Journal of Health Economics 37 (2014) 219–231
Contents lists available at ScienceDirect
Journal of Health Economics
journal homepage: www.elsevier.com/locate/econbase
Air pollution and infant mortality: A natural experiment from power
plant desulfurization夽
Simon Luechinger ∗
University of Lucerne and KOF Swiss Economic Institute, ETH Zurich, Switzerland
a r t i c l e
i n f o
Article history:
Received 28 September 2012
Received in revised form 6 June 2014
Accepted 17 June 2014
Available online 24 June 2014
JEL classification:
I12
Q53
J13
a b s t r a c t
The paper estimates the effect of SO2 pollution on infant mortality in Germany, 1985–2003. To avoid
endogeneity problems, I exploit the natural experiment created by the mandated desulfurization at power
plants and power plants’ location and prevailing wind directions, which together determine treatment
intensity for counties. Estimates translate into an elasticity of 0.07–0.13 and the observed reduction in
pollution implies an annual gain of 826–1460 infant lives. There is no evidence for disproportionate effects
on neonatal mortality, but for an increase in the number of infants with comparatively low birth weight
and length.
© 2014 Elsevier B.V. All rights reserved.
Keywords:
Health
Infants
Mortality
Infant mortality
Air pollution
1. Introduction
Health concerns are a primary rationale for air quality regulations such as the U.S. Clean Air Act and the German Federal
Immission Control Act. These regulations considerably improved
air quality. For example, Fig. 1 depicts the sulfur dioxide (SO2 ) concentration in Germany in 1985–2003. Other developed countries
夽 I thank Wolfgang Bräuniger, Andrea Minkos, and Wolfgang Müller from the
German Federal Environmental Agency for the pollution and power plant data, the
operating companies for giving confidential information on their generating units,
Hiltrud Bayer from the German Youth Institute and the statistical offices of the Länder for the mortality data, and Hans-Peter Mast and Stefan Weil from the Research
Data Centers of the Federal Statistical Office and the statistical offices of the Länder for help with remote access to the death and birth certificates. For comments
and suggestions, I thank Peter Nilsson, Shinsuke Tanaka, participants of the annual
meeting of the Swiss Society of Economics and Statistics 2009, the annual meeting
of the European Economic Association 2009, the annual meeting of the American
Economic Association 2010, the annual meeting of the German Economic Association 2010 and seminars at the ETH Zurich, the University of Berlin, the University of
Fribourg, the University of Mannheim, and the University of Rotterdam as well as
the editor Adriana Lleras-Muney and two anonymous referees.
∗ Correspondence to: University of Lucerne, Department of Economics,
P.O. Box 4466, 6002 Lucerne, Switzerland. Tel.: +41 41 229 5641.
E-mail address: [email protected]
http://dx.doi.org/10.1016/j.jhealeco.2014.06.009
0167-6296/© 2014 Elsevier B.V. All rights reserved.
experienced similar declines in SO2 concentrations. But this general trend masks considerable heterogeneity. Many people are still
exposed to high pollution levels and in developing countries air
pollution is often getting worse. Even in Europe, SO2 pollution is
bound to rise as coal-based power generation experiences a revival.
Therefore, knowledge of the health effects of previous air quality
regulations and corresponding improvements in air pollution is of
considerable interest.
Many studies investigate the effect of air pollution on adult
mortality. Epidemiologists typically assess acute effects with timeseries analyses and chronic effects of long-term exposure with
cross-section and cohort studies. For example, SO2 was significantly associated with mortality in a time-series analysis for
London in 1958–1972 (Schwartz and Marcus, 1990), in a crosssection of groups of U.S. counties in 1970 (Mendelsohn and Orcutt,
1979), or in a large group of adult Americans followed over the
period 1982–1998 (Pope et al., 2002). However, studies on infant
mortality have clearer implications regarding the number of lifeyears lost and suffer less from uncertainty regarding life-time
exposure (Chay and Greenstone, 2003b).
Omitted variables are an important concern in all these studies
since air pollution depends on economic activity and other unobserved factors with independent effects on mortality. The concern
220
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
Fig. 1. SO2 concentration and infant mortality in East and West Germany,
1985–2003.
Sources: Federal Statistical Office and Federal Environmental Agency.
is partially addressed in intervention analyses. Evidence of reduced
SO2 pollution and mortality following a ban on high sulfur fuel in
Hong Kong in 1990 is certainly more convincing than evidence
from other time-series analyses, but influences from other concurrent shocks cannot be ruled out without an adequate control
group (Hedley et al., 2002). Such concerns sparked a recent interest
among economists in natural experiments that allow researchers to
identify the effects of air pollution on infant mortality (see Section
2 for a review).
This paper estimates the health benefits of an air quality regulation and uses regulation-induced changes in air pollution to
identify chronic effects on infant mortality (similar to Chay and
Greenstone, 2003a). Specifically, it estimates the effect of the mandated installations of scrubbers at power plants and the resulting
reduction in SO2 pollution on infant health with data from Germany
in 1985–2003. Thereby, it contributes to the existing literature in
two respects.
First, the paper provides evidence on infant health effects of air
pollution and air quality regulations for another highly developed
country than the U.S. Thus, the results help us to understand if and
to what extent pollution-mortality relationships found in one context can the transferred to different contexts. While there are no
reasons to expect differences in biological dose-response relationships, effects of ambient air pollution also depend on medical care
consumption and avoidance behavior, which affects the relationship between ambient air pollution and individual exposure (Graff
Zivin and Neidell, 2013). In this regard, Germany and the U.S. may
differ along several relevant dimensions. For example, there may be
differences in medical care utilization. Pre-natal care can improve
infant health and thereby reduce susceptibility to air pollution,
good access to medical care after birth allows for timely measures
against health problems. Despite a huge increase in Medicaid eligibility of pregnant women in the U.S. in the 1980s, a large share of the
population is still uninsured and many eligible women enroll late in
their pregnancies (Currie and Gruber, 1996; Gruber, 1997). In contrast, health insurance coverage in Germany is near universal and
frequent screenings of infants are statutorily regulated. Similarly,
the extent to which individuals can avoid exposure to pollutants
may also differ across countries and will depend on such diverse
factors as building design with large regional differences in airtightness of houses or typical activity patterns of pregnant women
and parents (Ashmore and Dimitroulopoulou, 2009).
Second, the paper analyzes the effect of SO2 pollution. Of course,
different pollutants may be correlated (Lleras-Muney, 2010) and
the regulation may have affected several pollutants. However, as
I explain in more detail in Section 3, the German situation analyzed in this paper is particularly well-suited to specifically isolate
exogenous variation in SO2 pollution. A federal regulation mandated the installation of scrubbers at power plants and left local
authorities or operating companies little room for discretion. Power
plants are the main source of SO2 and the predominant scrubbing
technology removes SO2 but not other pollutants. Regulation of TSP
was already in place and new regulation of nitrogen oxides (NOx )
was generally not binding, affected different sources, and required
different compliance measures. Further, I find that the estimated
effect of desulfurization at power plants affected SO2 concentration but not NOx concentration. Therefore, I present not only
reduced form effects of the policy but also use it to instrument SO2
pollution.
Looking at SO2 is interesting for several reasons. First, toxicity
differs across pollutants. Second, pollutants differ in the extent to
which ambient air pollution translates into individual exposure.
For example, the correlation between outdoor and indoor air pollution is lower for SO2 than for other pollutants (Ashmore and
Dimitroulopoulou, 2009). Third, the existing evidence on the effects
of SO2 on infant mortality is inconclusive and suffers from omitted
variables (see Section 2 for a review). Given that SO2 pollution is
the focus of many air quality regulations and that there is considerable evidence for effects of SO2 on adult mortality and on adverse
pregnancy outcomes (for a review, see Šram et al., 2005), the lack
of convincing evidence on the effects of SO2 pollution on infant
mortality is regrettable and this paper aims to fill this void.
The most important finding is that the air quality regulation
had beneficial effects on infant health: Infant mortality decreases
with predicted reductions in SO2 concentration due to desulfurization at power plants. The sharp and simultaneous drops in SO2
pollution and in infant mortality between 1987 and 1988 in Fig. 1
anticipates this result. Assuming that the air quality regulation only
affected infant mortality through its effect on actual SO2 concentrations, I use the regulation-induced changes in SO2 concentration to
estimate the effect of SO2 pollution on infant mortality. According to fixed-effects regressions of infant mortality rates on SO2
concentration, 0.026 infant lives (per 1000 live births) are saved
for every 1 ␮g/m3 reduction in SO2 concentration. In instrumental
variable regressions with the predicted regulation-induced reductions in SO2 concentrations as an instrument, the effect amounts
to 0.045 infant lives. Since most of the variation in SO2 concentration is the result of the air quality regulation, the fixed-effects and
instrumental variable estimates are similar and both estimates are
informative about the health effects of the regulation.
The point estimates translate into an elasticity of 0.07–0.13.
The results are similar in subperiods and the West German subsample but not the East German subsample. The estimates are
robust to controls for local economic and demographic development, weather, TSP pollution, reunification effects, and rural/urban
trends. The instrumental variable estimates are also robust to the
inclusion of county-specific time trends, the fixed-effects estimates
less so. There is no evidence for strongly disproportionate effects of
SO2 on neonatal mortality, but evidence for effects on the number
of infants with comparatively low birth weight and, in particular, length. Thus, although poor fetal development due to exposure
during gestation seems to affect infant health, it is unlikely to be
the main biological mechanisms through which SO2 affects infant
mortality.
