Source: US State of New York Federal Reserve
Thank you, Professor Ho for that kind introduction and for the opportunity to talk to the Vassar community.1 I am happy to be back on campus. As a teenager in Washington, D.C., I had the very good fortune that a high school counselor pushed me to apply to Vassar College. I was accepted, and I earned my bachelor’s degree here. Attending Vassar opened a wider variety of opportunities to me than I would have otherwise had available. But I encountered one problem: Vassar did not offer any banking or business courses, which is what I wanted to study. So, I enrolled in an economics class, figuring it was the next best thing. I was hooked, and I have been studying economics ever since.
My time here as a student was transformative, and I was honored to have served on Vassar’s board from 2002 to 2022. Vassar is a vibrant intellectual community.
To motivate the topic of today’s speech, let me begin by sharing with you briefly my assessment of the current state of the U.S. economy. The performance of the U.S. economy has been quite strong overall.2 Last year, gross domestic product grew at a solid pace of 2.5 percent. I see the labor market as being in a solid position, with job creation steady and the unemployment rate at 4 percent in January. Inflation has come down a great deal over the past two and a half years but remains somewhat elevated relative to our 2 percent target. Based on recently released data, it is estimated that the 12-month change in the personal consumption expenditures price index was 2.4 percent in January. Progress toward our 2 percent objective has been slow in the past year. I expect the path of inflation to continue to be bumpy. While a cumulative cut in the policy rate by 100 basis points last year has brought the stance of monetary policy closer to a neutral setting, monetary policy continues to be restrictive. I believe that, with a strong economy and a solid labor market, we can take our time to assess the incoming data to make any further adjustments to our policy rate.
Household consumption grew by 3.2 percent over last year. Understanding the causes of the continued robustness in consumer spending is important because it accounts for two-thirds of overall economic activity. Therefore, any accurate forecast of future economic activity would need to get the growth in consumer spending right.
Today, I will discuss one important factor behind the recent strength in consumer spending: households’ balance sheets—that is, their assets, such as stocks, bank accounts, and houses, and their liabilities, such as mortgages, car loans, and other forms of borrowing. At first glance, households appear to be in a strong financial position. Overall, American households currently possess a very high level of wealth that is driven by elevated house values, relatively low overall debt levels, and a strong stock market.
Asset performance and the amount of debt, however, explain only part of the picture. The health of household finances also depends on the cost of new and existing debt and the availability of credit. Household balance sheets are an important factor behind the recent strength in consumer spending. That said, some households may have a difficult time weathering unexpected costs or economic shocks. Looking at a variety of indicators across the income distribution shows that, while, in aggregate, household balance sheets are indeed strong, low- and middle-income households, and those with lower credit scores, may be stretched.
The remainder of my talk is organized as follows. I will begin by discussing household wealth, both in aggregate and across the distribution of income. Then, I connect elevated wealth to recent spending patterns. After that, I discuss the assets side of household balance sheets. Then, I turn to liabilities, including the cost of servicing debt. Next, I discuss households’ ability to get new credit and the cost of such credit. Before concluding, I discuss the role of households’ balance sheets in the transmission of monetary policy.
Overall Household Wealth and Its Implications for SpendingLet me now turn to the overall picture of household wealth. Figure 1 shows a stylized household balance sheet, with assets on the left and liabilities on the right. Net worth, also called wealth, is the difference between the two sides of the balance sheet—assets less liabilities—and it is a key indicator of households’ financial health. Relative to income, households’ net worth is near its highest level in the past 30 years. Total net worth in the U.S. was over $50 trillion higher in the third quarter of last year than it was at the end of 2019. After one accounts for inflation, this accumulation represents an increase in overall wealth of about 20 percent for U.S. households, as shown by the solid black line in figure 2.
These recent gains in household net worth have been broad based across the income distribution. The net worth of low- and middle-income households—defined as the bottom 40 percent of the income distribution and shown by the dashed red line—has increased in line with aggregate net worth.3 Although these households account for 25 percent of total consumption, which is less than their population share, they are still key to the performance of the economy overall.
