Reader UserFriendly highlighted an important St. Louis Fed study, The Unequal Recovery: Measuring Financial Distress by ZIP Code by Ryan Mather and Juan M. Sánchez. It sheds light on a topic that that readers regularly debate: why are there so many signs of distress in a supposedly robust economy?
Some of the disconnect is due to rentier choke points, like rising housing costs, particularly in big cities, leading to more and more “affordable” housing being converted by gentrification or other means into upscale abodes, and ever-escalating medical costs, which creates medical bankruptcies, worry about seeking care, and too many people not getting treatments early. Another factor is misleading headline unemployment reports. Even though the press regularly brays about the strength of the job market, the high level of involuntary unemployment belies that.
The St. Louis Fed teases out another way in which rosy aggregate data masks shaky foundations. Since 2015, the lowest income households have become more vulnerable to shocks, taking on more debt as wealth (to the extent they have it) is even more concentrated in housing. The level of distress in lower-income households has also increase, defying the official story of increasing prosperity.
The authors drill into ZIP-code level data to show that even adjacent ZIPs have shown striking divergences in wealth accumulation (or its erosion) in the last few years. You can quibble with their methods (there isn’t good data on wealth) but if anything, they are almost certain to have understated how well the well off are doing, particularly given Gabriel Zucman’s findings on the extent of hidden wealth globally (6% to 8% of total wealth). And their measures of distress seems reasonable, for instance, using 80% or more of credit card limits.
The authors point out that before the crisis, analysts missed that a supposedly sound economy featured rising levels of household debt, too often funded by extraction of housing equity to preserve spending levels or at best pay down credit card debt. Falling housing prices exposed how fragile many borrowers were, taking the economy down with them. As the report says in its opening section:
Many narratives have been told for exactly why that recession was as bad as it was, but a common plot element is that declining house prices forced highly leveraged households to reduce consumption drastically. For example, economists Atif Mian, Kamalesh Rao and Amir Sufi estimated that for every dollar of housing wealth lost, households’ consumption decreased by 5 to 7 cents. While that may not seem like much out of any given dollar, the effect quickly becomes massive when added up across all home value losses suffered by all households.
I urge you to read this very accessible study in full. Late in the report, the authors set forth how easily high level data can mislead:
Given that there has been a wide dispersion in measures of wealth growth across ZIP codes since 2010, it seems fair to reconsider what the current distribution of households’ financial conditions means for financial stability. If it is the case that growth has been concentrated in the hands of wealthy ZIP codes with low leverage, then the poor and high-leverage ZIP codes that are more affected by wealth shocks may still be vulnerable. What’s more, trends in less affluent groups are masked in nationally aggregated statistics by groups with more wealth.
Imagine an economy with two people, one of whom has $1 of wealth and the other $99. Imagine further that the poorer individual’s wealth drops to nothing the next year, while the other’s remains unchanged. A nationally aggregated statistic will observe $100 of wealth in the first period and $99 in the next, which represents a 1 percent decrease in net wealth. The poorer individual, however, experienced a life-changing 100 percent decrease. Given how the top 1 percent in our country has around 40 percent of all wealth, this contrived example is not entirely unlike the real world. Life-changing shocks to net wealth at the lowest percentiles may be entirely invisible under near trivial changes at the highest percentiles.
Oddly, economists seem to regard it as news that people who are desperate find it hard to pull themselves up by their bootstraps:
Recent research by Fed economists Kartik Athreya, José Mustre-del-Río and Juan Sánchez suggests that for individual borrowers, financial distress is not a transitory phenomenon but rather a highly persistent one. To put it differently, while most people never have credit card payments over 120 days delinquent, they found that among those who at some point do, more than 30 percent spend at least a quarter of their time that way.
A high level explanation of the study’s approach:
The methodology—which is similar to that in Mian, Rao and Sufi—creates a data set of household balance sheets at the ZIP code level and examines whether the change since the beginning of the economic recovery in 2010 has been as positive as it seems at the aggregate level. ZIP codes, being nothing more than a collection of individuals within certain geographical boundaries, are thus used to represent individuals with certain characteristics.