The remainder of the paper is organized as follows. Section
2 briefly reviews the related literature. Section 3 introduces the
pollution data and the strategy to instrument SO2 concentrations.
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
Section 4 presents the mortality data, the baseline regressions and
robustness tests as well as the estimates for mortality for different
ages at death and for various birth outcomes. Section 5 concludes.
2. Related literature
This paper is closely related to a growing literature in economics
that exploits natural experiments to identify the effects of air pollution on infant mortality. Chay and Greenstone (2003a,b) exploit
regulation- and recession-induced changes in total suspended particulate (TSP) pollution in U.S. counties in the 1970s and 1980s
and find that air pollution increases infant mortality with elasticities between 0.35 and 0.5. Currie and Neidell (2005) and Currie
et al. (2009) use within zip-code month variation and find a positive effect of carbon monoxide (CO) with elasticities of 0.09 and
0.04, respectively, but no effects for particulate matter (PM) and
ground-level ozone (O3 ). Currie and Schmieder (2009) estimate
elasticities for emissions of toxic chemicals ranging from 1.82 for
heavy metals to 6.11 for volatile organic compounds and 6.49 for
chemicals known to affect child development. Further, there is
recent evidence for adverse effects of air pollution on infant survival
from developing countries with studies using changes in air pollution due to a voluntary pollution prevention program in Mexico
(Foster et al., 2009), wildfires in Indonesia (Jayachandran, 2009),
and air quality regulations in China (Tanaka, 2010).
In contrast to these well-identified effects of air pollution, the
existing evidence on the effects of SO2 on infant mortality suffers
from omitted variables. Further, the existing evidence is rather
inconclusive. Studies assessing acute effects of SO2 pollution on
infant mortality from all causes with time-series analyses consistently find positive effects near a volcano in Japan in 1978–1988
(Shinkura et al., 1999), in Seoul in 1955–1999 (Ha et al., 2003),
in São Paulo in 1998–2000 (Lin et al., 2004), and in ten English
cities in 1969–1973 (Hajat et al., 2007). The last study estimates
an elasticity of 0.03. SO2 pollution has been observed to increase
acute deaths from the sudden infant death syndrome (SIDS) in
twelve Canadian cities in 1993–2003 (Dales et al., 2004), but not
from respiratory causes (Hajat et al., 2007). Results from studies
investigating chronic effects from prolonged exposure with crosssection, pooled panel, or case–control studies generally discover
no significant effects of SO2 on all-cause infant mortality. This is
the case for studies using data across U.S. standard metropolitan
statistical areas in 1961–1964 (Hickey et al., 1976), U.S. county
groups in 1970 (Medelsohn and Orcutt 1979), counties and boroughs in England and Wales in 1969–1973 (Chinn et al., 1981),
Czech districts in 1986–1993 and in 1989–1991 (Bobak and Leon,
1992; 1999), or large U.S. counties in 1999–2002 (Woodruff et al.,
2008). An exception is the study by Joyce et al. (1989) with data
from U.S. counties in 1970, which finds SO2 pollution to increase
neonatal mortality with an elasticity of 0.02–0.05. The results from
studies researching the effects of prolonged exposure on mortality from specific causes are conflicting. Chinn et al. (1981) and
Woodruff et al. (2008) find no effect on pneumonia or respiratory
mortality, while Bobak and Leon (1992; 1999) find positive effects
on respiratory mortality. Lipfert et al. (2000) and Woodruff et al.
(2008) discover no effect on SIDS mortality. Crocker et al. (1979)
estimate mortality from early infant diseases to increase with
SO2 in a cross-section of 60 U.S. cities in 1970 with an elasticity
of 0.09.
3. Pollution: data, evolution and instrument
From the German Federal Environmental Agency (Umweltbundesamt; hereafter UBA for short) I have data on the annual mean SO2
221
concentration measured at air quality monitors for 1985–2003.1,2
There are data for between 196 monitors in 1985 and 416 monitors
in 1994, or 553 monitors in total. I interpolate the monitor readings
on a grid with cells of 1 km2 . For the interpolation, I use the method
of inverse distance weighting with the 9 nearest monitors (without
any distance cutoff) and the inverse cubed distance as weights.
The UBA determined the parameters on the basis of empirical
studies. However, both interpolated values and regression results
are very similar for slightly different parameters.3 Following the
approach suggested in Currie and Neidell (2005), I evaluate the
accuracy of the interpolation procedure by comparing at each
monitor actual readings with the concentration level that would
be estimated with the interpolation procedure if this particular
monitor was not there. The correlation of 0.87 implies that the
interpolated values are accurate. The results are very similar if
the sample is reduced to county-years with active monitors.4 The
number of monitors changes over time and the placement of the
monitors may be endogenous. However, if I use pollution measures
based on the 64 continuously operating monitors, the results get
even stronger but are broadly similar.5
To merge the air pollution data with the mortality data, I aggregate the interpolated pollution values to the county level by taking
the average value of all grid cells that fall within the borders of a
county.6 Fig. 2 depicts the mean SO2 concentration per county in
1985, 1990, 1995, and 2000.
Looking at the pattern and evolution of SO2 pollution, two
aspects are worth noting. First, in the mid-1980s, pollution levels
were high, especially at three hotspots, the Ruhr area in the west,
Northern Hesse in the center, and the area around Leipzig in the
east. Back then, these areas were important industrial centers and
coal mining areas. Second, air quality diminished substantially after
1985 and 1990 in West Germany and after 1990 in East Germany.
These improvements are largely the result of the large combustion
plant ordinance enacted in 1983. The ordinance required operating
companies to retrofit fossil fuel fired power plants with scrubbers
within three to nine years from 1986 on. Retrofitting deadlines
were statutorily fixed and depended on a plant’s capacity and
emissions. Thus, they were not chosen by regulators or operating
companies. The unification treaty of 1990 extended the regulation
to East Germany with East German power plants required to install
1
The pollution data and identification strategy have been previously used in
Luechinger (2009) to estimate the effect of air pollution on subjective well-being.
The description of the data and identification strategy in this section is based on the
description of Luechinger (2009).
2
The UBA calculates annual mean concentrations as simple averages based on
daily or hourly means. The temporal aggregation follows EU rules which require
minimum data capture of 50 percent for annual means (2001/752/EC and email
from Andrea Minkos, UBA, June 14, 2013).
3
The baseline estimates reported in column 1 of Table 2 for the cubed distance
and the nine nearest monitors are 0.026 (std. err.: 0.004) (fixed-effects estimate) and
0.044 (std. err.: 0.009) (instrumental variable estimate). Using the squared distance
and the six nearest monitors, the respective estimates are 0.026 (std. err.: 0.004)
and 0.044 (std. err.: 0.009), using the cubed distance and the six nearest monitors
0.030 (std. err.: 0.005) and 0.046 (std. err. 0.010) and using the squared distance and
nine nearest monitors 0.027 (std. err.: 0.004) and 0.044 (std. err. 0.009).
4
The baseline estimates reported in column 1 of Table 2 are 0.026 (std. err.: 0.004)
(fixed-effects estimate) and 0.044 (std. err.: 0.009) (instrumental variable estimate).
For the reduced sample the respective estimates are 0.021 (std. err.: 0.006) and 0.039
(std. err.: 0.012). Thus, despite of a reduction of the sample size by nearly 60 percent,
the results are quantitatively similar and still precisely estimated.
5
The baseline estimates reported in column 1 of Table 2 are 0.026 (std. err.: 0.004)
(fixed-effects estimate) and 0.044 (std. err.: 0.009) (instrumental variable estimate).
The respective estimates with pollution measures based on the 64 continuously
operating monitors are 0.046 (std. err.: 0.009) and 0.064 (std. err.: 0.014).
6
County mergers in East Germany caused the number of counties to fall from 543
in 1993 to 439 in 2001. The analysis in this paper is based on the 439 counties at
the end of the merging process.
222
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
the product of the operating status, pre-desulfurization or residual emissions, a distance decay function, f(Dcj ), and the frequency
county c lies downwind of power plant j, g(Rcj ). The contribution
from all power plants is the sum of the contributions from individual plants multiplied by an unknown parameter ˛1 converting
pollutant mass into concentration levels. Thus, the SO2 concentration in county c at time t can be described by Eq. (1):
Pct = ˛0 + ˛1
1(active)jt · Ej · (1 − ˛2 · 1(scrubber)jt )
j
·f (Dcj ) · g(Rcj ) + c + t + εct .
(1)
Slightly re-arranging terms yields Eq. (2):
Pct = ˛0 + ˛1
−˛1 ˛2
1(active)jt · Ej · f (Dcj ) · g(Rcj )
j
(2)
1(active)jt · Ej · 1(scrubber)jt · f (Dcj ) · g(Rcj ) + c + t + εct .
j
The third term in Eq. (2),
1(active)jt · Ej · 1(scrubber)jt · f (Dcj ) ·
j
g(Rcj ), is the estimated effect of desulfurization at power plants on
SO2 pollution, which I will use to identify the effects of air pollution.
