Childhood Medicaid Coverage and Later-Life Health Care Utilization

Abstract

Exploiting a discontinuity in childhood Medicaid eligibility based on date of birth, we find that more years of childhood eligibility are associated with fewer hospitalizations in adulthood. For blacks, we find a 7–15% decrease in hospitalizations and a suggestive 2–5% decrease in emergency department visits, but no similar effect for non-blacks. The effects are pronounced for utilization related to chronic illnesses and for patients living in low-income zip codes. Calculations suggest that lower rates of hospitalizations during one year in adulthood for blacks offset between 2 and 4 percent of the initial costs of expanding Medicaid for all children.

I. Introduction

One goal of publicly-subsidized health insurance is to improve the health of those without insurance, which may lead beneficiaries to need fewer expensive hospital and emergency department (ED) visits later in life. Such long-term reductions in utilization could potentially offset the initial cost of insurance provision. However, little empirical evidence exists on the relationship between publicly provided insurance and long-term health and utilization of care.

One limitation of existing research in this area is it tends to evaluate utilization changes over a relatively short time horizon. Most studies link health insurance to health contemporaneously, or for a few subsequent years. However, the health benefits of insurance may materialize only after a sustained period of insurance and regular use of medical care. In addition, certain types of medical care focus on protecting the patient from future health risks and the payoffs from these types of preventive services may not be evident until later in life. In both scenarios, shorter windows of analysis may not be adequate to identify the health benefits of insurance.

A second limitation of most studies evaluating the contemporaneous effects of public health insurance on hospitalizations and ED visits is that they are unable to isolate the potential health benefit of insurance from the increased access that results from insurance. Health insurance lowers out-of-pocket costs and induces greater use of health care, including hospitalizations. Therefore, even if an individual’s health improves as a result of public insurance coverage, the access effect may dominate in the short term, leading to higher utilization of medical services.

In this paper, we address these issues by examining whether the expansion of Medicaid in the late 1980s and early 1990s improved the later life health of those affected. Specifically, we exploit plausibly exogenous variation by birthdate in the cumulative number of years an individual was eligible for public health insurance coverage during childhood. To phase in the Medicaid expansions, Congress specified that several eligibility expansions for low-income children applied only to children born after September 30, 1983. As a result, children born before September 30, 1983 experienced lower rates of Medicaid eligibility and fewer Medicaid-eligible years in childhood than children born immediately following the cutoff. For example, a child in a family with income just under the Federal Poverty Level (FPL) gained approximately five additional years of Medicaid eligibility during childhood if she were born on October 1, 1983 rather than September 30, 1983. Black children were particularly likely to benefit from the Medicaid expansions, gaining on average more than twice the number of Medicaid-eligible years of white children. We exploit this policy discontinuity as a source of exogenous variation in Medicaid eligibility in order to evaluate the long-term effects on hospital and ED visits.

We focus on changes in health care utilization that occur following the period of coverage gain, when there are no longer policy-driven differences in Medicaid eligibility or out- of-pocket costs for cohorts on either side of the cutoff. This allows us to isolate the potential health benefits of insurance from any contemporaneous access effects resulting from increased health insurance coverage and reduced out-of-pocket costs for care. The gain in Medicaid eligibility for affected cohorts occurred between ages 8 and 14. We examine the effects of coverage one year after the additional Medicaid coverage ended (at age 15) and ten years later (at age 25). We selected these two points in time to provide information on both short- and longer-term effects. Given that it is not clear ex ante whether one would expect to observe larger effects on health immediately after the coverage gain or at later ages, this approach allows us to provide a first cut at answering this question.

We find no effect of the Medicaid expansions on health care utilization at age 15, one year after the period of additional Medicaid eligibility ended. However, we find sizeable effects of Medicaid eligibility in childhood on hospitalizations at age 25, and weaker evidence of reductions in ED visits. Black children born immediately after the September 1983 cutoff are estimated to have experienced approximately 7 to 15 percent fewer hospitalizations and 2 to 5 percent fewer ED visits at age 25 relative to those born just before the cutoff. Our results are particularly pronounced for hospitalizations and ED visits related to chronic illnesses and among patients from low-income zip codes. Examining diseases by category, we find that greater Medicaid eligibility was associated with decreases in hospitalizations for a broad range of conditions. We do not find reductions in the utilization of non-blacks who experienced smaller gains in Medicaid eligibility and take-up of coverage at the birth date cutoff and who have lower rates of illness than black cohorts.

Our results provide several insights that are relevant to current policy debates surrounding the provision of public health insurance and the role of government in expanding coverage. First, our estimates indicate that between 2 and 4 percent of the initial cost of the Medicaid expansions for all children were “offset” by lower hospitalizations among black patients at age 25 alone. If these effects persist at older ages, then the size of the cost offset is likely to be even greater. Second, we find no impact of Medicaid coverage in the short-term at age 15, but do find effects later in life at age 25. These findings suggest that the benefits of insurance may only materialize over a longer time horizon.

II. Background

Analyses of Medicaid eligibility expansions for children consistently show that Medicaid increases health care utilization, including hospitalizations, in the short term.¹ However, there are fewer studies of the effects of gaining Medicaid on children’s health and the evidence from this literature is mixed.² Thus, the effect of gaining health insurance on health remains an important, but unanswered question. In addition, one limitation of studies seeking to assess the effect of insurance on health is that they examine how coverage affects health, for example, as measured by hospital admissions, immediately after or within a few years of the coverage expansion. If the health benefits of insurance are realized later, then a contemporaneous or short-run analysis may miss much of the effect of insurance.³

An emerging literature on the longer-term effects of health insurance coverage in childhood on later life outcomes has begun to address this issue. Wherry and Meyer (2016) examine the later life mortality of cohorts born before and after the September 30, 1983 cutoff date specified in many expansions of Medicaid. They provide evidence linking an increase in childhood eligibility to a later decline in teenage mortality for black children who were more likely to gain eligibility under the expansions than white children. Brown, Kowalski, and Lurie (2015) use state-level variation in the timing of public health insurance expansions for children in the 1980s and 1990s to examine long-term effects and find that cohorts who gained coverage had lower mortality rates as adults. Boudreaux, Golberstein and McAlpine (2016) use state variation in the timing of the introduction of the Medicaid program in the 1960s to identify long-term effects among those who gained access to Medicaid early in childhood and find improvements in adult health as measured by the presence of a chronic health condition. Miller and Wherry (2015) find that cohorts whose mothers had higher eligibility rates for prenatal coverage in the 1980s and 1990s while the cohort was in utero had better health outcomes and fewer hospitalizations in adulthood related to chronic health conditions.⁴

In this paper, we add to this small literature by exploiting the discontinuity in Medicaid eligibility and coverage among those born around September 30, 1983 to study the effect of additional childhood Medicaid eligibility at ages 8 to 14 on hospitalizations and ED visits at ages 15 and 25. With the exception of Wherry and Meyer (2016), all previous papers on the long-term effects of Medicaid coverage use state and year level variation in Medicaid policy to examine long-term outcomes. Although this empirical approach (i.e., difference-in-differences) has been used many times in the literature, some authors have pointed out its limitations (e.g., the estimates tend to be sensitive to the inclusion of state-specific trends; see Dave et al. 2008). The regression discontinuity design we employ is arguably more credible than the difference-in- difference approach because it does not rely on using policy changes at the state level as the identifying variation (Lee and Lemieux 2010).

