor unions that are causing more injuries, but that more injuries (or more hazardous
workplaces) are causing workers to form unions in order to improve their working conditions.
Since labor unions are then to be found where working conditions are poor and injuries
high, empirical studies tend to …nd a positive association between more injuries and unions.
This explanation, therefore, suggests that there might be a simultaneity problem between
injuries and unions that does not allow to identify the causal e¤ect of unions on injuries. In
order to break the simultaneity, several instrumental variables and estimation techniques are
employed in this paper. The results clearly show that, after accounting for the simultaneity
3
problem, labor unions have a causal increasing impact on nonfatal injuries. The "causality"
explanation can also be ruled out.
The third explanation considered is called "wages for safety". This explanation emphasizes
that unions can only control certain aspects of employment, while management can react by
varying other decision variables. Thus, if unions were more concerned about wages than
about health and safety measures, they would prefer to increase wages. Since both higher
wages and better safety measures are costly, management might react by reducing investment
in safety measures, leading to an increase in occupational injuries. As a result, unionized
workers would be paid higher wages at the expense of having more injuries. However, the
most rigorous test that this paper o¤ers against this explanation is a regression of injuries on
union status that holds wages …xed and controls for wage and union endogeneity. The idea
is to isolate the causal impact of unions on injuries net of any in‡uences of unions on wages.
The results against this explanation are also very clear and con…rm that unions are having
a causal increasing impact on nonfatal injuries. The "wages for safety" explanation can also
be dismissed.
After having ruled out the three most important explanations from the literature, we are
in need of an alternative approach. In section 5, this paper introduces to the literature a new
explanation based on the theory of behavioral adaptation. According to this theory, individuals reduce their levels of vigilance and increase their voluntary exposure to danger if their
perceived risk is decreased. Behavioral adaptations can for example occur as a response to the
introduction of a safety measure. If individuals perceive this measure as a danger reduction,
their response is to balance this reduction by increasing their own risk exposure, diminishing
or even o¤setting the intended bene…ts of the safety measure. There is considerable evidence
of this theory in the areas of road safety and public health (see section 5). Moreover, this
type of phenomenon is not completely unknown to economists who have found a vast amount
of empirical support for a related concept called moral hazard.
This paper argues that the theory of behavioral adaptation can be applied to solve the
paradox. In particular, this theory explains why unionized workers are having more occupational injuries despite all the safety-enhancing e¤orts made by labor unions. The reason is
that the safety measures and, in general, the better protection that unions o¤er, lowers workers’ risk perception, who in turn react by reducing their own self-prevention activities. In
other words, the safety measures provided by unions might lower the number of occupational
injuries, but workers might more than o¤set these e¤orts if they reduce their self-protection
activities. The net impact is an increase in occupational injuries.
There is indirect and direct evidence supporting this new explanation. The indirect evidence is based on the fact that some occupational health and safety measures that have
been traditionally supported by labor unions have not always reduced occupational injuries
as expected. In fact, the evidence from the empirical literature shows that the introduction of
protective equipment, the passage of occupational health and safety legislation, and the introduction of the workers’compensation system have not always produced satisfactory results in
4
terms of reducing occupational injuries. For example, it is well-known from one body of the
empirical literature that the workers’ compensation system has encouraged more, not less,
nonfatal occupational injuries. This has been attributed to moral hazard, as noted above,
a concept closely related to behavioral adaptation. Another di¤erent body of the empirical
literature shows that unionized workers use the workers’ compensation system more intensively and more successfully. Combining these two bodies of research suggests that, unions
are, probably unintentionally, supporting a system that is delivering suboptimal results.
The direct evidence is based on three new empirical exercises using the NLSY79 data set.
Only with a panel data set at the individual level, like the NLSY79, it is possible to uncover
behavioral adaptation, since information is needed on the same individual for at least two
periods. None of the previous studies was able to investigate this because of lack of these
particular data. The …rst exercise presented estimates the union impact on injuries after
controlling for the workers’average industry risk. Using additional data from the Bureau of
Labor Statistics it was possible to assign most NLSY79 respondents to more than 200 industryrisk groups for every year considered. It is shown that unionization increases occupational
injuries even after holding …xed the workers’ average risk exposure. The second exercise is
an analysis based on workers’union status changes. The main results not only con…rm that
workers’injury probability is increased after joining a union, the results also clearly show that
this injury probability is decreased after a worker leaves the union. Moreover, it is shown that
the longer a worker stays unionized, the higher the worker’s injury probability is. Finally, the
third exercise is based on information on respondents’health and life insurance coverage. It is
found that a higher health and safety protection, in the form of health or life insurance or of
union services, increases workers’injury probability non-negligibly. All these three exercises
provide strong evidence of behavioral adaptation.
This paper extends the research on union e¤ects on occupational injuries in four important
ways: First, it provides a quantitative and qualitative summary of the literature, clarifying the
di¤erent impact of unions on fatal and on nonfatal occupational injuries. Second, it presents
the …rst panel estimates at the individual level, exploiting the advantages of this type of
data. Third, it studies in depth (and rules out) the three major explanations suggested in
the literature for the positive association between unions and nonfatal injuries. And fourth,
it introduces a new explanation for the paradox and provides direct and indirect evidence
supporting it.
2
Evidence from the empirical literature
This section surveys the empirical literature investigating the impact of labor unions on
occupational injuries. These studies usually estimate an equation of the form
IN JU RY =
U N ION + X0 + u;
5
(1)
6
7
where IN JU RY is some measure of the number, frequency, or severity of occupational injuries, U N ION is a variable indicating union status, X is a vector of control variables, and
u is the error term. The impact of unionism on occupational injuries is thus given by the
estimate of . Based on historical anecdotal evidence and on unions’activities, one should
expect unions to have a signi…cant impact in reducing injuries, that is, the coe¢ cient is
expected to be negative and signi…cant.
Table 1 summarizes 25 studies that have estimated some variation of the injury-union
regression (1). Of these studies, 10 have focused primarily on the impact of unions on injuries.
These are marked with an asterisk on column one. The other studies were interested in a
di¤erent question, but included a U N ION variable as one of the regressors. Table 1 also shows
the remarkable heterogeneity of these studies, encompassing di¤erent countries, industries,
years considered, data types, cross-sectional units, number of observations, and measures of
the U N ION and IN JU RY variables.
The most important result in this section, however, comes from column 9 of Table 1. This
column summarizes the type of IN JU RY variable used in each study and, in parenthesis, the
impact that the U N ION variable had on injuries. Only the estimates that used a measure
of fatal (FAT) or nonfatal injuries (NFI) for the IN JU RY variable were considered.5 Some
authors reported multiple estimates of . This is typically done to experiment with di¤erent
regression speci…cations, for sensitivity analysis, or when the dependent variable or the data
sets are di¤erent. For each di¤erent IN JU RY variable, the estimates that the author seemed
to judge as the best were chosen giving a total sample of 43 observations. Some key proportions
of the total sample are summarized in Table 2.
Table 2 shows that the 43 estimates pertain to only three countries, the USA being by far
the most prominent with 70% of the estimates, followed by the UK (26%), and Canada (5%).
5
All studies containing only estimates using a di¤erent measure like the severity of the injuries, workers’
compensation claims or bene…ts, or working conditions were excluded from the table.
8
Furthermore, more than half of the estimates are either for the coal mining industry (44%)
or the manufacturing (12%) industry. Other industries amount to 44% of the estimates.
