What to Expect When It Gets Hotter: The Impacts of Prenatal
Exposure to Extreme Heat on Maternal Health∗
Jiyoon Kim† Ajin Lee‡ Maya Rossin-Slater§
January 7, 2020

Abstract
We use temperature variation within narrowly-defined geographic and demographic cells to show
that exposure to extreme heat increases the risk of maternal hospitalization during pregnancy
for potentially life-threatening causes. We find that this effect is driven by women residing in
historically cooler rather than hotter counties, suggesting that adaptation plays a role in mitigating the health impacts of weather shocks. We also find that the heat-induced deterioration
in maternal pregnancy health is larger for black than for white mothers, suggesting that projected increases in extreme heat over the next century may further exacerbate the black-white
maternal health gap.

∗

We thank Alan Barreca, Janet Currie, Bhash Mazumder, Ciaran Phibbs, as well as participants at the 2019
American Society of Health Economists annual meeting and the 2020 Allied Social Science Associations (ASSA)
annual meeting. We use the State Inpatient Databases from the Healthcare Cost and Utilization Project (HCUP),
Agency for Healthcare Research and Quality, provided by the Arizona Department of Health Services, the New York
State Department of Health, and the Washington State Department of Health. We thank Jean Roth at the National
Bureau of Economic Research for assistance with the data.
†
Department of Economics, Elon University; E-mail: jkim14@elon.edu
‡
Department of Economics, Michigan State University. E-mail: leeajin@msu.edu.
§
Department of Medicine, Stanford University; NBER; IZA. E-mail: mrossin@stanford.edu.

1

 1

Introduction
The United States has experienced a deterioration in maternal pregnancy- and childbirth-related

health over the last several decades (Kassebaum et al., 2016), and the burden of health complications
is not borne equally by all mothers. For instance, black women are 3.3 times more likely to die from
a pregnancy-related cause than their white counterparts (Petersen et al., 2019). Most discussions
about maternal health have focused on the role of the health care system, but we know much
less about other—environmental —determinants of maternal health and the racial disparities in it.1
This paper studies the impact of an environmental factor that is becoming increasingly relevant
due to the growing consensus that climate change is contributing to a gradual warming of the earth
(NASA, 2013): exposure to extreme heat.
Specifically, we estimate the effect of exposure to extreme temperature during pregnancy on
maternal hospitalizations, using the universe of administrative inpatient discharge records from
three U.S. states: Arizona, New York, and Washington. In addition to providing us with rich
data on maternal health and health care utilization during pregnancy, at childbirth, and in the
postpartum period, these states vary in their historical weather patterns, allowing us to examine the
role of adaptation in mitigating the potential health impacts of temperature shocks. As individuals
in historically hotter places may adapt to high temperatures through the adoption of mitigating
technologies such as air conditioning and behavioral responses such as spending more time indoors
(Graff Zivin and Neidell, 2014), the health costs of extreme heat may be disproportionately borne
by individuals residing in historically cooler areas. Consistent with this notion, several studies have
documented such geographic heterogeneity in the relationship between temperature and elderly
mortality (Deschênes and Greenstone, 2011; Barreca et al., 2015; 2016; Carleton et al., 2018).
To identify the causal effects of extreme temperature, we leverage arguably exogenous temporal
variation within narrowly-defined geographic and demographic cells. Our preferred models control
for a full set of birth-county×birth-month×race fixed effects, birth-state×birth-year fixed effects,
and a quadratic time trend interacted with birth-county×birth-month indicators. As a concrete
example, consider a black woman giving birth in Queens county, New York, in August 2010 and a
black woman giving birth in the same county in August 2011. Our empirical strategy leverages the
difference between these women in the temperature deviation during their pregnancies from the
Queens-specific quadratic trend among all August births, while controlling for the average difference
in pregnancy temperature exposure between all New York state births in 2010 and 2011.
We find that exposure to extreme heat has adverse impacts on women’s health during pregnancy,
and that this health cost is not distributed equally across mothers. We estimate that an additional
1

For examples of these discussions in the press, see: https://www.vox.com/science-and-health/2017/6/26/
15872734/what-no-one-tells-new-moms-about-what-happens-after-childbirth
https://www.npr.org/2017/05/12/528098789/u-s-has-the-worst-rate-of-maternal-deaths-in-thedeveloped-world
https://www.npr.org/2017/05/12/527806002/focus-on-infants-during-childbirth-leaves-u-s-moms-indanger.

2

 day during the first trimester with an average temperature above 90◦ F increases the likelihood that
a woman is hospitalized during pregnancy by 0.03 percentage points, which represents a 0.8 percent
effect at the sample mean. This effect is driven by hospitalizations for emergency or urgent reasons,
suggesting that it represents a deterioration in underlying maternal health rather than a change in
women’s ability to access health care.
When we examine the timing of prenatal hospitalization, we find that extreme heat in the first
trimester has both immediate and latent impacts, as measured by heightened risks of hospitalization
both in the first and third trimesters. Analysis of diagnosis codes further indicates that this effect
is driven by hospitalizations due to a variety of pregnancy complications, including hemorrhage
in early pregnancy, antepartum hemorrhage, excessive vomiting, early or threatened labor, and
infectious and parasitic conditions. Several of these conditions are serious and potentially deadly—
Kuriya et al. (2016) report that hemorrhage is the third leading cause of maternal pregnancy-related
death, while infections can result in sepsis, which is the top cause of maternal pregnancy-related
death in the United States.
We next document that the aggregate effect on pregnancy hospitalizations is entirely driven by
women residing in historically cooler counties with below-median daily mean temperatures. For
these women, we observe a 0.1 percentage point increase in the likelihood of an emergency or
urgent hospitalization during pregnancy (4.4 percent at the sample mean). This pattern suggests
that because historically cooler places are likely less adapted to extreme heat than historically
hotter areas, mothers residing in cooler places bear a disproportionate cost to their pregnancy
health.2
We also show that the effects on prenatal hospitalizations are much more pronounced for black
than for white mothers. For black women, an additional day during the first trimester with average
temperature above 90◦ F increases the likelihood of first and third trimester hospitalization by
0.04 and 0.08 percentage points, respectively, representing 3.2 and 2.3 percent effect sizes at the
sample means. By contrast, for white women, the corresponding coefficients are much smaller and
statistically insignificant.
Lastly, we estimate that an additional day with above-90-degree heat in the first trimester raises
maternal length of hospital stay at the time of childbirth by 0.006 days (0.2 percent). Similar to the
findings on prenatal hospitalizations, the increase in length of stay at childbirth is greater in cooler
than in hotter counties. However, we find that the effect on length of hospital stay at childbirth
is driven entirely by white rather than black mothers. We further show that, for white mothers,
prenatal heat exposure reduces the likelihood of postpartum hospital readmission. These results
may reflect widely documented racial disparities in the types and quality of health services received
2
We have also considered modeling differences in effects based on air conditioning (AC) adoption rates. However,
AC data, available from the Residential Energy Consumption Survey (RECS), only exist in three years over our sample
period (2001, 2005, and 2009) and are aggregated to the Census region level. Given that we only use inpatient data
from three states in our analysis, we do not have sufficient variation to estimate heterogeneous effects based on AC
adoption rates.

3

 by women (Nelson, 2002; Hostetter and Klein, 2018)—compared to black mothers, white mothers
may be more successful in compensating for prenatal health shocks by staying longer at the hospital
when giving birth, thus averting future hospital readmissions in the postpartum period.
Our study contributes to a burgeoning literature, which has identified adverse short-term impacts of extreme temperature on several outcomes, including elderly mortality (Deschênes and
Moretti, 2009; Deschênes and Greenstone, 2011), population-level emergency department visits and
hospitalizations (Green et al., 2010; White, 2017), and cognitive performance (Cho, 2017; Garg et
al., 2018; Goodman et al., 2018; Graff Zivin, Hsiang, and Neidell, 2018). Multiple studies have
further documented negative effects of in utero heat exposure on birth outcomes—including birth
weight, gestation length, and the probability of stillbirth (e.g., Deschênes et al., 2009; Dadvand
et al., 2011; Schifano et al., 2016; Auger et al., 2017; Ha et al., 2017a,b; Kuehn and McCormick,
2017; Barreca and Schaller, 2019)—highlighting the sensitivity of the prenatal period to extreme
heat.3 To the best of our knowledge, only one prior study has analyzed the relationship between
prenatal heat exposure and maternal health, using information on mothers’ pregnancy risk factors
and labor/delivery complications reported on birth certificates (Cil and Cameron, 2017). However,
as multiple validation studies indicate that maternal pregnancy risk factors, obstetric procedures,
and complications of labor and delivery are heavily under-reported on birth certificates (Parrish
et al., 1993; Buescher et al., 1993; Piper et al, 1993; Dobie et al., 1998; Reichman and Hade,
2001; DiGiuseppe et al., 2002; Roohan et al., 2003; Lydon-Rochelle et al., 2005), and the degree of
under-reporting varies with maternal demographic characteristics and birth outcomes (Reichman
and Schwartz-Soicher, 2007), analyses of maternal health based on birth records data are likely
subject to bias from non-random measurement error. We address this issue by instead using inpatient discharge records that provide more accurate information on maternal health at each hospital
visit, and allow us to examine diagnoses and the timing of prenatal health insults.
Our findings suggest that, in the absence of mitigating interventions, the projected increase in
exposure to extreme heat over the next century may contribute to further worsening of maternal
health. This deterioration in maternal health is likely to be greater in historically cooler areas, which
have had less scope for adaptive responses. Moreover, since black women are both more likely to
be exposed to extreme heat (due to differences in residence locations and in access to mitigating
technologies such as air conditioning, see O’Neill et al., 2005; Gronlund, 2014) and experience larger
adverse impacts of heat exposure on pregnancy-related health, our estimates imply that climate
change could further exacerbate racial disparities in maternal health.