Let me now turn to the implications of household net worth for our understanding of the recent strength in spending. Figure 3 shows the saving rate, which measures the share of disposable income—that is, income after taxes and government transfers—that households save rather than spend. The saving rate has fluctuated widely over the past few years. It rose during the pandemic, as many households received supplementary income support from the government and some cut back on spending. Then, households spent some of the savings that they had accumulated during the pandemic, leading the saving rate to fall to a relatively low level in 2022. The saving rate has recovered somewhat since then. Now, it hovers around 1 to 1.5 percentage points below its level before the pandemic, indicating that households are still spending more of their income than usual. It seems likely that elevated household wealth helps explain this higher-than-usual spending.
Overall spending has been elevated, but how has high consumption been spread across the income distribution? Recent research shows that the spending of low- and middle-income households has lagged that of higher-income households over the past few years.4 As shown in figure 4, although real retail spending growth moved similarly for all households before the pandemic, it has diverged since the middle of 2021. Since then, spending for low-income households moved roughly sideways until the middle of last year, when it began to grow again. High-income households’ consumption, by contrast, has grown more consistently over this period.
AssetsHaving discussed net worth and its implications for spending, now I drill down into the two components of net worth—household assets and liabilities. With regard to the asset side, elevated net worth largely reflects gains in two important asset categories: stock market holdings and real estate. Each category accounts for roughly one-fourth of households’ assets. The stock market valuation has increased at a very rapid pace over the past five years, leading to a $20 trillion rise in the value of households’ stock portfolios. As house prices rose, the value of households’ real estate has also increased by about the same amount.
Real estate is a particularly important source of wealth for low- and middle-income households, comprising 40 percent of their net worth. Therefore, the growth in real estate wealth over the past five years accounts for a very significant share—over half—of the increase in these households’ overall wealth. That said, many low-income households do not own their home, and so they did not benefit from the growth in house prices. Equities comprise a smaller share of these households’ wealth, and so they account for only around 10 percent of the increase in their wealth.
Wealth allows households to weather unexpected shocks, such as the loss of a job or a surprise bill; however, not all forms of wealth are quickly and easily accessible in case of such emergencies. It can be expensive for households to access the equity that they have in their homes. Also, much of households’ stock holdings are in retirement accounts that are difficult to liquidate. So, to understand how resilient households’ financial situations are, I also pay close attention to the most liquid components of their net worth, which include bank deposits and money market mutual funds. As the solid black line in figure 5 shows, in aggregate, households hold about 20 percent more of these liquid assets than they did before the pandemic. As the dashed red line shows, in contrast to the aggregate, low- and middle-income households have a slightly smaller liquid asset buffer than they did before the pandemic. This smaller buffer suggests that some of these households may not be as equipped to handle economic shocks as they were five years ago. That said, low- and middle-income households still hold more of these assets than they did 10 years ago, when many of them were still recovering from the Great Recession.
On the whole, the asset side of households’ balance sheets paints a very healthy picture of their financial positions. Rising house and equity prices have increased net worth for households across the income distribution, and elevated asset valuations seem to help explain strong consumption growth last year.
LiabilitiesLet me now turn to household liabilities—what households owe to their lenders. Figure 6 plots three major categories of household debt relative to disposable personal income.5 You see home mortgages, the largest share, at the bottom in blue; consumer credit, which includes credit cards, auto loans, student loans in orange; and other consumer loans in beige.6
Total household debt rose through the 2000s and peaked around the time of the Global Financial Crisis of 2007 to 2009. It then began a slow decline as households “deleveraged.” The evolution of total debt is driven by mortgage debt, which currently accounts for about 60 percent of total household debt. Mortgage debt levels remain relatively subdued after rising somewhat during the COVID-19 pandemic, partly due to increasing home prices leading borrowers to take out larger loans.