Concerning the type of data, more than half (58%) of the estimates were based on crosssectional data, while panel data and time-series data were respectively used in 37% and 5%
of the cases. The injury-union regression has been estimated at di¤erent aggregation levels:
While most of the studies compared establishments (65%), other studies also estimated the
equation using data at the individual level (16%), industry level (7%), or regional level (12%).
Another interesting aspect is that the U N ION variable was in most cases measured using
information on either union membership (84%) or on union coverage (12%) and that the
IN JU RY variable used was either nonfatal injuries (74%) or fatal injuries (26%).
In order to better illustrate the union impact on injuries based on the estimates summarized in column 9 from Table 1, a signi…cance scale was constructed using the p-values, p,
associated to each coe¢ cient according to the following formula
sigscale =
1
(1
p if
p) if
>0
<0
Since the range of the p-value is between zero and one, the range of the signi…cance scale is
between -1 and 1. Remember that, loosely speaking, is the "more signi…cant", the lower the
p-value is. Moreover, the null hypothesis that is insigni…cant (i.e. = 0) is usually rejected
when p > 0:05. As a consequence, is negative and signi…cant if sigscale 2 [ 1; 0:95] and
is positive and signi…cant if sigscale 2 [0:95; 1]. Put simply, the lower the sigscale is, the more
negative and signi…cant is, and the higher the sigscale is, the more positive and signi…cant
is.
9
Figure 1 shows a histogram of the signi…cance scales based on all 43 estimates. The …rst
bar on the left shows that in 5 of the 43 regressions,
was negative and signi…cant. In
contrast to this, the last bar on the right shows that in 16 of the regressions, was positive
and signi…cant. All bars in between show that in 22 of the cases, was insigni…cant. What can
we conclude from this histogram? Since, based on anecdotal evidence and unions’activities,
we were expecting to be negative and signi…cant, the results are clearly disappointing. Only
in 5 of the 43 cases, labor unions were signi…cantly associated with fewer injuries.
A very interesting pattern, however, emerges in Figure 2 where the signi…cance scales for
fatal and nonfatal injuries are considered separately. The left histogram gives a picture that
is more according to expectations and suggests that labor unions are in most cases associated
with fewer fatalities. In contrast to this, the right histogram gives a very puzzling result
suggesting that in most cases labor unions are associated with more nonfatal injuries. In fact,
the right histogram indicates that labor unions are positively associated with more nonfatal
injuries in 84% of the estimates and this association is statistically signi…cant in 50% of the
estimates. More surprising is the fact that the negative and signi…cant association between
unions and nonfatal injuries that we were expecting based on anecdotal evidence and unions’
activities was found in only one single study! The weighted average of the signi…cance scales,
weighting by the total number of observations each estimate is based on, is -0.755 for fatalities
and 0.093 for nonfatal injuries. These numbers, taken at face value, imply that unions tend to
reduce fatalities, although not signi…cantly, and tend to increase nonfatal injuries, signi…cantly
at the 7% level.
10
The most important conclusion that can be drawn from the existing empirical literature
is, therefore, that the impact of unions on injuries appears to be di¤erent depending on the
type of injury studied. While the association between unions and nonfatal injuries is in most
cases positive, the association between unions and fatal injuries is less clear but it seems to
be more negative than positive. Our expectations regarding the impact of unions on injuries
are only (partially) con…rmed for fatal injuries. For nonfatal injuries, the empirical literature
clearly contradicts our expectations.
3
New evidence from panel data at the individual level
This section extends the empirical literature by providing estimates of the injury-union equation (1) using panel data at the individual level for the …rst time. The data come from the
National Longitudinal Survey of Youth 1979 (NLSY79). This survey was administered for
the …rst time in 1979, interviewing a representative sample of 12,686 American young men
and women aged between 14 and 22 years. Until 1994, the cohort was interviewed every
year. Since then, the survey has been conducted on a biennially basis. The analysis will
be restricted to the years for which information was available for all relevant IN JU RY and
U N ION variables. These years are 1988, 1989, 1990, 1992, 1993, 1994, 1996, 1998, and 2000,
corresponding to the period in which the respondents were aged between 23 and 44 years.
The major advantage of this survey is that it provides detailed data on occupational
injuries, on union status, and on an extensive set of questions on personal and job characteristics. The richness of the NLSY79 data makes possible to study the union impact on injuries
at a depth that has not been possible before using other data sets. There are at least three
reasons for this. First, as Table 1 shows, all previous estimations at the individual level where
based on cross-sectional data. This type of data has several limitations. In particular, it
only allows to make comparisons across individuals, and it is not possible to follow the same
person over time. As we will see in section 5, it is only with panel data at the individual
level that is possible to study behavioral adaptation issues. Second, in none of the data sets
used before there was information on both IN JU RY and U N ION variables. Researchers
were obliged to match injury rates at the industry level from another data source to each
individual for which they had information on their union status and other characteristics.
The NLSY79, however, allows us to calculate the probability of having an injury based on
each individual’s own experience and not on an average of the industry where they work.
Third, the NLSY79 data set is the only one that has information on both union membership
and on union coverage. This gives us two possibilities for measuring the U N ION variable.
Summary statistics for the IN JU RY and U N ION variables used to estimate the injuryunion equation (1) are reported (among other variables) in Table 3. The IN JU RY variable
is based on a question on nonfatal injuries (NFI). As Table 3 shows, on average, 6% of the
respondents reported having had a nonfatal work-related injury or illness in the period con11
sidered. The variables used to measure union status are union membership (MEMBERSHIP)
and union coverage (COVERAGE). In general, not all workers covered by a union contract
are members of a union. In fact, as Table 3 illustrates, an average of 18.7% of the respondents
was covered by a union contract while only 14.2% was member of a labor union.
Table 4 presents the estimates of the U N ION coe¢ cient from equation (1) using the two
union status measures (COVERAGE and MEMBERSHIP) and with nonfatal injuries (NFI)
as the dependent variable. The estimates reported include an extensive list of control variables
containing measurements of the individuals’ health, job satisfaction, tenure with employer
and its square, …rm size, hours per week worked, years of education, number of children, age,
and dummies for 8 years, marital status, type of residence, 3 regions, 11 industries, and 11
occupations, for a total of 44 control variables6 (see app. A.1 for complete de…nitions and
summary statistics). Only the estimates based on a linear probability model are reported in
this paper. Other models, like the logit, yield very similar results but are more di¢ cult to
interpret. All estimations are by …xed e¤ects.7
6
Gender and race dummies were excluded from the analysis since time-invariant variables are not identi…ed
by the estimation techniques used.
7
A Hausman test clearly rejects the random e¤ects model.
12
Table 4 gives a very clear picture of the impact of labor unions on nonfatal occupational
injuries (NFI). Irrespective of the U N ION measure used, unions are clearly associated with
more nonfatal injuries, after controlling for an extensive set of personal and job characteristics. The U N ION coe¢ cient is positive and highly signi…cant in both estimations, con…rming
and reinforcing the pattern of the empirical literature summarized in the previous section.
Turning to the interpretation of the estimates, the probability of having an occupational
injury increases by 0.016 for covered workers and by 0.02 for union members. These values
are not small. In fact, one way to put these values into perspective is by comparing them with
the predicted injury probabilities of the not-covered and of the nonunionized workers which
are 0.058 and 0.059, respectively. These predicted probabilities can then be used to compute
the injury probability di¤erential which is equal to 0:016=0:058 = 28% for the covered workers
and 0:02=0:059 = 34% for the unionized workers. In other words, the probability of having
an occupational injury increases by 28% for workers that change their union status from not
covered to covered and by 34% if they change from nonmember to member.