The temporal variation comes from two sources: The retrofitting
of power plants with scrubbers and changes in the power plant
population. Only the first is exogenous. Changes in power plant
population may be related to unobserved factors with independent
effects on infant mortality. Therefore, I include the effect of changes
in the
power plant population captured by the second term in Eq.
1(active)jt · Ej · f (Dcj ) · g(Rcj ), as a control variable. The geo-
(2),
j
Fig. 2. SO2 concentration in German counties; 1985, 1990, 1995 and 2000. Legend:
≤ 20 ␮g/m3 ,
20–40 ␮g/m3 ,
40–60 ␮g/m3 ,
60–80 ␮g/m3 ,
80–100 ␮g/m3 ,
100–125 ␮g/m3 ,
125–150 ␮g/m3 and > 150 ␮g/m3 ; cities:
D = Dortmund in the Ruhr area, K = Kassel in Northern Hesse, L = Leipzig and
B = Berlin.
Sources: See text.
scrubbers from 1993 on. Again, different retrofitting deadlines for
different categories of power plants were statutorily fixed.
Below I will use the estimated effect of the mandated installation of scrubbers at power plants to identify the effects of air
pollution on infant mortality. This effect is estimated with a simple model of air pollution, information on power plants’ operating
status, annual pre-desulfurization SO2 emissions, and retrofitting
status, and information on wind directions as well as distances and
directions between power plants and counties.
The SO2 concentration in county c at time t, Pct , comprises
contributions from power plant emissions and background pollution. Background pollution is captured by county effects, c ,
time effects, t , and a random component, εct . The contribution
of emissions from an individual plant j depends on the operating status and the retrofitting status of the power plant, 1(active)jt
and 1(scrubber)jt , pre-desulfurization emissions or residual emissions, Ej or Ej · (1 − ˛2 · 1(scrubber)jt ), and the distance and direction
between the plant and the county, Dcj and Rcj . 1(active)jt is a dummy
with value one if a power plant operates at time t and zero otherwise, 1(scrubber)jt is a dummy with value one if the power plant
has installed a scrubber at time t and zero otherwise, and ˛2 is
an unknown parameter reflecting average separation efficiency
of scrubbers. The contribution of an individual power plant j is
graphical variation also comes from two sources: Distance to power
plants and the distribution of wind directions.
The estimated effect of the installations of scrubbers on air pollution is the interaction of temporally and geographically varying
components and, thus, similar to a difference-in-difference term.
However, the term differs from a standard difference-in-difference
term in three respects. First, treatment and control group status is a matter of degree rather than one of kind and depends
on distance and wind direction frequencies. Second, I have to
include all power plants simultaneously. Therefore, my treatment
variable is a weighted sum of desulfurization at all plants with predesulfurization emissions as weights. Third, some power plants are
newly constructed, others taken offline. For this reason, I control
for changes in the power plant population.
For over 300 fossil fuel fired generating units with a capacity of
100 MW or more and active between 1985 and 2003, I have information on the starting year, the year the unit was taken offline, the
year of desulfurization, capacity, fuel and fuel efficiency. The information comes from the UBA, publications of operating companies
and the engineering literature, a questionnaire sent to operating companies, and statutory provisions (see Luechinger, 2009 for
details). Panel A of Fig. 3 pictures power plants’ locations. I use published emission factors (Bakkum et al., 1987) and plants physical
characteristics to estimate the pre-desulfurization SO2 emissions.7
7
Lacking data on utilization rates, I have to assume full utilization of capacities.
The emissions are then simply the product of the emission factor, the capacity,
inverse fuel efficiency, and the time period. This calculation may overstate emissions because the assumption of full utilization may not be plausible. However,
for my purpose, the absolute level of emissions is immaterial. I am only interested
in the relative size of pre-desulfurization emissions from different power plants,
which are mainly due to differences in fuels and plant size. Further, using actual
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
Fig. 3. Locations of fossil fuel fired power plants and wind stations.
Sources: See text.
I model distance decay with an exponential curve and a characteristic decay distance of 480 km, i.e. g(Dcj ) = exp(−2.1E-6 · Dcj ), as
proposed by field estimates (Schwartz, 1989; Summers and Fricke,
1989). Frequencies of wind directions in 12 30-degree sectors at
the nearest wind station characterize the wind situation at each
power plant (Traup and Kruse, 1996). Panel B of Fig. 3 depicts the
43 wind stations used in this paper. Finally, I use the distance and
direction between every power plant and every county to relate
plant- and county-level data.
To interpret the effect of the mandated installations of scrubbers at power plants as the causal effect of improved air quality, I
need to assume that the retrofitting of power plants affected infant
health only through its effect on air quality. In this context, it is
important to note that the statutory provisions were enacted prior
to the sample period. Thus, it is unlikely that the actual installation
of scrubbers is a response to concurrent demographic, economic
or political developments in faraway upwind or nearby downwind
regions. However, one might worry that counties lying downwind
in the vicinity of a retrofitted power plant experienced improvements in air quality and infant health before the plant installed
the scrubber. In this case, the difference-in-difference term in Eq.
(2) that is used as an instrument would partly capture these preexisting trends. In order to assess this issue, Fig. 4, panels A and
C, plots for each county the differences in SO2 concentration and
infant mortality between the first year of the desulfurization process and the last year before the process against the estimated effect
of desulfurization averaged over the whole sample period. The predesulfurization differences relate to the years 1986–1985 for West
Germany and the years 1993–1992 for East Germany. Fig. 4, panels B and D, also plots the differences in SO2 concentration and
infant mortality between the third and the first year of desulfurization process against averaged values of the estimated effect of
desulfurization. The post-desulfurization differences relate to the
years 1988–1986 for West Germany and the years 1995–1993 for
East Germany. These years mark the time frame during which the
most polluting power plants were required to install scrubbers.
Fig. 4 plots both actual values and Kernel-weighted local polynomial regression-smoothed values.
As we can see from Fig. 4, in the years before the desulfurization
process, average values of the estimated effect of desulfurization
utilization rates may be problematic because these rates are endogenous and potentially related to unobserved factors.
223
are not related to changes in SO2 concentration. The correlation
() is −0.05 (p = 0.294). In contrast, average values of the estimated
effect of desulfurization are strongly related to changes in SO2
concentration after the desulfurization process started ( = −0.49;
p < 0.0001). The same pattern holds for infant mortality rates even
though it may be less visible to the naked eye. The correlation
for pre-desulfurization is 0.05 (p = 0.276), post-desulfurization it
is −0.11 (p = 0.018).8
Another worry might be that that the retrofitting of power
plants lowered concentrations of other pollutants in addition to
SO2 . However, the German situation analyzed in this paper is wellsuited to specifically isolate exogenous variation in SO2 pollution.
Air quality regulation in other countries, notably the U.S., often
leaves authorities at subnational levels substantial leeway in formulating their own implementation plans and in deciding how
pollution targets are reached. Similarly, changes in economic activity are likely to affect several pollutants. In contrast, the large
combustion plant ordinance specified not only targets but also
measures and it left local authorities or operating companies little room for discretion. The regulation of TSP had a long tradition
in Germany before the large combustion plant ordinance. However, in addition to emission limits for SO2 , the large combustion
plant ordinance did establish for the first time emission limits for
nitrogen oxides (NOx ), though the limits were preliminary and
generally not binding. Further, for four reasons my instrument is
unlikely to capture changes in NOx pollution. First, power plants
are the most important source of SO2 , but not of NOx pollution.
For example, in 1990 power plants accounted for 60 percent of
all SO2 emissions but only for 20 percent of NOx emissions; NOx
emissions are primarily caused by road traffic (UBA, 2012). Second,
SO2 emission factors of coal and oil fired power plants are two to
five times as large as NOx emission factors. Conversely, gas fired
power plants are important emitters of NOx , but not SO2 (Bakkum
et al., 1987). Third, 93 percent of the installed scrubber capacity
was wet scrubbers with a water-gypsum-limestone/lime-slurry as
scrubbing liquid (Mittelbach, 1991). This scrubbing technology is
aimed at removing SO2 , not NOx . Fourth, and most importantly, for
the years 1990–2003, the years for which NOx data are available,
there is a strong first stage for SO2 (t-value: −4.29) but none for
NOx (t-value: 0.45).
4. Effect of SO2 pollution on infant mortality
4.1. Data and empirical strategy
The state statistical agencies collect data on births and deaths in
standardized way in accordance with federal laws and then publish infant mortality rates at the county level in state reports or
make them available upon request. The German Youth Institute, an
independent children and family research institute, compiled the
data for years since 1986 and generously shared the data with me.
The 1985 data come directly from the statistical agencies.9 Infant
8
It is important to note that the actual development stacks the deck against
finding the instrument to be valid. Even though power plants were not statutorily required to install scrubbers before 1986 and 1993 in West and East Germany,
respectively, individual power plants (<6 percent) were retrofitted before these
dates to test desulfurization technologies or in connection with regular overhauls.
