1 Introduction

Modern research has documented rising despair and declining psychological well-being among large sections of the US population. A particular focus has been low-education White Americans. Some of this important empirical work has documented deaths caused by opioids and suicide.Footnote 1 A further group of writings has concentrated on measures of psychological ill-health and indirect markers of distress (Cherlin 2018; Gaydosh et al. 2019; Goldman et al. 2018; Glei and Weinstein 2019; Graham 2017; Muennig et al. 2018; Graham et al. 2018).

This paper focuses on a different, and routinely excluded, group of men and women: Native Americans. In 2019, there were just over four million citizens who identified solely as Native American and nearly seven million in combination with one or more races (Appendix Table 13). Yet their experiences are rarely included in work by economists and social scientists (Ferguson 2016; Falleti 2020). As one example, a search on “Native Americans” on the entire history of the Web of Science produces 47 journal articles that are categorized as Economics. The total number of published articles in the Web of Science categorized as Economics is approximately 700,000. A search on the simple term “Americans” produces approximately 16,000 articles in Economics.

This lack of literature among social scientists is somewhat perplexing. Despite the general emphasis on deaths of despair among White Americans, Shiels et al. (2017) has shown that between 1999 and 2016, premature mortality among Native Americans increased at almost twice the rate of White Americans, with the primary causes being due to “deaths of despair” such as drug poisonings, chronic liver disease and suicide. Woolf et al. (2018) confirm the increase in premature deaths during midlife from similar causes for the same period. Han et al. (2021) show that non-Hispanic American Indian or Alaska Natives had the highest rates of increase for methamphetamine overdose in the USA for the period 2011–2018, rising from 5.6 to 26.4 per 100,000 among men and from 3.6 to 15.6 for women. Non-Hispanic white men saw increases from 2.2 to 12.6 and white women from 1.1 to 6.2, with all the other racial gender groups having lower rates. Spillane et al. found that from 2000to 2016 the age-standardized rate of alcohol-induced deaths among Natives was markedly higher than for any other race or ethnicity (Spillane et al. 2020). The largest increase in the rate of alcohol-induced deaths was observed among American Indian and Alaska Native men (3.3%) and women (4.2%) versus 1.4% and 3.1% overall. Recent data also show that suicide rates have been highest among Native peoples (Curtin and Hedegaard 2019), and drug poisoning deaths per 100,000 population in 2019 were higher among Native people than any other group.Footnote 2

These differences are long standing (Tower 1989; Christian et al. 1989; Jones 2006) but may be worsening: Best et al. (2018) estimates that from 2017 to 2030, all-cause premature death rates among Native Americans and Alaska natives will increase by 10%. However, to our knowledge, a simultaneous rise in physiological pressures that may drive these trends in deaths of despair among Native people and the economic forces that may underlie them has not been demonstrated in the literature. This is the primary contribution of this paper.Footnote 3

We draw upon a mixture of economic and psychological data and create a measure of persistent, or chronically poor, mental health, which we call chronic distress, and then chart its level through time among particular groups of US citizens and its association with economic factors. To allow comparisons with modern research on White and Black Americans, we pool data from 27 years of the Behavioral Risk Factor Surveillance System (BRFSS) survey organized by the Centers for Disease Control and also use recent information available from surveys organized by the Gallup Corporation. These data are most comprehensive survey data available on the mental well-being of Native Americans in the USA.