4
Explanations from the literature
The results from the previous two sections provide clear evidence of the positive association
between labor unions and nonfatal injuries. In this section, we will try to understand why.
As the last column in Table 1 shows, the literature has suggested several explanations for
this paradoxical result. There are, in particular, three explanations that appear to be gaining
some consensus among researchers. From the 25 studies summarized in Table 1, "reporting"
was mentioned in 11 studies, "causality" in 7 studies, and "wages for safety" in 4 studies.
This section will explore these three explanations in turn.
4.1
Reporting
The explanation most often mentioned in Table 1 is reporting (REP). According to this
explanation, unions are believed to reduce the number of actual nonfatal injuries but also to
increase the number of injuries that are reported. Since most data sources are not based on
13
actual but on reported injuries, unions appear to be associated with more injuries in most of
the cases.
There are at least two reasons why unions might increase the number of reported injuries.
First, at the establishment level, unions might better monitor the reporting of injuries by
employers. Second, at the individual level, unionized workers might simply report more
injuries because they are more knowledgeable of potential work hazards and because they
might be less fearful of management retaliation.
One interesting aspect of the reporting explanation is that it also seems to be consistent
with the likely negative association between fatalities (FAT) and unions presented in Figure 2.
Unions might reduce both actual fatalities and nonfatal injuries. However, since fatalities are
more di¢ cult to conceal and employers might be more inclined to report them, the di¤erence
between reported and actual fatalities might be presumably very small. Therefore, if unions
reduce actual fatalities, this will also translate into a reduction in reported fatalities. As a
consequence, the union-fatalities association will be negative as suggested by Figure 2.
Despite all these considerations, the evidence that can be provided against this explanation
is simple but conclusive. Contrary to the IN JU RY variables of all studies summarized in
Table 1, the NLSY79 IN JU RY variable is not based on reported but on actual injuries.8
In fact, respondents were asked questions on occupational injuries privately, and there is no
apparent reason for them to give inaccurate information. This means that the estimates of the
injury-union equation in Table 4 indicate a positive association between unions and actual,
not reported, injuries. The reporting explanation can be clearly dismissed.
4.2
Causality
As the literature summary in Table 1 shows, the second most important explanation after
reporting (REP) is causality (CAUS). The proponents of this explanation argue that it is
not unions that are causing more injuries, but more injuries (or more hazardous working
conditions) that are causing workers to join unions. In more technical terms, this means that
causation might be running in the opposite direction: not from U N ION to IN JU RY but
from IN JU RY to U N ION . Thus, according to this explanation, the positive association
between unions and injuries is due to the fact that unions are more likely to be found were
injuries are high and not because unions are causing more injuries.
One possibility to check the empirical validity of this argument is by estimating an equation
that investigates the other side of the causality, that is, the impact of having had an injury
on the probability of becoming a union member:
U N ION =
UI
IN JU RY + X0 + u:
8
(2)
The only exception is one of the two estimates with nonfatal injuries (NFI) as a dependent variable from
Worral and Butler (1983). They use for this a measure of actual, not reported, NFI.
14
The anecdotal evidence suggests that workers have traditionally favored union membership
as a mechanism to reduce workplace hazards. According to this evidence, U I is expected to
be positive and signi…cant. The estimates based on a linear probability model by …xed e¤ects
and including the full set of control variables are given in Table 5. As the table shows, all
estimates are positive and highly signi…cant, suggesting that the hazardousness of a workplace
might indeed be decisive when choosing to join a labor union or not. These results are also
supported by other empirical studies. Hirsch and Berger (1984), for example, …nd that higher
average industry injury rates signi…cantly increase the likelihood that a worker is a union
member. Duncan and Sta¤ord (1980) …nd the same pattern but not for injury rates but
for working conditions in general. Moreover, Robinson (1988, 1990) provides evidence that
individuals working under hazardous conditions are signi…cantly more likely to vote for union
representation.
The causality explanation and the results from Tables 4 and 5 suggest that the relationship
between unions and injuries might be a simultaneous one, that is, unionization in‡uences the
amount of injuries and the amount of injuries in‡uences unionization as well. In such a case,
the model can be speci…ed as a system of two equations using (1) and (2). This de facto
acknowledges the endogeneity of the U N ION variable in (1).
According to one of the de…nitions of endogeneity, the U N ION variable is endogenous
if it is correlated with the error term u in equation (1). This error term can be viewed as
having two components, one time-variant "t and one time-invariant , so that ut = + "t ,
where t indexes time. The …xed e¤ects estimation approach that we are using to estimate (1)
already controls for union endogeneity if U N ION is correlated only with the time-invariant
component of the error.9 In other words, if U N ION is only correlated with , the estimates
presented in Table 4 are indeed giving the size of the causal union impact on injuries and
the causality explanation could be ruled out.
However, what if U N ION is correlated with the time-variant component of the error? A
stricter approach that controls for this type of union endogeneity is based on instrumental
variable techniques. In fact, one way to break the simultaneity and to …nd the causal impact
9
See Wooldridge (2002) and Cameron and Trivedi (2005) for more details on this and on the estimation
techniques used in this paper.
15
of union on injuries is by estimating only (1) but by using an instrument for the U N ION
variable. A valid instrument has to satisfy two conditions. First, the instrument itself has to
be uncorrelated with the error term u in (1). And second, the instrument must be partially
correlated with the U N ION variable after controlling for the remaining exogenous regressors.
This type of exercise has been performed before in three of the empirical studies from the
literature in Table 1, but in all cases using not panel but cross-sectional data. Moreover, all
three studies employed British data at the establishment level. Fenn and Ashby (2004, pp.
475-6) and Robinson and Smallman (2006, p. 94) worked with the same British data set and
used the same instrument to test for possible endogeneity of the U N ION variable. Their
tests failed to reject the null hypothesis of union exogeneity, suggesting that their estimates
without controlling for union endogeneity were valid and unions were indeed signi…cantly
causing more nonfatal injuries. Contrary to this, Nichols et al. (2007, p. 218) found that
endogeneity was present. Their results after controlling for endogeneity were also positive but
insigni…cant.10
The NLSY79 data set o¤ers at least two possibilities to instrument for the U N ION
variable. The …rst possibility is based on empirical evidence suggesting that unionized workers
receive better fringe bene…ts than their nonunionized counterparts (Freeman and Medo¤ 1984,
ch. 4). Fortunately, the NLSY79 has detailed information on fringe bene…ts which can be
used as instruments. Three candidates from the NLSY79 that potentially ful…ll the two
requirements of a valid instrument are RETIREMENT, MATERNITY, and DENTALINS.
Summary statistics and de…nitions of these variables are provided in Table 3.
The second possibility is based on the panel data nature of the NLSY79. In fact, with
panel data it is possible to use exogenous regressors in other time periods as instruments for
endogenous regressors in the current time period. In particular, assuming that past U N ION
status is exogenous, we can use lagged levels and lagged di¤erences of the U N ION variable
to instrument for the current endogenous U N ION variable.