9
Data on infant mortality are available for 326 Western counties and West
Berlin/Berlin for 19 years (1985–2003) and for 112 Eastern counties over 14 years
(1990–2003) or for a maximum of 7781 county-years. I drop observations when
the infant mortality rates are decreasing with increasing life span, e.g. if the mortality rate for deaths within the first 28 days is larger than the mortality rate for
deaths within the first year, because this is indicative of coding errors. This reduces
the number of observations by 73. Further, 3 observations for one county are
224
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
A. Pre-desulfurization pollution differences
B. Post-desulfurization pollution differences
20
20
0
0
-20
-20
-40
-40
-60
-60
Average estimated effect of desulfurization per county
C. Pre-desulfurization mortality differences
Average estimated effect of desulfurization per county
D. Post-desulfurization mortality differences
20
20
10
10
0
0
-10
-10
-20
-20
Average estimated effect of desulfurization per county
Average estimated effect of desulfurization per county
Fig. 4. Pre- and post-desulfurization pollution and infant mortality differences and average estimated effect of desulfurization per county. Notes: (1) Pre-desulfurization
differences relate to the years 1986–1985 for West Germany and the years 1993–1992 for East Germany, post-desulfurization differences to the years 1988–1986 for West
Germany and 1995–1993 for East Germany. (2) The panels show the actual and Kernel-weighted local polynomial regression-smoothed values. Stata’s default options are
used for Kernel smoothing, i.e. an Epanechnikov kernel function, zero degree polynomial and rule-of-thumb bandwidth.
mortality is reported as the number of deaths within the first year
per 1000 live births. According to the summary statistics in Table 1,
5.76 infants per 1000 live births died on average in the sample
period. For the years 1986–1999, I have disaggregated data on the
number of deaths within 1 day and 28 days. Due to changes in federal laws, states no longer reported detailed information on early
infant deaths in later years.
For the analysis on disaggregated mortality rates and birth
outcomes, I also use microdata stemming from the death and
birth certificates for the years 1991–2003.10 I combine the birth
and the mortality data by aggregating both to the level of
county × year × sex × legitimacy-cells. For confidentiality reasons,
the microdata are hosted at the Research Data Centers of the Federal Statistical Office and the statistical offices of the Länder and
have to be analyzed by remote access. In 1991–2003, 5.33 infants
died on average per 1000 live births.
The explanatory variable of interest is the annual mean concentration of SO2 introduced in Section 3. The mean concentration
missing in the original data. 117 observations for Eastern counties (incl. Berlin)
before 1991 are dropped because of missing control variables. However, dropping these 117 observations has little effect on the results. The baseline estimates
reported in column 1 of Table 2 are 0.026 (std. err.: 0.004) (fixed-effects estimate)
and 0.044 (std. err.: 0.009) (instrumental variable estimate). For the larger sample
with 7705 observations, the respective estimates are 0.023 (std. err.: 0.003) and
0.041 (std. err.: 0.008). The original data refer to infant mortality rates in counties
that existed at that particular point in time. I account for county mergers in East
Germany (see footnote 8) by taking simple averages of merging counties since I lack
data on births and since counties did not always merge integrally.
10
Microdata for Saarland in 1991 and Mecklenburg-Vorpommern in 1991–1994
are missing.
is 16.03 ␮g/m3 in the 1985–2003 sample and 12.18 ␮g/m3 in the
1991–2003 sample.
A first set of control variables includes basic economic and
demographic variables, namely GDP per capita, employment and
population. The data are from Cambridge Econometrics. A second
set of control variables includes weather variables. Weather and,
in particular, extreme weather conditions affect both pollution and
mortality (Deschênes and Greenstone, 2011). Therefore, I control
for the mean temperature in the coldest month, the mean temperature in the hottest month, the mean precipitation in the driest
month, and the mean precipitation in the wettest month in some
specifications. To construct these variables, I use for each county
daily data from the nearest of 35 weather stations from the German meteorological office (Deutscher Wetterdienst) or the European
Climate Assessment and Data with continuous readings.
In the extensions with the microdata aggregated to the
county × year × sex × legitimacy-level, I control for sex and legitimacy status of infants as well as for three characteristics of mothers
at the cell-level: the average age of mothers, the percentage of
mothers with German citizenship, and the percentage of working
mothers.
In the following, I estimate different variants of the following
empirical model:
IMRct = ˇ0 + ˇ1 Pct + ˇ2 Zct + c + t + εct ,
(3)
where IMRct is the infant mortality rate in county c in year t,
Pct the SO2 concentration in this county and year, Zct a vector of
control variables, c and t county and year effects and εct an
error term. The relationship between pollution and mortality is
modeled linearly since F-tests reject models with second or third
order polynomials in the present case.
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
225
Table 1
Summary statistics.
Obs
A. County × year-level data, 1985–2003
Infant mortality rate (w/in 1 year; per 1000 live births)
SO2 (␮g/m3 )
B. County × year-level data, 1986–1999
Infant mortality rate (w/in 1 year; per 1000 live births)
Infant mortality rate (w/in 1 day; per 1000 live births)
Infant mortality rate (w/in 28 days; per 1000 live births)
SO2 (␮g/m3 )
C. County × year × sex × legitimacy-level data, 1991–2003
Infant mortality rate (w/in 1 year; per 1000 live births)
Infant mortality rate (w/in 1 day; per 1000 live births)
Infant mortality rate (w/in 28 days; per 1000 live births)
Stillbirths (per 1000 births)
Weight
Small for gestational age, weight (per 1000 live births)
Length
Small for gestational age, length (per 1000 live births)
SO2 (␮g/m3 )
Mean
Median
Std. Dev.
Min
Max
7588
7588
5.76
16.03
5.40
10.50
2.91
16.13
0.00
1.54
23.30
174.07
5481
5481
5481
5481
6.03
1.45
3.35
18.14
5.70
1.30
3.10
13.10
2.86
1.25
1.99
16.03
0.00
0.00
0.00
1.67
23.30
9.40
20.40
174.07
22,500
22,500
22,500
22,500
22,500
22,500
22,500
22,500
22,500
5.33
1.44
3.10
4.37
3331
36.45
50.98
34.34
12.18
4.01
0.00
1.07
2.98
3330
33.71
51.02
30.97
7.52
6.67
3.30
4.95
5.79
102
18.11
0.69
18.77
15.08
0.00
0.00
0.00
0.00
2876
0.00
47.73
0.00
1.54
142.86
48.78
83.33
114.29
3756
208.33
53.56
193.55
174.07
Sources: (1) German Youth Institute; statistical offices of the Länder; (2) Research Data Centres of the Federal Statistical Office and the statistical offices of the Länder,
statistics of births and statistics of deaths, 1991–2003; (3) Federal Environmental Agency; own analysis.
Notes: Small for gestational age is here defined as birth weight or length more than two standard deviations below the mean of the birth cohort.
To account for serial and spatial correlations, I cluster standard
errors at the level of 92 analytical regions defined by the Federal
Office for Building and Regional Planning. These regions, determined on the basis of commuter flows, encompass an economic
center and its periphery and are often congruent with sates’ planning regions. Thus, it is likely that counties within these regions are
exposed to common shocks.
4.2. Basic results
Table 2 presents the basic results for the whole sample. It
presents the results from OLS (panel A), fixed-effects (panel B), and
instrumental variable (panel E) regressions as well as the reduced
form (panel C) and first stage (panel D) related to the instrumental
variable regressions. All estimates are intended to assess the health
benefits of the large combustion plant ordinance, which accounts
for a large part of the reduction in SO2 pollution in Germany.
Though most of the variation in SO2 pollution is the result of
this specific regulation, highly polluted counties and counties with
large regulation-induced improvements in air quality will differ
from other counties along several, potentially unobservable dimensions. Fixed-effects in panels B-E account for this cross-sectional
heterogeneity across counties. In addition, the instrumental variable estimates correct for classical measurement errors and for
unobserved factors that may have accompanied non-regulationinduced changes in pollution. Table 2 presents the results for three
different specifications: Column 1 presents estimates with only
year effects as controls, column 2 adds economic variables, and
column 3 weather conditions.
The results reported in Table 2 suggest that infant mortality
increases with air pollution. According to the OLS estimates, 0.032
infant lives are lost per 1000 live births for every 1 ␮g/m3 increase
in SO2 concentration (column 3). Accounting for cross-sectional
heterogeneity by including fixed-effects only marginally affects
the results. In column 3, the point estimate is slightly reduced
to 0.026 infant deaths per 1000 live births. Using the exogenous
variation introduced by the large combustion plant ordinance, we
can see that infant mortality decreases with increasing reductions in air pollution due to flue gas desulfurization at power
plants. Thus, the environmental regulation had the intended health
effects. The first stage results show that actual SO2 concentration
is lower, the larger the predicted reduction in SO2 concentration
is. As explained in Section 3, the respective variable is akin to a
difference-in-difference term with the temporal variation coming
from the staggered installation of scrubbers at power plants and the
geographical variation from the distance and direction between a
county and power plants, whereby the direction determines the
frequency the county lies downwind of the plants. If I construct
a similar term without taking wind frequency into account, the
resulting instrumental variable estimates are very similar to the
ones reported in Table 2 (differences of less than 6 percent). Thus,
distance to power plants seems to be more important for the differential effects of desulfurization across counties than direction.
The size of the reduced form coefficients is difficult to interpret.
If I scale the coefficients by the respective first stage estimates, the
resulting instrumental variable estimates imply a marginal effect of
SO2 on infant mortality of 0.045 (column 3). Thus, the instrumental
variable estimates are slightly larger, but the difference is small
and Hausman tests reject the null hypotheses of equality of fixedeffects and instrumental variable estimates at the 5 percent level
at best. This is consistent with the notion that most of the variation
in pollution is the result of the large combustion plant ordinance.