III. The Policy Discontinuity

Discontinuity in Eligibility

Prior to the 1980s, eligibility for Medicaid for non-disabled children was primarily limited to children in families receiving cash welfare under the Aid to Families with Dependent Children (AFDC) program. Beginning in the mid-1980s, Congress took steps to expand eligibility for Medicaid to children not participating in AFDC. In an effort to phase in changes in Medicaid eligibility, Congress specified that many of the legislative changes applied only to children born after September 30, 1983. This provision meant that children born just before and after this birthdate cutoff faced very different eligibility criteria for Medicaid during their childhood years.

Given the nature of the expansions, the discontinuity in eligibility was largest for children with family incomes below the poverty line and above AFDC income levels. For children in families with incomes between 75 and 100 percent of the poverty line, those born in October 1983 gained, on average, an additional 4.6 years of eligibility during childhood compared to children born one month earlier. Children with incomes between 50 and 75 percent of poverty and those with incomes between 25 and 50 percent of poverty also experienced sizeable gains with an additional 3.4 and 2.0 years of eligibility, respectively (see Appendix Figure 1).

The gain in eligibility was primarily concentrated at ages 8 to 14 for children born immediately after the birthdate cutoff. Eligibility levels were similar for children born in September and October 1983 prior to age 8 and again starting at age 15 (see Appendix Figure 2).⁵ We examine health care utilization for cohorts born just before and after September 30, 1983 at the end of the differential gain in Medicaid eligibility - first at age 15 and then 10 years later at age 25.

We also examine differential effects of the expansions by race. Black children were more likely to gain eligibility under the expansions (Table 1) due to their family incomes. On average, black children born in October versus September 1983 were 17 percentage points more likely to gain Medicaid eligibility. Among those who were made Medicaid-eligible, the average gain in eligibility throughout childhood was 4.9 years. The average years of eligibility gained by black children (0.9 eligible years) is over twice the gain for non-black children (0.4 eligible years), who experienced a 9 percentage point gain in eligibility across the birth date threshold that led, on average, to 4.4 additional Medicaid-eligible years throughout childhood.

Table 1.

Average Childhood Medicaid Eligibility Gain for Children Born in October vs. September 1983 by Race Group

	Percent Gaining Eligibility	Average Gain (in Years) for Children Gaining Eligibility	Average Gain (in Years) for Total Child Population
All Races	10.00	4.54	0.48
Blacks	17.25	4.91	0.87
Non-Blacks	8.71	4.41	0.41

Notes: Weighted averages calculated from simulation of childhood eligibility if born in September vs. October 1983 for children ages 0–17 in the pooled 1981–1988 March CPS. See appendix for additional discussion of these estimates and how they differ from Wherry and Meyer (2016).

IV. Data

We begin by examining changes in childhood health insurance coverage and utilization during the period of expanded Medicaid eligibility (ages 8–14) using the 1992–1996 National Health Insurance Survey (NHIS) Health Insurance Supplements.⁷ The NHIS provides information on respondent quarter of birth, which allows us to estimate discontinuities at the September 30, 1983 birth date. Our analysis is of a similar spirit to Card and Shore-Sheppard (2004), who first estimated a significant increase in coverage and doctor visits for children born after this date.

To examine changes in later life health care utilization (age 15 and 25), we use discharge- level hospital data from three sources. First, we use hospitalization data from the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases. We acquired all state databases available from HCUP for the 1999 and 2009 years that contained information on the patient’s date of birth. These data provide information on all inpatient hospitalizations that occurred in acute care hospitals in 1999 in Arizona, Hawaii, Iowa, New York, Oregon, and Wisconsin, and in 2009, in Arizona, Arkansas, Colorado, Hawaii, Iowa, Kentucky, Maryland, Michigan, Nebraska, New York, North Carolina, Oregon, South Dakota, Utah, Vermont, and Wisconsin.⁸ We supplement these data with the census of hospital discharges that occurred in Texas and California in 1999 and 2009, obtained from the Texas Department of State Health Services and the California Health and Human Services Agency, resulting in the complete census of hospital discharges for 8 states in 1999 and for 19 states in 2009.⁹ To our knowledge, this represents all available state hospital discharge data with patient’s date of birth information for these years.

In addition to hospital discharge data, we use data from all State Emergency Discharge Databases available from HCUP in 2009 that include patient’s date of birth information. These databases provide the universe of outpatient ED visits that occurred in Arizona, Hawaii, Iowa, Kentucky, New Jersey, New York, Utah and Wisconsin (obtained from HCUP) and California (obtained from the California Health and Human Services Agency) in 2009. These data cover all visits for which a patient was treated in an emergency department and released the same day, rather than being admitted to the hospital. State ED data are available for 2009, but not for 1999.

Both the hospital discharge and emergency department data contain information on the diagnoses associated with each visit, total charges, and patient demographics including race and birth month and year. In 2009 (but not in 1999), we observe whether the patient is from a low-income zip code defined as a zip code with median income below $39,999.¹⁰ We classify primary diagnoses as relating to “chronic” or “non-chronic” conditions using the Chronic Condition Indicator software distributed by HCUP.¹¹ We exclude hospitalizations and ED visits for diagnoses related to pregnancy and delivery. Condition Indicator software distributed by HCUP.¹¹ We exclude hospitalizations and ED visits for diagnoses related to pregnancy and delivery.

Combined, our hospitalization data include 689,546 hospital discharge-level observations and 3.9 million ED visits in 2009 for patients born between 1979 and 1987. Our hospitalization sample covers approximately 36 percent of the national population in 1999 and 50 percent of the national population in 2009; our ED visit sample covers 29 percent of the national population.¹²

The bottom rows of Tables 3–5 present hospital and ED utilization rates for our sample.¹³ Table 3 displays hospitalization rates (per 10,000 individuals) for 15 year-olds in 1999, the first year for which we have data. Hospitalization rates at this age were higher for blacks. About 53 percent of all hospitalizations were for chronic illnesses overall; for blacks, chronic illnesses represented about 60 percent of total hospitalizations. For 15-year-olds, the most common of these chronic illnesses are mental disorders, followed by asthma and diabetes.

Table 3.