Table 6 reports the COVERAGE and MEMBERSHIP estimates using panel instrumental
variables (IV) methods and adjusting for the full set of controls. Columns (1) to (3) show
the …xed e¤ects IV estimates, each respectively using one of the instruments RETIREMENT,
MATERNITY, or DENTALINS, while the estimates in column (4) use all these three instruments and are by …xed e¤ects two-stage least squares (2SLS). Column (5) reports the
estimates by the so-called di¤erence Generalized Method of Moments (GMM) using lagged
levels of COVERAGE and MEMBERSHIP as instruments. Finally, the estimates in column (6) are by the so-called system GMM and use lagged levels and lagged di¤erences of
10
Boal (2008, p. 35) also controlled for union endogeneity but the dependent variable was fatalities and
not nonfatal injuries. He found, without performing any endogeneity test and after recognizing the poor
quality of his instruments, that the impact of unions on fatalities was insigni…cant after controlling for union
endogeneity.
16
COVERAGE and MEMBERSHIP as instruments.
The di¤erent instruments produce di¤erent U N ION estimates, probably re‡ecting the different strength of each instrument. However, in terms of the sign of the impact and its
signi…cance, Table 6 gives a very clear picture. Irrespective of the union measure, instruments, or estimation technique used, all estimates are positive and signi…cant, although the
di¤erence GMM estimates are only signi…cant at the 10% level. These results, and those
from the three studies mentioned above, provide compelling evidence against the causality
explanation. The results thus show that unions are indeed causing more injuries.11
What can we conclude from the causality explanation? Anecdotal and empirical evidence suggest that individuals that have had an injury (or that are working under hazardous
conditions) are indeed more likely to join a union. However, after controlling for union endogeneity, the results remained positive and signi…cant. In particular, the estimates in Table
6 give strong evidence that the positive relationship between unions and injuries is not only
an association, but that there is also causation from unions to injuries. In other words, the
evidence provided in this section does not reject the idea that one of the reasons why workers
join unions is hazardous working conditions. What is been rejected here is that this explanation accounts for the result that unionization is also causing more injuries. The causality
explanation can be ruled out.
11
Several formal tests performed to verify the quality of these estimates gave the following results. First,
the endogeneity of the union variable was con…rmed. Second, the Hansen test for overidenti…ed restrictions
after 2SLS and GMM led to the conclusion that the instruments were valid. And third, the null hypothesis
of weak instruments after 2SLS was rejected based on a Wald F-statistic.
17
4.3
Wages for safety
Wages for safety (WFS) is the third most important explanation in Table 1. This explanation
might be attributed to Duncan and Sta¤ord (1980) who did not consider occupational injuries
but working conditions in general. The idea of this explanation is that there are di¤erent goals
that unions can pursue. Two of them are higher wages and better working conditions. Since
both objectives are costly from the management’s perspective, often unions have to focus
their energy on only one of them. If labor unions increase wages, management might react
by deteriorating working conditions in order to reduce costs. The explanation thus suggests
that unions are associated with more injuries because they favor higher wages at the expense
of better working conditions. If working conditions are poor, the number of occupational
injuries is high.
In fact, some authors have even suggested that occupational health and safety has not been
one of unions’main priorities. Bacow (1980, p. 101), for example, a¢ rms that "[h]ealth and
safety issues do not command a high position on union bargaining agendas because there is
little political return on cleaning up the workplace; changes are often not recognized for years
and the individuals most likely to bene…t tend to be underrepresented." Nelking and Brown
(1984, p. 117) a¢ rm that "[w]orkers are often frustrated by the limited union in‡uence over
hazardous conditions. Preoccupied with bread and butter issues, some local o¢ cers regard
health hazards as secondary." And …nally, Fishback (1986, p. 290) argues that "the [United
Mine Workers of America] may have devoted more of their e¤orts to improving wages and
organizing nonunion districts than to improving safety."
In order to test the wages for safety explanation empirically, we …rst have to check if unions
are really increasing wages in the …rst place. Using a question on wages from the NLSY79,
we can estimate an equation of the form
W AGE =
WU
U N ION + X0 + u:
(3)
The de…nition and summary statistics of the dependent WAGE variable are given in Table
3. Fixed e¤ect (FE) estimates of W U adjusting for the full set of control variables X are
presented in Table 7. Since the WAGE variable is expressed in the log form (see de…nition
in Table 3), the interpretation of W U is that of a semi-elasticity. For example, the U N ION
coe¢ cient on column one means that covered workers receive wages that are 9.5% higher
than those of noncovered workers, while for union members, in column three, the wages are
11.5% higher than for nonmembers. Columns (2) and (4) report the 2SLS estimates that
account for U N ION endogeneity using RETIREMENT, MATERNITY, and DENTALINS
as instruments. All estimates are positive and highly signi…cant suggesting that unions are
indeed e¤ective in increasing wages.
The results on Table 7 are, however, not new and already in the 1980’s there was a huge
literature con…rming them (see Lewis 1986). In fact, in this literature the question is not on
whether unions increase wages but on how much. What is remarkable about these results
18
is that, taken together, Tables 6 and 7 seem to support the idea that the same unions that
are increasing wages are also increasing the injury probability, giving some support for the
"wages for safety" explanation.12
However, the most rigorous test this explanation still has to pass is if unionization increases
the injury probability even after holding wages …xed. In other words, we would like to know
the causal impact of unions on injuries after purging it from any e¤ects of unions on wages.
The new regression extends (1) by including the W AGE variable as one of its regressors:
IN JU RY =
U N ION +
IW W AGE
+ X0 + u:
(4)
Moreover, since the W AGE and U N ION variables are potentially endogenous, instrumental
variable methods might be needed to estimate (4).
Three studies from the empirical literature in Table 1 have estimated an injury-union
regression holding wages …xed as in (4). All of them used cross-sectional data at the establishment level. Thomason and Pozzebon (2002, Table 6: “Total Sample”) and Chelius (1974,
pp. 717-21, 728) …nd a positive and signi…cant union impact on injuries, but only Chelius
controls for wage endogeneity. Fenn and Ashby (2004), after rejecting the endogeneity of
wages (pp. 473-5), …nd that the union impact is positive and signi…cant for illnesses and
positive but insigni…cant for injuries (Table 2). In all three studies the W AGE coe¢ cient,
13
IW , was insigni…cant.
Using the NLSY79 it is also possible to control for W AGE endogeneity. To instrument
for W AGE, a question is used that asks each respondent to estimate the market value of all
12
Some authors have suggested that part of the wage premium that unionized workers are receiving is a
compensation for their higher injury probability (see Duncan and Sta¤ord 1980). This seems to be supported
by the empirical literature (see Schurman et al., 1998, p. 131, and the references therein).
13
Fishback (1986) also estimated an injury-union regression holding wages …xed but his dependent variable
was a measure of fatalities and not of nonfatal injuries.
19
vehicles owned by himself or his spouse (CARVALUE). Summary statistics of this variable
are provided in Table 3. Table 8 reports the estimates. The …rst and fourth columns give the
…xed e¤ects estimates without controlling for U N ION or W AGE endogeneity. Columns two
and …ve report the IV …xed e¤ects estimates after controlling only for W AGE endogeneity,
while columns three and six also control for U N ION endogeneity.
Let us …rst consider the impact of wages on injuries. According to the "wages for safety"
explanation, an increase in wages should be followed by an increase in the injury probability.
This idea seems to be con…rmed by the results in columns one and four, since the W AGE
coe¢ cients are positive and highly signi…cant. However, after controlling for U N ION and
W AGE endogeneity, the sign of the coe¢ cient is reversed and becomes insigni…cant. The
W AGE coe¢ cient in all three studies mentioned above was also insigni…cant.
