A growing socioeconomic divide: Effects of the Great Recession on perceived economic distress in the United States

Abstract

We demonstrate widening socioeconomic disparities in perceived economic distress among Americans, characterized by increasing distress at the bottom and improved perceptions at the top of the socioeconomic ladder. We then assess the extent to which hardships related to the Great Recession account for the growing social disparity in economic distress. Based on the concept of loss aversion, we also test whether the psychological pain associated with a financial loss is greater than the perceived benefit of an equivalent gain. Analyses are based on longitudinal survey data from the Midlife Development in the US study. Results suggest that widening social disparities in perceived economic distress between the mid-2000s and mid-2010s are explained in part by differential exposure to hardships related to the Great Recession, the effects of which have lingered even four to five years after the recession officially ended. Yet, auxiliary analyses show that the socioeconomic disparities in economic distress widened by nearly as much (if not more) during the period from 1995-96 to 2004-05 as they did during the period in which the recession occurred, which suggests that the factors driving these trends may have already been in motion prior to the recession. Consistent with the loss aversion hypothesis, perceptions of financial strain appear to be somewhat more strongly affected by losses in income/assets than by gains, but the magnitude of the differentials are small and the results are not robust. Our findings paint a dismal portrait of a growing socioeconomic divide in economic distress throughout the period from the mid-1990s to the mid-2010s, although we cannot say whether these trends afflict all regions of the US equally. Spatial analysis of aggregate-level mortality and objective economic indicators could provide indirect evidence, but ultimately economic "despair" must be measured subjectively by asking people how they perceive their financial situations.

Fig 1. Predicted change in current financial strain between M2 (2004–05) and M3 (2013–14) by relative SES.

Demographic-adjusted scores are based on Model 2 (Table 3), while the remaining scores are based on Model 4. Predicted scores are computed using the coefficients from the linear regression model and setting the covariates to the following values: financial strain at M2 (2004–05) is fixed at the mean; relative SES at 0 (bottom 1%) or 1 (top 1%); indicators for exposure to Great Recession-related hardships as specified; and all other covariates (sex, age, minority, marital status) are fixed at the mean for our sample. Thus, scores represent the estimated change in current financial strain between M2 and M3. Error bars indicate the 95% confidence interval for each estimate.

Fig 2. Predicted change in current financial strain between M2 (2004–05) and M3 (2013–14) by changes in assets.

Predicted scores are computed using the coefficients from Model 5 (Table 3) and fixing financial strain at M2 (2004–05) at the mean; the level of assets at M2 at the median ($120,000 based on absolute values); the change in assets between M2 and M3 at the specified values, and all other covariates (sex, age, minority, marital status, and household income at M2) are based on the observed distribution in the sample. Thus, scores represent the estimated change in current financial strain between M2 and M3. The difference (M3—M2) in log assets is equivalent to the log of the ratio of assets at M3 relative to M2. Thus, a 0.7 decrease in log assets represents a loss of 50% (24% of the observed sample exhibited a loss of that magnitude or greater), whereas a 0.7 increase corresponds with an approximate doubling of assets (28% of the sample reported a gain of that magnitude or greater). A decrease of 5 in log assets would be equivalent to a decline from $40,000 at M2 to $270 at M3 (10% of the sample reported a loss of similar or greater magnitude, 99% of whom ended up with no assets at M3), while an increase in log assets of 5 would be equivalent to a gain from $500 at M2 to $74,000 at M3 (9% of the sample exhibited a gain of similar or greater magnitude, 97% of whom had no assets at M2).