Imagine tracking two numbers on your banking app: the interest rate that your savings earn and the interest rate that your credit card or overdraft charges. The gap between them – what’s known as the consumer credit spread – is easy to ignore in good times and painful to notice in bad ones.
In a study written with four colleagues (Renato Faccini, Seungcheol Lee, Ralph Luetticke and Tobias Ranking) and published recently in the American Economic Review, we argue that movements in this spread affects who can smooth their spending, who gets stuck living payday to payday, and how deep recessions feel for different households. The paper is unique in connecting household finance with macroeconomic theory – and it draws conclusions for monetary policy and financial regulation (Faccini et al, 2026).
Analysing data on the finances of millions of Danish households
We connect a very rich set of Danish administrative records – 18 million household-year observations for the period 2003-18 – to bank‑level interest rates. This link allows us to build, for each household, an annual household‑specific measure of the consumer credit spread: the average loan rate at the household’s main loan bank minus the average deposit rate at its main deposit bank.
Because of Danish institutional features, the spread captures non‑mortgage borrowing such as credit cards, overdrafts and bank loans. The administrative data contain detailed information on household assets and liabilities, income and features such as education, gender, age and number of children. This allows us to connect asset dynamics to movements in the interest rate spread, and also to impute consumption from tax‑verified incomes, pension contributions and changes in assets.
Around 9% of households hover in a state of ‘zero net wealth’: their net assets are within roughly two weeks of median income at any point in time. The number of households moving in and out of this zero net wealth state depends on the interest rate spread and on household income.
When the spread widens, households near zero wealth are more likely to stay stuck there or fall back into it, while the very rich and the very indebted barely budge. This is exactly what you’d expect if the ‘price of borrowing’ rises just when incomes are soft and cushions are thin.
How the consumer credit spread shapes household spending
There is also an intimate link between household consumer spending and the consumer credit spread. Although the evidence does not necessarily indicate causality, it shows some notable links:
- Higher income and higher consumption go hand-in-hand for the wealthy and the poor, but most strongly so for poorer families.
- Higher spreads are related with lower spending for indebted and low‑wealth households, but higher spending for the wealthy.
- The income-consumption link tightens when spreads are high, especially for those near zero wealth.
Aggregating these responses into a time‑varying elasticity of consumption in response to income, we show that it is volatile and countercyclical. Crucially, this pattern strengthens once spread movements are included.
In the model that we build, this statistic closely tracks the economy’s marginal propensity to consume (MPC) – the proportion of additional income spent on consumption rather than being saved. Taken together, the evidence points to a countercyclical MPC: it rises when the economy is in a downturn and falls when the economy is doing well, and movements in credit spreads are important for this.
The consumer credit spread and macroeconomic outcomes
To turn these facts into a coherent story, we build a HANK (heterogeneous agent new Keynesian) model with an optimising banking sector. Banks take deposits from savers, and provide loans to firms and to households that need credit. Households are subject to income risk that is sometimes idiosyncratic and at other times common to the whole economy. No matter their source, households use bank deposits and loans to smooth their consumption over time.
Frictions in the banking sector imply that the interest on loans comes with a time-varying premium over the deposit rate. When banks’ net worth falls, this spread – the margin made by banks – needs to increase. In the HANK model, movements in the spread have three powerful effects:
- They make borrowing costly for poor households.
- They create a ‘kink’ in household budgets at zero liquid wealth: small savers earn a low rate, while small borrowers pay a high one. That kink builds up a mass of households living right at the edge and gives them high MPCs – small income changes show up quickly in spending.
- They imply that indebted households face different interest rates to those faced by wealthy savers, which affects their savings and consumption choices, as well as inequality.
Calibrating the model to Danish data reproduces the big empirical features: countercyclical spreads; pro-cyclical and volatile consumer credit; and a countercyclical aggregate MPC that moves with the measured elasticity of consumption in response to income.
When shocks hit, the spread decides who hurts – and how much
We use the HANK model to show that movements in the spread in response to economy-wide shocks amplify their effects. The basic mechanism here is a ‘financial accelerator’ with consumer spending being central to this mechanism. This differs from much other research in this area, which stresses investment in the amplification mechanism. In our analysis, contractionary shocks damage banking sector net worth and are amplified through the rise in the spread through its impact on consumption and savings.
Consider, for example, an interest rate hike. As policy tightens, activity cools and banks’ equity values fall, which implies that the borrowing spread widens. On impact, indebted households face both lower incomes and higher borrowing costs, so they cut spending significantly; wealthy savers may briefly increase spending as deposit returns improve.
The distributional footprint of monetary policy – who tightens their belts and who doesn’t – runs straight through the spread. Thus, when shocks hit banks, banks hit spreads, spreads hit the kink and the kink hits the MPC. That is why the same aggregate shock can feel so different across neighbourhoods.
The policy plot twist: calmer cycles, riskier lives?
What if regulators force banks to be safer by effectively lowering leverage? In our model, that corresponds to tightening the bankers’ incentive constraint. On the macro side, the level of the consumer spread goes up, but its cyclicality goes down. The classic financial accelerator is weaker; the volatility of output and investment falls (by roughly 6% and 13% respectively).
Because the spread goes up, households also respond by saving more for precautionary reasons and by increasing labour supply. Both of these responses lead to higher aggregate activity. That looks like a clear win: business cycle stabilisation and little output loss.
But on the micro side, the higher spread makes consumption insurance costlier. More households end up near the kink; idiosyncratic consumption volatility rises across the wealth distribution; and a formal welfare calculation shows losses across most households.
In other words, safer banks can mean calmer business cycles but riskier individual lives, especially for those who rely on credit to buffer shocks. That is the macro‑versus‑micro trade‑off at the core of our study.
Does adding housing‑like ‘illiquid wealth’ change the story?
Yes and no. When we introduce housing or business capital (illiquid assets) into the analysis, the model naturally creates ‘rich hand‑to‑mouth’ behaviour: people with plenty of net worth but little cash. This tweak helps to explain why, in the data, higher spreads can raise consumption for wealthy households via a stronger wealth effect.
But the core mechanism survives: changes in spreads have different impacts across the wealth distribution. The regulation trade‑off remains too: macroeconomic stabilisation still arrives with more day‑to‑day risk for households, now with especially large losses for those holding lots of liquid assets as their returns fall.
Wider implications
While our empirical analysis is focused on Denmark (due to availability of data), there is every reason to believe that the basic results hold true in other economies and perhaps to be even starker – after all, the Scandinavian ‘flexicurity’ model provides income insurance that is lacking elsewhere. Households in many other countries, including the UK, are likely to be even more sensitive to movements in spreads.
In a world with heterogeneity, the consumer credit spread is not a footnote; it reallocates intertemporal prices across people and across time. Macroprudential tools are useful for stabilisation of the economy and for handling financial risks, but distributional costs should also be considered.