Our primary measure of chronic distress, following Blanchflower and Oswald (2020), is whether someone reported that all 30 days out of the last 30 that their mental health was not good, including stress, depression and problems with emotions. We use this term given recent evidence of the long-run consequences of chronic physical and mental pain (Blanchflower and Bryson 2022, 2021). This definition differs from that used by Gagné et al. (2021) who used 15 days or more out of 30; however, we show below that regardless of the precise cutoff of 15 days or later, Native Americans face consistently higher rates of distress past. Obviously, this measure of chronic distress is not the directly the same as “despair” popularized by the work of Case and Deaton (2015, 2017). Shanahan et al. (2019) note that the deaths of despair literature “has neither defined nor empirically assessed its central concept, despair” and argue that it requires “multidimensional conceptual mapping” (p. 854). We view the measure used here and grounded in the work of Blanchflower and Oswald (2020) as part of this mapping. While the concept of despair arguably implies something subtle about expectations of the future and our question focuses on experiences in the present, it is plausible that current, persistent feelings of distress will impact beliefs and expectations of the future. This measure, meaningful in its own right, thus also captures some dimensions of the experience of despair, even if imperfectly.

We show that the upward trend in chronically poor mental health among Native Americans is more pronounced than for White Americans. Those who are classified by the BRFSS as “other race” have seen a similar rise. This general increase of chronic distress among Native Americans during the 1990s and 2000s may be surprising given the large increase in economic opportunities provided by on-reservation Casino gaming during that period. The literature on the rise of casino gaming has found it has had significant positive effects for the regions and people it affects, both physiologically and materially (Akee et al. 2010; Evans and Topoleski 2002).Footnote 4 We will be unable to answer whether the upward trend in distress exists primarily for Native peoples in, near,or outside of reservations given limited data on small-area geography. However, less than half of Native peoples live on reservations, so the patterns we observe may be due to people outside of regions that experienced income gains due to gaming (Akee and Taylor 2014). A prior literature has documented differences in the economic conditions and health conditions of Native Americans (Gracey and King 2009; Walls and Whitbeck 2011; Barnes et al. 2010). However, to our knowledge, the widening and consistently high levels of chronic distress in the latest era have never been shown previously, nor their correlation with local economic change.

This paper is not intended as part of a so-called deficit narrative (Tuck 2009) nor to diminish the successes of many Native Nations.Footnote 5 Rather, it is an attempt to constructively refocus research and policy conversations to understand the social forces that have resulted in the rise in distress by highlighting the fact that White Americans are not the most intensely affected. We also provide evidence on what factors may mitigate chronic distress for Native peoples: we offer evidence that US states with largest proportion of Native Americans as a fraction of their population exhibit lower levels of extreme distress.Footnote 6

2 Background and data

We make use of data from two main sources. First, we pool twenty-seven years of 1993–2019 annual survey data files from the CDC's Behavioral Risk Factor Surveillance System (BRFSS).Footnote 7 Second, we examine information from the Gallup Daily Tracker Polls of 2017 and 2018. We focus on these two surveys specifically because they are the only nationally representative, large publicly available data sets in the USA that separately identifies Native American respondents and also has information on mental health. It is important to note that both of these data sets only allow us to identify Native Americans who claim only Native American racial identities. Native Americans who report multiple identities are classified as “other race” or “multi-race” along with all other racial groups reporting multiple identities or a race other than White, Black, Asian, or Native American. In addition, these data are cross-sectional and do not allow for linking individuals over time.

The primary data we use are taken from the Behavioral Risk Factor Surveillance System (BRFSS) from 1993 to 2020: there are 9,089,155 observations. Information in the survey is collected monthly, by telephone, with a standardized questionnaire for the Center for Disease Control (CDC). The Behavioral Risk Factor Surveillance System (BRFSS) is the nation’s premier system of health-related telephone surveys that collect state data about US residents regarding their health-related risk behaviors, chronic health conditions and use of preventive services. Established in 1984 with 15 states, BRFSS now collects data in all 50 states as well as the District of Columbia and three US territories. BRFSS completes more than 400,000 adult interviews each year, making it the largest continuously conducted health survey system in the world.Footnote 8 The CDC regularly publishes nationally representative estimates for the fifty states in its Morbidity and Mortality Weekly reports, for example see Gamble et al (2017)