Estimates of Effect of Childhood Medicaid Eligibility on Log Hospitalizations at Age 15 (1999)

	(1) All	Blacks (2) Chronic	(3) Non-Chronic	(4) All	Non-Blacks (5) Chronic	(6) Non-Chronic
Global polynomial model
4-Year window (N=96)	−0.025	−0.071	0.039	0.042**	0.042	0.041*
	(−0.142, 0.092)	(−0.246, 0.105)	(−0.077, 0.156)	(0.009, 0.074)	(−0.013, 0.096)	(−0.002, 0.084)
3-Year window (N = 72)	0.020	0.028	0.022	0.026	0.028	0.025
	(−0.125, 0.166)	(−0.174, 0.230)	(−0.127, 0.170)	(−0.011, 0.064)	(−0.039, 0.095)	(−0.025, 0.074)
2-Year window (N = 48)	0.034	−0.031	0.135	0.060***	0.087**	0.038
	(−0.150, 0.217)	(−0.319, 0.257)	(−0.043, 0.314)	(0.020, 0.100)	(0.017, 0.157)	(−0.030, 0.107)
Local linear regression
IK Bandwidth Selector	−0.013	−0.005	0.014	0.018	0.027	0.004
	(−0.121, 0.094)	(−0.165, 0.155)	(−0.104, 0.133)	(−0.012, 0.049)	(−0.016, 0.069)	(−0.040, 0.048)
CCT Bandwidth Selector	−0.021	−0.027	0.009	0.032	0.020	0.038
	(−0.163, 0.122)	(−0.281, 0.227)	(−0.114, 0.132)	(−0.014, 0.078)	(−0.032, 0.072)	(−0.034, 0.110)
Rate of hospitalization (ner 10,000)	323.14	193.63	129.52	253.35	131.50	121.85

Notes: Sample includes birth-month observations from pooled AZ, CA, HI, IA, NY, TX, and WI data. All global regression models include birth month fixed effects and a quadratic function in birth month cohort interacted with an indicator that the birth month cohort is October 1983 or later. 95% confidence intervals reported in parentheses;

^***p<0.01,

^**p<0.05,

^*p<0.1.

Table 5.

Estimates of Effect of Childhood Medicaid Eligibility on Log Emergency Department Visits at Age 25 (2009)

	(1) All	Blacks (2) Chronic	(3) Non-Chronic	(4) All	Non-Blacks (5) Chronic	(6) Non-Chronic
Global polynomial model
4-Year window (N=96)	−0.037** (−0.070, −0.004)	−0.102* (−0.211, 0.007)	−0.026* (−0.056, 0.003)	−0.002 (−0.024, 0.019)	0.000 (−0.027, 0.028)	−0.003 (−0.026, 0.020)
3-Year window (N = 72)	−0.049** (−0.087, −0.011)	−0.152** (−0.279, −0.026)	−0.031* (−0.064, 0.002)	0.001 (−0.023, 0.024)	0.005 (−0.028, 0.038)	0.000 (−0.025, 0.025)
2-Year window (N = 48)	−0.030 (−0.077, 0.018)	−0.147* (−0.305, 0.012)	−0.009 (−0.053, 0.034)	0.025* (−0.001, 0.051)	0.046** (0.008, 0.085)	0.022 (−0.007, 0.052)
Local linear regression
IK Bandwidth Selector	−0.023** (−0.045, −0.000)	−0.123*** (−0.212, −0.034)	−0.009 (−0.030, 0.011)	0.021 (−0.005, 0.046)	0.013 (−0.010, 0.036)	0.020 (−0.007, 0.048)
CCT Bandwidth Selector	−0.022* (−0.044, 0.000)	−0.144*** (−0.242, −0.047)	−0.002 (−0.024, 0.020)	0.026* (−0.004, 0.055)	0.012 (−0.012, 0.035)	0.028 (−0.006, 0.062)
Rate of ED utilization (per 10,000)	5,714.95	795.63	4,919.33	2,890.76	335.67	2,555.09

Notes: Sample includes birth-month observations from pooled AZ, CA, HI, IA, KY, NJ, NY, UT, and WI ED data. All global regression models include birth month fixed effects and a quadratic function in birth month cohort interacted with an indicator that the birth month cohort is October 1983 or later. 95% confidence intervals reported in parentheses;

^***p<0.01,

^**p<0.05

^*p<0.1.

Table 4 displays hospitalization rates for 25 year-olds in 2009. Hospitalization rates are more common for this age group. We observe a striking difference in hospital utilization rates across race groups, particularly for chronic illnesses: blacks at age 25 have a chronic illness hospitalization rate more than twice that of non-blacks. About 57 percent of hospitalizations of black patients and about 45 percent of those for non-black patients were for chronic conditions. The most common chronic condition for this age group is also mental disorders. The second most common is diabetes and the third most common is asthma.

Table 4.

Estimates of Effect of Childhood Medicaid Eligibility on Log Hospitalizations at Age 25 (2009)

	(1) All	Blacks (2) Chronic	(3) Non-Chronic	(4) All	Non-Blacks (5) Chronic	(6) Non-Chronic
Global polynomial model
4-Year window (N=96)	−0.071*	−0.106**	−0.025	0.016	0.038*	−0.001
	(−0.144, 0.003)	(−0.198, −0.014)	(−0.108, 0.058)	(−0.012, 0.045)	(−0.004, 0.080)	(−0.037, 0.035)
3-Year window (N = 72)	−0.095**	−0.120**	−0.061	0.022	0.046*	0.001
	(−0.176, −0.013)	(−0.223, −0.017)	(−0.158, 0.036)	(−0.011, 0.054)	(−0.005, 0.097)	(−0.036, 0.038)
2-Year window (N = 48)	−0.144***	−0.167**	−0.113*	0.054***	0.081***	0.031
	(−0.247, −0.040)	(−0.301, −0.034)	(−0.232, 0.006)	(0.020, 0.089)	(0.031, 0.132)	(−0.013, 0.075)
Local linear regression
IK Bandwidth Selector	−0.126***	−0.145***	−0.081**	0.016	0.031	0.002
	(−0.183, −0.068)	(−0.216, −0.073)	(−0.151, −0.012)	(−0.012, 0.043)	(−0.011, 0.072)	(−0.027, 0.031)
CCT Bandwidth Selector	−0.154***	−0.184***	−0.114***	0.013	0.032	0.037
	(−0.228, −0.081)	(−0.278, −0.091)	(−0.179, −0.049)	(−0.026, 0.051)	(−0.013, 0.077)	(−0.017, 0.091)
Rate of hospitalization (per 10,000)	517.14	293.15	223.99	303.71	137.03	166.67

Notes: Sample includes birth-month observations from pooled AR, AZ, CO, CA, HI, IA, KY, MD, MI, NJ, NY, OR, SD, TX, UT, VT and WI hospital data. All global regression models include birth month fixed effects and a quadratic function in birth month cohort interacted with an indicator that the birth month cohort is October 1983 or later. 95% confidence intervals reported in parentheses;

^***p<0.01,

^**p<0.05,

^*p<0.1.