Figure 7 zooms in on revolving credit—largely, credit card balances—which is part of the previous “consumer credit” category.7 Balances were at about 7 percent of disposable income until the COVID-19 pandemic. Households reduced their spending—decreasing the need for credit card debt—and in part used income support programs to pay down existing credit card debt. The result was a nearly 3 percentage point drop in revolving credit relative to disposable personal income. As consumer spending rose and households began to take on more credit card debt, this ratio began to rebound in 2021 but remains about 1 percentage point below its pre-pandemic levels.
Although levels of debt may be low, how costly is it for households to remain current on that debt? Figure 8 plots the debt service ratio, which is the amount of required debt payments relative to disposable personal income.8 Along with the fall in debt to which I just referred, this ratio plummeted during the initial stages of the COVID-19 pandemic. It has since risen, but it remains about 1 percentage point below its pre-pandemic level. That said, interest payments on revolving debt, which excludes mortgages, have risen over the past few years. The share of disposable personal income going to pay this interest rate is now slightly higher than it was just before the pandemic.
Credit Availability and CostsSo far, I have discussed households’ current debt liabilities and how households are able to manage their current debt payments. Even households with elevated levels of assets may wish to obtain new credit. Policymakers and economists often ask, how easy is it for households, in general, to increase their borrowing, and at what cost?
Lenders consider a range of factors in determining whether to supply credit and how much credit to extend. One key factor is the borrower’s “credit risk score.” These scores, which are calculated by private companies, use information on individuals’ past payment behavior and a variety of other factors to create a number that is predictive of their ability to repay debt.
Figure 9 plots the fraction of individuals with credit risk scores in the subprime, near-prime, and prime categories since 2014. There has been a gradual increase in the fraction of borrowers with prime scores, in part reflecting the deleveraging that I referred to earlier, which is mirrored by the decline in the fraction with subprime scores. As you can see, the fraction of subprime scores took a sharp turn downward at the start of the COVID-19 pandemic. At that time, many people were able to use the pandemic-era income support programs to become current on their debt and otherwise boost their scores into near-prime and prime categories. This “credit score migration” helped many individuals obtain credit.9
Before obtaining new credit, people may first turn to lines of credit that they already have—for example, credit cards. Figure 10 plots “utilization rates”—the ratio of credit card balances to credit limits—for subprime, near-prime, and prime consumers. Utilization rates fell for all three groups at the beginning of the pandemic but have risen since then and are now somewhat above their pre-pandemic levels for both subprime and near-prime borrowers. These groups may be reluctant to draw down their credit lines further.
It can be challenging to determine the availability of new credit. While the total amount of credit that people have and their new borrowing can be observed, these quantities are determined both by lenders’ willingness to supply credit and borrowers’ demand for credit. Borrowers taking out fewer new loans may be due to a reduced supply of credit, lower demand for credit, or a combination of the two. Sometimes, however, one of these factors can be identified. For example, during the COVID-19 pandemic, reductions in household spending and increases in income support programs likely reduced the demand for credit, contributing to the decline in debt levels during that period.
A more systematic method that we have used at the Federal Reserve to help disentangle credit supply from demand has involved questions in our Senior Loan Officer Opinion Survey, or SLOOS.10 This quarterly survey asks officials who oversee bank lending practices for their institutions about how they have changed loan underwriting standards over the past quarter for a variety of loan categories. “Loan underwriting standards,” also known more simply as lending standards, refers to the requirements that banks impose before extending a loan. For example, banks may establish minimum credit risk scores for potential borrowers to qualify for certain kinds of consumer borrowing. Banks that raise minimum credit scores are said to have “tightened” standards and those that lower them to have “eased” standards. Tightening standards likely reduces the supply of credit.11
Because the SLOOS surveys commercial banks, its results are most informative for those loan categories for which banks do a substantial amount of lending. Hence, figure 11 shows survey results for consumer loans (credit card and auto loans), averaged together, weighting by balance sheet size.12 Banks make almost all credit card loans, and about one-third of auto loans. The figure plots the fraction of banks that have reported tightening less the fraction that have reported easing each quarter, weighted by the bank’s loan portfolio—so that plus-100 percent would indicate that all banks tightened, and minus-100 percent would indicate that all banks eased standards. For both credit cards and auto loans, banks eased standards in the early days of the pandemic but began to tighten them in 2022. More recent responses suggest that banks continued to tighten standards over 2024, making it more difficult for borrowers to obtain new loans. Although this tightening could limit growth in spending by those households that would need more credit cards to do so, recall that higher-credit-score borrowers are not close to exhausting their credit lines. In the most recent survey, banks have eased standards, which could support spending.