In the BRFSS sample, we have 136,125 Native American observations on despair once missing values are omitted, which is 1.02% of the sample of 9,089,155. This is a higher proportion than suggested by Census Bureau estimates, including children. Based upon the 2019 American Community Survey, the Census Bureau estimates that there are 2,847,336 American Indians and Alaskan Natives out of a population including children, of 328,239,523 or 0.87%.Footnote 9 The percentage of Native respondents in the BRFSS stays relatively constant over time: 0.86% in 1993–1997, 0.99% in 1998–2002, 1.04% in 2003–2007, 1.05% in 2008–2012, 0.98% in 2013–2017 and 1.2% 2018–2021 with potentially a slight rise. This is potentially relevant since the number of people identifying as Native American in surveys such as the Census has been increasing over time.Footnote 10 We discuss this issue in more detail when discussing our results.

The BRFSS question we use is the following: “Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” We code people who give the answer “30 days out of 30” as being in extreme distress. We make use of data files from 1993 because that is the year a key question on “bad mental-health days” was first asked. Sample sizes increased from around 100,000 annually in 1993 to 450,000 by the end of the years. We also make use of data on various well-being measures from the Gallup US Daily Tracker polls of 2017 and 2018.

Table 1 reports the weighted distribution of responses to the question by five non-Hispanic groups—White, Black, Asian/Pacific Islanders, Native Americans and those identified in the BRFSS as “other race,” plus Hispanics. We present the data for the full sample of respondents for all years of data, from 1993 to 2021. The percentages presented in the first row show two-thirds of respondents answer that zero of the past thirty days have they experienced “not good” mental health. The percentage is below sixty for Native Americans and other races. However, it is also clear from the second to last row of Table 1, that the percentage of people responding that their mental health was “not good” all thirty of the previous thirty days is especially high among Natives.

Table 1 Distribution of bad mental health days, 1993–2021
Full size table

Table 2 reports the rate of chronic distress of the weighted samples by racial group. It averages 5.1% over the period 1993–2020 and is higher among Native people overall, as well as for those with no or some college education with the distress markedly higher in the former case (10.4%) than in the latter (7.8%). The same is true for the prime age (35–54) and especially so for the prime age without any college education. As we will show below the gap between natives and other groups has increased over time for the less educated.

Table 2 Distribution of chronic distress by race, 1993–2021
Full size table

Table 3 shows changes in extreme distress over time. It is notable that for Native Americans the percentage reaches double figures after the Great Recession and stays there subsequently. There are increases over time for all groups, with the smallest increase for Hispanics.

Table 3 Time-series changes in despair by race, 1993–2021
Full size table

3 Econometric results

The paper’s principal analytical findings are reported in time-series plots in Figs. 1–3 and in regression equations in Tables 4–9 using the BRFSS 1993–2021. We add to that supporting evidence from Gallup’s US Daily Tracker polls of 2017 and 2018 in Tables 9–11. Our principal finding is that Natives have higher levels of chronic distress than whites, blacks, Hispanics and Asians controlling for a host of variables as well as cohort and income. We also find they have lower levels of general health as well as higher numbers of bad physical health days.

Fig. 1
figure 1

The incidence of chronic distress by race, 1993–2021

Full size image
Table 4 Chronic distress equations, 1993–2021
Full size table
Table 5 Chronic distress equations with cohort effects, 1993–2021
Full size table
Table 6 Chronic distress equations with adjustment for household income, 1993–2021
Full size table
Table 7 Chronic distress equations by age, 1993–2021
Full size table
Table 8 Chronic distress equations, 1993–2021, for Natives
Full size table
Table 9 OLS general health and number of bad physical health days, BRFSS, 1993–2021
Full size table

Figure 1 uses the full-sample data from 1993 to 2021. The graph separates citizens into five non-Hispanic racial groups—Whites, Blacks, Asians and Pacific Islanders, Native Americans, those identified as “other race,” plus Hispanics. Perhaps the most striking feature of Fig. 1 is that chronic distress is far greater among Native peoples and other races, and the gap has risen over time. By 2019, one in eight Native American citizens said that every day of the last month was a bad day for mental health. The proportion was far above the incidence observed among White or Black. The small differences in distress between White and Black Americans are consistent with the existing literature on mental health and stress in these populations: Despite the socioeconomic and historical differences between the two groups, Black Americans have relatively comparable or better good mental health than White Americans (Neighbors et al. 1983; Riolo et al. 2005; Keyes 2009, Erving et al. 2019). As we note below once controls are included for education or labor market status Black have significantly lower levels of persistently poor mental health than White.