There are several ways to put the results into perspective. In
the following, I will compute the implied elasticity and compare
them to previous estimates, calculate the number of infant lives
saved by the improvement in air quality, estimate the contribution of improvements in air quality to the overall decline in infant
mortality rates, and monetize the reductions in lost infant lives.
The marginal effects of 0.026–0.045 translate into an elasticity
in the range of 0.07–0.13, which is at the lower end of the elasticities reported in the recent economics literature. It is lower than
the elasticities for chronic effects of toxic chemicals of 1.82–6.49
(Currie and Schmieder, 2009) and for chronic effects of TSP of
0.35–0.5 (Chay and Greenstone, 2003a,b), but larger than the
elasticities for acute effects of CO of 0.04–0.09 (Currie and Neidell,
2005; Currie et al., 2009). It is larger than an estimated elasticity
for acute effects of SO2 of 0.03 (Hajat et al., 2007). There are
several reasons for these differences. First, pollutants differ in their
toxicity and the degree to which exposure to them can be avoided.
The wide range of elasticities reported by Currie and Schmieder
(2009) for different toxic chemicals nicely illustrates this point.
Second, acute effects may differ from chronic effects. For example,
chronic effects of particulate pollution on infant mortality are
generally larger than acute effects (Lacasaña et al., 2005). Finally, as
226
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
Table 2
Basic results: Effect of SO2 pollution on infant mortality, 1985–2003.
(1)
(2)
(3)
0.029**
(0.005)
0.25
0.030**
(0.005)
0.25
0.032**
(0.005)
0.26
0.026**
(0.004)
0.27
0.026**
(0.004)
0.27
0.026**
(0.004)
0.27
−0.224**
(0.051)
0.26
−0.217**
(0.051)
0.26
−0.226**
(0.052)
0.27
−5.135**
(0.502)
0.72
104.49
−4.949**
(0.468)
0.73
111.89
−4.995**
(0.484)
0.74
106.39
R2
Hausman test p-value
0.044**
(0.009)
0.27
0.073
0.044**
(0.009)
0.27
0.068
0.045**
(0.009)
0.27
0.047
Control variables
Year effects
Economic variables
Climate variables
No. of observations
No. of counties
No. of clusters
Y
N
N
7588
439
92
Y
Y
N
7588
439
92
Y
Y
Y
7588
439
92
A. OLS estimates
Dependent variable
Infant mortality rate
Pollution
SO2
R2
B. Fixed-effects estimates
Dependent variable
Infant mortality rate
Pollution
SO2
R2
C. Reduced form estimates
Dependent variable
Infant mortality rate
Pollution
Predicted SO2
R2
D. First stage estimates
Dependent variable
SO2
Excluded instrument
Predicted SO2
R2
F-stat excl. instrument
E. IV estimates
Dependent variable
Infant mortality rate
Pollution
SO2
Notes: (1) Panel A reports OLS regressions; panels B-E report regressions with county
fixed effects. The instrumental variable is the estimated reduction in SO2 pollution
due to desulfurization at power plants for each county-year. It is estimated by summing over all power plants the product of a dummy variable with value one if a
power plant operates at time t and zero otherwise, estimated pre-desulfurization
emissions, a dummy variable with value one if a power plant has installed a scrubber
at time t and zero otherwise, a distance decay function, and the frequency a county
lies downwind of a power plant. The instrumental variable is included together
with a control variable capturing the estimated effect of changes in the power plant
population on SO2 pollution. It is estimated in a similar way as the instrumental
variable but without the dummy variable for the installation of scrubbers. (2) The
unit of observation is the county-year. (3) Cluster robust standard errors in parentheses allow for clustering at the level of 92 analytical regions encompassing several
counties. (4) ** is significant at the 99 percent level.
discussed in the introduction, differences across countries regarding health care utilization and many other aspects may influence
the relationship between air pollution and infant mortality.
The average decrease in SO2 concentration in counties between
the first and the last year that they are in the sample is around
43 ␮g/m3 . With around 746,000 live births each in year in Germany,
the number of infant lives saved amounts to around 826 for the
fixed-effects estimates and to around 1460 for the instrumental
variable estimates.
The average decrease in infant mortality in counties over the
sample period is 4.40 per 1000 live births. Thus, by multiplying
the average reduction in SO2 pollution of 43 ␮g/m3 with the
coefficients in column 3 and then relating the resulting product to
the average decrease in infant mortality rates, I find that between
25 and 44 percent of the decrease in infant mortality over the
sample period is due to the improvement in air quality.
If we are prepared to monetize these benefits in terms of infant
health, we can compare them to the costs of regulation. Chay and
Greenstone (2003a,b) use a value of a statistical life (VSL) estimate
of $1.7 million (2000 US$) based on a study by Ashenfelter and
Greenstone (2004); Currie and Neidell (2005) use EPA’s estimate of
$5.0 million (2000 US$); the median VSL estimate for prime aged
workers in the U.S. in the meta-analysis of Viscusi and Aldy (2003)
is $6.7 million (2000 US$). Hence, the estimates differ by an order of
magnitude and so will the estimates of infant health benefits. The
lowest annual benefit estimate for each of the around 27,793,000
households (and 727,199 births) in the year 1989/1990 in West
Germany based on the fixed-effects estimate and the VSL estimate
of Ashenfelter and Greenstone (2004) is around $50 (2000 US$), the
highest estimate based on the instrumental variable estimate and
the VSL estimate of Viscusi and Aldy (2003) is around $343 (2000
US$).
These benefit estimates compare favorably to rough compliance
cost estimates for West Germany in the range of between $33 and
$165 per year and household (2000 US$).11 Of course, the large
combustion plant ordinance is not the sole reason for improvements in air quality. At the same time, the reduction in infant
mortality may well reflect much broader health benefits. Further,
a potentially important cost of air pollution not reflected in the
estimates are the defensive steps individuals take in order to avoid
exposure to air pollution or adverse health effects (Bresnahan et al.,
1997; Neidell, 2009; Graff Zivin and Neidell, 2009; Moretti and
Neidell, 2011; Deschênes et al., 2012). Thus, it is very likely that
overall benefits of cleaner air exceed the health benefits presented
here.
4.3. Robustness tests
The analysis so far has been based on the annual mean SO2 concentration in counties. In the present context, this is the appropriate
pollution measure for three reasons. First, I am interested in chronic
effects resulting from prolonged exposure to air pollution, not in
acute effects to extreme events. Second, the temporal resolution of
the infant mortality data is inherently annual. Third, lacking temporally more disaggregated data on power plants, the instrumental
variable also has an annual resolution. Nevertheless, I also estimate
the effects for three alternative measures of SO2 concentration, the
annual maximum concentration and the number of days the German and the former U.S. 24-h air quality standards of 125 ␮g/m3
and 365 ␮g/m3 are exceeded.12 With the exception for the
11
Schärer and Haug (1990) estimate costs of the desulfurization at West German
power plants at DM 14.2 billion (1988 DM). I double this value to crudely consider
operating costs, assume a real long-term interest rate of 5 percent, and divide by
27,793,000 households (living in Germany in 1989). This results in costs of $33 (2000
US$) per household and year. At the other end, Schulz’ (1985) most pessimistic cost
forecast is DM 9 per person and month (1984 DM assumed) or $165 (2000 US$) per
household and year.
12
Germany has since 2002 ambient air pollution standards at the federal level.
The 24-h SO2 standard of 125 ␮g/m3 is not to be exceeded more than three times
a year (22nd BImSchV, September 11, 2002). In the U.S. in 1971–2010, the primary SO2 standard was the 24-hour standard of 365 ␮g/m3 , which was not to be
exceeded more than once per year (see, e.g., Federal Register, “National ambient
air quality standards for sulfur oxides (sulfur dioxide) – Final decision,” May 22,
1996, p. 25568). Since 2010, there is only a 1-h standard in the U.S. for SO2 (Federal
Register, “Primary National Ambient Air Quality Standard for sulfur dioxide,”
June 22, 2010). For computational reasons, daily monitor readings are not first
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
227
Table 3
Robustness tests: Subsamples.
Dependent variable
Infant mortality rate
Fixed-effects estimates
SO2
R2
IV estimates
SO2
2
R
F-stat excl. instrument
Control variables
Year effects
County effects
Economic variables
Climate variables
No. of observations
No. of counties
No. of clusters
Total
(1)
West
(2)
East
(3)
1985–1990
(4)
1991–2003
(5)
0.026**
(0.004)
0.27
0.019*
(0.007)
0.31
−0.003
(0.009)
0.21
0.023*
(0.010)
0.09
0.021**
(0.005)
0.12
0.045**
(0.009)
0.27
106.39
0.040**
(0.012)
0.31
133.81
0.101
(0.115)
0.09
0.60
0.042*
(0.017)
0.09
75.13
0.047**
(0.014)
0.11
16.21
Y
Y
Y
Y
7588
439
92
Y
Y
Y
Y
6148
326
71
Y
Y
Y
Y
1440
113
21
Y
Y
Y
Y
1956
326
71
Y
Y
Y
Y
5632
439
92
Notes: (1) OLS fixed-effects and IV fixed-effects regressions. The instrumental variable is the estimated reduction in SO2 pollution due to desulfurization at power plants for
each county-year. It is estimated by summing over all power plants the product of a dummy variable with value one if a power plant operates at time t and zero otherwise,
estimated pre-desulfurization emissions, a dummy variable with value one if a power plant has installed a scrubber at time t and zero otherwise, a distance decay function,
and the frequency a county lies downwind of a power plant. The instrumental variable is included together with a control variable capturing the estimated effect of changes
in the power plant population on SO2 pollution. It is estimated in a similar way as the instrumental variable but without the dummy variable for the installation of scrubbers.