Monetary Policy TransmissionNow, before I conclude, let me say a few things about how the Federal Reserve’s monetary policy has been affecting the cost of borrowing for households. The primary tool that the Federal Reserve uses to influence the economy is the federal funds rate. The Federal Open Market Committee (FOMC) meets eight times a year to discuss the appropriate setting of the committee’s target range for the federal funds rate. The FOMC’s objective when setting this range is to achieve its congressionally mandated goals of maximum employment and price stability. Changes in the FOMC’s target for the federal funds rate affect overall financial conditions through various channels, including its effect on interest rates that matter for consumers’ decisions to purchase houses and cars or borrow on their credit cards. For example, when the FOMC eases monetary policy—that is, reduces its target for the federal funds rate—the resulting lower interest rates on consumer loans elicit greater spending on goods and services. Higher spending can, in turn, lead prices to rise. Lower mortgage rates make buying a house more affordable and encourage existing homeowners to refinance their mortgages. Of course, the rates charged on longer-term loans, such as mortgages, are also affected by expectations of how monetary policy and the broader economy will evolve over the duration of the loans, not just by the current level of the federal funds rate.
With respect to lending costs, the reductions in the target range for the federal funds rate last year have begun to pass through to rates on consumer borrowing. In the credit card market, interest rates are floating and are set as a fixed markup over the prime rate. By convention, the prime rate is equal to the upper end of the target range the FOMC sets for the federal funds rate, plus 3 percentage points.13 As seen in figure 12, auto loan and credit card rates have fallen in recent months, with the decline in the prime rate. Rates on auto loans are also influenced by the interest rates on shorter-maturity Treasury securities and risk spreads lenders assess to account for delinquencies and defaults. Auto loan rates have declined, thus far largely because of falls in risk spreads.
In the U.S., mortgages are generally fixed rate and have a longer duration than most other forms of consumer borrowing. Consequently, rates on new and existing loans can differ substantially. As shown by the solid blue line in figure 13, the majority of households still have mortgages with rates below 4 percent that were set some time ago. But rates on new mortgages are elevated compared with the ranges observed since the 2007–09 financial crisis, with the current average 30-year fixed rate around 7 percent. As I noted earlier, mortgages’ long duration means their rates are driven more by longer-term interest rates, which are in turn determined by many factors beyond just monetary policy. Households who recently became homeowners or moved must bear the cost of paying elevated mortgage rates. As a result, many are not moving.14
Overall, interest rates for many forms of consumer credit—with the notable exception of mortgages—have declined in recent months, starting to show the effects of the recent fall in shorter-term interest rates. Nonetheless, available data suggest that while new credit is available for households with higher credit scores and income levels, those households with lower credit scores and income levels are finding it relatively more difficult to obtain credit.
ConclusionLet’s return to the title question: How strong are households’ balance sheets? Generally, households appear to be in a good position: Asset holdings are high across the income distribution, driven by high house and equity prices, and debt levels are subdued. Interest rates on some forms of debt have begun to come down, and required debt service is low as a share of income. That said, some households appear to be stretched. Lower-credit-score households’ utilization rates are elevated, and banks have tightened loan underwriting standards on some forms of credit. And even though, as a group, low- and middle-income households possess elevated levels of overall wealth, they have less of a buffer of liquid assets than they did before the pandemic. These indicators suggest that certain groups of households may have a hard time weathering unexpected costs or economic shocks.