Figure 2 depicts equivalent data for subsample of individuals much discussed in the literature on despair, namely those without a college education. The graph reveals a broadly similar pattern to that of Fig. 1. Once again, chronic distress is markedly higher for Native Americans, and there is a secular trend upwards for Black, White, Native and other race Americans but the series is flat for Hispanic and Asian Americans. A particularly sharp rate of increase is visible among less-educated Native Americans and other races. By the year 2019, Fig. 2 shows that approximately one in eight exhibited persistently poor mental health. Figure 3 turns to US citizens with college education, and again, the incidence of persistently poor mental health is markedly higher among Native peoples and other races who have seen a rise in the incidence of extreme distress. The rise is much less marked for the other groups.

Fig. 2
figure 2

The incidence of chronic distress by race—non college, 1993–2021

Full size image
Fig. 3
figure 3

The incidence of chronic distress by race, some college, 1993–2021

Full size image

Figures 1–3 demonstrate a consistently high large difference between the proportion of Native people experiencing chronic distress and a general rise in over time.Footnote 11 Given the shifts in who identifies as Native American over time,Footnote 12 how to interpret the rise in persistently poor mental health is an open question: is the rise in distress as due to an increase among people who would have always identified Native or due to people who have more recently connect with their Native ancestry being more likely to have chronically poor mental health? We believe that the former interpretation is more likely than the later for at least two reasons. First, we focus on Native Americans who report only a single race and those that report a different race between censuses are much more likely to have multi-racial identities so are more likely to end up in the other race category in the BBRFS then the sole Native American racial identity category (Liebler et al. 2016). Second, proportionately, highly educated adults are more likely to report being Native American when they had not done so previously (Liebler and Ortyl 2014), and much of the growth in the Native population in the past decade has been among older cohorts (see Appendix Table 13 and Liebler and Ortyl (2014)).

If, on average, individuals who experience chronic distress than the Native population in the 1990s begin to enter the Native population category in survey data, trends in distress will potentially underestimate any rise in chronic distress among Native people who would always identify as Native. Since both the more highly educated and those over the age of 64 experience less chronic distress on average, we would expect changes in identification would mitigate the rise observed among Native Americans.

From Fig. 3, we can see that those with at least some college have a slightly flatter trend among Native people than those with less education, possibly suggesting the upward trends are muted if there is greater ethnic mobility among the more highly educated. However, regardless of how the trends are interpreted, those who historically and currently identify as Native American have greater levels of average chronic distress than White Americans.

The time-series graphs in Figs. 1–3 simply plot the raw data through time. Thus, they do not correct for the potentially different characteristics of different groups of individuals (as shown in Tables 4–8). A first form of such adjustment is provided in Table 4. Here we present results of estimating pooled cross-section time-series regression equations with the dependent variable being a dummy variable for whether a respondent reported thirty of the last thirty days were bad mental health days. Sample size is 8.8 million. Each specification includes dummy variables for Native, Black, Asians/Pacific Islander and Hispanic people as well as for those classified as “other race” with the base category being White. A quadratic in age is included, plus a dummy variable for gender and a dummy variable for whether the survey was taken using a cellphone.Footnote 13 We present the coefficients on each of these variables with and without controlling personal circumstance and characteristic variables.

