Borrowing in response to windfalls

by Michaela Pagel & Arna Olafsson
Posted on June 11, 2020

A lot of research has been done to find out how households respond to unexpected, temporary, positive income shocks, the so-called financial windfalls. Think of finding money in a long-forgotten jacket, inheriting a small sum from an aunt you didn't even know you had, or finally winning the lottery. We know by now that, contrary to standard economic theory, consumers do not react as we would expect: they don't save every penny, but spend it all - and even more.

Yet the accumulation of consumer debt in response to temporary income shocks has been less widely studied, and relatively little is known about what drives households' choices of debt. In our TFI research project, we examine the way consumers accumulate high-interest unsecured debt in response to windfalls. From 1945 to the second quarter of 2009, the amount of debt owed by households increased substantially in all developed countries. Understanding the underlying mechanisms of consumer debt accumulation and paydown is important to understand the way households could better manage their finances.

Lottery winnings analysed

For our research, we took data from a personal finance platform in Iceland. This app that aggregates all financial accounts contains comprehensive transaction-level information on individual spending, income, account balances, and credit limits. This data source overcomes limitations in accuracy, scope, and frequency that have plagued consumption data from traditional sources.

Our Icelandic data has five main advantages versus - much more commonly used - US data:

1. Icelandic consumers almost exclusively use electronic means of payments, which pretty much solves the only remaining problem of app data: the absence of cash transactions.

2. The Icelandic app is marketed through all Icelandic banks, covering a broad fraction of the Icelandic population.

3. The spending and income data is pre-categorized, allowing accurate predictions about the responses of different spending categories (think of groceries, restaurants and leisure spending).

4. We observe all balances and credit limits of all accounts.

5. Individuals within households can link each other to get an overview over the household budget, but all accounts are personal.

To analyse temporary income shocks, we observe payments from lotteries that are very commonly played in Iceland. We looked at both small and large unexpected gains in income, also referred to as financial windfalls. Most interestingly, the marginal propensity to consume (MPC) out of small windfalls turns out to be larger than 1, meaning that individuals consume more than they received from their windfall income.

The larger-than-1 MPCs are financed by expensive short-term unsecured debt, such as overdrafts, which is then rolled over for a considerable period of time. It seems that individuals use the windfall payment as an excuse to buy something larger, resulting in additional consumer debt. Figure 1 shows the averages of the percentage of people with overdrafts in their checking accounts at least once a month, 5 months before and after their financial windfall.


Figure 1: Likelihood of overdrawing the checking account in the five months before and after small and large windfalls: Raw data binned averages split by the median windfall amount of approximately $100.


Consumers clearly increase their borrowing in response to small windfalls. This conclusion has to be looked at carefully though, because of other hidden things happening in the data. What if, for instance, only a small group of people play the lottery? Or what if individuals play the lottery more often around Christmas, when they need money for their Christmas shopping, leading to more consumer debt?

However, we made sure that our results are not driven by seasonality, trends or lottery expenses (how often they play). Lotteries are very common and widespread in Iceland, with 75% of our observed individuals winning at least once. Additionally, we compared large and small lottery winners to ensure that if one group won a larger sum, it's just because of sheer luck. Also, the data tells us that 75% of lottery expenses are part of subscription payments, resulting in the fact that monthly personal events don't affect whether individuals take part in a lottery or not.


Our findings suggest that individuals consume so much out of windfalls not because they were not able to borrow money before the windfall as our analyses show: people in our database clearly don’t face borrowing constraints. Therefore, our study suggests an alternative explanation for why people increase borrowing when experiencing a positive income shock: perhaps they feel 'richer' and hence free to 'splurge' as soon as they receive a lottery payment.

The 2015 American Household Credit Card Debt Study estimates the total credit card debt owed by an average U.S. household to be $15,762, which amounts to a total of $733 billion. The average Icelandic household's amount borrowed is of similar magnitude. Such large high-interest debt holdings over longer periods of time are very hard to rationalize in standard economic models.

We document an important discrepancy between theoretical and empirical results. The borrowing response we see in our data is clearly at odds with the predictions of standard economic models, which believe credit demand to be countercyclical and strongly negatively correlated with income shocks. The next step in our research is to further dissect our empirical findings, in order to better understand when and why individuals accumulate high-interest unsecured debt.

Michaela Pagel is the Roderick H. Cushman Associate Professor of Business at Columbia Business School, USA

Arna Olafsson is Assistant Professor of Finance at Copenhagen Business School, Denmark