(2) The unit of observation is the county-year. (3) Cluster robust standard errors in parentheses allow for clustering at the level of 92 analytical regions encompassing several
counties. (4) * is significant at the 95 percent level and ** at the 99 percent level.
fixed-effects estimate based on the number of days the former U.S.
24-h standard is exceeded, I find for all these alternative measures statistically significant positive effects on infant mortality.
However, the estimated effects are smaller with elasticities of 0.04
(fixed-effects estimate) and 0.09 (instrumental variable estimate)
for the annual maximum concentration, 0.01 and 0.03 for the number of days the German 24-h standard is exceeded, and 0.002
and 0.02 for the number of days the former U.S. 24-h standard is
exceeded.13
To further test the robustness of the results, I estimate the baseline regressions for several subsamples, namely for West Germany,
East Germany, the 1985–1990 period, and the 1991–2003 period,14
and I control for additional confounding variables.
Table 3 presents the results for the different subsamples. The
estimates are comparable across subsamples with the exception of
the East German subsample. In the East German subsample, SO2
pollution is unrelated to infant mortality. In case of the instrumental variable estimate, actual differences to other subsamples
are more pronounced than it appears from the point estimate,
since both the reduced form and the first stage regressions have
the wrong sign and are imprecisely estimated. Therefore, neither measurements nor predicted changes of SO2 concentration
have the expected effect on infant mortality in East Germany. To
interpret this result, it is important to remember that the large
improvements in air quality in East Germany coincided with large
interpolated on a grid and then aggregated to the county level but rather directly
interpolated to county centroids.
13
The means and estimated effects for the annual maximum concentration are
123.84 ␮g/m3 (mean), 0.002 (std. err.: 0.001) (fixed-effects estimate), and 0.004 (std.
err.: 0.001) (instrumental variable estimate). For the number of days in exceedance
of the German standard the respective figures are 4.93 days, 0.014 (std. err.: 0.007),
and 0.036 (std. err.: 0.009). For the number of days in exceedance of the former U.S.
24-h standard they are 0.36 days, 0.025 (std. err.: 0.035), and 0.255 (std. err.: 0.086).
14
Two reasons explain the uneven partition of the time periods. First, this partition roughly separates the main period of desulfurization in the West from the
main period of desulfurization is the East. Second, the subsample for the period
1991–2003 corresponds to the sample of the robustness analysis based on microdata
reported below.
economic and socio-demographic changes in the aftermath of
reunification.
An alternative way to look at spatial heterogeneity is by estimating differential effects of SO2 on infant mortality with distance
to the former East-West border. Fig. 5 presents the resulting estimates together with the baseline estimates reported in column 3
of Table 2. As can be seen, I find no evidence for differences in the
fixed-effects estimates with distance to the inner German border.
However, the instrumental variable estimate seems to be larger
in the West compared to the East, although the 95 percent confidence interval encompasses the baseline estimate except for the
western-most parts of Germany.
Table 4 presents the results from regressions with additional
control variables. For comparison, column 1 shows again the baseline estimates corresponding to column 3 of Table 2. In column 2, I
control for the TSP concentration. SO2 and TSP concentrations are
often correlated. Hence, without controlling for the TSP concentration, SO2 concentration may stand for air pollution at large. In
Bobak and Leon (1999), the inclusion of TSP concentration renders
the effect of SO2 concentration insignificant. This decrease in the
effect of SO2 does not imply that SO2 pollution is irrelevant. SO2 is
a precursor of TSP and, thus, the two pollutants are causally linked.
However, as can be seen from Table 4, adding TSP concentration
leaves the SO2 coefficient virtually unaffected.
In column 3, I augment the models with year specific
distance-to-city polynomials and year specific close-to-theEast-West-German-border effects. Year specific distance-to-city
polynomials should control for urban/rural trends, year specific
close-to-the-East-West-German-border effects for reunification
related developments. I follow Redding and Sturm (2008) and
define closeness as <75 km. They show that West German cities
within this distance to the East-West German border suffered
a sudden drop in population growth compared to other cities
with the breakup of Germany after the Second World War but
regained ground with the German reunification. Including these
additional control variables affects the fixed-effects estimates little
but slightly increases the instrumental variable estimate.
In column 4, I add county specific linear time-trends. These
time-trends flexibly account for changes in socio-demographic
228
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
Table 4
Robustness tests: Additional controls.
.04
R2
IV estimates
SO2
.02
Fixed-effects estimates
SO2
.03
Dependent variable
Infant mortality rate
2
0
.01
Marginal effect of SO2
.05
A. Fixed-effects estimates
-400
-200
0
200
Distance to border; negative for West Germany
(2)
(3)
(4)
0.026**
(0.004)
0.27
0.026**
(0.004)
0.27
0.026**
(0.005)
0.27
0.009
(0.008)
0.35
0.045**
(0.009)
0.27
106.39
0.046**
(0.010)
0.27
104.59
0.058**
(0.010)
0.27
73.38
0.071**
(0.027)
0.33
39.49
Y
Y
Y
Y
N
N
N
N
7588
439
92
Y
Y
Y
Y
Y
N
N
N
7588
439
92
Y
Y
Y
Y
N
Y
Y
N
7588
439
92
Y
Y
Y
Y
N
N
N
Y
7588
439
92
.05
.1
Notes: (1) OLS fixed-effects and IV fixed-effects regressions. The instrumental variable is the estimated reduction in SO2 pollution due to desulfurization at power
plants for each county-year. It is estimated by summing over all power plants the
product of a dummy variable with value one if a power plant operates at time t and
zero otherwise, estimated pre-desulfurization emissions, a dummy variable with
value one if a power plant has installed a scrubber at time t and zero otherwise, a
distance decay function, and the frequency a county lies downwind of a power plant.
The instrumental variable is included together with a control variable capturing the
estimated effect of changes in the power plant population on SO2 pollution. It is
estimated in a similar way as the instrumental variable but without the dummy
variable for the installation of scrubbers. (2) The unit of observation is the countyyear. (3) Cluster robust standard errors in parentheses allow for clustering at the
level of 92 analytical regions encompassing several counties. (4) ** is significant at
the 99 percent level.
0
Marginal effect of SO2
.15
B. Instrumental variable estimate
R
F-stat excl. instrument
Control variables
Year effects
County effects
Economic variables
Climate variables
TSP
Year spec. distance to city
Year spec. close to E-W border
County spec. time trends
No. of observations
No. of counties
No. of clusters
(1)
-400
-200
0
200
Distance to border; negative for West Germany
Fig. 5. Differential effects of SO2 on infant mortality with distance from former
East–West border. Notes: The solid black lines depict differential effects of SO2 on
infant mortality with distance from former East–West border, which are estimated
in models with an interaction term of SO2 × distance from border but otherwise
identical to the models in column 3 of Table 2. Red lines indicate the size of the
baseline effects reported in column 3 of Table 2. Dashed lines denote the 95 percent
confidence intervals for a standard normal distribution. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of
this article.)
factors, environmental policies, and other factors affecting
infant mortality. The inclusion of these time trends affects the
fixed-effects and instrumental variable estimates in different
directions: While the fixed-effects estimate is reduced by around
65 percent and becomes statistically insignificant, the instrumental variable estimate increases. There are two potential
explanation for the decrease in the fixed-effects estimate caused
by the inclusion of county specific time-trends. On the one hand,
the decrease may indicate that the baseline results are biased by
omitted variables. On the other hand, county specific time-trends
may eliminate much of the identifying variation in the fixedeffects estimates and, thereby, render measurement errors more
important.
4.4. Extensions
As discussed by Chay and Greenstone (2003a,b), one mechanism
through which air pollution affects infant mortality are adverse
effects on fetal development. To assess the importance of this
mechanism, they suggest looking at the effects of air pollution on
neonatal mortality because deaths due to poor fetal development
are likely to occur in the neonatal period. In this spirit, Table 5
presents the estimated effects of SO2 concentration on infant mortality within the first day (columns 1 and 4) and within the first
28 days (columns 2 and 5) after birth. Since I only have disaggregated county x year-level data for the period 1986–1999 and since
microdata are only available for the period 1991–2003, Table 5 also
reports the estimated effect for infant mortality within the first year
for the respective periods (columns 3 and 6).15
The fixed-effects estimates suggest that the effect of air pollution on infant mortality is not larger in the neonatal period and that
the effect even may be slightly smaller. There is no effect on deaths
within the first day in either period and on deaths within the first 28
days in 1986–1999. Deaths within the first 28 days increase with an
elasticity of 0.05 in 1991–2003, which corresponds to the elasticity
of 0.05 for deaths within the first year found both in 1986–1999
and in 1991–2003. Similarly, the instrumental variable estimates
too provide no evidence for a strongly disproportionate effect on
mortality in the neonatal period. The respective elasticities are 0.05
in 1986–1999 and 0.09 in 1991–2003 for deaths within the first
day, 0.10 in 1985–1999 and 0.11 in 1991–2003 for deaths within
15
The infant mortality data are inherently annual. Similarly, lacking more temporally disaggregated data on power plants, this is also true for the instrumental
variable. Thus, all estimates are based on estimates of contemporaneous effects of
air pollution on infant mortality and pregnancy outcomes. Using the contemporaneous level is also appropriate for assessing the effects of poor fetal development
due to exposure during pregnancy since air pollution during the last trimester has
been shown to be important for infant health (Currie et al., 2009).