3

Fetuses and infants are sensitive to extreme heat due to their developing thermoregulatory and sympathetic
nervous systems; see Young (2002); Knobel and Holditch-Davis (2007); Xu et al. (2012). Two recent studies have
also shown that early life heat exposure has lasting negative effects on long-term cognitive ability (Hu and Li, 2019)
and adult earnings (Isen, Rossin-Slater, and Walker, 2017).

4

 2

Data
Our data comes from the State Inpatient Databases (SID) from the Healthcare Cost and Uti-

lization Project (HCUP). The SID are state-specific files that contain the universe of inpatient
records from participating states. Since the availability of variables varies across states and years,
we focus on three states that contain all three of the key variables necessary for our analysis: (1)
patient county of residence, (2) admission month, and (3) encrypted person identifiers to track
patients over time in the same state. Our resulting sample consists of 2.72 million inpatient records
of 2.24 million mothers from Arizona (2003 to 2007), New York (2003 to 2013), and Washington
(2003 to 2013).
We use diagnosis codes to identify inpatient visits associated with childbirth.4 Since less than
two percent of all births occur outside of hospitals during our analysis time period, we observe the
near-universe of all mothers giving birth in our analysis states.5 We also identify maternal hospitalizations during pregnancy (i.e., those occurring in the 9 months before delivery) and postpartum
hospital re-admissions using patient identifiers.
To measure temperature exposure, we obtain data from the National Oceanic and Atmospheric
Administration (NOAA). We have information on the mean, maximum, and minimum daily ground
temperature and precipitation levels for every county and year-month during our analysis time
frame. We then merge these data to the maternal inpatient records, using information on the
mother’s county of residence at the time of delivery. We use the mother’s year and month of
delivery to assign exposure to temperature during pregnancy by assuming a 40-week pregnancy
duration for all observations.6

Distribution of Temperature Exposure.

Figure 1 shows the distribution of daily average

temperature in Arizona, New York, and Washington during our sample period. We compute the
average number of days per year falling into each of the 10 temperature bins expressed in Fahrenheit
degrees. When we measure temperature exposure during pregnancy for each woman, we find that
five percent of observations in our data have non-zero exposure to above-90-degree heat.

Summary Statistics.

Panel A of Table 1 shows the average number of days per year with

mean temperature falling in different bins in each of our three analysis states. Arizona on average
4
5

We use DRG 370-375 or 765-768 & 774-775, depending on the version of DRG.
See https://www.cdc.gov/nchs/products/databriefs/db144.htm for statistics on out-of-hospital births in the

U.S.
6

We have information on gestational age for only about 10 percent of our HCUP sample, which comes from
diagnosis codes. It appears that gestational age is only recorded in cases where there are health complications,
and we find that children with gestational age information have lower birth weight, longer length of stay, and
higher likelihoods of readmission and death than those without gestational age information. Moreover, using actual
pregnancy duration to assign exposure can be problematic due to the possible endogeneity of gestational age with
respect to the in utero shock (Currie and Rossin-Slater, 2013).

5

 experiences 17 days per year with mean temperatures above 90◦ F. By contrast, New York and
Washington have substantially fewer days with above 90◦ F mean temperatures. These differences
underscore the importance of examining heterogeneity across local areas with different historical
climates.
Panel B of Table 1 provides means of maternal health outcomes that we analyze (expressed as
rates per 100 mothers). Approximately four percent of women get hospitalized during pregnancy,
with the most common diagnosis being a pregnancy-related complication. Overall, 0.5, 1.2, and 2.6
percent of women are hospitalized in the first, second, and third trimesters, respectively. There are
some meaningful differences in the maternal health outcomes across the three states, highlighting
an additional reason for including state×year fixed effects in all our regression models, which we
describe in more detail next.

3

Empirical Strategy
A robust medical literature discusses the biological mechanisms through which extreme heat

could be damaging to human health, and highlights that exposure to extreme temperature can be
particularly risky for pregnant women. The underlying issue is that pregnant women are not able
to regulate temperature as efficiently as non-pregnant individuals due to the physiologic changes
they undergo during gestation (Schifano et al., 2016), which means that elevated body temperature
during pregnancy can lead to various complications. Heat exposure can alter placental blood flow
patterns, which can reduce the integrity of the placenta and increase the chance of abruption (He
et al., 2018). Heat could further raise the likelihood of other serious pregnancy complications,
including hypertension, preeclampsia, and prolonged premature rupture of membranes (Beltran et
al., 2014, Yackerson et al., 2007). In addition, elevated temperature can increase the fetal heart rate
and lead to uterine contractions (Vaha-Eskeli and Erkkola, 1991). All of these issues can translate
into women needing to be hospitalized during pregnancy and experiencing complications at and
even after childbirth.
The goal of this paper is to quantify the causal relationship between extreme heat and maternal health. A central challenge is that exposure to hot days is not randomly assigned. For
instance, several studies have documented differences in the health and human capital outcomes
of children born in different months of the year due to selection into conception based on parental
characteristics and differential exposure to seasonal factors such as the influenza virus (Buckles
and Hungerman, 2013; Currie and Schwandt, 2013). In addition, there is non-random sorting of
families into hotter and colder regions of the country based on incomes, preferences, and other
characteristics, suggesting that cross-sectional comparisons between mothers residing in different
regions are unlikely to isolate the causal effects of temperature exposure from the influences of
other factors.
To address this challenge, we follow the prior literature by leveraging temperature variation
6

 within narrowly defined geographic and demographic cells, and flexibly accounting for local outcome
trends. We first collapse our data into cells defined by all possible combinations between the
mother’s county of residence at delivery, the year-month of childbirth, and race/ethnicity categories
(White, Black, Hispanic, Asian American, and other). We then use the following regression model
to estimate the effects of exposure to extreme temperature during pregnancy:

Yc,y,m,r = α+

3
10
X
X

βt,j T empt,j
c,y,m +

t=1 j=1,j6=7

3
X

f (P reciptc,y,m )+θc,m,r +ηy,s(c) +δc,m ×f (y)+ c,y,m,r (1)

t=1

Yc,y,m,r is an outcome for a mother residing in county c, giving birth in year y and month m, of
race/ethnicity r, and we rescale the outcomes by multiplying by 100 (e.g., the number of mothers
admitted to the hospital during pregnancy per 100 mothers). The variables T empt,j
c,y,m represent the
number of days in trimester t falling into each (j) of the 10 bins of temperature, ranging from less
than 10◦ F to 90◦ F or more, as illustrated in Figure 1.7 The bin representing temperatures in the
[60o F , 70o F ) range is omitted as the reference group. Thus, the βt,j coefficients can be interpreted
as estimates of the impact of an additional day in a given temperature range (j) relative to a day in
the [60o F , 70o F ) range in trimester t. We are particularly interested in the coefficient βt,10 , which
measures the effect of an additional above-90-degree day in each trimester t.
We control for indicators for the bottom and the top terciles of mean precipitation in each
trimester, f (P reciptc,y,m ). θc,m,r are fixed effects for every birth-county×birth-month×race cell.
ηy,s(c) are birth-state×birth-year fixed effects, which account for differential outcome trends across
states, any state time-varying policies, and the fact that we observe states in different sets of years
in the HCUP data. δc,m × f (y) are county-by-calendar-month-specific trends (e.g., Queens-Countyby-August-specific trends), which we model with a quadratic polynomial. To further account for
differential population sorting based on temperature, we control for the average number of mothers
per 100 residing in zip codes in different quartiles of the median income distribution. We weight
all regressions by cell size.8 Because weather is highly spatially correlated, we cluster our standard
errors on the commuting zone level.9

Identifying Assumption.

Our model identifies the effects of extreme heat exposure using year-

to-year deviations in temperature from the county-month trend within each cell. Thus, our estimates of βt,j represent causal effects of pregnancy exposure to temperature under the assumption
that the within-cell variation in temperature (conditional on birth-state×birth-year fixed effects and
7

In some specifications, we examine the effect of the number
entire period of pregP3 Pof10 days during the
t,j
nancy falling into each temperature bin. That is, we replace
β
T
emp
t,j
c,y,m in equation (1) with
t=1
j=1,j6=7
P10
j
β
T
emp
.
c,y,m
j=1,j6=7 j
8
Results based on collapsed data with cell size weights are identical to those using the underlying individual-level
data, since we do not have any other individual-level controls.
9
Our results are also robust to using an alternative adjustment of standard errors to reflect spatial dependence,
as modeled by Conley (1999) and implemented by Hsiang (2010). Results available upon request.

