September 2025

The article at a glance: In theory, a direct way through which monetary policy is able to affect demand is through consumer credit rates. When a rise in the policy rate is “passed-through” to consumer credit rates, taking out a loan becomes more expensive and aggregate demand is suppressed. In new research, we use Brazilian credit registry data to examine whether this pass-through of monetary policy to consumer interest rates happens equally across all borrowers. Preliminary results suggest that monetary policy disproportionately affects credit costs for lower-income borrowers such that it increases the interest rate spread across the income distribution. 

Does monetary policy affect individuals’ interest rates equally? Evidence from Brazil.

By Antonia Tsang, CERF Scholar, Faculty of Economics, University of Cambridge

What are the mechanisms through which monetary policy may affect inequality? A growing literature in economics explores the relationship between monetary policy and inequality, the empirical component of which documents that monetary policy may increase inequality. In preliminary new research, we contribute to this literature by suggesting that monetary policy disproportionately affects credits costs for lower-income borrowers such that it increases the interest rate spread across the income distribution. Using data from Brazil, we document that the interest rates on new loans are more correlated with the monetary policy rate for lower-income individuals, and that this higher pass-through is driven by the (post-Covid) period of monetary tightening. These results indicating that the effects of monetary policy on credit rates are not equal across borrowers thus also imply a potential transmission channel from monetary policy to consumption inequality through household debt.

This investigation into the heterogeneous effects of monetary policy on consumer debt is fundamentally facilitated by data provided by the Brazilian credit registry maintained by the Central Bank of Brazil.  Our dataset tracks 750,000 individuals between 2018 and 2022 randomly sampled from the universe of formal Brazilian consumer credit. For each individual, we obtain information on the personal characteristics, the lender financial institution and the terms of each loan in each individual's consumer credit portfolio. Crucially, using information on the composition and characteristics of individuals’ liabilities (rather than total debt or net asset positions) is what allows us to isolate heterogeneity in monetary policy pass-through from effects due to differences in loan portfolios.

Why heterogeneity in monetary policy pass-through matters

The most obvious motivation for examining heterogeneity in the pass-through of monetary policy to consumer interest rates is the distributional implications it entails. By changing household credit costs, monetary policy is able to directly affect consumer demand through changes in disposable income. Heterogeneity in pass-through to credit rates is therefore a direct channel through which monetary policy can impact inequality -- if the weight of rising borrowing costs is disproportionately borne by low-income borrowers, then the effect of a monetary contraction directly exacerbates inequality along the income distribution. These distributional effects are particularly important in our context – while one of the world’s biggest economies, Brazil is simultaneously one of the most unequal, with a Gini coefficient ranking around the 8th highest globally.

From a macroeconomic perspective, heterogeneity in pass-through could also have implications for the aggregate effects of monetary policy. If the lower-income borrowers who see the largest increases in credit costs have higher marginal propensities to consume (MPC), then heterogeneous pass-through to consumer credit could amplify the effects of monetary tightening on aggregate consumption and output. Furthermore, asymmetry in pass-though heterogeneity across hiking and cutting cycles could generate parallel asymmetry in aggregate monetary transmission. If the pass-through of monetary policy expansions is weaker for high-MPC, low-income borrowers, this would dampen the effects of expansionary monetary policy. This asymmetry implies that, in the credit channel of monetary policy transmission, policy tightening is more effective at inducing contractions than loosening is in generating expansions (echoing the idea of “pushing on a string”).

Do the direct effects of monetary policy on consumer credit rates affect high- and low-income borrowers equally?

We first examine whether the pass-through of monetary policy to consumers’ average credit rate is heterogeneous along the income distribution – that is, whether the relationship between an individual’s (size-weighted) average interest rate on new loans made in each period and the monetary policy rate differs across borrowers of different incomes. We document that the pass-through of monetary policy to interest rates on an individual's consumer credit portfolio is stronger for lower-income borrowers than for higher-income borrowers, implying that a change in the policy rate is associated with a larger change in borrowing costs for lower earners. For a 100 basis point increase in the overnight rate, our point estimates suggest that pass-through is roughly 12 basis points weaker for every additional decile of the income distribution.