ED visits are more common than hospitalizations and tend to treat less severe conditions. Rates of ED use (per 10,000 individuals) are described in Table 5. The ED visit rate is again higher for blacks. ED visits tend to be for acute conditions with a lower share for chronic illnesses, relative to hospitalizations, for both race groups. We again observe dramatic differences across races related to chronic illnesses, as blacks experience more than twice the rate of ED utilization for chronic illnesses as non-blacks.

v. Empirical Strategy

To estimate the impact of childhood Medicaid eligibility, we use a regression discontinuity approach and compare outcomes for cohorts born just before and after the September 30, 1983 birthdate cutoff. We rely on both a parametric specification (i.e., polynomial) and local linear regression to estimate the discontinuity in outcomes at the birth date cutoff point. These complementary methods offer tradeoffs in terms of bias and variance and are presented together to assess the stability of results (Lee and Lemieux 2010). For our measures of later life utilization, we use the log number of hospitalizations or ED visits as the dependent variable, rather than using a rate that relies on a hard to measure population denominator. Estimates of the RD are interpreted as the proportionate change in the rate of hospitalizations or ED visits.^14,15

We first estimate a second-order polynomial regression model that uses observations from monthly cohorts born within a specified window of the cutoff date. The cohorts are indexed by c, where c=0 for the first cohort born after the cutoff (October 1983). We present results for our main specification that relies on a 4-year window of birth month cohorts on either side of the cutoff (October 1979 through September 1987, c ϵ [−48, 47]). Additionally, we show alternative analyses that use three year (c ϵ [−36, 35]) and two year (c ϵ [−24, 23]) windows of birth month cohorts.

The regression specification used to examine later life utilization is given by¹⁶

$$\log (y_c)=\alpha +\beta D_c+{\gamma }_{0}c+{\gamma }_1{c}^2+{\gamma }_2{D}_c\cdot c+{\gamma }_3{D}_c\cdot c^2+{\delta }_m{M}_c+{\epsilon }_c$$ (1)

where y_c represents the number of hospitalizations or ED visits for a given birth cohort and D_c is an indicator for cohorts born after September 30, 1983 (c ≥ 0). We include a quadratic function in birth month cohort c that is allowed to differ on the two sides of the cutoff point by including an interaction term for those cohorts born after the cutoff. In addition, we include calendar month dummies M_c to control for variation in outcomes related to the link between timing of birth and family characteristics (Buckles and Hungerman 2012). The inclusion of these dummies will also net out the effects of policies that may differentially affect cohorts born in certain months (for example, school entry dates). For each outcome, we present visual evidence by plotting in two-month bins the residuals for birth month cohorts in the 4-year observation window from a regression on the set of calendar month dummies. The fitted line is the result of regressing the residuals on a quadratic function in birth month cohort interacted with the post-September 30, 1983 dummy.

We also use local linear regression to estimate the discontinuity in outcomes at the cutoff point. The estimation is conducted with a triangular kernel and we present results that employ two different optimal bandwidth selector procedures proposed by Imbens and Kalyanaraman (2012) and Calonico, Cattaneo, and Titiunik (2014).

These methods estimate the effect of the eligibility expansions averaged across the full sample of children at the cutoff. This is an example of a “fuzzy” RD design because factors other than date of birth determine eligibility for and take-up of public health insurance. In the NHIS data, we examine outcomes separately by race and by the poverty status of the household. In our analyses of later life health care utilization, we do not have information in our data on whether individuals were eligible for or took up public health insurance during childhood. However, we are able to examine outcomes for certain subsamples that were more likely to be affected by the change in Medicaid policy. In particular, we examine outcomes separately by race and then conduct additional analyses for low-income areas and by state, taking into account variation in the magnitude of the expansions across geographic areas.

VI. Results

Discontinuity in Contemporaneous Coverage and Utilization

We begin by measuring discontinuities in childhood health insurance coverage to supplement our analysis of the discontinuity in eligibility at the September 30, 1983 birth date. Figure 1 plots reported levels of Medicaid coverage for each birth month cohort within a 4-year window on each side of the birthdate cutoff (centered at October 1983). The lines are fitted values from a quadratic regression function in birth month cohort interacted with a dummy variable for cohorts born after September 30, 1983. The graph for blacks and, to a lesser extent, for all races, shows evidence of an increase in Medicaid coverage at the cutoff. When we look separately at children in households with incomes below the poverty line, we see additional evidence of a discontinuity in coverage.

Figure 1:

Medicaid Coverage in Childhood, Ages 8 to 13, NHIS

Source: Authors’ calculations from the National Health Interview Survey, 1992–1996. Cohorts born in 1983 are between the ages of 8 and 13 in these figures. The trend is estimated using children between the ages of 4 and 17.

Table 2 presents the corresponding regression estimates. For all races, we see some evidence of an increase in Medicaid coverage of between 1 and 2 percentage points; however, for the most part, the estimates are not statistically significant. We see strong evidence of an increase in Medicaid coverage for blacks of between 5 and 8 percentage points depending on the specification. The estimates are significantly different from zero at the .05 level in four of five specifications and at the .10 level in the remaining case. Given our previous estimate that 17 percent of black children gained eligibility at the cutoff, these estimates represent a take-up rate on the order of approximately 29 to 47 percent. For non-blacks, we do not find a significant increase in Medicaid coverage and the point estimates are much smaller, indicating less than a one-percentage point change in coverage and implying a take-up rate of at most six percent. Examining children with families below the poverty line only, we find an increase in Medicaid coverage of between 6 and 9 percentage points.¹⁷ The estimates are significant at the .01 level in three of five specifications and at the .10 level in the remaining two cases. We find no change among children in families with incomes above the poverty line.

Table 2.