7

 county×calendar-month trends) is uncorrelated with other determinants of maternal health. While
this assumption is inherently untestable, we present some indirect tests to assess its plausibility.
First, we check whether there is any systematic relationship between temperature variation
and population demographic characteristics. We collapse our data to the birth-county×birthyear×birth-month level, and estimate a version of equation (1), excluding controls for demographic
characteristics and zip code income quartiles. As outcomes, we consider the number of mothers
who are of different races/ethnicities and the numbers of mothers residing in zip codes in different
quartiles of the median income distribution per 100.
Panel A of Appendix Table A.1 shows that our measure of extreme heat exposure is not correlated with the shares of mothers who are white, black, or Asian. However, we do find a positive
correlation between the number of mothers who are Hispanic and the number of days above 90
degrees during pregnancy, suggesting the importance of examining the effects of heat exposure
within cells defined by different race/ethnicity subgroups.10
In panel B of Appendix Table A.1, we find a marginally significant positive correlation between
heat exposure during pregnancy and the share of mothers residing in zip codes in the third quartile
of the median income distribution (but not with the shares of mothers in other quartiles). To
address the concern that differential trends in exposure to heat are correlated with income, we
include controls for zip code level income quartiles in all of our regression models.
Second, we test the robustness of our results to including hypothetical exposure to temperature
assuming a mother gave birth two years prior to her actual delivery year-month. As we show below,
we find that our main effects of exposure during pregnancy remain strong and significant even when
we add two-year leads in temperature exposure.

4

Results
Table 2 and Figure 2(a)-(c) show that extreme heat exposure during the first trimester raises the

likelihood that a mother is hospitalized during pregnancy. Specifically, we find that an additional
day with above-90-degree heat during the first trimester raises the likelihood that a mother is
hospitalized during pregnancy by 0.03 percentage points, which translates into a 0.8 percent effect
size when evaluated at the sample mean.11 In column (2) of Table 2, we show that the increase
10

In supplementary analyses, we have also examined the relationship between extreme heat and the sex ratio at
birth, finding no significant effects (results available upon request). The lack of relationship between extreme heat
exposure and infant sex suggests that there is no significant effect on miscarriages, as changes in the sex ratio at birth
are often used as proxies for changes in miscarriage rates (e.g., Sanders and Stoecker, 2015; Halla and Zweimüller,
2013).
11
Appendix Table A.2 shows how the estimates change as we add in different fixed effects and trends. Adding in
race×birth-county×birth-month fixed effects substantially changes the estimates on heat exposure during the first
trimester, highlighting the importance of controlling for differences in maternal outcomes across counties and birth
months. Further adding in state×year fixed effects and quadratic county×month trends increases the precision and
the magnitudes of the estimates slightly. However, the estimates on heat exposure during the first trimester in

8

 in prenatal hospitalizations is driven entirely by visits for emergency or urgent reasons rather than
scheduled appointments, which implies a deterioration in underlying maternal health as opposed
to an improvement in health care access or utilization.
In Figure 3, we present estimates and 95% confidence intervals from regression models that
use indicators for various diagnoses codes associated with prenatal hospitalization as outcomes.
We find that the increase in maternal hospitalizations in response to extreme heat during the first
trimester is driven by a range of pregnancy complications (ICD-9 codes 640-649). Specifically, these
include hospitalizations due to hemorrhage in early pregnancy (ICD 640), antepartum hemorrhage
(ICD 641), excessive vomiting (ICD 643), early or threatened labor (ICD 644), and infectious and
parasitic conditions (ICD 647). Several of these conditions can be life-threatening—hemorrhage is
the third leading cause of maternal pregnancy-related death, while infections can result in sepsis,
which is the number one cause of maternal pregnancy-related death (Kuriya et al., 2016).
Next, we examine heterogeneity in the effect on maternal hospitalizations during pregnancy by
geography, timing of the hospitalization, and maternal race.

Adaptation and Heterogeneity Across Historically Cooler and Hotter Counties. To
examine the role of adaptation to extreme heat, we study differences between mothers residing
in counties with below- and above-median daily mean temperatures averaged over the whole data
period. Table 3 and Figure 2(d)-(i) show that the effect of extreme heat on maternal pregnancy
hospitalization is driven entirely by women residing in cooler rather than hotter counties. Specifically, an additional day with above-90-degree temperature increases the likelihood of an emergency
or urgent hospitalization during pregnancy by 0.11 percentage points (or 4.4 percent) for mothers in
below-median counties. For mothers in above-median counties, the corresponding estimate is much
smaller and statistically insignificant. Moreover, the difference in the effects on emergency/urgent
hospitalizations between mothers in below-median and above-median counties is statistically significant (p-value: 0.009). Further, in Appendix Figure A.1, we show that the effect sizes for different
diagnosis categories are larger in cooler than in hotter counties.

Timing of Hospitalization and Differences by Maternal Race.

We investigate the timing

of prenatal hospitalization in Table 4 and find that extreme heat exposure during the first trimester
has both immediate and latent effects on prenatal hospitalization for mothers. Specifically, Panel
B of Table 4 suggests that additional day with above-90-degree heat in the first trimester increases
the likelihood of hospitalization in the first trimester by 0.01 percentage points and hospitalization
in the third trimester by 0.02 percentage points.
Further, we find that the effect of exposure to extreme heat is much more pronounced for black
than for white mothers. Table 5 shows that an additional day with above-90-degree heat increases
columns (3) and (4) are within the confidence interval of our main estimate in column (5), [0.009, 0.053], suggesting
that our main results are not driven by a particular choice of fixed effects and trends.

9

 first trimester hospitalizations by 0.04 percentage points (or 3.2 percent) and third trimester hospitalizations by 0.08 percentage points (2.3 percent) for black mothers. By contrast, we find no
significant relationship between heat exposure and prenatal hospitalizations in any trimester for
white mothers. The differences in effects are statistically significant (p-values are 0.014 and 0.077,
respectively, for first and third trimester hospitalizations).12 Appendix Figure A.2 also shows that
the coefficient magnitudes for effects on various diagnosis categories are larger for black than for
white mothers (although the differences here are not always statistically significant, due to reduced
power when focusing on specific diagnosis codes).
Lastly, Table 6 demonstrates that the increases in prenatal hospitalizations for black mothers
are much larger in historically cooler counties for all three trimesters, highlighting once again the
importance of adaptation. The differences in estimated coefficients are statistically significant with
p-values close to zero.
On the whole, these results suggest that temperature exposure may be an important determinant
of the widely documented black-white gap in maternal pregnancy-related health. In particular, as
black mothers are on average exposed to more days with extreme heat than white mothers, our
estimates imply that disparities in both the levels of extreme heat exposure and the magnitudes of
the effects of exposure could help explain the racial gap in maternal health.

Maternal Health at and after Childbirth.

Table 7 presents results for maternal length of

hospital stay at the time of childbirth and readmission to the hospital after childbirth. Column (1)
of Table 7 shows that an additional day with above-90-degree heat in the first trimester leads to a
significant 0.006 day increase in the average length of stay (0.2 percent). Consistent with our results
on prenatal hospitalizations, Table 8 shows that the increase in maternal length of stay is larger
in historically cooler than in hotter counties, and the difference is marginally significant (p-value:
0.063). However, unlike the results for pregnancy hospitalizations, Table 9 shows that the effect
on maternal length of stay is larger for white than for black mothers (although the difference is
not statistically significant at conventional levels). This pattern provides suggestive evidence that
white mothers may be better able to compensate for adverse health effects by staying longer in the
hospital at childbirth, reflecting racial disparities in women’s ability to access health care resources
(Nelson, 2002; Hostetter and Klein, 2018).
We do not find any evidence that prenatal heat exposure raises the likelihood that a mother
is readmitted to the hospital in the postpartum period. If anything, Table 7 shows that third
trimester exposure to extreme heat reduces the risk of postpartum readmission on average. That
said, the negative effect on postpartum readmission is driven entirely by white mothers (Table
9), while the coefficient for black mothers is positive (albeit insignificant). This pattern is again
consistent with the idea that white mothers are able to compensate for prenatal health insults by
12

When we estimate our models separately for black and white mothers, we drop counties that have fewer than
50 black or white mothers. This sample restriction allows us to identify the effects for each subgroup by providing
sufficient variation in temperature exposure conditional on a large set of fixed effects and trends.

10

 staying longer at the hospital at the time of childbirth, which may avert the need for postpartum
hospital readmission.

4.1

Additional Results

Placebo Temperature Exposure. To assess the possibility of bias due to differential trends
in temperature exposure that are not controlled for in our main regression models, we test the
robustness of our results to including two-year leads of temperature exposure. In particular, for
every birth-county×birth-year-month, we calculate the hypothetical exposure to temperature assuming that the child had been born two years prior. We use a two-year (instead of a one-year) lead
to avoid confounding our estimates with possible effects of temperature on conception or fertility
(Lam et al., 1994; Barreca et al., 2015; Wilde et al., 2017). Appendix Table A.3 shows that our
main results are robust to the inclusion of this placebo control.

Controlling for Air Pollution.

Further, since prior research shows that pollution is highly

correlated with weather and affects population health (e.g., Ye et al., 2012), we estimate our main
models, controlling for the air quality index (AQI) as measured by the Environmental Protection
Agency. Since AQI is not available for all counties and year/months in our analysis sample, we also
re-run our main specifications using a subsample of the data with non-missing AQI measures. We
find that our estimates are similar and robust to including pollution controls (see Appendix Table
A.4).