And key findings:
In Table 1, the economic recovery since 2010 is divided into two periods…. the first, lasting until 2015, the Federal Reserve pushed interest rates near zero to stimulate the economy. Then, beginning in December 2015, the Federal Reserve has been lifting interest rates…..
While both periods saw similarly robust growth in terms of net wealth (7.4 percent for 2010-2015 and 6.2 percent for 2015-2018), the composition of that growth is quite different. From 2010 to 2015, financial wealth was the strongest component of growth (6.9 percent), and debt accumulation was very low (0.5 percent).
Beginning in 2015, however, U.S. housing wealth posted the largest gains (6.1 percent) and brought with it faster debt accumulation as well (2.7 percent). Should house prices drop again, households may find themselves more highly leveraged and vulnerable than they were at the beginning of 2015.
The rest of the table shows the dispersion of these growth rates across ZIP codes in our sample, ranked from lowest to highest in each category. Over the years 2010-2015, for example, ZIP codes at the 90th percentile in terms of debt accumulation saw their debt grow by 6.2 percent annually, well above the national average of 0.5 percent annually. The dispersion is even wider from 2015 to 2018.
So the Fed’s super low interest rates initially, in aggregate, did more to goose financial asset prices than help that bulwark of consumer wealth, the housing market recover.
However, one effect of super-low interest rates was that homeowners with mortgages who had reasonably good credit could refi at lower rates. A dirty secret is that the benefit of refis goes significantly to financial intermediaries. However, for at least some borrowers, the effect of lowered payment would be increased savings…which would contribute to the increase in financial assets documented in this study.
A geographical look is also sobering:
The report elaborates:
The national trend is immediately apparent in both maps: While financial distress along our measure improved across most of the country from 2010 to 2015, it deteriorated with similar yearly magnitude from 2015 to 2018.
At the same time, it is equally apparent that this national trend masks a large amount of variation within states and even within counties. To give some perspective on these numbers, the national weighted average of borrowers reaching at least 80 percent of their credit limit in our sample was 16.5 percent, 14.7 percent and 16.1 percent in the years 2010, 2015 and 2018, respectively. Those ZIP codes in the deepest shade of red, then, were deteriorating each year by around an eighth or more of the national average. Compare that to the darkest shade of blue, which marks ZIP codes that were improving each year by around an eighth or more of the national average.
That so many areas show these two extremes directly adjacent to one another points to how conditions of financial distress can diverge rapidly across neighborhoods. This effect is particularly pronounced in major population centers where ZIP codes parcel out smaller areas of land, such that they are impossible to distinguish in the national graphs of Figures 1 and 2.
Consider, for example, Hennepin County in Minnesota and, within that, the city of Eden Prairie, which is composed of three mutually adjacent ZIP codes: 55344 to the east, 55346 to the west, and 55347 to the south. The eastern ZIP code experienced almost no change in net wealth from 2010 until 2015 but a slight increase in financial distress, while the western and southern ZIP codes experienced sizable increases in net wealth and slight decreases in financial distress over the same period.
After these changes, in 2015, the share of residents in all three ZIP codes that had used at least 80 percent of their credit limit was nearly identical: about 10.6 percent. During the period from 2015 until 2018, however, the eastern and southern ZIP codes each experienced increases in financial distress of about 6 percentage points, putting them near the national mean of 16.1 percent in 2018. By contrast, financial distress in the western ZIP code remained nearly unchanged over the same time period.
Clearly, the recovery experiences of these three ZIP codes were very different, even though all of them are in the same city; they share the same community center, send their children to the same public high school, and have but one McDonald’s restaurant.
I’m not certain that income and wealth disparities in a community are that unusual; think of the tony versus the bad areas near where you live. But the fact that the differences are now extreme enough to show up in ZIP code level cuts seems significant.
And more generally, this study confirms what many readers see or sense. There are more left out of this supposedly robust economy than the pols and most economists want to believe. This study shows that averages can conceal a lot of pain. But the fact that this isn’t widely enough examined, much the less publicized, mirrors the cloistered view of the top 10% and their allies in the press.