Estimates of Effect of Childhood Medicaid Eligibility on Medicaid Coverage at Ages 8–13

	All Races (1)	Blacks (2)	Non-Blacks (3)	Households in Poverty (4)	Households Not in Poverty (5)
Global polynomial model
4-Year window	0.010 (−0.008, 0.027)	0.054** (0.002, 0.105)	−0.003 (−0.022, 0.016)	0.089*** (0.044, 0.133)	−0.001 (−0.013, 0.011)
3-Year window	0.017 (−0.004, 0.038)	0.071** (0.012, 0.129)	−0.001 (−0.024, 0.022)	0.091*** (0.038, 0.145)	0.000 (−0.013, 0.013)
2-Year window	0.014 (−0.010, 0.038)	0.045* (−0.004, 0.095)	0.001 (−0.026, 0.029)	0.057* (−0.005, 0.118)	−0.000 (−0.015, 0.014)
Local linear regression
IK Bandwidth Selector	0.014* (−0.001, 0.028)	0.051** (0.003, 0.099)	0.004 (−0.012, 0.019)	0.086*** (0.032, 0.141)	0.001 (−0.010, 0.013)
CCT Bandwidth Selector	0.023** (0.001, 0.045)	0.077** (0.009, 0.145)	0.006 (−0.014, 0.027)	0.073* (−0.004, 0.150)	0.003 (−0.011, 0.017)
Baseline mean
N	0.145 54410	0.319 9,000	0.111 45410	0.503 10609	0.042 40,258

Notes: Data from 1992–1996 National Health Interview Survey Health Insurance Supplements. All global regression models include birth month fixed effects and a quadratic function in birth month cohort interacted with an indicator that the birth month cohort is October 1983 or later. 95% confidence intervals reported in parentheses,

^***p<0.01,

^**p<0.05,

^*p<0.1.

In additional analyses, we examined changes in reports of any insurance coverage (see Appendix Table 2). The change in overall insurance coverage for black children is not statistically significant, but the coefficients suggest between a 2 and 5-percentage point increase among those born just after the cutoff. The fact that the change in Medicaid coverage is larger than the change in overall insurance coverage suggests that these expansions were associated with some crowd-out (i.e., that some children enrolled in Medicaid who would otherwise have enrolled in private insurance). We also examined changes in the probability of a doctor visit in the last 12 months, as well as any short-stay hospital visit (not related to delivery). We find some evidence of an increase in doctor visits among black children.¹⁸ Although it is imprecisely estimated, this positive effect is consistent with Card and Shore-Sheppard (2004) and other research documenting that expanded Medicaid coverage leads to higher rates of utilization.¹⁹

Changes in Medicaid eligibility and coverage documented in Figure 1, Tables 1–2, and Appendix Figures 1–2 indicate that there is clear variation in treatment by race. Accordingly, if Medicaid coverage has long-term effects on health and use of medical care such as hospitalization, it is plausible to expect that effects will vary in a way consistent with the variation in treatment. In the analyses that follow, we present results separately for blacks and non-blacks. Results for all races may be found in Appendix Table 3 and Appendix Figures 6–7.

Discontinuity in Later Life Hospitalizations and ED Visits

Figure 2 presents the profile of log hospitalizations by birth month cohort in 1999, when the cohorts born just on either side of the cutoff are approximately 15 years of age. Visually, the figure reveals little evidence of a discontinuity in outcomes at the September 30, 1983 threshold. Estimates of the discontinuity from the regression analyses reported in Table 3 support this conclusion. Overall, we do not find a consistent pattern of changes in hospital utilization in this year.

Figure 2:

Hospitalizations at Age 15, 1999

Figures plot the residuals from a regression of the log of hospitalizations by birth month on calendar month of birth fixed effects. Results presented using two-month bins.

The top half of Figure 3 displays hospitalizations in 2009 when cohorts born around the cutoff were approximately 25 years old. Table 4 presents the corresponding discontinuity estimates. For blacks, there is a notable drop in hospitalizations visible at the cutoff. The regression analysis indicates a reduction in hospitalizations of between 7 and 15 percent for those cohorts born just after the September 30, 1983 date, depending on the specification. Our estimates also indicate declines in hospitalizations for chronic illnesses on the order of 11 to 18 percent across specifications. For hospitalizations related to non-chronic illness, the estimated decline is smaller—3 to 11 percent—and only significant when using local linear regression methods. We do not find any evidence of a similar improvement for non-blacks.

Figure 3:

Hospitalizations and Emergency Department Visits at Age 25, 2009

Figures plot the residuals from a regression of the log of hospitalizations or ED visits by birth month on calendar month of birth fixed effects. Results presented using two-month bins.

Since data were available for more states in 2009 than in 1999, the sample composition is different for these two analyses. However, we find similar results when we estimate our model with the 2009 data for only those states that were available in 1999. This suggests that the inclusion of additional states is not driving the difference in results between 1999 and 2009. These results are reported in Appendix Table 4.

The bottom half of Figure 3 and Table 5 present results for ED visits. We find a reduction in rates of ED visits of between 2 and 5 percent among black cohorts born immediately after the birth date cutoff, although the estimates are not significant across all windows of observation. When we examine ED visits by their relation to chronic illness, we find evidence of a sizeable decline in visits related to chronic illness (of between 10 and 15 percent). For non-blacks, we find no significant evidence of a similar reduction in ED visits.

Assuming similar effect sizes and hospitalization rates across other states, our point estimates imply that, nationally, there were approximately 2,200 to 4,800 fewer inpatient hospitalizations among black cohorts born during the first year after the cutoff at age 25.²⁰ The change in the probability of gaining eligibility across the birth date cutoff was about 17 percentage points for blacks. Our point estimates from our specifications therefore imply that there were between 2.1 and 4.6 fewer hospitalizations at age 25 for every 100 black children who were made eligible for (on average) 4.8 additional years of Medicaid eligibility as a result of the expansions. This reduction for eligible children is large relative to the average rate of hospitalization among all 25-year-old blacks, representing 41 to 88 percent fewer hospitalizations relative to that average. However, because the children that were affected by these expansions were in poor households, and because the poor tend to be in worse health than the general population, it is likely that their baseline hospitalization rates would be higher than that of a typical black 25 year old (see, e.g., Case, Lubotsky and Paxson 2002). For the particular cohorts that we study, the presence of a chronic condition during childhood was significantly higher for those individuals enrolled in Medicaid compared to those who were not.²¹

We can further scale these estimates by take-up rates to arrive at the effect of Medicaid coverage, rather than Medicaid eligibility, on hospitalizations later in life. Based on the estimates presented in Tables 1 and 2, we estimated take-up rates on the order of 29 to 47 percent. A take-up rate of 29 percent implies that for every 100 black children who newly enrolled in Medicaid, there were between 7.2 and 15.9 fewer hospitalizations at age 25. Meanwhile, a take-up rate of 47 percent implies between 4.5 and 9.8 fewer hospitalizations for every new 100 black children who enrolled in Medicaid. However, these estimates are likely to overstate the magnitude of the effects among children who took up Medicaid since counts of Medicaid enrollment reported in survey data are often well below administrative record counts.²² This measurement error in reported coverage is likely to lead to an underestimate of changes in coverage associated with the Medicaid expansions.²³ In addition, since parents are more likely to enroll children who are most likely to benefit from coverage (e.g. who are sick or have high medical needs, see Marton and Yelowitz 2014), the effect of coverage for this group is likely larger than the effect would be for all children who gained Medicaid eligibility if they took up coverage.