Alternative Relative Temperature Exposure Measure.

Lastly, as an alternative way of

examining the role of adaptation to extreme heat, we estimate models that analyze the relationship
between maternal hospitalizations and temperature deviations from the historical county-month
mean. We calculate the average temperature for every county-month (e.g., July in Queens county,
NY), using data from all available years. Then, for every month in all county-year combinations
(e.g., July 2012 in Queens county, NY), we calculate the difference between the given month’s mean
temperature and the overall average for that county-month, and divide by the standard deviation
(SD). We thus obtain a z-score that allows us to classify each month in any given county-year
based on its deviation from the overall county-month average. We then estimate a regression model
analogous to equation (1), except that instead of measuring the number of days that fall into each
of the ten bins of temperature in absolute terms (◦ F), we use eight bins of SDs of temperature from
the county-month average, ranging from less than −3 SDs to at least 3 SDs or more. We report the
coefficient on the number of days with “above-3-SD” heat during pregnancy in Appendix Table A.5.
If anything, using the relative measure of temperature exposure strengthens our results. We find
that an additional day with “above-3-SD” heat is associated with a larger increase in the likelihood

11

 of maternal hospitalization during pregnancy than an additional day with above-90◦ F.13 These
results underscore the role of adaptation and indicate that extreme heat is particularly damaging
when it is a relatively rare event.

5

Conclusion
Scientists predict that global average temperatures will rise over the next 50 to 100 years, mostly

due to a shift to the right in the upper tail of the temperature distribution. For instance, the number
of days with mean temperature above 90◦ F in the average American county is forecasted to increase
from about 1 to approximately 43 per year by 2070-2099 (Intergovernmental Panel on Climate
Change, 2014). Understanding the health consequences of this increase in extreme heat is critical
for informing discussions about the costs of climate change and the possible benefits of mitigating
policies. Moreover, the growing literature that demonstrates heterogeneity in effects of heat across
regions with different average temperatures and the importance of adaptation (Deschênes and
Greenstone, 2011; Graff Zivin et al., 2014; Barreca et al., 2015; Barreca et al., 2016; Carleton et
al., 2018) suggests that extreme deviations from typical weather may be particularly damaging.
In this paper, we contribute to the evidence on the costs of exposure to extreme heat by
documenting maternal health impacts. We use the universe of inpatient discharge records from three
states and find that exposure to extreme heat in the first trimester of pregnancy leads to an increase
in women’s emergency and urgent prenatal hospitalizations for pregnancy-related complications
that are potentially life-threatening. The fact that the increase in hospitalizations during pregnancy
is larger in historically cooler than hotter counties highlights the importance of adaptation, and
the larger effects for black than for white mothers suggest that climate change may exacerbate the
already substantial racial gap in maternal health.
We further find that prenatal exposure to extreme heat raises maternal length of hospital stay
at the time of childbirth and reduces the likelihood of postpartum hospital readmission, which may
in part represent a compensatory response for prenatal health insults. The fact that these effects
are only observed for white and not black mothers is consistent with the widely documented racial
disparities in the amount and quality of health care services received by patients, possibly due to
factors including implicit bias and structural racism (Hostetter and Klein, 2018).
One limitation of our study is that we are not able to measure health impacts not captured by
the hospitalizations data. Just like measures of maternal health in birth records may miss effects
on other aspects of health that we do measure, our estimates based on hospitalizations cannot
capture potential impacts on more minor health insults that do not lead to hospital encounters.
Future research may expand our understanding of these effects with better data on other health
13

Appendix Table A.6 summarizes the temperature cutoffs for our relative measure of extreme heat (i.e., above3-SD heat). It shows that the relative measure covers a larger range of temperature than the absolute measure of
above 90◦ F, which explains the discrepancy in the estimates between the two measures.

12

 conditions.

13

References
Auger, N., W. D. Fraser, A. Smargiassi, M. Bilodeau-Bertrand, and T. Kotasky (2017). Elevated
Outdoor Temperatures and Risk of Stillbirth. Int. J. Epidemiol 46(1), 200–208
Barreca, A., K. Clay, O. Deschenes, M. Greenstone, and J. Shapiro (2015). Convergence in Adaptation to Climate Change: Evidence from High Temperatures and Mortality, 19002004. American
Economic Review: Papers & Proceedings 105(5), 247–251
Barreca, A., K. Clay, O. Deschenes, M. Greenstone, and J. Shapiro (2016). Adapting to Climate Change: The Remarkable Decline in the US Temperature-Mortality Relationship over
the Twentieth Century. Journal of Political Economy 124(1),105–159
Barreca, A. and J. Schaller (2019). The Impact of High Ambient Temperatures on Delivery Timing
and Gestational Lengths. Nature Climate Change, 1–6
Beltran, A., J. Wu, and O. Laurent (2014). Associations of Meteorology with Adverse Pregnancy
Outcomes: a Systematic Review of Preeclampsia, Preterm Birth and Birth Weight. Chronic
Dis. Can 11, 91–172
Buckles, K. and D. Hungerman (2013). Season of Birth and Later Outcomes: Old Questions, New
Answers. Review of Economics and Statistics 95(3), 711–724
Buescher, P. A., Taylor, K. P., Davis, M. H., and J. M. Bowling. (1993). The Quality of the New
Birth Certificate Data: a Validation Study in North Carolina. American Journal of Public
Health 83(8), 1163–1165
Carleton, T., M. Delgado, M. Greenstone, T. Houser, S. Hsiang, A. Hultgren, A. Jina, R. Kopp,
K. McCusker, I. Nath, J. Rising, A. Rode, H. Seo, J. Simcock, A. Viaene, J. Yuan, and A.
Zhang (2018). Valuing the Global Mortality Consequences of Climate Change Accounting for
Adaptation Costs and Benefits. University of Chicago, Becker Friedman Institute for Economics
Working Paper No. 2018-51.
Cho, H. (2017). Effect of Summer Heat on Test Scores: A Cohort Analysis. Journal of Environmental Economics and Management 83, 185–196
Cil, G. and T. A. Cameron. (2017). Potential Climate Change Health Risks from Increases in Heat
Waves: Abnormal Birth Outcomes and Adverse Maternal Health Conditions. Risk Analysis
37(11), 2066–2079
Conley, T.G. (1999). GMM Estimation with Cross Sectional Dependence. Journal of Econometrics
92(1), 1–45
Currie, J. and M. Rossin-Slater (2013). Weathering the Storm: Hurricanes and Birth Outcomes.
Journal of Health Economics 32, 487–503
Currie, J. and H. Schwandt (2013). Within-mother Analysis of Seasonal Patterns in Health at

14

 Birth. PNAS 110(30), 12265–12270
Dadvand, P., X. Basagana, and C. Sartini (2011). Climate Extremes and the Length of Gestation
Environ. Health Perspect. 119 (10), 1449–1453
Deschenes, O., M. Greenstone, and J. Guryan (2009). Climate Change and Birth Weight. American
Economic Review 99(2), 211–217
Deschenes, O., and M. Greenstone (2011). Climate Change, Mortality, and Adaptation: Evidence
from Annual Fluctuations in Weather in the US. American Economic Journal: Applied Economics 3, 152–185
Deschenes, O., and E. Moretti (2009). Extreme Weather Events, Mortality, and Migration. The
Review of Economics and Statistics 91(4), 659–681.
DiGiuseppe, D.L., D. C. Aron, L. Ranbom, D. L. Harper and G. E. Rosenthal (2002). Reliability of
Birth Certificate Data: a Multi-Hospital Comparison to Medical Records Information. Maternal
and Child Health Journal 6(3), 169–179
Dobie, S.A., L. M. Baldwin, R. A. Rosenblatt, M. A. Fordyce, C. H. A. Andrilla and L. G. Hart
(1998). How Well Do Birth Certificates Describe the Pregnancies They Report? The Washington State Experience with Low-risk Pregnancies. Maternal and Child Health Journal 2(3),
145–154
Garg, T., M. Jagnani, and V. Taraz (2018). Temperature and Human Capital in India. Available
at SSRN: https://ssrn.com/abstract=2941049 or http://dx.doi.org/10.2139/ssrn.2941049
Goodman, J., M. Hurwitz, J. Park, and J. Smith (2018). Heat and Learning. NBER Working
Paper No. 24639
Graff Zivin, J., and M. Neidell (2014). Temperature and the Allocation of Time: Implications for
Climate Change. Journal of Labor Economics 32(1), 1–26
Graff Zivin, J., S. M. Hsiang, and M. Neidell (2018). Temperature and Human Capital in the Short
and Long Run. Journal of the Association of Environmental and Resource Economists 5(1),
77–105
Green, R., R. Basu, B. Malig, R. Broadwin, J. Kim, and B. Ostro (2010). The Effect of Temperature
on Hospital Admissions in Nine California Counties. Int. J. Public Health 55, 113–121
Gronlund, C. J. (2014). Racial and Socioeconomic Disparities in Heat-Related Health Effects and
Their Mechanisms: a Review. Current Epidemiology Reports 1(3), 165–173
Ha, S., D. Liu, Y. Zhu, S. Kim, S. Sherman, K. Grantz, and P. Mendola (2017). Ambient Temperature and Stillbirth: a Multi-center Retrospective Cohort Study. Environ. Health Perspect.
125, Article 067011
Ha, S., D. Liu, Y. Zhu, S. Kim, S. Sherman, and P. Mendola (2017). Ambient Temperature and
Early Delivery of Singleton Pregnancies. Environ. Health Perspect. 125, 453–459
15

 Halla, Martin, and Martina Zweimüller (2013). Parental Response to Early Human Capital Shocks:
Evidence from the Chernobyl Accident. IZA Discussion Paper 7968.
He, S., T. Kosatsky, A. Smargiassi, M. Bilodeau-Bertrand, and N. Auger (2018).

