Not Just “Stimulus” Checks: The Marginal Tendency to Repay Debt – Liberty Street Economics – Liberty Street Economics –

Gizem Kosar, Davide Melcangi, Laura Pilossoph and David Wiczer

Households often use stimulus checks to pay off existing debt. In this post, we discuss the empirical evidence on this marginal propensity to repay debt (MPRD), and present new findings using the consumer expectations study. We find that households with low net worth-to-income ratios were more likely to use transfers from the March 2020 CARES Act to service debt. Next, we show that standard models of consumption-saving behavior can be reconciled with these empirical findings if borrower interest rates rise with debt. Our model suggests that fiscal policy may face a trade-off between increasing aggregate consumption today and helping those with the greatest debts.

How have households used their stimulus payments?

We study the CARES Act, a major stimulus package that was passed by the US federal government on March 27, 2020. As part of this package, all eligible adults received a one-time transfer of up to $1,200, with $500 per additional child. To document how households used these payments, we use a special module that is part of the New York Fed’s Survey of Consumer Expectations (SCE). In this module, respondents who had already received their CARES Act payments reported whether they used them to spend or donate, save or invest, or pay off debt.

Using these answers, we define marginal propensity to consume (MPC) as the share of the stimulus payment a household spent, marginal propensity to save (MPS) as the share a household saved, and marginal propensity to pay back debt to pay (MPRD) if the stock used to pay off their debt. With these measures we document three important facts. First, households used a third of their remittances to pay off debt. This is higher than the average marginal propensity to consume (MPC), which is usually central. Other studies (see Coibion, Gorodnichenko, and Weber (2020) and Sahm, Shapiro, and Slemrod (2010)) have also found large MPRDs, suggesting that the use of fiscal transfers to service debt is not special to the COVID-19 crisis. 19 pandemic. Second, households with a low net liquid assets-to-income ratio are more likely to pay off debt and to adjust their net asset positions sooner. Third, and related, households with lower net liquid assets to income ratios have lower MPCs. We show these last two facts in the graph below. Note that we define net liquid assets as the sum of liquid assets minus non-household debts. For the income we use the annual income of the household.

MPRDs decrease while MPCs increase with the net liquid assets to income ratio

Source: Authors’ calculations using the June Special Module of the New York Fed’s Consumer Expectations Survey.
Notes: slope= -4.507** (left panel), slope=3.186** (right panel); *p<0.1, **p<0.05, ***p<0.01.

The MPRD and the Effects of Tax Transfers

The household behavior we document can be explained by a model that contains a simple observation: interest rates on loans rise with household debt. For example, borrowers with higher debt often have lower credit scores, leading to higher interest rates. This prompts households to use at least part of their check to pay off debt, to reduce their debt load and thus support higher consumption in the future. We formally describe this mechanism in detail in our working paper. In addition, we provide evidence for the quantitative form of debt price schedules necessary for the model to be consistent with our empirical findings. It turns out that the resulting loan interest rates are also consistent with what would result from models with endogenous default and delinquent motives.

We find that accounting for debt-sensitive interest rates changes the effects of fiscal policy. We do three exercises to demonstrate this. First, when interest rates are debt sensitive, stimulus and insurance motives move in opposite directions between households. The insurance motive here refers to household savings to protect their spending capacity in the event of a fall in income. We illustrate this in the graph below, which shows the two effects of a tax transfer on the net wealth-income distribution of our model.

Stimulus and insurance motives of fiscal policy move in opposite directions in households

Source: Authors’ calculations using model simulations.

In the chart, we group borrowers in the model by ten deciles of their net liquid assets/income ratio on the horizontal axis. Each dot therefore contains the same share of households. For each quantile, we calculate the average share of each discount dollar that households spend upon receipt of the check, in red, and the lifetime wealth gain after receipt of the transfer, in blue. The graph underlines a clear inverse relationship between these variables: households that today make the greatest consumption gains through transfers will have the lowest wealth gains throughout their lives.

This result implies that policymakers may face a trade-off when designing fiscal policies, depending on whether they want to maximize aggregate spending or aggregate wealth. This trade-off is also apparent when comparing short- and long-term fiscal multipliers: debt-sensitive interest rates reduce the consumption responses of the poorest households to the impact, while these consumption responses become more persistent over time. Conversely, they increase the MPCs of low-indebted households, but make their spending responses less persistent.

In our second exercise, we examine how these two effects stack up in the economy and unfold over time. We show that in the medium term an increasing debt-price schedule reinforces the consumption effects of fiscal policy. For every dollar of tax transfer paid to households, a reduction in debt service after seven years leads to an additional spending effect of 8 percentage points.

Finally, we calculate the average welfare and consumption gains resulting from the payments with economic impact by allocating the payments in the model as they were allocated in the CARES Act in 2020. In our calibrated model, welfare increases by 0.52 percent, while 21 cents per discount dollars are spent in the first quarter, which, otherwise equal, is about 1.5 percent of nominal total consumption at the time of the CARES Act.

In conclusion, we have provided new empirical evidence for an undervalued fact: most households, especially those with the lowest net liquid assets, use tax transfers to service debt. As a result, indebted households will see better interest rates and thus be able to consume more in the future. This underscores an additional insurance motive for what has often been too narrowly labeled “incentive” checks. However, incentive and insurance motives are not well aligned in households. As such, policymakers can maximize both immediate spending and longer-term welfare gains by targeting different households.

Laura Pilossoph is an assistant professor of economics at Duke University.

David Wiczer is a research economist and consultant at the Federal Reserve Bank of Atlanta.

How to quote this message:
Gizem Koşar, Davide Melcangi, Laura Pilossoph, and David Wiczer, “Not Just ‘Stimulus’ Checks: The Marginal Tendency to Repay Debt,” Federal Reserve Bank of New York Liberty Street Economy27 June 2023, https://libertystreeteconomics.newyorkfed.org/2023/06/not-just-stimulus-checks-the-marginal-propensity-to-repay-debt/.

Adblock test (Why?)