Similarly, our point estimates imply that, nationally, there were approximately 7,100 to 17,600 fewer annual ED visits experienced at age 25 by blacks born the first year after the cutoff. This estimate implies that there were between 6.7 and 16.8 fewer ED visits at this age for every 100 black children made eligible as a result of the expansions. This suggests that gaining an average of 4.8 additional years of Medicaid eligibility in childhood lowers ED use by 12 to 29 percent at age 25. When scaled by our estimate of 29 percent take-up, this implies that there were between 23.1 and 57.9 fewer ED visits at age 25 for every 100 black children newly enrolled in Medicaid. Using our higher take-up estimate (47 percent) implies between 14.3 and 35.7 fewer ED visits for every 100 new enrollees. However, as we noted previously, baseline ED use among adults who grew up in low-income families (and took up Medicaid coverage) is likely higher than average ED use in the population.

We did not find evidence of similar reductions in utilization among non-black children; the difference in the estimated effects between blacks and non-blacks is plausible for two reasons. First, black children were much more likely to benefit from these expansions than other children (as seen in Table 1). In addition, we did not observe a similar take-up of Medicaid coverage for non-blacks as compared with blacks (Table 2). Second, blacks experience a rate of chronic illness hospitalization and chronic illness ED use that is more than twice that of non-blacks (see Table 5). As blacks are at a substantially higher risk of using these types of medical services, they may have the most to gain from expanded access to health care in childhood.

Finally, we also explore the effects of the discontinuity in eligibility on hospitalizations and ED visits for black patients using more narrow disease categories based on the International Classification of Diseases, Ninth Revision standard. Although these estimates tend to be imprecisely estimated due to the small number of hospitalizations that fall within each category, our estimates suggest that Medicaid eligibility was associated with a reduction in hospitalizations for a broad range of disease groups. Further discussion is provided in the Appendix and the results are reported in Appendix Tables 5 through 8.

VII. Exploiting Variation in Extent of Treatment

Low-income Zipcodes

Motivated by estimates in Appendix Figure 1 and Table 2 that show particularly large changes in Medicaid eligibility and coverage for children in low-income households, we next examine changes in hospitalizations and ED visits in 2009 for patients living in low-income zip codes (Tables 6–7 and Figure 4). If the information were available in our data, we would ideally examine the later health care utilization of individuals who were poor (and eligible for the Medicaid expansions) during childhood. However, if children who grew up in poor families still reside in low-income zip codes, we might expect to see larger changes for patients from these zip codes.

Table 6.

Estimates of the Effect of Childhood Medicaid Eligibility on Log Hospitalizations in Low-Income Zipcodes at Age 25 (2009)

	(1) All	Blacks (2) Chronic	(3) Non-Chronic	(4) All	Non-Blacks (5) Chronic	(6) Non-Chronic
Global polynomial model
4-Year window (N=96)	−0.100	−0.148**	−0.032	0.052*	0.070*	0.038
	(−0.225, 0.024)	(−0.293, −0.003)	(−0.181, 0.118)	(−0.001, 0.106)	(−0.011, 0.151)	(−0.021, 0.096)
3-Year window (N = 72)	−0.150**	−0.188**	−0.092	0.046	0.062	0.032
	(−0.294, −0.005)	(−0.353, −0.023)	(−0.266, 0.082)	(−0.023, 0.114)	(−0.038, 0.162)	(−0.039, 0.103)
2-Year window (N = 48)	−0.227**	−0.280***	−0.150	0.078*	0.080*	0.075
	(−0.412, −0.043)	(−0.467, −0.094)	(−0.376, 0.075)	(−0.005, 0.161)	(−0.015, 0.174)	(−0.020, 0.171)
Local linear regression
IK Bandwidth Selector	−0.169**	−0.235***	−0.081	0.036	0.037	0.035
	(−0.306, −0.032)	(−0.386, −0.084)	(−0.235, 0.073)	(−0.017, 0.089)	(−0.034, 0.108)	(−0.019, 0.090)
CCT Bandwidth Selector	−0.170**	−0.257***	−0.080	0.041	0.038	0.068*
	(−0.308, −0.031)	(−0.433, −0.081)	(−0.245, 0.085)	(−0.023, 0.106)	(−0.049, 0.125)	(−0.010, 0.146)

Notes: Sample includes birth-month observations from pooled AR, AZ, CA, CO, HI, IA, KY, MD, MI, NJ, NY, OR, SD, TX, UT, VT, and WI hospital data. Models with all races also include NC and NE. All global regression models include birth month fixed effects and a quadratic function in birth month cohort interacted with an indicator that the birth month cohort is October 1983 or later. 95% confidence intervals reported in parentheses

^***p<0.01

^**p<0.05

^*p<0.1.

Table 7.

Estimates of Effect of Childhood Medicaid Eligibility on Log Emergency Department Visits in Low-Income Zipcodes at Age 25 (2009)

	(4) All	Blacks (5) Chronic	(6) Non-Chronic	(7) All	Non-Blacks (8) Chronic	(9) Non-Chronic
Global polynomial model
4-Year window (N=96)	−0.056**	−0.129*	−0.043**	−0.014	−0.012	−0.014
	(−0.105, −0.007)	(−0.283, 0.025)	(−0.084, −0.002)	(−0.053, 0.025)	(−0.066, 0.043)	(−0.055, 0.028)
3-Year window (N = 72)	−0.064**	−0.134	−0.052**	−0.027	−0.033	−0.027
	(−0.123, −0.005)	(−0.307, 0.039)	(−0.102, −0.003)	(−0.075, 0.020)	(−0.098, 0.032)	(−0.076, 0.023)
2-Year window (N = 48)	−0.038	−0.165	−0.015	0.001	0.012	−0.001
	(−0.115, 0.039)	(−0.376, 0.046)	(−0.080, 0.049)	(−0.050, 0.052)	(−0.064, 0.087)	(−0.054, 0.052)
Local linear regression
IK Bandwidth Selector	−0.035*	−0.119	−0.024*	−0.012	−0.018	−0.011
	(−0.072, 0.002)	(−0.266, 0.029)	(−0.051, 0.004)	(−0.047, 0.023)	(−0.061, 0.025)	(−0.047, 0.025)
CCT Bandwidth Selector	−0.036*	−0.136*	−0.013	−0.001	−0.017	0.001
	(−0.078, 0.006)	(−0.292, 0.019)	(−0.043, 0.017)	(−0.051, 0.049)	(−0.071, 0.037)	(−0.052, 0.054)

Notes: Sample includes birth-month observations from pooled AZ, CA, HI, IA, KY, NJ, NY, UT, and WI ED data. All global regression models include birth month fixed effects and a quadratic function in birth month cohort interacted with an indicator that the birth month cohort is October 1983 or later. 95% confidence intervals reported in parentheses

^***p<0.01

^**p<0.05

^*p<0.1.