In the first column, the coefficient of the Native and “other race” variables is around 0.037 in both cases, suggesting that Native peoples and those who race is not captured in the main race categories of the BRFSS have a 3.7 percentage point higher rate of chronic distress conditional on age, gender and method the survey was taken. The second column adjusts for a set of controls for labor-market status, level of education, marital status and state and year dummy variables. Because of the large sample sizes in Table 4, the coefficients are estimated precisely with t-statistics for the overall sample that typically exceed 50. In the second column of Table 4, the highly statistically significant coefficient on Native Americans is 0.0096 with a t-statistic of more than 15. As the base category is Whites, this implies that after adjustment for the independent variables the level of extreme distress is approximately 1 percentage point higher among Native Americans than among White Americans. The coefficients on the Black and Hispanic variables both switch sign to significantly negative in column 2 from positive in column 1: Asians are negative in both columns. In columns three and four, the equations are re-estimated for the earlier period 1993–2000 and then for 2001–2021, and both the Native and other race coefficients rise slightly.

An important implication emerges from Table 4. In Table 1 and Fig. 1, it was seen that the raw gap between Native Americans and Whites in the incidence of chronically poor mental well-being was slightly over 4 percentage points (that is, comparing 9.3% with 4.9%). Yet in the adjusted estimates of Table 4 it is only 1 percentage point. This implies that the majority of the raw difference is accounted for by the different underlying circumstances of Native respondents when compared to White respondents. If Native Americans had the same economic and social circumstances as White Americans, the solid line in Fig. 1 would in principle lie only slightly above the two dotted lines for other American citizens.

We have also varied our definition of chronic distress as a robustness exercise. Below we report the coefficients and t-values on the Native American variable using the specification in column 2 of Table 4 with personal controls. We vary the number of days so that half of the days are counted as not good mental health days. We then do so another four times as a smaller proportion of the 30 days are included as despair. The results tightly cluster around a coefficient of 0.1 so it appears that our results are largely insensitive to changes in definition.Footnote 14

Table 5 repeats the same econometric exercise but adds decade-of-birth cohort dummy variables with the full sample. It does this to try to probe whether successive generations of Native people are enduring lower or higher rates of chronic distress (with no personal controls in column 1 and with them in column 2). The results are largely unchanged, which suggests that any general cohort effects in mental well-being do not explain why Native peoples have higher rates of chronic distress on average. Table 6 does a further check on a sample from 1993 to 2021 by adding income controls. Column 1 presents the results with and without personal controls. Again, Native Americans have significantly higher rates of chronic distress than all other groups. In both of these tables, there is a hump-shaped age pattern (plotted in Appendix, Fig. 4), which also holds within each race, consistent with work on representative samples of people in many nations’ data, such as recently in Blanchflower (2020, 2021).Footnote 15

Table 7 produces separate estimates for four age groups. There are large positive and significant coefficients for the three age groups: 35–54, 55–69 and 70 + but not for the youngest age-group. For those under thirty, the coefficient is significantly lower than for whites although the coefficients on the Black, Hispanic and Asian race variables are all significantly smaller, that is more negative, than for Natives. The other race variable though is significantly positive as it is for all four age groups and for all four age groups is significantly higher and hence have higher levels of distress than for Natives.Footnote 16

Modern research on despair (captured partially by what we measure here), shaped originally by Case and Deaton’s work, has been especially interested in midlife individuals with low levels of education (Case and Deaton 2015, 2017).Footnote 17 A commonly discussed hypothesis is that poor employment chances have been an important driver of this growth of chronically poor mental health in the USA. Is that true of Native Americans?