Turning to the union impact on injuries, Table 8 gives a very clear picture: All estimates
of the U N ION coe¢ cient are positive and signi…cant, even after holding wages …xed and
controlling for U N ION and W AGE endogeneity. These results, together with the estimates
from the three studies mentioned above, provide solid evidence against the "wages for safety"
explanation.
5
The new explanation: Behavioral adaptation
The goal of the previous section was to try to understand the paradoxical e¤ect of unions
on nonfatal injuries. The three major explanations from the literature were explored using
econometric techniques, to a great extent, but the results presented led us to reject all explanations. There is, therefore, a need for alternative approaches. The objective of this section is
20
to introduce to the literature a new explanation based on the theory of behavioral adaptation
and to provide direct and indirect empirical evidence supporting it.
5.1
Explaining behavioral adaptation
Anecdotal evidence and unions’ activities suggest that unions have made workplaces safer.
The question is why has this enhanced safety not translated into less occupational injuries?
One explanation that …ts the evidence very well is based on the theory of behavioral adaptation
(also called risk compensation).
According to this theory, individuals modify their behavior in response to their perceived
changes in risk. If the perceived risk increases, individuals behave more cautious. If the
perceived risk decreases, a less cautious behavior is adopted. There is thus a permanent
adaptation process in the individuals’levels of vigilance and voluntary exposure to danger as
a reaction to perceived risk. One consequence of this theory is that individuals can o¤set the
bene…ts of a safety measure if they react by increasing their risky behavior. A safety measure
intended to reduce the number of accidents, can even lead to an increase in the accident rate
if individuals behave less cautious.
There is considerable evidence supporting this theory in the area of road safety. The
most in‡uential paper by an economist is probably that of Peltzman (1975). He argues
that the introduction of auto safety measures (like seat belts or dual braking system) has not
reduced highway death rates as intended due to behavioral adaptation. His explanation is that
safety measures make drivers feel safer. As a reaction, drivers adapt their behavior and drive
faster or more carelessly than they would do without the safety measures, diminishing and
maybe even o¤setting any positive e¤ects of regulation. Several other studies have con…rmed
Peltzman’s result. In fact, one study that is often cited in this context is an OECD report by
an international scienti…c expert group. After reviewing the existing scienti…c literature, the
report concludes that “behavioural adaptation exists [...] and does reduce the e¤ectiveness of
road safety programmes in a number of cases" (OECD 1990, p. 7).
There is also evidence of behavioral adaptation in the area of public health. Viscusi
(1984), for example, notes that the introduction of child-resistant packaging on aspirin and
other drugs did not achieve the expected decrease in poisoning rates. He attributed this to a
"general reduction in parental caution with respect to such medicines" (p. 324). Furthermore,
several researchers have argued that behavioral adaptation may also help explain the limited
e¤ect of promoting promising health innovations on HIV rates. It has been for example
noted that despite the clear evidence showing that condoms decrease the e¢ ciency of HIV
transmission, promoting the protective impact of condoms reduces the perceptions of risk of
the population, which leads to a general increase in risky sexual behavior (see Cassell et al.
2006 and the references therein)14 . Behavioral adaptation has also been used to explain the
14
See also Richens et al. 2003, Auld 2003, and Lakdawlla et al. 2006.
21
positive association between sunscreen use and skin cancer (Autier et al. 1998).
Moreover, the theory of behavioral adaptation is closely related to the concept of moral
hazard for which considerable support has been found in di¤erent …elds of economics. Originally, moral hazard was a term used by insurance companies to refer to the idea that individuals that are insured know that they are protected and take more risks than they would
do without the insurance. The consequence is that insurers have to bear the additional costs
of the individuals’ behavioral adaptation. Thus, behavioral adaptation is only one of the
several aspects of moral hazard. In fact, in moral hazard it is not also important who pays
the insurance premium but also who bears the costs of the behavioral adaptation. Since
its introduction to the economic literature by Arrow (1963) and Pauly (1968), moral hazard
has been applied to a large number of problems in the areas of public and private …nance,
management, labor contracts, health economics, and many others. In all these applications,
substantial evidence of the behavioral adaptation aspect of moral hazard has been provided.
5.2
Indirect evidence of unions and behavioral adaptation
The argument that is advanced in this paper is that behavioral adaptation can also explain
why unionized workers are having more occupational injuries. The reason is that the increased
safety and protection that unions provide enhance workers’feeling of safety, leading workers
to adapt their behavior, for example, by working faster, becoming bolder, or by taking less
safety precautions. This riskier behavior more than o¤sets the union safety e¤orts. As a
consequence, the net impact is an increase in occupational injuries.
There is some indirect evidence supporting this argument. For example, as it was mentioned in the introduction, it is well-known that one of the safety-enhancing activities of labor
unions is to provide workers with protective equipment. Klen (1997), however, …nds that the
introduction of personal protectors (like safety helmets, eye protectors, or ear caps) did not
reduce accident injuries among Finnish loggers as intended, partially because of behavioral
adaptation.
Another safety-enhancing union activity is to lobby and support the passage of government
legislation and to monitor its enforcement. Labor unions, for example, were not only crucial
in the passage of the Occupational Safety and Health Act (OHSA) in 1970, but the empirical
evidence also suggests that the OHSA is better enforced in union establishments (Weil 1991,
1992). In spite of this, the impact of the OHSA on reducing occupational injuries has been
labeled by many as ine¤ective (Viscusi 1986). Viscusi (1979) argues that the reason for this
might lie in the idea that safety regulations that increase enterprises’ investment in work
quality might be o¤set if workers react by diminishing their own safety-enhancing actions.
Yet another safety-enhancing activity of unions is to aid and represent workers in accident
compensation claims. Borba and Appel (1987), for example, …nd that workers’compensation
claimants who are union members are more likely to be represented by attorneys than are
nonunion workers (see also Latta and Lewis 1974). Probably because of this and other types
22
of support, unionized workers seem to be more successful in the compensation process. In fact,
Hirsch et al. (1997), conclude in their empirical study that "[u]nion workers are far more likely
than nonunion workers, other things equal, to receive bene…ts from workers’ compensation
[...]" (p. 233). Other empirical studies have con…rmed the positive impact of unionization on
workers’ compensation claims and bene…ts (see, for example, Butler and Worrall 1983 and
Chelius 1974, p. 729). However, a well-established result from the workers’ compensation
literature is the presence of moral hazard, which as mentioned above is related to behavioral
adaptation. In fact, the evidence from this literature shows that workers’ compensation is
encouraging more, not less, nonfatal injuries but seems to be reducing fatalities (Moore and
Viscusi 1990, ch. 9, Ruser 1993). This is exactly the pattern found in Figure 2. In other
words, unions might be facilitating a more extensive use of workers’compensation and thus,
probably unwillingly, magnifying the moral hazard problems of this system.
5.3
Direct evidence of unions and behavioral adaptation
This section presents three empirical exercises using the NLSY79 data set that provide direct
evidence in favor of the "behavioral adaptation" explanation.
5.3.1
Controlling for industry risk
One way to test the behavioral adaptation explanation is to check if the injury probability of
unionized workers is higher after controlling for the workers’average industry risk. In other
words, we would like to test if workers that are exposed to the same average industry risk
behave di¤erently (for example take more risks) when they join unions. The idea here is to
try to isolate the impact of unions on workers self-protection activities. This type of analysis
is similar to that of Viscusi and Hersch (2001, pp. 278-9) but in a di¤erent context and using
panel data.