Figure 4:

Hospitalizations and ED Visits, Patients with Low-Income Zipcodes, 2009

Figures plot the residuals from a regression of the log of hospitalizations or ED visits by birth month on calendar month of birth fixed effects. Results presented using two-month bins.

Estimates in Table 6 indicate a reduction in total hospitalizations of between 10 and 23 percent among black cohorts in low-income zip codes born just after September 30, 1983, and the coefficient estimate is statistically significant at the .05 level in four of our five specifications. ²⁵ In addition, the decline appears to be concentrated among hospitalizations related to chronic illness, where we see reductions of between 15 and 28 percent that are statistically significant in all specifications. There is no significant reduction in hospitalizations for non-chronic illnesses.

Similarly, we find evidence of a larger decrease in ED visits for blacks in low-income zip codes than for all blacks with estimates ranging between 4 and 6 percent, although the estimates are only significant in certain specifications. There is some evidence of reductions in chronic and non-chronic illness related ED visits as well, but the estimates are not consistently significant. Finally, as in our previous analysis, we find no evidence of a significant reduction in hospitalizations or ED visits among non-black persons from low-income zip codes. However, in most models we are unable to rule out small reductions in utilization for non-blacks. In addition, the use of current low-income zip code as a proxy for childhood poverty status may be less accurate for non-blacks given a higher rate of mobility over time for these cohorts compared to the black cohorts we study.²⁶

State Variation in Treatment

In this section, we use estimated differences across states in the average size of the discontinuity in eligibility-years for cohorts born at the cutoff as an additional source of variation in our analysis. The gain in Medicaid eligibility for children born after the cutoff varied by state due primarily to differences in Medicaid policies in place before the expansions and differences in state socioeconomic characteristics. For example, the impact of the requirement that states cover all poor children would depend on both prior eligibility levels determined by the state’s AFDC eligibility threshold (e.g. 14% FPL vs. 79% FPL) and the concentration of children in the state with family incomes between AFDC eligibility thresholds and the poverty line.

Appendix Table 1 presents estimates of the average eligibility gain at the cutoff for each state in our study. These estimates were calculated using state-specific samples of children from the Current Population Survey (CPS) and therefore capture the average magnitude of the discontinuity at the cutoff given a state’s eligibility rules and distribution of family characteristics. The size of the discontinuity in eligibility varies from 0.05 years of eligibility in California to 1.33 years of eligibility in Arkansas.²⁷

Given this variation across states in the estimated discontinuity in childhood eligibility, one might expect the discontinuity in utilization at the cutoff to be larger in states where the change in eligibility across that threshold is greater. A potential concern with this source of variation, however, is that pre-expansion state eligibility rules and state socioeconomic characteristics may be correlated with additional types of heterogeneity that could affect either the implementation of the Medicaid expansions, access to care, or trends in the underlying health of the population. For this reason, this source of variation in expanded Medicaid eligibility is not as clearly exogenous as birthdate. There also is likely measurement error in the assigned discontinuity in Medicaid eligibility, particularly in the later period, as individuals may have moved out of their childhood state. For these reasons we are cautious about placing too much weight on the results from the state analyses.

To explore this heterogeneity, we estimate equation (1) separately for each state using 2009 data. We then plot the estimate for each state against its estimated discontinuity in eligibility-years with 95% confidence intervals. Appendix Figures 8 and 9 present the resulting scatterplots for each outcome by race group.²⁸ For the most part, the individual state-level estimates are not statistically distinguishable from zero. However, the figures do suggest a negative relationship between the change in utilization at the cutoff and the size of the discontinuity in Medicaid-eligibility years at the cutoff.

In the appendix, we formalize this analysis in a regression framework by interacting the eligibility discontinuity size with the cutoff indicator. Although not always statistically significant, the point estimates indicate a negative relationship between utilization and the size of the discontinuity for almost all model specifications and race groups. We also report results using hospitalizations in 1999, when birth cohorts born around the cutoff birth date are 15 years old. As in our main RD analysis, we do not find systematic evidence that those born immediately after the cutoff had fewer hospitalizations at this age.

VIII. Sensitivity Analyses

Hospitalizations for Acute Conditions

As a sensitivity analysis, we estimate changes in hospitalizations for appendicitis and injuries for all patients and the low-income sample in 2009 by race group. Both appendicitis and injury are acute conditions that are less likely to be sensitive to medical care received in the past. For that reason, we believe it is less plausible that coverage in childhood would substantially influence hospitalizations for these conditions. If we find effects on these types of hospitalizations, it may indicate that the assumptions of our RD model are incorrect.

The results of these analyses are reported in Appendix Table 21. The first panel shows the results for hospitalizations in 2009 for patients from all zip codes stratified by race. The second panel shows similar results for patients from low-income zip codes. In both panels, we find point estimates that are close to zero, none of which are statistically significant. Overall, this suggests there was little impact of the policy on these types of visits, consistent with our expectation that these visits should not be affected by access to medical care in childhood.

Placebo Estimates at Non-Discontinuity Points

We conduct a second type of placebo test using data on cohorts born prior to the actual eligibility cutoff. We place an artificial “cut off” date in the center of each two-, three-, and four-year window beginning with cohorts born in January 1965. We then estimate models that mimic our “global” polynomial specifications using these window sizes and the “local” specifications that employ our two optimal bandwidth selection procedures. The last month used in the estimation of these placebo effects is September 1983, immediately before the actual change in Medicaid eligibility occurs. We perform this analysis using the sample of black patients for ED visits and hospitalizations and those related to chronic illnesses.

We then compare the distribution of these placebo estimates to the effect estimated at the “true” cutoff (see Appendix). The fraction of placebo estimates that indicate an effect larger in magnitude than our true estimate provides a p-value.²⁹ We find that the placebo estimates for hospitalizations for blacks are only larger than the true estimate 10.8 percent (4-year bandwidth, also not statistically significant using standard inference), 4.5 percent (3-year bandwidth), and less than 1 percent (2-year bandwidth) of the time under the global polynomial specification. Under the local linear specifications, less than 1 percent of the placebo estimates exceed the true estimate for both bandwidth selectors. For chronic illness related hospitalizations, only 5.4 percent, 5.2 percent, and 5.6 percent of the placebo estimates in the global model using 4-, 3-, and 2-year bandwidths and only 3.1 percent and 2.3 percent of the placebo estimates in the local models exceed the true estimate.

Conducting a similar analysis on the low-income sample, only 3.8 of placebo estimates are larger in magnitude than the true estimate in the 4-year global model and none is larger in the 3-year and 2-year global models or in the local model that uses the IK bandwidth selector. In the local model that uses the CCT bandwidth selector, the true effect exceeds all but 4.6 percent of the placebo effects. When we conduct this analysis for chronic illness related hospitalizations, we find that the true effect exceeds at least 99% of the placebo estimates in all 5 specifications.