To investigate this, Table 8 moves to a fuller set of regression equations for Native Americans only who constitute 1% of our weighted sample with approximately 130,000 observations. We present four sets of results restricting the sample to be just Native Americans. The first column is for all Native people; then, as we move to the right in column 2 we restrict the sample to high school dropouts, high school graduates and then those with some college. In each column of Table 8, large effect sizes are visible from being out of work, from being unable to work and from a person’s level of education. Other background covariates of evident importance in the equations of Table 8 are: being a student, being retired, being divorced or separated, being widowed and being female. All are associated with a higher probability of chronic distress. There is again an inverted U-shape in age maximizing at age 40 consistent with the findings of Blanchflower (2021b).Footnote 18

Finally, extreme distress among less-educated Native peoples is lowest in states that have the highest proportion of Natives Americans. We examined which states had the highest percentage of Native Americans in our data file. The seven states that had the greatest proportion of Native peoples among their population, pooled across the years, were: Alaska 13.0% (with chronic distress 4.4%); Arizona 2.8% (7.4%); Montana 4.5% (7.0%); New Mexico 7.4 (6.9%); North Dakota 3.6% (7.1%); South Dakota 5.4% (6.7%); and Oklahoma 6.9% (8.8%). Here the numbers in parentheses give the proportions of people who reported being in extreme distress among the least educated, with an overall mean of 10.4%. Of particular note is the fact that living in one of these states seems to be the most protective for those who are midlife high school dropouts. For example, while living in Alaska is associated a roughly 6 percentage points lower rate of chronic distress for Native Americans on average relative to the excluded state of Alabama (as seen in column 1), living in Alaska is associated with an 8 percentage point lower rate of chronic distress among other high school dropouts relative to lower proportion population states. This pattern, which is suggestive, is consistent with a feeling of belonging or importance of being culturally connected to a community being psychologically protective.Footnote 19Footnote 20

Table 9 reports equations using two other variables from BRFSS: Q2–Q3, relating to general health and the number of days in the past thirty when their physical health was not good. The questions used were as follows:

  • Q2. Would you say that in general your health is — excellent (5; very good (4); good (3); fair (2) or poor (1)? With our coding in parentheses.

  • Q3. Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?

The results are similar to those using chronic distress. Natives have poorer general health and have a higher number of bad physical health days, plus are more likely to say that every day is a bad physical health day than whites, Blacks, Hispanics and Asians, and other groups. Separate results are provided with limited controls and then adding personal controls for education, labor force and marital status. This is consistent with results reported in Appendix Table 14 based on data from the National Health Interview Survey of 2018, which shows that a much lower proportion of American Indians report excellent health than any other group. All our results to this point indicate low well-being, among a greater proportion of Native Americans than for other populations in the USA.

As a further robustness check, therefore, we turn to using Gallup data for 2017 and 2018 from the US Daily Tracker Survey. Here the overall sample size, in Table 10 using 2017 data, is approximately 150,000 individuals. Because the survey questions are different from those in BRFSS, the table is not able to estimate exactly the same form of our chronically poor mental health models as in earlier tables. Its value is that we show similar results with a variety of other well-being measures. Instead of examining extreme distress, Table 10 reports regression equations for four other negative affect dependent variables—pain, stress, worry and depression—along with a positive affect variable on enjoyment. These are (1,0) dependent variables relating to “yesterday.” The question wordings are: Q5. Did you experience physical pain/worry/enjoyment/stress yesterday? Yes/No; Q6. Depression – do you currently have, or are you currently being treated for depression? Yes/No?

Table 10 OLS well-being models, US Gallup Daily Tracker, 2017
Full size table

We also consider another positive affect variable, a life satisfaction question on a scale of 0–10, the Cantril ladder with the question wording as follows: Q7. Please imagine a ladder with steps numbered from zero at the bottom to ten at the top. The top of the ladder represents the best possible life for you and the bottom of the ladder represents the worst possible life for you. On which step of the ladder would you say you personally feel you stand at this time?

The substantive implications of Table 10 are consistent with the paper’s earlier patterns. The proportion of people with chronic distress are found to be far greater in general among Native American citizens than other individuals in the USA. In each case, Table 10 reveals that, after controlling for a large set of personal variables, the poor mental health proxy coefficients for American Indians are large and significantly higher than whites. Except for enjoyment, the other racial and ethnic groups have higher levels of mental health than whites.