Additionally, since more dangerous industries are more likely to be unionized, a worker
that changes to a job to a more dangerous industry and at the same time joins a union
in the new job might have a higher injury probability. It is not clear if the higher injury
probability is because the new industry is more hazardous or because the worker joined the
union. The 11 industry dummies variables that have been included in all regressions so far
partially control for this. However, the following type of analysis is more precise since the
number of industry-risk groups is considerably higher.
The data for the industry-risk variable are based on the incidence rates from the Bureau of
Labor Statistics (BLS) Survey of Occupational Injuries and Illnesses that can be downloaded
at ftp://ftp.bls.gov/pub/time.series/sh/ and at ftp://ftp.bls.gov/pub/time.series/hs/. The
incidence rates are de…ned as the number of nonfatal occupational injury and illness cases
per 100 full-time workers. These rates were transformed by multiplying them by 100 and
by taking the log in order to obtain the …nal IN DU ST RY RISK variable (see Table 3 for
23
de…nition and summary statistics).
Following this, each NLSY79 respondent was matched to the (transformed) BLS incidence
rates based on the respondents’reported industry code at the most precise level of industry
breakdown that the data sets allowed to. In many cases, this was at the three-digit level.
Due to data limitations, however, it was not possible to assign an industry-risk group to every
NLSY79 respondent. For example, the BLS Survey does not provide incidence rates for the
public administration sector. Despite these limitations, it was possible to construct more than
200 industry-risk groups for every year. The BLS data are based on the Standard Industrial
Classi…cation (SIC) System from 1972 and 1987, while the NLSY79 respondents are coded
using the 1970 and the 1980 industry classi…cation system of the Census of Population. The
two data sets were merged using concordance tables that relate the classi…cation systems to
each other.
The estimated equation was the following
IN JU RY =
U N ION +
IW W AGE
+
II
IN DU ST RY RISK + X0 + u:
(5)
Table 9 summarizes the estimates. The …rst and fourth columns report the …xed e¤ects
estimates. Columns two and …ve control for wage endogeneity, while columns three and
four control for both wage and union endogeneity. As expected, the results indicate that
workers in risky industries are more likely to have an occupational injury, although the
IN DU ST RY RISK variable is only signi…cant at the 10% level for the …xed e¤ects estimates. However, the main interest lies on the estimates of the U N ION variable, which are
positive and signi…cant in all speci…cations. Unionized workers thus have a signi…cantly higher
injury probability, even after controlling for the average industry risk level.
24
5.3.2
Within variation and union joiners-leavers analysis
When only cross-sectional data is available, it is only possible to estimate the union impact on
injuries by comparing the group of unionized with the group of nonunionized workers in one
single period of time. If it is found that unionized workers have more occupational injuries
than nonunionized workers (after controlling for other variables), then it is concluded that
unions have an increasing impact on injuries.
However, in order to …nd evidence of behavioral adaptation, we still need to establish if
the same worker is having more injuries after joining a union. The question is if there is
an increase in the injury probability of a worker that in period one was not unionized and in
period two joins a union. Has joining a union made any di¤erence for this worker in terms
of injury probability? Or, in other words, is this worker adapting his safety behavior after
joining a union? This type of analysis can only be performed with panel data at the individual
level, like the NLSY79, since only this type of data has information on the same person for
two or more periods. None of the previous studies from the literature was able to perform
such an analysis because of data limitations.
In order to perform such an analysis, all regressions in this paper have been estimated
by …xed e¤ects, a panel-data estimation method. The …xed e¤ects estimator only relies on
the so-called within variation, that is, the variation over time of a given individual (see fn.
9). As a consequence, all estimates presented so far already provide evidence of behavioral
adaptation since they imply that the injury probability of the same worker is increased when
the worker changes status from nonunion to union.
However, in order to take this analysis to a deeper level, the model in (5) was extended to
a less restrictive model that allows di¤erent changes in union status to have di¤erent impacts
on the injury probability. The estimated model was
IN JU RY =
un U N
+
nn N N
+
nu N U
+ X0 + u;
(6)
where W AGE and IN DU ST RY RISK were absorbed in the vector of control variables X
and N U , U U , U N , and N N are zero-one dummy variables and U U was used as the base
category. In particular, N U is equal to one if the worker is not unionized in period t 1
and unionized in t; U U is equal to one if unionized in both periods; U N is equal to one if
unionized in t 1 and not unionized in t; and N N is equal to one if not unionized in both
periods. Thus, the coe¢ cients un , nn , and nu respectively measure the relative di¤erence
in injury probabilities of union leavers (U N ), nonunion stayers (N N ), and union joiners (N U )
with respect to the omitted base category of union stayers (U U ). This model is, therefore,
less restrictive than (5) in the sense that does not have to be symmetrical: An increase in
the injury probability of union joiners needs not to be equivalent to a decrease in the injury
probability of union leavers, that is, nu might be di¤erent from
un .
Since the proportion of workers who change their union status is in general relatively
low, large data sets, like the NLSY79, with su¢ cient observations are needed to provide a
25
meaningful analysis. Summary statistics of the union status changes variables are relegated
to the appendix. Simply note that, just considering two consecutive periods and only based
on union membership (not coverage), 80.7% of the respondents remained nonunion, 11.2%
remained unionized, 3.7% joined unions, and 4.4% left a union.
The injury probabilities that resulted from the estimation of the model (see app. A.2) are
illustrated in Figure 3, which is only based on the membership measure of union status.15 The
graphic on the left shows the injury probabilities using the two-period union status dummies.
The graphic on the right extends the analysis to three-period union status dummies.
There are three conclusions that can be drawn from Figure 3 that give further support to
the behavioral adaptation explanation. First of all, when a worker joins a union, the injury
probability increases. This only con…rms our results from the previous sections. Second, the
…gure suggests that the longer a worker remains unionized, the higher the injury probability
is. This might be because workers need some time to fully adapt their behavior to the new
protection provided by unions. Obviously, the injury probability will not increase in…nitely
with the duration of unionization and the …gure suggests that an upper limit might be reached
after two consecutive periods. Third, when a worker leaves a union, the injury probability is
reduced. This is in particular interesting since it shows that there is also behavioral adaptation
in the other direction, that is, when the unions’safety protection is removed, workers increase
their self-protection activities and their injury probability is reduced.
5.3.3
Health and life insurance
In the United States, workers who are injured in the course of employment receive bene…ts
from their employer under the workers’ compensation system. These bene…ts range from
medical to total disability bene…ts for nonfatal injuries and extend to burial and survival
15
The methodology employed here has never been used before to study the impact of changes in union
status on injury probabilities. There are, however, several studies that use this approach to study the impact
of changes in union status on wages (see Lewis 1986, ch. 5, and the references therein).
26
bene…ts for fatal injuries. Parallel to the workers’ compensation system, employers might
provide workers with health or with life insurance covering for incidents not connected to the
job.
Since health and life insurance do not cover for o¤-the-job incidents, there is in principle
no reason why these types of insurance might have any impact on occupational injuries.
However, the theory of behavioral adaptation suggests that if workers start perceiving that
their general risk is lower after signing the insurance contract, the insured workers might
adopt a "riskier" lifestyle, including how careful they work. In other words, health and life
insurance might also have an increasing impact on the workers’injury probability.
This hypothesis can also be tested using the NLSY79 data set. There are in particular
two questions asking respondents if employers made available to them health (HEALTHINS)
or life insurance (LIFEINS) that covers injuries, illnesses or death o¤ the job. The de…nitions
and summary statistics of these variables are presented in Table 3. In our sample, around 76%
of the respondents had an employer-provided health insurance and around 65% an employerprovided life insurance.