Heat and

Pregnancy-related Emergencies: Risk of Placental Abruption During Hot Weather. Environment International 111, 295–300
Hostetter, M., and S. Klein (2018). In Focus: Reducing Racial Disparities in Health Care by Confronting Racism. The Commonwealth Fund, available at: https://www.commonwealthfund.
org/publications/newsletter-article/2018/sep/focus-reducing-racial-disparitieshealth-care-confronting.
Hsiang, S.M (2010). Temperatures and Cyclones Strongly Associated with Economic Production in
the Caribbean and Central America. Proceedings of the National Academy of Sciences 107(35),
15367–15372
Hu, Z., and T. Li (2019). Too Hot to Handle: The Effects of High Temperatures During Pregnancy
on Adult Welfare Outcomes. Journal of Environmental Economics and Management 94, 236–
253
Intergovernmental Panel on Climate Change (2014). Climate Change 2014: Synthesis Report.
Isen, A., M. Rossin-Slater, and R. Walker (2017). Relationship Between Seasons of Birth, Temperature Exposure, and Later Life Wellbeing. PNAS 114(51), 13447–13452
Kassebaum, N. J., R. M. Barber, Z. A. Bhutta, L. Dandona, P. W. Gething, S. I. Hay, Y. Kinfu, H.
J. Larson, X. Liang, and S. S. Lim (2016). Global, Regional, and National Levels of Maternal
Mortality, 1990–2015: A Systematic Analysis for the Global Burden of Disease Study 2015.
The Lancet 388 (10053), 17751812.
Knobel, R. and D. Holditch-Davis (2007). Thermoregulation and Heat Loss Prevention after Birth
and During Neonatal Intensive-care Unit Stabilization of Extremely Low-Birthweight Infants.
Journal of Obstetric, Gynecologic, and Neonatal Nursing 36, 280–287.
Kuehn, L., and S. McCormick (2017). Heat Exposure and Maternal Health in the Face of Climate
Change. International Journal of Environmental Research and Public Health 14(8): 853.
Kuriya, A., Piedimonte, S., Spence, A. R., CzuzojShulman, N., Kezouh, A., and Abenhaim, H.
A. (2016). Incidence and Causes of Maternal Mortality in the USA. Journal of Obstetrics and
Gynaecology Research, 42(6), 661–668.
Lam, D. A., J. A. Miron, and A. Riley (1996). The Effects of Temperature on Human Fertility.
Demography 33(3), 291–305
Lydon-Rochelle, M.T., V. L. Holt, J. C. Nelson, V. Crdenas, C. Gardella, T. R. Easterling and W.
M. Callaghan (2005). Accuracy of Reporting Maternal Inhospital Diagnoses and Intrapartum
Procedures in Washington State Linked Birth Records. Paediatric and Perinatal Epidemiology

16

 19(6), 460–471
NASA (2013). More Extreme Weather Events Forecast. https://www.nasa.gov/centers/langley/
science/climate_assessment_2012.html
Nelson, A. (2002). Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care.
Journal of the National Medical Association 94(8): 666.
O’Neill, M. S., A. Zanobetti and J. Schwartz (2005). Disparities by Race in Heat-related Mortality
in Four US Cities: the Role of Air Conditioning Prevalence. Journal of Urban Health 82(2),
191–197
Parrish, K.M., V. L. Holt, F. A. Connell, B. Williams and J. P. LoGerfo (1993). Variations in the
Accuracy of Obstetric Procedures and Diagnoses on Birth Records in Washington State, 1989.
American Journal of Epidemiology 138(2), 119–127
Petersen, E. E., N. L. Davis, D. Goodman, S. Cox, N. Mayes, E. Johnston, C. Syverson, K. Seed,
C. K. Shapiro-Mendoza, W. M. Callaghan and W. Barfield (2019). Vital Signs: PregnancyRelated Deaths, United States, 2011–2015, and Strategies for Prevention, 13 States, 2013–2017.
Morbidity and Mortality Weekly Report 68(18), 423–429
Piper, J.M., E. F. Mitchel Jr, M. Snowden, C. Hall, M. Adams and P. Taylor (1993). Validation of
1989 Tennessee Birth Certificates Using Maternal and Newborn Hospital Records. American
Journal of Epidemiology 137(7), 758–768
Reichman, N.E. and E. M. Hade (2001). Validation of Birth Certificate Data: a Study of Women
in New Jersey’s HealthStart Program. Annals of epidemiology 11(3), 186–193
Reichman, N. E. and O. Schwartz-Soicher (2007). Accuracy of birth certificate data by risk factors and outcomes: analysis of data from New Jersey. American Journal of Obstetrics and
Gynecology, 197(1): 32-e1.
Roohan, P.J., R. E. Josberger, J. Acar, P. Dabir, H. M. Feder, and P. J. Gagliano (2003). Validation
of Birth Certificate Data in New York State. Journal of Community Health 28(5), 335–346
Sanders, Nicholas J., and Charles Stoecker (2015). Where Have All the Young Men Gone? Using
Sex Ratios to Measure Fetal Death Rates. Journal of Health Economics 41: 30–45.
Schifano, P., F. Asta, P. Dadvand, M. Davoli, X. Basagana, and P. Michelozzi (2016). Heat and Air
Pollution Exposure as Triggers of Delivery: a Survival Analysis of Population-based Pregnancy
Cohorts in Rome and Barcelona. Environment International 88, 153–159
Veha-Eskeli, K., and R. Erkkola (1991). The Effect of Short-term Heat Stress on Uterine Contractility, Fetal Heart Rate and Fetal Movements at Late Pregnancy. Eur. J. Obstet. Gynecol.
Reprod. Biol. 38, 9–14
White, C. (2017). The Dynamic Relationship Between Temperature and Morbidity. Journal of the
Association of Environmental and Resource Economists 4(4), 1155–1198

17

 Wilde, A., B. H. Apouey, and T. Jung (2017). The Effect of Ambient Temperature Shocks during
Conception and Early Pregnancy on Later Life Outcomes. European Economic Review 97(C),
87–107
Xu, Z., R. A. Etzel, H. Su, C. Huang, Y. Guo, and S. Tong (2012). Impact of Ambient Temperature
on Children’s Health: A Systematic Review. Environmental Research 117, 120–131
Yackerson, N. S., B. Piura, and M. Friger (2007). The Influence of Weather State on the Incidence
of Preeclampsia and Placental Abruption in Semi-arid Areas. Clin. Exp. Obstet. Gynecol. 34,
27–30
Ye, X., R. Wolff, W. Yu, P. Vaneckova, X. Pan, and S. Tong (2012). Ambient Temperature and
Morbidity: A Review of Epidemiological Evidence. Environmental Health Perspectives 120 (1),
p. 19.
Young, J. B. (2002). Programming of Sympathoadrenal Function. Trends Endocrinol Metab 13,
381–385

18

 Figures

0

10

Average number of days per year
20
30
40
50
60

70

6

<10

10-19

20-29

30-39

40-49

50-59

60-69

70-79

80-89

>90

Temperature bins

Figure 1: Distribution of Daily Average Temperature
Sources: NOAA weather data.
Notes: This figure shows the overall average number of days per year falling into each of the temperature bins (◦ F)
denoted on the x−axis. We compute daily average temperature by taking the average of minimum and maximum
temperature in a given day measured at weather stations in Arizona 2003 to 2007, New York 2003 to 2013, and
Washington 2003 to 2013.

19

 .05

.05

.1

.1

.06
.04

0

.02

−.1

−.05

−.05

0

0
−.02
−.04
<10

10−19

20−29

30−39

40−49

50−59

60−69

70−79

80−89

>90

<10

10−19

20−29

30−39

Temperature bins

40−49

50−59

60−69

70−79

80−89

>90

30−39

40−49

50−59

30−39

60−69

70−79

80−89

>90

50−59

60−69

70−79

80−89

>90

.4
.2
0
−.2
<10

10−19

20−29

30−39

Temperature bins

40−49

50−59

60−69

70−79

80−89

>90

<10

10−19

20−29

30−39

Temperature bins

40−49

50−59

60−69

70−79

80−89

>90

Temperature bins

(e) Hospitalization during
pregnancy, trimester 2 exposure,
colder counties

(f) Hospitalization during
pregnancy, trimester 3 exposure,
colder counties

<10

10−19

20−29

30−39

40−49

50−59

60−69

70−79

80−89

Temperature bins

>90

<10

10−19

20−29

30−39

40−49

50−59

60−69

70−79

80−89

Temperature bins

>90

−.1

−.1

−.15

−.05

−.1

−.05

0

−.05

0

.05

0

.05

.1

.05

.1

(d) Hospitalization during
pregnancy, trimester 1 exposure,
colder counties

40−49

.6

.1
−.1
−.2
−.3
20−29

20−29

(c) Hospitalization during
pregnancy, trimester 3 exposure

0

.1
0
−.1
−.2

10−19

10−19

Temperature bins

(b) Hospitalization during
pregnancy, trimester 2 exposure

.2

(a) Hospitalization during
pregnancy, trimester 1 exposure

<10

<10

Temperature bins

<10

10−19

20−29

30−39

40−49

50−59

60−69

70−79

80−89

>90

Temperature bins

(g) Hospitalization during
pregnancy, trimester 1 exposure,
hotter counties

(h) Hospitalization during
pregnancy, trimester 2 exposure,
hotter counties

(i) Hospitalization during pregnancy,
trimester 3 exposure, hotter counties

Figure 2: Effects of Temperature During Pregnancy on Any Prenatal Hospitalization
Notes: The figures plot regression coefficients, βt,j , from equation (1) for each temperature bin (j) for each trimester
(t) with 95% confidence intervals. Outcome is rescaled by multiplying by 100. Standard errors are clustered by the
commuting zone level. All regressions control for mother’s race/ethnicity×birth-county×birth-month fixed effects,
zip code level income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month
level, and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used.