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
229
Table 5
Infant mortality within 1 day, 28 days and 1 year and male and female infant mortality.
Dependent variable
Infant mortality rate
Fixed-effects estimates
SO2
R2
IV estimates
SO2
R2
Control variables
Individual and maternal variables
Year effects
County effects
Economic variables
Weather variables
No. of observations
No. of counties
No. of clusters
A. County × year-level data, 1986–1999
B. County × year × sex × legitimacy-level data, 1991–2003
W/in 1 day
(1)
W/in 28 days
(2)
W/in 1 year
(3)
W/in 1 day
(4)
W/in 28 days
(5)
W/in 1 year
(6)
Males
(7)
Females
(8)
−0.001
(0.002)
0.04
0.001
(0.004)
0.09
0.018**
(0.006)
0.22
0.002
(0.002)
0.05
0.014**
(0.004)
0.07
0.020**
(0.005)
0.12
0.024**
(0.006)
0.13
0.016**
(0.005)
0.10
0.004
(0.004)
0.04
0.018*
(0.008)
0.08
0.040**
(0.011)
0.22
0.011
(0.008)
0.05
0.027*
(0.012)
0.07
0.040*
(0.017)
0.12
0.042*
(0.019)
0.13
0.040*
(0.019)
0.10
–
Y
Y
Y
Y
5481
439
92
–
Y
Y
Y
Y
5481
439
92
–
Y
Y
Y
Y
5481
439
92
Y
Y
Y
Y
Y
22,500
439
92
Y
Y
Y
Y
Y
22,500
439
92
Y
Y
Y
Y
Y
22,500
439
92
Y
Y
Y
Y
Y
11,250
439
92
Y
Y
Y
Y
Y
11,250
439
92
Notes: (1) OLS fixed-effects and IV fixed-effects regressions. The instrumental variable is the estimated reduction in SO2 pollution due to desulfurization at power plants for
each county-year. It is estimated by summing over all power plants the product of a dummy variable with value one if a power plant operates at time t and zero otherwise,
estimated pre-desulfurization emissions, a dummy variable with value one if a power plant has installed a scrubber at time t and zero otherwise, a distance decay function,
and the frequency a county lies downwind of a power plant. The instrumental variable is included together with a control variable capturing the estimated effect of changes
in the power plant population on SO2 pollution. It is estimated in a similar way as the instrumental variable but without the dummy variable for the installation of scrubbers.
(2) The unit of observation is the county-year in Panel A and the county × year × sex × legitimacy-cell in Panel B. Regressions in Panel B are weighted by the number of live
births in each cell. (3) Cluster robust standard errors in parentheses allow for clustering at the level of 92 analytical regions encompassing several counties. (4) * is significant
at the 95 percent level and ** at the 99 percent level. Source of mortality data in Panel B: Research Data Centres of the Federal Statistical Office and the statistical offices of the
Länder, statistics of births and statistics of deaths, 1991–2003; own analysis. See text for other sources.
the first month, and 0.12 in 1985–1999 and 0.09 in 1991–2003 for
deaths within the first year. Overall, these estimates do not suggest
that poor fetal development caused by exposure during gestation
is an important biological mechanism.
As in other studies (Jayachandran, 2009; Tanaka, 2010), gender
differences are relatively small. For males the estimated elasticities
amount to 0.05 (fixed-effects estimate) and 0.09 (instrumental
variable estimate), for females to 0.04 (fixed-effects estimate) and
0.10 (instrumental variable estimate).
In order to further explore potential effects of exposure to air
pollution during pregnancy, Table 6 presents estimates on the
effect of SO2 on the number of stillbirths (per 1000 births) (column
Table 6
Birth outcomes.
Dependent variable
Fixed-effects estimates
SO2
R2
IV estimates
SO2
R2
Control variables
Individual and maternal variables
Year effects
County effects
Economic variables
Weather variables
No. of observations
No. of counties
No. of clusters
Stillbirths
(1)
Weight
(2)
SGA, weight
(3)
Length
(4)
SGA, length
(5)
0.005( *)
(0.003)
0.08
−0.104
(0.143)
0.86
0.024
(0.016)
0.23
0.001
(0.001)
0.89
0.061*
(0.026)
0.33
−0.006
(0.011)
0.08
−0.540
(0.415)
0.86
0.093( *)
(0.051)
0.22
−0.001
(0.002)
0.89
0.195**
(0.060)
0.33
Y
Y
Y
Y
Y
22,500
439
92
Y
Y
Y
Y
Y
22,500
439
92
Y
Y
Y
Y
Y
22,500
439
92
Y
Y
Y
Y
Y
22,500
439
92
Y
Y
Y
Y
Y
22,500
439
92
Source of birth variables: Research Data Centres of the Federal Statistical Office and the statistical offices of the Länder, statistics of births and statistics of deaths 1991–2003;
own analysis. See text for other sources.
Notes: (1) OLS fixed-effects and IV fixed-effects regressions. The instrumental variable is the estimated reduction in SO2 pollution due to desulfurization at power plants for
each county-year. It is estimated by summing over all power plants the product of a dummy variable with value one if a power plant operates at time t and zero otherwise,
estimated pre-desulfurization emissions, a dummy variable with value one if a power plant has installed a scrubber at time t and zero otherwise, a distance decay function,
and the frequency a county lies downwind of a power plant. The instrumental variable is included together with a control variable capturing the estimated effect of changes
in the power plant population on SO2 pollution. It is estimated in a similar way as the instrumental variable but without the dummy variable for the installation of scrubbers.
(2) SGA stands for small for gestational age. The variables here are similar in spirit to this measure but gestational age is unknown. Instead, the variables are defined here
as birth weight or length more than two standard deviations below the mean of the birth cohort. (3) The unit of observation is the county × year × sex × legitimacy-cell.
Regressions are weighted by the number of births in each cell in column 1 and by the number of live births in each cell in columns 2–5. (4) Cluster robust standard errors in
parentheses allow for clustering at the level of 92 analytical regions encompassing several counties. (5) ( *) is significant at the 90 percent level, * at the 95 percent level, and
** at the 99 percent level.
230
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
1), average birth weight (column 2), the number of infants with
a birth weight of more than two standard deviations below the
mean of their birth cohort (per 1000 live births) (column 3), average length at birth (column 4), and the number of infants whose
length is more than two standard deviations below the mean of
their birth cohort (per 1000 live births) (column 5).
According to the fixed-effects estimate the frequency of stillbirths increases with SO2 pollution, but this finding does not
carry over to the instrumental variable estimate. Conversely,
the instrumental variable estimate but not the fixed-effects estimate suggests that SO2 pollution increases probability of infants
with low weight, though the effect is imprecisely estimated.
The strongest effects are found for the chance that infants have
comparatively short length at birth. Interestingly, this finding is
consistent with a study on air pollution and infant health for
Germany by Coneus and Spiess (2012). Combining survey data and
high-frequency pollution data for 2002–2007, they find that SO2
pollution negatively affects length at birth but not birth weight.
In sum, the evidence suggests that exposure to SO2 during gestation affects fetal development and infant health at birth, but that
poor fetal development is not the main mechanisms through which
SO2 affects infant mortality.
5. Conclusion
The results of this paper suggest that the desulfurization at
power plants entailed considerable benefits in terms of infant
health. Thus, they are directly relevant for the retrospective evaluation of this policy in Germany and, arguably, policies in other
developed countries such as the emission trading program in the
U.S. They also point to potentially large gains from cleaning up the
air in developing countries with often very high levels of air pollution. This is all the more true as improved infant health is only one
benefit of air quality among others.
The estimated elasticities between SO2 pollution and infant
mortality are at the lower end of published elasticities between pollution and infant mortality. Still, the very large reductions in SO2
pollution in developed countries in the last century imply large
cumulative effects. In future research it would be interesting to
disentangle the various reasons behind differences in estimated
effects of pollution. Potential reasons range from differences in
the toxicity of pollutants, differences in how ambient air pollution
translates to individual exposure, to contextual factors that differ
across countries and influence individuals’ exposure and susceptibility to air pollution.
References
Ashenfelter, O., Greenstone, M., 2004. Using mandated speed limits to measure the
value of a statistical life. Journal of Political Economy 112 (S1), S226–S267.
Ashmore, M.R., Dimitroulopoulou, S., 2009. Personal exposure of children to air
pollution. Atmospheric Environment 43 (1), 128–141.
Bakkum, A., Bartelds, H., Duiser, J.A., Veldt, C., 1987. Handbook of Emission Factors,
Part 3. Stationary Combustion Sources. Ministry of Housing, Physical Planning
and Environment, The Hague, NL.
Bobak, M., Leon, D.A., 1992. Air pollution and infant mortality in the Czech Republic,
1986–88. Lancet 340 (8826), 1010–1014.
Bobak, M., Leon, D.A., 1999. The effect of air pollution on infant mortality appears
specific for respiratory causes in the postneonatal period. Epidemiology 10 (6),
666–670.