The estimates (not shown) of HEALTHINS and LIFEINS, including one of these variables
at the time, in a regression such as (5) were all positive and highly signi…cant, giving support
for the hypothesis. An even more interesting exercise can be performed by interacting the
HEALTHINS and LIFEINS variables with the union status variables. In particular, it is
possible to create dummies for every one of the four possible combinations between union
worker (yes or no) and insured worker (yes or no). Following this, an equation such as (5) can
be estimated but with the U N ION variable replaced by the dummies created and using the
"no-union-no-insured" dummy as the base category. The injury probabilities that resulted
from this exercise are reported in Table 10.
It is clear from the table that, regardless of the union status measure or insurance type
considered, the lowest injury probabilities are for the uninsured, nonunion workers. According
to the behavioral adaptations theory, these are the workers that should perceive the highest
risk and that should take the most safety precautions in and outside their workplace, leading
to lower injury probabilities. In fact, as the table shows, providing workers with more "protection" in the form of either an insurance or of union services increases their injury probability
27
non-negligibly. Moreover, the highest injury probabilities are for the workers that have the
highest "protection": the union and insured workers.
Another interesting result that reinforces this idea and that is not only apparent from
table 10 but from all estimates shown in this paper is that union members exhibit higher
injury probabilities than workers covered by a union contract. This gives further support
to the behavioral adaptation theory since membership gives workers more protection than
coverage.
6
Conclusions
This paper begins by presenting a quantitative analysis based on 25 empirical studies investigating the impact of labor unions on occupational injuries. The …rst result of this analysis
suggests that labor unions seem to be e¤ective in reducing fatalities, but the second result
shows that unions are clearly associated with more nonfatal occupational injuries. This second
result is very puzzling since it clearly contradicts expectations based on anecdotal evidence
and on unions’ safety-enhancing activities. Moreover, this second result is con…rmed and
reinforced by the new panel estimates at the individual level reported in this paper.
The task of this paper was then to attempt to solve this paradox. For this purpose, the
most prominent explanations from the literature were identi…ed and were tested. Surprisingly, all three major explanations from the literature could be dismissed using (for the most
part) econometric techniques. It was, therefore, necessary to look for alternative approaches.
Accordingly, this paper then introduced a novel explanation based on the theory of behavioral adaptation. The new explanation appears to …t the anecdotal and empirical evidence
remarkably. In particular, it explains why unionized workers have more nonfatal occupational
injuries in spite of all unions’safety-enhancing activities. The reason is that the introduction
of health and safety measures, and in general the additional protection o¤ered by unions,
make workers feel safer. As a reaction, workers adapt their behavior and reduce their own
self-prevention activities, more than o¤setting any injury-reducing e¤ects intended by unions.
The direct and indirect evidence presented in this paper gives solid support for this explanation. Moreover, this type of behavioral adaptation is similar to the idea of moral hazard, a
phenomenon for which economists have already found considerably supporting evidence.
Further research should overcome some of the limitations of this paper. First of all, it
should establish if the paradoxical relationship between unions and nonfatal injuries also extends to other countries not considered here. Second, more detailed data should be collected
that allows to identify workers’ behavioral adaptation after the introduction of speci…c occupational health and safety measures. Finally, further research should establish why unions
have not su¢ ciently considered the potential negative implications of behavioral adaptation
resulting from their safety-enhancing activities.
28
A
A.1
Appendix
De…nitions and summary statistics of the control variables
29
A.2
Union joiners-leavers analysis
In order to maximize the number of observations, the union status changes variables were
constructed also using information on membership and coverage from the years 1986 (for
the three-period dummies) and 1987. The year 1991 was excluded from the analysis since
there is no information on injuries for this year. In 1986 and 1987 only the question on
coverage was asked. Therefore, only for those two years, it was assumed that MEMBERSHIP=COVERAGE. The following two tables respectively present the summary statistics of
these variables and report the estimates of equation (6) and of an extended version of (6)
using the three-period dummies.
30
References
Appleton, W. C., and J. G. Baker (1984): “The e¤ect of unionization on safety in
bituminous deep mines,”Journal of Labor Research, 5(2), 139–47.
(1985): “The e¤ect of unionization on safety in bituminous deep mines: Reply,”
Journal of Labor Research, 6(2), 217–20.
Arrow, K. J. (1963): “Uncertainty and the welfare economics of medical care,” American
Economic Review, 53(5), 941–73.
Auld, M. C. (2003): “Choices, beliefs, and infectious disease dynamics,” Journal of Health
Economics, 22(3), 361–77.
Autier, P., J.-F. Doré, M. S. Cattaruzza, F. Renard, and et al. (1998): “Sunscreen
use, wearing clothes, and number of nevi in 6- to 7-year-old European children,” Journal
of the National Cancer Institute, 90(24), 1873–80.
Beaumont, P. B. (1983): Safety at work and the unions. Croom Helm.
Boal, W. M. (2008): “The e¤ect of unionism on accidents in U.S. coal mining, 1897-1929,”
Discussion paper.
(2009): “The e¤ect of unionism on accidents in U.S. coal mining, 1897-1929,” Industrial Relations Journal, 48(1), 97–120.
Boden, L. I. (1977): “Underground coal mining accidents and government enforcement of
safety regulations,”Ph.D. thesis, Massachusetts Institute of Technology.
(1985): “Government regulation of occupational safety: Underground coal mine
accidents 1973-75,”American Journal of Public Health, 5(5), 497–501.
Borba, P. S., and D. Appel (1987): “The propensity of permanently disabled workers to
hire lawyers,”Industrial and Labor Relations Review, 40(3), 418–29.
Botsch, R. E. (1993): Organizing the breathless. Cotton dust, Southern politics and the
Brown Lung Association. The University Press of Kentucky, Lexington, Kentucky.
Budd, J. W., and B. P. McCall (1997): “The e¤ect of unions on the receipt of unemployment insurance bene…ts,”Industrial and Labor Relations Review, 50(3), 478–92.
(2004): “Unions and unemployment insurance bene…ts receipt: Evicence from the
Current Population Survey,”Industrial Relations, 43(2), 339–55.
31
Butler, R. J., and J. D. Worrall (1983): “Workers’compensation: Bene…t and injury
claims rates in the seventies,”Review of Economics & Statistics, 65(4), 580–9.
Cameron, A. C., and P. K. Trivedi (2005): Microeconometrics: Methods and applications.
Cambridge University Press.
Cassell, M. M., D. T. Halperin, J. D. Shelton, and D. Stanton (2006): “Risk compensation: The Achilles’heel of innovations in HIV prevention?,”British Medical Journal,
332, 605–7.
Chelius, R. J. (1974): “The control of industrial accidents: Economic theory and empirical
evidence,”Law and Contemporary Problems, 38, 700–29.
Clark, C. (1997): Radium girls. The University of North Carolina Press, Chapel Hill, NC.
Dorman, P. (1996): Markets and mortality: Economics, dangerous work, and the value of
human life. Cambridge University Press.
Duncan, G. J., and F. P. Stafford (1980): “Do union members receive compensating
wage di¤erentials?,”American Economic Review, 70(3), 355–71.
Eaton, A. E., and T. Nocerino (2000): “The e¤ectiveness of health and safety committees:
Results of a survey of public-sector workplaces,”Industrial Relations, 39(2), 265–90.
Fairris, D. (1992): “Compensating payments and hazardous work in union and nonunion
settings,”Journal of Labor Research, 13(2), 205–21.
