The Big Idea
Liquidity across the income scale
Stephen Stanley | August 18, 2023
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors.
One of the lodestars of my economic outlook during pandemic and after has been the extent to which a massive run-up in household liquid assets has boosted consumer spending. This remains an important factor in the ongoing robust performance of consumers, which has consistently exceeded consensus expectations by a wide margin. The Federal Reserve publishes the household numbers with a lag by income quintile, offering a detailed look at finances across the income scale. The latest numbers reveal why families at the lower end of the scale have begun to feel stress, as inflation has eroded the purchasing power of their savings. But there are reasons to be optimistic that household payment performance will remain favorable by historical standards.
Federal Reserve balance sheet data
Most economists continue to base the bulk of their analysis of household finances on complicated model-driven calculations derived from the personal savings rate. This is a needlessly complicated approach, as the Federal Reserve publishes detailed data on household balance sheets quarterly, a dataset that I have highlighted repeatedly in recent years. A month or two after the Fed releases aggregate numbers each quarter, it publishes more comprehensive data that breaks down the results in several different ways, including by income. These data are called the “Distributional Financial Accounts,” and they offer an interesting view of household finances across the income spectrum.
The Fed divides households into six groups by income:
- The top 1%
- 80% to 99%
- 60% to 80%
- 40% to 60%
- 20% to 40%, and
- 0% to 20%
This piece examines the household liquid assets series—bank deposits plus money market fund balances—as a conservative proxy for households’ spending power for each of these groups.
Liquid assets by income quintiles
The following six charts show the household liquid assets for each of the groups noted above.
Exhibit 1: Liquid assets – Top 1% by income
Exhibit 2: Liquid assets – 80% to 99% income
Exhibit 3: Liquid assets – 60% to 80% income
Exhibit 4: Liquid Assets – 40% to 60% Income
Exhibit 5: Liquid Assets – 20% to 40% income
Exhibit 6: Liquid Assets – 0% to 20% Income
These pictures reveal a few key observations. First, the predominant narrative regarding household balance sheets, derived from analysis of the savings rate, is wrong. Households have not been furiously spending down their savings to maintain unsustainable consumption patterns. For most quintiles, the level of liquid assets has barely come off from its highs. In fact, for the 20% to 40% quintile, the latest reading, for the first quarter of this year, set a new high. The lone exception is the middle quintile, but even for this group, the series jumped by $340 billion from the end of 2019 through the first quarter of last year and has only retraced about $80 billion of that over the next four quarters, about a quarter of the Covid windfall.
However, while the nominal level of household liquid assets has remained elevated, inflation has eaten away at families’ purchasing power, especially for the bottom 40% of the income scale. Liquid assets after stripping out the impact of inflation are lower than nominal liquid assets (Exhibit 7). The bottom 40% of households are basically back to their pre-Covid levels of liquid assets in real terms.
Exhibit 7: Nominal and real liquid assets increases by income cohort
This likely helps to explain the rebound in delinquency rates on various loan types in recent quarters. It should be noted that delinquency rates, at least so far, have merely returned to 2019 levels, which are historically low.
It is worth noting that to the extent that lower-income households have seen their Covid savings depleted, they are not walking away empty-handed. The bottom and second income quintile have seen the value of their real estate holdings rise by 32% and 27%, respectively, compared to 36% for all households. And the value of their consumer durables increase by 42% and 39%, respectively, compared to 35% for all households. Of course, households only need so many couches and TVs, so the softness in these categories of retail sales in 2023 may persist for a while longer, as the flurry of spending on goods during 2020 and 2021 may tide households over for the time being.
There is one bit of good news that suggests that the end of the Covid windfall in real terms for lower-income households does not herald a consumer recession. A highly resilient labor market has generated strong job gains and robust income growth. Over the past 12 months, real disposable income—after inflation and taxes—has risen by 4.7%, roughly double the corresponding rise in real consumer spending. It seems doubtful that consumers and lower-income households in particular will falter in large numbers unless or until the labor market weakens substantially. The consensus has been predicting an imminent surge in unemployment for the better part of the year, but the unemployment rate has been relatively steady, within a tenth or so of 3.5%, for over a year. I look for the jobless rate to remain in that range through the rest of 2023, but I do expect to see at least a modest backup in 2024. Only then would I expect to see the consumer come under more stress.
A second source of data
Behind the Fed data, my second favorite go-to source of information on the details of the consumer is the Bank of America Institute. Researchers there leverage their access to Bank of America’s massive reach among households to offer unique insights on the state of household finances. The Institute typically publishes a monthly update in advance of the Census Bureau’s retail sales release.
The BoA Institute’s recent reports offer several additional observations on the state of household finances and the consumer.
First, the Bank of America data offer a notably brighter picture than the Fed figures on household liquid assets. For the subset of BoA customers who have held a consumer deposit account (checking and/or savings) constantly since the beginning of 2019, median deposit balances are up sharply across every income cohort. In fact, the income group with the smallest gain since 2019 is the highest cohort ($250,000 and up). Their median balances are up about 30% from 2019, while the corresponding figures for lower cohorts are in the range of 40% to 60%. These data extend through July and are therefore considerably more timely than the quarterly Fed numbers. I would tend to rely more heavily on the Fed figures, but these data offer another set of hard data that the prevailing narrative of households exhausting their Covid-era savings, derived from models, is wrong.
The Bank of America Institute offers additional bits of good news for lower-income households. One frequently cited problem looming is the resumption of federal student loan payments in October. This will undoubtedly be a drain on many households’ finances. However, the actual situation is different than the popular conception. For every young person who took out large loans to get an impractical degree and now does not have the income to repay the debt, there are multiple people who borrowed heavily to get a medical or law degree. Based on Federal Reserve data, the BoA Institute estimates that over half of all student debt is held by the two upper income quintiles. They also cite a breakdown from Brookings showing that over half of all student debt is owed by households where the highest education level is a Master’s or Doctoral degree.
Among their customers, Bank of America researchers found that households earning $100,000 and up who stopped making monthly student loan payments in 2020 had noticeably higher median deposit balances than those with similar incomes who did not. There will, of course, be many individual cases where the resumption of student loan obligations causes a cutback in consumer spending, but it appears that a number of affected households have prepared for this day and will not necessarily have to change their outlay patterns by much.
Finally, it appears that the labor market is currently favoring those with modest incomes. Institute researchers found that the dollar amount of unemployment checks being deposited into BoA customer accounts were rising fastest for (previously) high-income households ($125,000 and up). This is consistent with the anecdotal observation that job cutbacks over the past year or so have been coming mainly in white-collar and upper-level positions. Similarly, the year-over-year growth in paycheck deposits into BoA accounts has been running considerably weaker for high-income households than for lower-income households.
The typical dynamic when an economy is sliding into recession is that lower-income households suffer first and worst. In contrast, in the current situation, at a time when growth is proving more robust than expected, the BoA data suggest that lower-income households are faring reasonably well, if anything better on a relative basis than wealthier families.
All in all, the Fed and BoA data offer a mixed picture for lower-income households. The Covid windfall may be gone on an inflation-adjusted basis, but a robust labor market appears to be disproportionately benefitting households with modest and middle class incomes. While the unprecedented credit performance on various types of consumer debt during the pandemic appears to be behind us, a buoyant labor market is, for now, helping to offset the punishing effects of inflation and make keep debt charge-offs from rising significantly beyond 2019 levels.