20

 ICD 640−649
ICD 640
ICD 641
ICD 642
ICD 643
ICD 644
ICD 645
ICD 646
ICD 647
ICD 648
ICD 649
−.02

0

.02
Parameter estimate

.04

.06

Figure 3: Effects of Temperature Above 90 Degrees During the First Trimester on
Diagnoses at Prenatal Hospitalization
Notes: The figure plots separate regression coefficients, β1,10 , from equation (1) for temperature above 90-degrees for
the first trimester with 95% confidence intervals for each diagnosis category. Outcomes are rescaled by multiplying
by 100. ICD codes 640-649 indicate “complications mainly related to pregnancy.” The definition of each subcategory is as follows. ICD 640: Hemorrhage in early pregnancy; ICD 641: Antepartum hemorrhage abruptio
placentae and placenta previa; ICD 642: Hypertension complicating pregnancy childbirth and the puerperium; ICD
643: Excessive vomiting in pregnancy; ICD 644: Early or threatened labor; ICD 645: Late pregnancy; ICD 646:
Other complications of pregnancy not elsewhere classified; ICD 647: Infectious and parasitic conditions in the mother
classifiable elsewhere but complicating pregnancy childbirth or the puerperium; ICD 648: Other current conditions in
the mother classifiable elsewhere but complicating pregnancy childbirth or the puerperium; ICD 649: Other conditions
or status of the mother complicating pregnancy, childbirth, or the puerperium. Standard errors are clustered by the
commuting zone level. All regressions control for mother’s race/ethnicity×birth-county×birth-month fixed effects,
zip code level income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month
level, and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used.

21

 7

Tables
Table 1: Summary Statistics
(1)

(2)

(3)

(4)

Combining
three states

Arizona

New York

Washington

5.206
1.178

38.907
16.533

3.324
0.046

1.612
0.003

Any hospitalization during pregnancy
Emergency/urgent hospitalization during pregnancy

3.995
2.571

3.645
2.945

4.032
2.601

4.022
2.335

Diagnoses at prenatal hospitalization
Pregnancy-related complication (ICD 640-649)
Hemorrhage in early pregnancy (ICD 640)
Antepartum hemorrhage (ICD 641)
Hypertension complications (ICD 642)
Excessive vomiting in pregnancy (ICD 643)
Early or threatened labor (ICD 644)
Late pregnancy (ICD 645)
Other complications (ICD 646)
Infectious and parasitic conditions (ICD 647)
Other current conditions (ICD 648)
Other conditions (ICD 649)

3.722
0.048
0.284
0.489
0.227
1.451
0.239
0.869
0.158
2.031
0.273

3.501
0.043
0.270
0.526
0.154
1.654
0.105
0.961
0.112
1.778
0.032

3.771
0.053
0.294
0.464
0.251
1.437
0.255
0.855
0.151
2.129
0.279

3.659
0.035
0.257
0.551
0.184
1.415
0.243
0.876
0.197
1.837
0.351

Timing of prenatal hospitalization
Trimester 1
Trimester 2
Trimester 3

0.546
1.212
2.562

0.279
0.880
2.685

0.606
1.261
2.505

0.468
1.195
2.683

Maternal outcomes at and after childbirth
Length of stay at childbirth
Any readmission
Readmission within 28 days

2.691
1.979
1.144

2.568
1.518
0.908

2.819
1.973
1.157

2.354
2.174
1.198

Observations

44349

3902

30347

10100

A. Exposure to temperature extremes
Annual days with mean temperature
[80o F, 90o F )
≥ 90o F
B. Maternal health outcomes (per 100 mothers)

Sources: NOAA weather data and HCUP databases.
Notes: We use the data collapsed at the race×birth-county×birth-year-month level.

22

 Table 2: Effects of Exposure to Above-90-Degree Heat on Prenatal Hospitalization
(1)

(2)

Prenatal hospitalization
Any

Emergency/urgent

Panel A. Exposure during pregnancy
# Days above-90-degree during pregnancy

0.021
(0.021)

0.010
(0.014)

Observations
Adjusted R2
Mean

44336
0.465
3.995

44336
0.489
2.571

# Days above-90-degree in trimester 1

0.031∗∗∗
(0.011)

0.032∗∗∗
(0.008)

# Days above-90-degree in trimester 2

0.018
(0.029)

0.001
(0.024)

# Days above-90-degree in trimester 3

0.004
(0.030)

-0.007
(0.023)

Observations
Adjusted R2
Mean

44342
0.466
3.995

44342
0.490
2.571

Panel B. Exposure separately by each trimester

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting zone, are in parentheses. Each outcome is rescaled
by multiplying by 100. All regressions control for mother’s race/ethnicity×birthcounty×birth-month fixed effects, zip code level income quartiles, birth-state×birthyear fixed effect, a quadratic time at the county×calendar month level, and a series of
indicators for terciles of precipitation in each trimester. We use the data collapsed at
the race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10,
** p<0.05, *** p<0.01.

23

 Table 3: Effects of Exposure to Above-90-Degree Heat on Prenatal Hospitalization, by Historic
Average Daily Mean Temperature
(1)

(2)

Prenatal hospitalization
Any

Emergency/urgent

Panel A. Below median counties
# Days above-90-degree during pregnancy

0.073
(0.051)

0.109∗∗∗
(0.038)

Observations
Adjusted R2
Mean

21816
0.204
3.876

21816
0.181
2.476

# Days above-90-degree during pregnancy

0.019
(0.021)

0.006
(0.013)

Observations
Adjusted R2
Mean

22520
0.573
4.111

22520
0.595
2.663

P-value from testing the difference

0.297

0.009

Panel B. Above median counties

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting zone, are in parentheses. Each outcome is rescaled
by multiplying by 100. All regressions control for mother’s race/ethnicity×birthcounty×birth-month fixed effects, zip code level income quartiles, birth-state×birthyear fixed effect, a quadratic time at the county×calendar month level, and a series of
indicators for terciles of precipitation in each trimester. We use the data collapsed at
the race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10,
** p<0.05, *** p<0.01.

24

 Table 4: Effects of Exposure to Above-90-Degree Heat on the Timing of Prenatal Hospitalization
(1)
Trimester 1

(2)
Trimester 2

(3)
Trimester 3

# Days above-90-degree during pregnancy

0.001
(0.005)

0.007
(0.013)

0.017∗∗
(0.008)

Observations
Adjusted R2
Mean

44336
0.225
0.546

44336
0.327
1.212

44336
0.324
2.562

# Days above-90-degree in trimester 1

0.009∗∗
(0.005)

0.007
(0.008)

0.023∗∗
(0.009)

# Days above-90-degree in trimester 2

0.003
(0.010)

-0.001
(0.021)

0.015
(0.011)

# Days above-90-degree in trimester 3

-0.006
(0.007)

0.004
(0.019)

0.009
(0.007)

Observations
Adjusted R2
Mean

44342
0.225
0.546

44342
0.327
1.212

44342
0.324
2.562

Panel A. Exposure during pregnancy

Panel B. Exposure separately by each trimester

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust
standard errors, clustered by commuting zone, are in parentheses. Each outcome is
rescaled by multiplying by 100. All regressions control for mother’s race/ethnicity×birthcounty×birth-month fixed effects, zip code level income quartiles, birth-state×birth-year
fixed effect, a quadratic time at the county×calendar month level, and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10, **
p<0.05, *** p<0.01.

Table 5: Effects of Exposure to Above-90-Degree Heat on the Timing of Prenatal Hospitalization,
by Race
(1)
Trimester 1

(2)
Trimester 2

(3)
Trimester 3

0.007
(0.007)

-0.004
(0.010)

0.030
(0.018)

9835
0.242
0.514

9835
0.315
1.146

9835
0.328
2.489

0.035∗∗∗
(0.009)

0.055
(0.066)

0.084∗
(0.043)

Observations
Adjusted R2
Mean

4923
0.162
1.093

4923
0.369
2.246

4923
0.273
3.637

P-value from testing the difference

0.014

0.288

0.077

Panel A. White mothers
# Days above-90-degree during pregnancy
Observations
Adjusted R2
Mean
Panel B. Black mothers
# Days above-90-degree during pregnancy

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust
standard errors, clustered by commuting zone, are in parentheses. Each outcome is
rescaled by multiplying by 100. All regressions control for mother’s race/ethnicity×birthcounty×birth-month fixed effects, zip code level income quartiles, birth-state×birth-year
fixed effect, a quadratic time at the county×calendar month level, and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10, **
p<0.05, *** p<0.01.

