Economic inequality between white and non-white Americans remains pervasive. This column analyses how health inequality relates to disparities in other key economic outcomes. At age 55, Black men and women have the frailty (i.e. biological age) of white men and women 13 and 20 years older, respectively, while Hispanic men and women exhibit frailty akin to white men and women five and six years older. These health disparities explain a large portion of the differences in economic outcomes such as disability, length of working life, nursing home entry, and overall lifespan.
Economic inequality between white and non-white Americans is pervasive. Black and Hispanic Americans earn less than white Americans. For instance, in 1995, Black and Hispanic men’s hourly wages were about two-thirds those of white men (Altonji and Blank 1999). Although racial earnings gaps have improved over time, the median annual earnings of full-time Black workers in 2009 were only 80% those of full-time white workers (Lang and Lehman 2012). Non-white Americans also hold less wealth than their white counterparts. Between 1984 and 1995, Black couples held about a quarter of the wealth held by white couples (Altonji and Doraszelski 2005). Between 1989 and 2016, the median net worth of white families was almost seven times larger than that of Black families (Ashman and Neumuller 2020). Black Americans also invest significantly less than white Americans in the stock market and face a higher risk of unemployment (Derenoncourt et al. 2024).
Racial inequality in the US is not confined to economic outcomes. In Russo et al. (2024), we show that the US is plagued by extensive health inequality by race, ethnicity, and gender. We also show that racial health disparities can explain a large portion of the differences in economic outcomes such as disability, length of working life, nursing home entry, time spent in bad health, and overall lifespan.
We use data from the Health and Retirement Survey (HRS) on Americans aged 51 and older between 1996 and 2018. Borrowing from the medical literature (Searle et al. 2008), we construct a frailty index, which measures the share of health deficits (diagnosed diseases, difficulties with carrying out basic tasks, etc.) a person has at every age. In particular, our frailty index includes 35 deficits, encompassing physical and mental impairments and medical diagnoses. We study the prevalence of these deficits between white, Black, and Hispanic men and women and find evidence of large differences by race, ethnicity, and gender. Most deficits are more prevalent among Black and Hispanic people. Though Hispanic and Black men and women between the ages of 55 and 59 suffer more from diabetes and obesity than their white counterparts, we find that the medical diagnosis of various conditions is less frequent among non-white people.
Motivated by our findings and by the medical literature on racial disparities in diagnosis and treatment (Geiger 2003, Kim et al. 2018, Lin et al. 2021), we develop a new health measure that addresses underdiagnosis. We call this measure ‘potential frailty’ and construct it by imputing diagnoses for Black and Hispanic individuals under the assumption that there is no underdiagnosis for white individuals and that the relationship between health impairments and diagnosis is the same by race and ethnicity. We find that potential diagnosed deficits are substantially more prevalent than the observed deficits (Figure 1). For instance, potential lung disease is 161.5% more prevalent among Black men than its observed counterpart. We also find that underdiagnosis is more widespread among Black than Hispanic people and is worse for men than for women.
Figure 1 Potential health deficits prevalent at ages 55–59, women (top) and men (bottom)
As economists, an important dimension of a good measure of health is its ability to help predict key economic outcomes, including future health, mortality, disability, retirement, nursing home entry, and so on. We find that frailty is an important predictor of these health outcomes, including by race, ethnicity, and gender; excepting retirement, it is somewhat more predictive than self-reported health status, an alternative, widely available, and commonly used measure of health. In addition, frailty has a desirable quantitative interpretation. We thus adopt it as our measure of health inequality.
In our analysis of health inequality, we find that, on average, white men and women experience lower levels of frailty than their Black and Hispanic counterparts (Figure 2). For instance, a 55-year-old Black man has the same frailty (or biological age) as a Hispanic man who is five years older (age 60) and a white man who is 13 years older (age 68). Similarly, a 55-year-old Black woman has the frailty of a Hispanic woman who is six years older (age 61) and a white woman who is 20 years older.
Figure 2 Average frailty by age, women (top) and men (bottom)
Because of our previous results on the underdiagnosis of diseases for Black and Hispanic Americans, we also document inequality in potential frailty. Average potential frailty is consistently higher than observed frailty (Figure 3). In particular, between the ages of 51 and 90, average potential frailty exceeds average frailty by 15.8%, 12.1%, 6.0%, and 4.2% for Black men, Black women, Hispanic men, and Hispanic women, respectively. Moreover, potential frailty amplifies the differences in biological age between non-white and white people. For example, a 55-year-old Black woman has the same frailty as a 75-year-old Hispanic woman (a 20-year gap) and the same potential frailty as an 80-year-old white woman (a 25-year gap). Similarly, a 55-year-old Black man has the same frailty as a 69-year-old white man (a 14-year gap) and the same potential frailty as a 76-year-old white man (a 21-year gap).
Figure 3 Average potential frailty by age, women (top) and men (bottom)
Having found evidence of such large health inequalities, we quantify to what extent health differences at age 55 can affect disparities in life expectancy, disability, retirement, and the duration of nursing home residency. We do so by estimating a statistical model that captures the dynamic evolution of health, mortality, and our economic outcomes of interest. We start by estimating how health and these outcomes change over time. Next, we simulate these outcomes to create simulated histories. Then, we simulate these outcomes by assigning Black and Hispanic men and women the same initial health distribution at age 55 as white men and women. We find that health inequality across racial and ethnic groups at age 55 can explain about half of the gaps observed between Black and white people in every economic outcome we consider (Figure 4). For instance, assigning white women’s health at 55 to Hispanic and Black women reduces the gap in the fraction of time spent in bad health by 40% and 63%, respectively. Similarly, health at 55 accounts for 35.3% of the life expectancy gap between Black and white men, and 48% of the gap between Black and white women. Our simulation results show that assigning 55-year-old non-white people the frailty of their white counterparts vastly reduces gaps in our outcomes of interest. Therefore, policies to reduce health gaps could also reduce gaps in other important economic outcomes.
Figure 4 Effects of health inequality on economic disparities
Notes: These graphs report averages by demographic group. The percentage of time in bad health is computed as the fraction of remaining life spent in one of the two lowest health states (‘poor’ and ‘fair’ health or frailty quintiles) conditional on remaining alive. Working years are defined as the years of not receiving Social Security or disability benefits.
Source : VOXeu