Some other results on Gallup data are given in Table 11 using 100,000 observations from the 2018 US Gallup Tracker Poll. This is based on question Q5 above relating to pain, stress and worry. Plus, a new question on worry over money scored from one to five: Q8. In the last seven days you have worried about money – strongly disagree (1) to strongly agree disagree (5).

Table 11 Further evidence: OLS Equations, US Gallup Daily Tracker, 2018, age < 70
Full size table

The results in Table 11 suggest once again that Native Americans have significantly higher rates of pain, general worry, worry about money and stress than White Americans even conditional on age, education, labor force and marital status.

Table 12 uses additional variables from the 2018 Gallup Tracker data with 110,000 observations. We specifically report results for the Cantril ladder variable from above and three additional dependent variables. These variables are created using the following questions: Q10. You always feel safe and secure – strongly disagree (1) to strongly agree disagree (5) (variable hwb23); Q11. You like what you do every day – strongly disagree (1) to strongly agree disagree (5) (variable hwb1) and; Q12. Are you are satisfied with your standard of living – agree (1) disagree (0)?

Table 12 OLS positive affect well-being equations, US Gallup Daily Tracker, 2018
Full size table

The results in Table 12 demonstrate that Native peoples are, as in the 2017 survey, less likely to rate themselves highly on the Cantril ladder, significantly less likely to feel safe and secure, less likely to report they agree with the statement they like what they do every day and less likely to be satisfied with their standard of living than White Americans, even conditional on age, gender, state of residence, education, labor force status and marital status.

Even though a large number of independent variables are held constant in the empirical analysis and various statistical checks are performed, this paper uses observational data. It finds clear patterns, but that does not necessarily identify deep causal mechanisms behind the higher distress among Native peoples. Economic and social circumstances seem, as a matter of statistical association, to explain much of the observed statistical differences by group, but not all of those differences. We view the paper as a contribution toward the longer-run ideal of establishing the true driving forces of distress and perhaps eventually aiding an understanding of the structural changes potentially required in order to address the observed differences in extreme-distress levels within the modern USA.

4 Conclusions

This paper is the first to compare the much-discussed worsening mental well-being of White American citizens with what was experienced by Native American citizens over the same period. The paper draws upon BRFSS data to calculate a measure of chronic distress, from 1993 to 2020, among a randomly selected group of 8 million US citizens. Complementary data from recent Gallup surveys are used.

The main results are as follows.

  1. (i)

    In every year since 1993, the level of chronic distress among Native Americans was substantially higher than among White or Black Americans.Footnote 21 This was apparent for Native Americans ages 30 and over but not for the young.

  2. (ii)

    Chronic distress was also relatively high among those who identified themselves as having other racial backgrounds. This is important given the rise over time in the proportion of Americans who report they are Native American combined with some other race.

  3. (iii)

    The chronic distress gap (illustrated graphically in Fig. 1 in the paper) between the groups widened over time. Levels of persistently poor, and chronic, mental health increased fastest among Native peoples.

  4. (iv)

    Chronic distress is lower among Native American people living in states with the highest proportion of Native Americans generally. We see this as consistent with the prior literature that suggests the cultural connectedness and community are important for psychological well-being.

  5. (v)

    Modern data from Gallup paint an equivalent statistical picture, using other kinds of mental ill-being indicators, including information on feelings of pain and worry.

There are limits to what we can learn from the BRFSS and Gallup Survey’s about the diverse experiences of Native peoples and Tribal Nations. Nevertheless, we believe this paper is a step forward in the economics literature. The broader American public and many academics are often unaware of the experiences of Native peoples. Native American indicators are often excluded from standard data sets or are merely categorized as “other race” with the excuse that sample sizes are too small. However, there are millions of Native Americans and data sets exist, both public and confidential, that can be used by researchers to contribute to sharing the experiences of Native peoples.