25

 Table 6: Effects of Exposure to Above-90-Degree Heat on the Timing of Prenatal Hospitalization,
by Historic Average Daily Mean Temperature & Race
(1)
Trimester 1

(2)
Trimester 2

(3)
Trimester 3

-0.062∗∗
(0.021)

0.046
(0.038)

0.067
(0.054)

5602
0.220
0.546

5602
0.173
1.111

5602
0.225
2.426

0.011
(0.009)

-0.008
(0.006)

0.030
(0.021)

Observations
Adjusted R2
Mean

4233
0.275
0.471

4233
0.451
1.193

4233
0.399
2.574

P-value from testing the difference

0.003

0.160

0.513

Panel A1. Below median counties, white mothers
# Days above-90-degree during pregnancy
Observations
Adjusted R2
Mean

Panel A2. Above median counties, white mothers
# Days above-90-degree during pregnancy

Panel B1. Below median counties, black mothers
# Days above-90-degree during pregnancy

0.560∗∗
(0.192)

0.703∗∗∗
(0.220)

0.892∗∗∗
(0.079)

Observations
Adjusted R2
Mean

2250
-0.001
1.156

2250
0.154
2.081

2250
0.137
3.106

0.033∗∗∗
(0.009)

0.053
(0.066)

0.086∗
(0.048)

Observations
Adjusted R2
Mean

2673
0.270
1.040

2673
0.461
2.384

2673
0.316
4.083

P-value from testing the difference

0.008

0.007

0.000

Panel B2. Above median counties, black mothers
# Days above-90-degree during pregnancy

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust
standard errors, clustered by commuting zone, are in parentheses. Each outcome is
rescaled by multiplying by 100. All regressions control for mother’s race/ethnicity×birthcounty×birth-month fixed effects, zip code level income quartiles, birth-state×birth-year
fixed effect, a quadratic time at the county×calendar month level, and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10, **
p<0.05, *** p<0.01.

26

 Table 7: Effects of Exposure to Above-90-Degree Heat on Maternal Health at and after Childbirth
(1)

(2)

(3)

Length of stay at
childbirth

Any readmission

Readmission
within 28 days

# Days above-90-degree during pregnancy

0.003∗∗
(0.002)

-0.009
(0.007)

-0.008∗∗∗
(0.003)

Observations
Adjusted R2
Mean

44336
0.551
2.691

44336
0.086
1.979

44336
0.056
1.144

# Days above-90-degree in trimester 1

0.006∗∗∗
(0.001)

-0.008
(0.007)

-0.007
(0.005)

# Days above-90-degree in trimester 2

0.005
(0.003)

-0.005
(0.009)

-0.009
(0.006)

# Days above-90-degree in trimester 3

-0.001
(0.002)

-0.020∗∗∗
(0.007)

-0.010∗∗∗
(0.002)

Observations
Adjusted R2
Mean

44342
0.551
2.691

44342
0.086
1.979

44342
0.055
1.144

Panel A. Exposure during pregnancy

Panel B. Exposure separately by each trimester

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting zone, are in parentheses. Each binary outcome is rescaled by multiplying by 100.
All regressions control for mother’s race/ethnicity×birth county×birth month fixed effects, zip code level
income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month level,
and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10, ** p<0.05, *** p<0.01.

Table 8: Effects of Exposure to Above-90-Degree Heat on Maternal Health at and after Childbirth,
by Historic Average Daily Mean Temperature
(1)

(2)

(3)

Length of stay at
childbirth

Any readmission

Readmission
within 28 days

# Days above-90-degree during pregnancy

0.016∗∗
(0.007)

-0.059
(0.044)

-0.004
(0.030)

Observations
Adjusted R2
Mean

21816
0.280
2.691

21816
0.040
1.979

21816
0.015
1.155

# Days above-90-degree during pregnancy

0.003∗∗
(0.001)

-0.009
(0.007)

-0.007∗∗
(0.003)

Observations
Adjusted R2
Mean

22520
0.590
2.691

22520
0.129
1.979

22520
0.092
1.134

P-value from testing the difference

0.063

0.229

0.908

Panel A. Below median counties

Panel B. Above median counties

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting zone, are in parentheses. Each binary outcome is rescaled by multiplying by 100.
All regressions control for mother’s race/ethnicity×birth county×birth month fixed effects, zip code level
income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month level,
and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10, ** p<0.05, *** p<0.01.

27

 Table 9: Effects of Exposure to Above-90-Degree Heat on Maternal Health at and after Childbirth,
by Race
(1)

(2)

(3)

Length of stay at
childbirth

Any readmission

Readmission
within 28 days

0.005∗∗
(0.002)

-0.019∗∗∗
(0.005)

-0.014∗∗
(0.006)

9835
0.719
2.565

9835
0.079
1.894

9835
0.036
1.068

-0.001
(0.005)

0.014
(0.019)

0.004
(0.027)

Observations
Adjusted R2
Mean

4923
0.324
2.879

4923
-0.009
3.069

4923
0.028
1.900

P-value from testing the difference

0.352

0.096

0.386

Panel A. White mothers
# Days above-90-degree during pregnancy
Observations
Adjusted R2
Mean
Panel B. Black mothers
# Days above-90-degree during pregnancy

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting zone, are in parentheses. Each binary outcome is rescaled by multiplying by 100.
All regressions control for mother’s race/ethnicity×birth county×birth month fixed effects, zip code level
income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month level,
and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used. * p<0.10, ** p<0.05, *** p<0.01.

28

 Online Appendix
Appendix A. Appendix Figures and Tables

ICD 640−649

ICD 640−649

ICD 640

ICD 640

ICD 641

ICD 641

ICD 642

ICD 642

ICD 643

ICD 643

ICD 644

ICD 644

ICD 645

ICD 645

ICD 646

ICD 646

ICD 647

ICD 647

ICD 648

ICD 648

ICD 649

ICD 649
−.05

0
.05
.1
Estimates, below−median counties

.15

−.05

(a) Below median counties

0
.05
.1
Estimates, above−median counties

.15

(b) Above median counties

Figure A.1: Effects of Temperature Above 90 Degrees During Pregnancy on Diagnoses
at Prenatal Hospitalization by Historic Temperature
Notes: The figure plots separate regression coefficients, β1,10 , from equation (1) for temperature above 90-degrees for
the first trimester with 95% confidence intervals for each diagnosis category. Outcomes are rescaled by multiplying
by 100. ICD codes 640-649 indicate “complications mainly related to pregnancy.” The definition of each subcategory is as follows. ICD 640: Hemorrhage in early pregnancy; ICD 641: Antepartum hemorrhage abruptio
placentae and placenta previa; ICD 642: Hypertension complicating pregnancy childbirth and the puerperium; ICD
643: Excessive vomiting in pregnancy; ICD 644: Early or threatened labor; ICD 645: Late pregnancy; ICD 646:
Other complications of pregnancy not elsewhere classified; ICD 647: Infectious and parasitic conditions in the mother
classifiable elsewhere but complicating pregnancy childbirth or the puerperium; ICD 648: Other current conditions in
the mother classifiable elsewhere but complicating pregnancy childbirth or the puerperium; ICD 649: Other conditions
or status of the mother complicating pregnancy, childbirth, or the puerperium. Standard errors are clustered by the
commuting zone level. All regressions control for mother’s race/ethnicity×birth-county×birth-month fixed effects,
zip code level income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month
level, and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used.

29

 ICD 640−649

ICD 640−649

ICD 640

ICD 640

ICD 641

ICD 641

ICD 642

ICD 642

ICD 643

ICD 643

ICD 644

ICD 644

ICD 645

ICD 645

ICD 646

ICD 646

ICD 647

ICD 647

ICD 648

ICD 648

ICD 649

ICD 649
−.2

0
.2
Estimates, white mothers

.4

−.2

(a) White mothers

0
.2
Estimates, black mothers

.4

(b) Black mothers

Figure A.2: Effects of Temperature Above 90 Degrees During Pregnancy on Diagnoses
at Prenatal Hospitalization by Race
Notes: The figure plots separate regression coefficients, β1,10 , from equation (1) for temperature above 90-degrees for
the first trimester with 95% confidence intervals for each diagnosis category. Outcomes are rescaled by multiplying
by 100. ICD codes 640-649 indicate “complications mainly related to pregnancy.” The definition of each subcategory is as follows. ICD 640: Hemorrhage in early pregnancy; ICD 641: Antepartum hemorrhage abruptio
placentae and placenta previa; ICD 642: Hypertension complicating pregnancy childbirth and the puerperium; ICD
643: Excessive vomiting in pregnancy; ICD 644: Early or threatened labor; ICD 645: Late pregnancy; ICD 646:
Other complications of pregnancy not elsewhere classified; ICD 647: Infectious and parasitic conditions in the mother
classifiable elsewhere but complicating pregnancy childbirth or the puerperium; ICD 648: Other current conditions in
the mother classifiable elsewhere but complicating pregnancy childbirth or the puerperium; ICD 649: Other conditions
or status of the mother complicating pregnancy, childbirth, or the puerperium. Standard errors are clustered by the
commuting zone level. All regressions control for mother’s race/ethnicity×birth-county×birth-month fixed effects,
zip code level income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month
level, and a series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the
race×birth-county×birth-year-month level. Cell size weights are used.