Bresnahan, B.W., Dickie, M., Gerking, S., 1997. Averting behavior and urban air pollution. Land Economics 73 (3), 340–357.
Chay, K.Y., Greenstone, M., NBER Working Papers No. 10053 2003a. Air quality, infant
mortality, and the Clean Air Act of 1970. National Bureau of Economic Research,
Cambridge, MA.
Chay, K.Y., Greenstone, M., 2003b. The impact of air pollution on infant mortality:
evidence from geographic variation in pollution shocks induced by a recession.
Quarterly Journal of Economics 118 (3), 1121–1167.
Chinn, S., du, C., Florey, V., Baldwin, I.G., Gorgol, M., 1981. The relation of mortality
in England and Wales 1969–73 to measurements of air pollution. Journal of
Epidemiology and Community Health 35 (3), 174–179.
Coneus, K.C., Spiess, K., 2012. Pollution exposure and child health: evidence for
infants and toddlers in Germany. Journal of Health Economics 31 (1), 180–196.
Crocker, T.D., Schulze, W., Ben-David, S., Kneese, A.V., 1979. Methods development
for assessing air pollution control benefits. Experiments in the Economics of Air
Pollution Epidemiology, vol. I. Environmental Protection Agency, Washington.
Currie, J., Gruber, J., 1996. Saving babies: the efficacy and cost of recent changes in
the Medicaid eligibility of pregnant women. Journal of Political Economy 104
(6), 1263–1296.
Currie, J., Neidell, M., 2005. Air pollution and infant health: what can we learn
from California’s recent experience? Quarterly Journal of Economics 120 (3),
1003–1030.
Currie, J., Neidell, M., Schmieder, J.F., 2009. Air pollution and infant health: lessons
from New Jersey. Journal of Health Economics 28 (3), 688–703.
Currie, J., Schmieder, J.F., 2009. Fetal exposures to toxic releases and infant health.
American Economic Review. Papers and Proceedings 99 (2), 177–183.
Dales, R., Burnett, R.T., Smith-Doiron, M., Stieb, D.M., Brook, J.R., 2004. Air pollution
and sudden infant death syndrome. Pediatrics 113 (6), e628–e631.
Deschênes, O., Greenstone, M., 2011. Climate change, mortality, and adaptation:
evidence from annual fluctuations in weather in the US. American Economic
Journal: Applied Economics 3 (4), 152–185.
Deschênes, O., Greenstone, M., Shapiro, J.S., NBER Working Paper No. 18267 2012.
Defensive Investments and the Demand for Air Quality: Evidence from the NOx
budget program and ozone reductions. National Bureau of Economic Research,
Cambridge, MA.
Foster, A., Gutierrez, E., Kumar, N., 2009. Voluntary compliance, pollution levels, and
infant mortality in Mexico. American Economic Review. Papers and Proceedings
99 (2), 191–197.
Graff Zivin, J., Neidell, M., 2009. Days of haze: environmental information disclosure
and intertemporal avoidance behavior. Journal of Environmental Economics and
Management 58 (2), 119–128.
Graff Zivin, J.S., Neidell, M., 2013. Environment, health, and human capital. Journal
of Economic Literature 51 (3), 689–730.
Gruber, J., 1997. Health insurance for poor women and children in the U.S: lessons
from the past decade. In: James Poterba (Ed.), Tax Policy and the Economy, vol.
11. MIT Press, Cambridge, MA, pp. 169–211.
Ha, E.-H., Lee, J.-T., Kim, H., Hong, Y.-C., Lee, B.-E., Park, H.-S., Christiani, D.C., 2003.
Infant susceptibility of mortality to air pollution in Seoul, South Korea. Pediatrics
111 (2), 284–290.
Hajat, S., Armstrong, B., Wilkinson, P., Busby, A., Dolk, H., 2007. Outdoor air pollution and infant mortality: analysis of daily time-series data in 10 English cities.
Journal of Epidemiology and Community Health 61 (8), 719–722.
Hedley, A.J., Wong, C.-M., Tach, T.Q., Ma, S., Tai-Hing Lam, H., Anderson, R., 2002.
Cardiorespiratory and all-cause mortality after restrictions on sulphur content
of fuel in Hong Kong: an intervention study. Lancet 360 (9346), 1646–1652.
Hickey, R.J., Boyce, D.E., Clelland, R.C., Bowers, E.J., Slater, P.B., Technical Report
No. 15 1976. Demographic and chemical variables related to chronic disease
mortality in man. Department of Statistics, The Wharton School, University of
Pennsylvania, Philadelphia.
Jayachandran, S., 2009. Air quality and early-life mortality: evidence from
Indonesia’s wildfires. Journal of Human Resources 44 (4), 916–954.
Joyce, T.J., Grossman, M., Goldman, F., 1989. An assessment of the benefits of air
pollution control: the case of infant health. Journal of Urban Economics 25 (1),
32–51.
Lacasaña, M., Esplugues, A., Ballester, F., 2005. Exposure to ambient air pollution and
prenatal and early childhood health effects. European Journal of Epidemiology
20 (2), 183–199.
Lin, C.A., Pereira, L.A.A., Nishioka, D.C., Conceição, G.M.S., Braga, A.L.F., Saldiva, P.H.N.,
2004. Air pollution and neonatal deaths in São Paulo, Brazil. Brazilian Journal of
Medical and Biological Research 37 (5), 765–770.
Lipfert, F.W., Zhang, J., Wyzga, R.E., 2000. Infant mortality and air pollution: a comprehensive analysis of U.S. data for 1990. Journal of the Air & Waste Management
Association 50 (8), 1350–1366.
Lleras-Muney, A., 2010. The needs of the Army. Using compulsory relocation in the
military to estimate the effect of air pollutants on children’s health. Journal of
Human Resources 45 (3), 549–590.
Luechinger, S., 2009. Valuing air quality using the life satisfaction approach. Economic Journal 119 (536), 482–515.
Mendelsohn, R., Orcutt, G., 1979. An empirical analysis of air pollution dose-response
curves. Journal of Environmental Economics and Management 6 (2), 85–106.
Mittelbach, G., 1991. Desulfurization of flue gases on the basis of lime or limestone
scrubbing. In: Daniel van Velzen (Ed.), Sulphur Dioxide and Nitrogen Oxides
in Industrial Waste Gases: Emission, Legislation and Abatement. Springer, Dordrecht, pp. 93–109.
Moretti, E., Neidell, M., 2011. Pollution, health, and avoidance behavior: evidence
from the ports of Los Angeles. Journal of Human Resources 46 (1), 154–175.
Neidell, M., 2009. Information, avoidance behavior, and health: the effect of
ozone on asthma hospitalizations. Journal of Human Resources 44 (2),
450–478.
Pope, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K., Thurston, G.D.,
2002. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine
particulate air pollution. Journal of the American Medical Association 287 (9),
1132–1141.
S. Luechinger / Journal of Health Economics 37 (2014) 219–231
Redding, S.J., Sturm, D.M., 2008. The costs of remoteness: evidence from German
division and reunification. American Economic Review 98 (5), 1766–1797.
Schärer, B., Haug, N., 1990. Bilanz der Großfeuerungsanlagen-Verordnung. Staub.
Reinhaltung der Luft 50, 139–144.
Schulz, W., 1985. Der monetäre Wert besserer Luft: Eine empirische Analyse individueller Zahlungsbereitschaft und ihrer Determinanten auf der Basis von
Repräsentativumfragen. Lang, Frankfurt a.M., Bern/New York.
Schwartz, J., Marcus, A., 1990. Mortality and air pollution in London: a time series
analysis. American Journal of Epidemiology 131 (1), 185–194.
Schwartz, Stephen E., 1989. Acid deposition: unraveling a regional phenomenon.
Science 243 (4892), 753–763.
Shinkura, R., Fujiyama, C., Akiba, S., 1999. Relationship between ambient sulfur dioxide levels and neonatal mortality near the Mt. Sakurajima volcano in Japan.
Journal of Epidemiology 9 (5), 344–349.
Šram, R.J., Binkova, B., Dejmek, J., Bobak, M., 2005. Ambient air pollution and pregnancy outcomes: a review of the literature. Environmental Health Perspectives
113 (4), 375–382.
231
Summers, P.W., Fricke, W., 1989. Atmospheric decay distances and times for sulphur
and nitrogen oxides estimated from air and precipitation monitoring in Eastern
Canada. Tellus 41B (3), 286–295.
Tanaka, S., 2010. Air Pollution, Infant Mortality, and the Environmental Regulations
in China. Boston University, Mimeo.
Traup, S., Kruse, B., 1996. Wind und Windenergiepotentiale in Deutschland. Winddaten für Windenergienutzer. Deutscher Wetterdienst, Offenbach am Main,
DE.
UBA, 2012. National trend tables for the German atmospheric emission reporting, 1990–2011. http://www.umweltbundesamt.de/themen/luft/emissionenvon-luftschadstoffen
Viscusi, W.K., Aldy, J.E., 2003. The value of a statistical life: a critical review of market estimates throughout the world. Journal of Risk and Uncertainty 27 (1),
5–76.
Woodruff, T.J., Darrow, L.A., Parker, J.D., 2008. Air pollution and postneonatal infant
mortality in the United States, 1999–2002. Environmental Health Perspectives
116 (1), 110–115.