Fenn, P., and S. Ashby (2004): “Workplace risk, establishment size and union density,”
British Journal of Industrial Relations, 42(3), 461–80.
Fishback, P. V. (1986): “Workplace safety during the progressive era: Fatal accidents in
bituminous coal mining, 1912-1923,”Explorations in Economic History, 23(3), 269–98.
(1987): “Liability rules and accident prevention in the workplace: Empirical evidence
from the early twentieth century,”Journal of Legal Studies, 16(2), 305–28.
Freeman, R. B., and J. L. Medoff (1984): What do Unions do? Basic Books.
Garen, J. (1988): “Compensating wage di¤erentials and the endogeneity of job riskiness,”
Review of Economics & Statistics, 70(1), 9–16.
Gray, W. B., and J. M. Mendeloff (2005): “The declining e¤ects of OSHA inspections on
manufacturing injuries, 1979-1998,”Industrial and Labor Relations Review, 58(4), 571–87.
32
Hillage, J., B. Kersley, P. Bates, and J. Rick (2000): Workplace consultation on
health and safety. Health and Safety Executive.
Hirsch, B. T., and M. C. Berger (1984): “Union membership determination and industry
characteristics,”Southern Economic Journal, 50(3), 665–79.
Hirsch, B. T., D. A. Macpherson, and J. M. Dumond (1997): “Workers’compensation
recipiency in union and nonunion workplaces,” Industrial and Labor Relations Review,
50(2), 213–36.
Klen, T. (1997): “Personal protectors and working behaviour of loggers,” Safety Science,
25(1-3), 89–103.
Lakdawalla, D., N. Sood, and D. Goldman (2006): “HIV breakthroughs and risky
sexual behavior,”Quarterly Journal of Economics, 121(3), 1063–102.
Lanoie, P. (1992): “The impact of occupational safety and health regulation on the risk of
workplace accidents : Quebec, 1983-87,”The Journal of Human Resources, 27(4), 643–60.
Latta, G., and R. Lewis (1974): “Trade union legal services,”British Journal of Industrial
Relations, 12(1), 56–70.
Leigh, J. P. (1982): “Are unionized blue collar jobs more hazardous than nonunionized blue
collar jobs?,”Journal of Labor Research, 3(3), 349–57.
Lewis, H. G. (1986): Union relative wage e¤ects: A survey. The University of Chicago Press.
Litwin, A. S. (2000): “Trade unions and industrial injury in Great Britain,”CEP Discussion
Papers 0468, Centre for Economic Performance, LSE.
Moore, M. J., and W. K. Viscusi (1990): Compensation mechanisms for job risks: Wages,
workers’compensation, and product liability. Princeton University Press.
Nichols, T., A. Dennis, and W. Guy (1995): “Size of employment unit and injury rates
in British manufacturing: A secondary analysis of WIRS 1990 data,” Industrial Relations
Journal, 26(1), 45–56.
Nichols, T., D. Walters, and A. C. Tasiran (2007): “Trade unions, institutional mediation and industrial safety: Evidence from the UK,”Journal of Industrial Relations, 49(2),
211–25.
OECD (1990): Behavioural adaptations to changes in the road transport system. Organisation
for Economic Co-Operation and Development.
33
Pauly, M. V. (1968): “The economics of moral hazard: Comment,” American Economic
Review, 58(3), 531–7.
Peltzman, S. (1975): “The e¤ects of automobile safety regulation,” Journal of Political
Economy, 83(4), 677–725.
Reardon, J. (1996): “The e¤ect of the United Mine Workers of America on the probability
of severe injury in underground coal mines,”Journal of Labor Research, 27(2), 239–52.
Reilly, B., P. Paci, and P. Holl (1995): “Unions, safety committees and workplace
injuries,”British Journal of Industrial Relations, 33(2), 275–88.
Richens, J., J. Imrie, and H. Weiss (2003): “Human immunode…ciency virus risk: Is
it possible to dissuade people from having unsafe sex?,” Journal of the Royal Statistical
Society: Series A, 166(2), 207–15.
Robinson, A. M., and C. Smallman (2006): “The contemporary British workplace: A
safer and healthier place?,”Work, Employment & Society, 20(1), 87–107.
Robinson, J. C. (1988): “Workplace hazards and workers’desires for union representation,”
Journal of Labor Research, 9(3), 237–49.
(1990): “Worker responses to occupational risk of cancer,” Review of Economics &
Statistics, 72(3), 536–41.
(1991): Toil and toxics: Workplace struggles and political strategies for occupational
health. University of California Press, Berkeley.
Rosner, D., and G. Markowitz (1991): Deadly dust - Silcosis and the politics of occupational disease in twentieth-century America. Princeton University Press, Princeton, New
Jersey.
Ruser, J. W. (1993): “Workers’compensation and the distribution of occupational injuries,”
Journal of Human Resources, 28(3), 593–617.
Schurman, S. J., D. Weil, P. Landsbergis, and B. A. Israel (1998): “The role of
unions and collective bargaining in preventing work-related disability,”in New approaches
to disability in the workplace, ed. by T. Thomason, J. F. Burton, and D. E. Hyatt. Industrial
Relations Research Association.
Smith, B. E. (1987): Digging our own graves. Coal miners and the struggle over Black Lung
Disease. Temple University Press, Philadelphia.
34
Smitha, M. W., K. A. Kirk, K. R. Oestenstad, K. C. Brown, and S.-D. Lee (2001):
“E¤ect of state workplace safety laws on occupational injury rates,”Journal of Occupational
and Environmental Medicine, 43(12), 1001–10.
Thomason, T., and S. Pozzebon (2002): “Determinants of …rm workplace health and
safety and claims management practices,” Industrial and Labor Relations Review, 55(2),
286–307.
Trejo, S. J. (1991): “The e¤ects of overtime pay regulation on worker compensation,”
American Economic Review, 81(4), 719–40.
(1993): “Overtime Pay, Overtime Hours, and Labor Unions,” Journal of Labor
Economics, 11(2), 253–78.
Viscusi, W. K. (1979a): Employment Hazards: An Investigation of Market Performance.
Harvard University Press.
(1979b): “The impact of occupational safety and health regulation,”Bell Journal of
Economics, 10(1), 117–40.
(1984): “The lulling e¤ect: The impact of child-resistant packaging on aspirin and
analgesic ingestions,”American Economic Review, 74(2), 324–7.
(1986): “Reforming OSHA regulation of workplace risks,” in Regulatory reform:
What actually happened, ed. by L. W. Weiss, and M. W. Klass, pp. 234–68. Little, Brown
and Co, Boston.
Viscusi, W. K., and J. Hersch (2001): “Cigarette smokers as job risk takers,” Review of
Economics & Statistics, 83(2), 269–80.
Wallace, M. (1987): “Dying for coal: The struggle for health and safety conditions in
American coal mining, 1930-82,”Social Forces, 66(2), 335–64.
Weil, D. (1991): “Enforcing OSHA: The role of labor unions,” Industrial Relations, 30(1),
20–36.
(1992): “Building safety: The role of construction unions in the enforcement of
OSHA,”Journal of Labor Research, 13(1), 121–32.
Wooldridge, J. M. (2002): Econometric Analysis of Cross Section and Panel Data. MIT
Press.
Worrall, J. D., and R. J. Butler (1983): “Health conditions and job hazards: Union
and nonunion jobs,”Journal of Labor Research, 4(4), 339–47.
35