30

 Table A.1: Placebo Outcome: Race and Zip-Code-Level Income
(1)
White

(2)
Black

(3)
Hispanic

(4)
Asian

# Days above-90-degree during pregnancy

-0.025
(0.049)

-0.012
(0.013)

0.165∗∗
(0.064)

-0.008
(0.028)

Observations
Adjusted R2
Mean

10121
0.962
76.614

10121
0.972
4.547

10121
0.934
11.286

10121
0.889
2.305

Q1

Q2

Q3

Q4

# Days above-90-degree during pregnancy

-0.223
(0.387)

0.144
(0.367)

0.291∗
(0.147)

-0.212
(0.138)

Observations
Adjusted R2
Mean

8978
0.958
25.033

8978
0.919
41.301

8978
0.915
22.978

8978
0.985
10.688

A. Race

B. Zip-code-level income quartiles

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors,
clustered by commuting zone, are in parentheses. All regressions control for birth-county×birthmonth fixed effects, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month
level, and a series of indicators for terciles of precipitation. We use the data collapsed at the birthcounty×birth-month level. Cell size weights are used. * p<0.10, ** p<0.05, *** p<0.01.

Table A.2: Effects of Exposure to Above-90-Degree Heat on Any Prenatal Hospitalization
(1)

(2)

(3)

(4)

(5)

-0.015∗∗
(0.006)

-0.010
(0.006)

0.009
(0.011)

-0.002
(0.014)

0.021
(0.021)

44337
0.039
3.995

44337
0.073

44336
0.442

44336
0.445

44336
0.465

# Days above-90-degree in trimester 1

-0.031∗∗∗
(0.007)

-0.025∗∗∗
(0.006)

0.024∗
(0.012)

0.016∗
(0.009)

0.031∗∗∗
(0.011)

# Days above-90-degree in trimester 2

0.001
(0.010)

0.010
(0.009)

-0.007
(0.009)

-0.018
(0.014)

0.018
(0.029)

# Days above-90-degree in trimester 3

-0.016∗
(0.009)

-0.012
(0.008)

0.005
(0.009)

-0.012
(0.018)

0.004
(0.030)

Observations
Adjusted R2
Mean

44343
0.145
3.995

44343
0.175

44342
0.443

44342
0.446

44342
0.466

Y

Y
Y

Y
Y
Y

Y
Y
Y
Y

Y
Y
Y
Y
Y

Panel A. Exposure during pregnancy
# Days above-90-degree during pregnancy
Observations
Adjusted R2
Mean
Panel B. Exposure separately by each trimester

Precipitation
Zip code level income quartiles
Race×birth-county×birth-month fixed effects
Birth-state×birth-year fixed effects
Quadratic birth-county×birth-month trends

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting
zone, are in parentheses. Each outcome is rescaled by multiplying by 100. The bottom rows of the table shows the list of
control variables included in each regression in addition to the temperature exposure variables. ‘Precipitation’ includes a
series of indicators for terciles of precipitation in each trimester. We use the data collapsed at the race×birth-county×birthyear-month level. Cell size weights are used. * p<0.10, ** p<0.05, *** p<0.01.

31

 Table A.3: Robustness to Including Two-Year Leads in Temperature Exposure
(1)

(2)

Prenatal hospitalization
Any

Emergency/urgent

Panel A. Main specification in the subsample with two-year leads
0.108∗∗∗
(0.019)

0.069∗∗∗
(0.021)

35775
0.448
3.995

35775
0.476
2.571

# Days above-90-degree during pregnancy

0.150∗∗∗
(0.021)

0.100∗∗∗
(0.021)

# Days above-90-degree during pregnancy (placebo)

0.046
(0.035)

0.068∗∗∗
(0.012)

Observations
Adjusted R2
Mean

35775
0.448
3.995

35775
0.477
2.571

# Days above-90-degree during pregnancy
Observations
Adjusted R2
Mean
Panel B. Adding two-year leads

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting zone, are in parentheses. Each outcome is rescaled by multiplying
by 100. All regressions controls for mother’s race/ethnicity×birth-county×birth-month fixed effects, zip code level income quartiles, birth-state×birth-year fixed effect, a quadratic time at the
county×calendar month level, and a series of indicators for terciles of precipitation. We use the
data collapsed at the race×birth-county×birth-year-month level. Cell size weights are used. *
p<0.10, ** p<0.05, *** p<0.01.

Table A.4: Effects of Exposure to Above-90-Degree Heat on Prenatal Hospitalization, Robustness
to AQI Controls
(1)

(2)
Full sample

Any

(3)

(4)

Colder counties

Emergency/urgent

Any

Emergency/urgent

(5)

(6)

Hotter counties
Any

Emergency/urgent

Panel A. Main specification in the subsample with AQI measures
# Days above-90-degree during pregnancy

0.022
(0.020)

0.011
(0.014)

0.129∗∗∗
(0.043)

0.137∗∗∗
(0.042)

0.021
(0.020)

0.008
(0.012)

Observations
Adjusted R2
Mean

28717
0.523
4.113

28717
0.561
2.603

12450
0.229
4.049

12450
0.211
2.519

16267
0.616
4.162

16267
0.642
2.668

# Days above-90-degree during pregnancy

0.023
(0.019)

0.007
(0.013)

0.130∗∗∗
(0.041)

0.140∗∗∗
(0.042)

0.024
(0.019)

0.004
(0.011)

Observations
Adjusted R2
Mean

28717
0.523
4.113

28717
0.561
2.603

12450
0.229
4.049

12450
0.211
2.519

16267
0.617
4.162

16267
0.643
2.668

Panel B. Adding AQI measures

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients, βt,10 , from equation (1). Robust standard errors, clustered by commuting zone, are in parentheses.
Each outcome is rescaled by multiplying by 100. All regressions control for mother’s race/ethnicity×birth-county×birth-month fixed effects, zip code
level income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar month level, and a series of indicators for terciles of
precipitation in each trimester. In panel B, we include a series of indicators for AQI categories (“good,” “moderate,” “unhealthy for sensitive groups,”
“very unhealthy,” with “hazardous” as a reference group) separately for each trimester. We use the data collapsed at the race×birth-county×birth-yearmonth level. Cell size weights are used. * p<0.10, ** p<0.05, *** p<0.01.

32

 Table A.5: Effects of Exposure to Extreme Heat on Prenatal Hospitalization, Using a Relative
Temperature Measure

# Days above-3-SD heat during pregnancy

(1)

(2)

Any

Emergency/urgent

0.088∗∗∗
(0.031)

0.056∗∗
(0.024)

44336
0.466
3.995

44336
0.489
2.571

Observations
Adjusted R2
Mean

Source: HCUP SID merged with NOAA weather data
Notes: This table reports regression coefficients from a model analogous to equation (1),
except that instead of measuring the number of days that fall into each of the ten bins
of temperature in absolute terms (◦F), we use eight bins of standard deviations (SDs) of
temperature from the county-month average, ranging from less than −3 SDs to at least 3
SDs or more. The table reports the effect of the number of days at least 3 SDs above the
county-month mean temperature. Robust standard errors, clustered by commuting zone,
are in parentheses. Each outcome is rescaled by multiplying by 100. All regressions control for mother’s race/ethnicity×birth-county×birth-month fixed effects, zip code level income quartiles, birth-state×birth-year fixed effect, a quadratic time at the county×calendar
month level, and a series of indicators for terciles of precipitation in each trimester. We use
the data collapsed at the race×birth-county×birth-year-month level. Cell size weights are
used. * p<0.10, ** p<0.05, *** p<0.01.

Table A.6: Temperature Cutoffs for Extreme Heat Exposure (o F )
(1)

(2)

(3)

Arizona

New York

Washington

A. Average cutoff for 2-SD above the county-month averages
January
February
March
April
May
June
July
August
September
October
November
December

46.3
49.3
53.8
57.8
68.0
96.8
74.5
70.5
72.2
60.8
63.4
44.6

41.8
37.5
48.5
57
67.8
74.5
75.5
75.5
71
62.5
52
43.2

43.2
41
45.9
50.7
60.3
64.5
68
66.5
64.3
55
46.3
37.9

B. Average cutoff for 3-SD above the county-month averages
January
February
March
April
May
June
July
August
September
October
November
December

.
.
.
.
.
.
.
.
.
.
.
.

56.1
.
61.7
67.5
80
.
84.1
83
82.5
.
.
56

52.1
.
53.7
58.4
65.4
69
71.8
70.3
68
65
.
52.5

Source: NOAA weather data
Notes: For each state, we calculate average temperature cutoffs for our measures
of extreme heat, 2 or 3 standard deviations above the overall mean temperature
for a given county and month. Arizona experiences no exposure to above-3-SD
heat during our study period. New York and Washington also do not experience
above-3-SD heat in some months.

33