In closing, let me reiterate that it is important to monitor closely the strength of household balance sheets, which inform forecasts of overall economic activity. Strong balance sheets help support consumption spending, which in turn can help deliver the economic growth that puts the Federal Reserve in the best position to achieve its policy goals of maximum employment and price stability.
ReferencesAladangady, Aditya, Jacob Krimmel, and Tess Scharlemann (2024). “Locked In: Rate Hikes, Housing Markets, and Mobility,” Finance and Economics Discussion Series 2024-088. Washington: Board of Governors of the Federal Reserve System, November.
Bassett, William F., Mary Beth Chosak, John C. Driscoll, and Egon Zakrajšek (2014). “Changes in Bank Lending Standards and the Macroeconomy,” Journal of Monetary Economics, vol. 62 (March), pp. 23–40.
Driscoll, John C., Jessica N. Flagg, Bradley Katcher, and Kamila Sommer (2024). “The Effects of Credit Score Migration on Subprime Auto Loan and Credit Card Delinquencies,” FEDS Notes. Washington: Board of Governors of the Federal Reserve System, January 12.
English, William B. (2021). “The ‘Marketization’ of Bank Business Loans in the United States.” Working Paper, Yale School of Management, October.
Goodman, Sarena, Geng Li, Alvaro Mezza, and Lucas Nathe (2021). “Developments in the Credit Score Distribution over 2020,” FEDS Notes. Washington: Board of Governors of the Federal Reserve System, April 30.
Hacıoğlu Hoke, Sinem, Leo Feler, and Jack Chylak (2024). “A Better Way of Understanding the US Consumer: Decomposing Retail Spending by Household Income,” FEDS Notes. Washington: Board of Governors of the Federal Reserve System, October 11.
1. The views expressed here are my own and are not necessarily those of my colleagues on the Federal Reserve Board or the Federal Open Market Committee. Return to text
2. For a detailed discussion on my recent views on inflation, see Philip N. Jefferson (2025), “U.S. Economic Outlook and Monetary Policy,” speech delivered at the Economics Department Special Lecture, Lafayette College, Easton, Pennsylvania, February 4; and for my recent views on the labor market, see Philip N. Jefferson (2025), “Do Non-inflationary Economic Expansions Promote Shared Prosperity? Evidence from the U.S. Labor Market,” speech delivered at Swarthmore College, Swarthmore, Pennsylvania, February 5. Return to text
3. See Board of Governors of the Federal Reserve System (2024), “DFA: Distributional Financial Accounts,” webpage. These data provide quarterly estimates of the distribution of a comprehensive measure of U.S. household wealth. Return to text
4. For more details, see Hacıoğlu Hoke, Feler, and Chylak (2024). Return to text
5. Data are taken from Board of Governors of the Federal Reserve System (2024), Statistical Release Z.1, “Financial Accounts of the United States”. Return to text
6. See Board of Governors of the Federal Reserve System (2024), Statistical Release Z.1, “Financial Accounts of the United States”. Return to text
7. Data are taken from Board of Governors of the Federal Reserve System (2025), Statistical Release G.19, “Consumer Credit”. Return to text
8. For the series and information on how it is computed, see Board of Governors of the Federal Reserve System (2024), “Household Debt Service Ratios”. Return to text
9. For more discussion, see Goodman and others (2021) and Driscoll and others (2024). Return to text
10. See Board of Governors of the Federal Reserve System (2025), “Senior Loan Officer Opinion Survey on Bank Lending Practices”. Return to text
11. For an example of use of the SLOOS to help disentangle loan supply and demand, see Bassett and others (2014). Return to text
12. The SLOOS results reported here are based on banks’ responses weighted by each bank’s outstanding loans in the respective loan category and might therefore differ from the results reported in the published SLOOS, which are based on banks’ unweighted responses. Return to text
13. Before the establishment in 2008 of a range for the federal fund rate, the convention was to use the target for the federal funds rate plus 3 percentage points. See English (2021) for more discussion. Return to text
14. See Aladangady, Krimmel, and Scharlemann (2024). Return to text
MIL OSI USA News –