When examining ED visits, our results conform less to the original inference conducted in Tables 5 and 7. Most model specifications exhibit p-values substantially higher than conventional significance levels when looking at all ED visits. We find somewhat more encouraging results for chronic illness related ED visits: using the global model, we find that 10.8 percent (4-year window), 2.6 percent (3-year window), and 7.3 percent (2-year window) of the placebo effects are larger in magnitude than the true estimate. In the local models, we find that 5.4 percent of placebo effects (IK bandwidth selector) and 2.3 percent of placebo effects (CCT bandwidth selector) are greater than the true estimate. However, in the low income sample, we find much higher p-values when using the placebo distribution, although these estimates were for the most part not statistically significant when using traditional methods of inference (see Table 7).

Overall, the placebo simulations conducted in this section indicate that the estimated effects of the Medicaid policy on hospitalizations among black cohorts is larger than estimates that we might observe due to chance. The simulations provide particularly convincing evidence supporting our results for the low-income subsample of blacks. On the other hand, the placebo tests for ED visits suggest that the evidence for Medicaid effects is not as strong as suggested by the conventional test statistics. Although the effect of Medicaid on chronic illness ED visits estimated at the true discontinuity exceeds the majority of placebo estimates, many of the placebo estimates are larger in absolute value than the “true” effect in other models.

Endogenous Mobility

In our analyses, we assume that population levels and characteristics trend smoothly across birth month cohorts. While interstate migration is substantial, it should not affect the estimates in our main model unless there is differential mobility at the cutoff date. One concern might be, for instance, if individuals move out of state in response to receiving additional coverage, this mobility may bias the estimates. To further investigate this concern, we use restricted access NHIS data for the years 2004–2013 with information on both state of birth and state of residence for the cohorts in our sample. We re-estimate equation (1) with a dependent variable indicating whether the respondent resides in a state different from their state of birth at age 25. The results are reported in Appendix Table 22. We find no evidence of a change in mobility for cohorts born after the September 30, 1983 cutoff.³⁰

IX. Was the Upfront Cost of the Medicaid Expansions Offset by Lower Utilization Later in Life?

The results presented in this paper provide compelling evidence that expanding Medicaid coverage to children lowers future health care costs by improving health and reducing later life utilization among those who gain eligibility. In this section, we provide “back of the envelope” calculations on the magnitude of these cost savings relative to the upfront cost of expanding Medicaid. We focus on savings related to reduced hospitalizations given the strong evidence for this finding but similar analyses for ED visits may be found in the Appendix.

To conduct this analysis, we estimate a model similar to that reported in Table 4 but using the log of total hospital costs by birth month cohort as the dependent variable. We calculate total hospital costs by applying HCUP “cost to charge” ratios to the discharge-level data on total charges.³¹ These “cost to charge” ratios are designed to estimate the resource cost of a hospital visit using data from accounting reports collected by the Centers for Medicare and Medicaid Services. We then sum total costs at the birth month cohort level and estimate models as described in equation (1). The results of this analysis are reported in Appendix Table 23 and Appendix Figure 16.

We find that hospital costs among black cohorts fell by between 9 and 14 percent for those born immediately after the birth date cutoff. These reductions in costs are highly significant in the local linear models, but only marginally significant or not significant in the global polynomial specifications. Using the point estimates, assuming that the results apply to all states, and using the total hospitalization costs of those born the year before the cutoff as a baseline, our analysis implies that the Medicaid expansions reduced total hospital costs at age 25 by between $20.8 and $32.4 million for black cohorts born the year following the cutoff.³²

To arrive at an estimate of the original cost of expanding Medicaid, we rely on the average spending per non-disabled child enrolled in Medicaid in 1991, which was $902 per child (in 1991 dollars) (Congressional Research Service 1993).³³ Multiplying this amount by the average gain in years enrolled in Medicaid using information from Tables 1 and 2 and assuming a 3 percent discount rate, this implies that the total cost of the eligibility expansions for all children born during the year following the September 30, 1983 cutoff was approximately $910 million dollars in 2009.³⁴ The cost offsets from childhood Medicaid expansions generated by reductions in hospitalizations among black cohorts at age 25 therefore represent between 2 and 4 percent of the total cost of the expansions to children of all races.³⁵ If the reduction in utilization we observe at age 25 persists for several years, the cost offsets associated with these expansions will be even larger. In additional analyses reported in the Appendix, we find evidence suggesting that a portion of cost savings accrued to the government in the form of reductions in publicly-subsidized health care utilization.

These calculations indicate that the long-run cost savings from the Medicaid expansions may be quite substantial. Considering other research on the long run effects of these expansions on other outcomes, the true cost offsets of the Medicaid expansions might be larger still.

X. Conclusion

Policies that expand public health insurance coverage tend to increase utilization and, thus, the total resources devoted to health care spending in the economy in the short term. However, there may be longer-term consequences that do not materialize until years later if increased access to care improves health and productivity over the lifecourse. While these long-term effects are often cited in policy discussions, very little credible evidence exists on the magnitude of these effects, or even if they are present at all. Long-term health improvements may translate into important cost offsets if they reduce the need for future spending on expensive medical care, which could represent a substantial, but previously unaccounted-for, benefit of public insurance expansions.

In this paper, we provide evidence of such effects by exploiting a discontinuity in the number of years a child is eligible for Medicaid based on his or her date of birth. Because several of the early Medicaid coverage expansions to poor children applied only to children born after September 30, 1983, children born immediately after this cutoff received more years of Medicaid eligibility throughout childhood. Among blacks, who were most likely to be affected by these expansions, we find that those born immediately after the cutoff had a significant reduction in hospitalizations at age 25 compared to those born immediately before the cutoff. The effect is particularly pronounced for chronic illness related hospitalizations, among patients in low-income neighborhoods, and in states where the size of the eligibility discontinuity was large. A back of the envelope calculation based on our point estimates suggests that these reductions in utilization for the cohorts of black children born one year after the birth date cutoff offset between 2 and 4 percent of the total cost of the expansions for this cohort. Additionally, we find suggestive evidence that those born immediately after the cutoff experienced fewer ED visits at age 25, although these results are not consistently statistically significant. This evidence suggests that health interventions in the pre-teen and early teen years for disadvantaged populations can provide long-term health benefits.

An important limitation to this work is that we only observe hospitalizations and ED visits at two points in time for the affected cohorts. We examine these outcomes at age 15, one year following the period of elevated Medicaid coverage, and then 10 years later at age 25. While we observe a sizeable reduction in health care utilization for black cohorts at age 25, we are unable to pinpoint at what age this effect emerges and whether it is sustained. An additional limitation is that we are unable to identify the specific mechanisms leading to the observed changes. Future opportunities to analyze the health care utilization and health of these cohorts over time will be critical to establishing a more complete picture of the long-term effects of childhood Medicaid coverage.