Could this be because lower-income borrowers are fundamentally “different”?

We emphasise that we find this negative correlation between pass-through and borrower income even when our baseline regression includes controls for individual characteristics such as age, gender, total indebtedness, existing payments in arrears and measures of banking relationships. The stronger pass-through of monetary policy to lower-income borrowers is therefore not because lower-income individuals are more indebted or have defaulted more, and is separate from differences in pass-through due to relationship lending. The baseline regression also accounts for potential differences in geographic location across low- and high-income borrowers, and differences in which banks different individuals choose to take out loans from. The observed heterogeneity in pass-through along the income distribution is therefore neither caused by lower-income individuals taking out loans in areas of higher pass-through or from banks that are more responsive to the policy rate.

The role of the “direct heterogeneity effect”

Aside from individual characteristic differences, one way in which borrowers differ across the income distribution is in the composition of their loan portfolios. Using a simple analytical decomposition, we show that heterogeneity in pass-through can be because of two reasons. Firstly, we identify a “direct heterogeneity effect”, whereby pass-through is stronger for lower-income individuals relative to higher-income individuals even when comparing like-for-like loans. This corresponds to the interest rate on on a loan with identical terms (loan type, volume, maturity) being more responsive to the policy rate for lower-income individuals (even when individuals are otherwise identical). Secondly, heterogeneity in pass-through could be driven by “loan composition effects”, where there are differences in the composition of loan portfolios across low- and high-income individuals, or in the responses of portfolio compositions to monetary policy. Intuitively, a rise in the policy rate could lead to a larger increase in interest rates for low-income individuals if these borrowers disproportionately increase their share of high interest rate loan types after a monetary policy hike. Alternatively, if low-income individuals disproportionately borrow in loan types that are more sensitive to the policy rate, then the responsiveness of their average interest rate on credit to monetary policy will also be higher.

To test the existence of the “direct heterogeneity effect”, we re-estimate our baseline specification by loan type. This allows us to isolate the role of the “direct pass-through” effect from the portfolio composition channels by only comparing loans of the same type, as well as controlling for loan volume and maturity. We find that the “direct heterogeneity” channel is operational, so heterogeneity in pass-through is still seen even when comparing like-for-like loans. In particular, our results show that the pass-through of monetary policy is stronger for low-income borrowers when restricting the sample to personal and payroll loans, the two largest consumer credit types in Brazil which account for over 75% of credit in our observed loan portfolios.

Asymmetry in pass-through heterogeneity across monetary policy hikes and cuts

Our results so far suggest that the interest rates of lower-income borrowers are more responsive to the monetary policy rate, and that this isn’t because of differences in loan portfolios or individual characteristics. However, note that the distributional implications of higher pass-through to lower earners differs across monetary tightening and loosening cycles. While stronger pass-through of monetary policy hikes to low-income borrowers increases (credit cost) inequality, stronger pass-through of interest rate cuts alleviates this inequality by reducing the debt service costs for lower-income individuals.

To investigate these distributional considerations, we adjust our specification to allow for asymmetry in pass-through heterogeneity between the post-Covid tightening cycle and the previous loosening cycle. Our results show that while the pass-through of monetary contractions during the hiking cycle was stronger for lower-income borrowers, the pass-through of cuts during the loosening cycle was weaker for low-income borrowers than for high-income borrowers. That is, low-income borrowers saw the largest increases in borrowing costs following a hike, but received the smallest reductions in interest rates following a cut, suggesting that monetary policy can exacerbate credit spreads across the income distribution during both hiking and loosening cycles.

Conclusion

Taken together, our results suggest that the direct effect of monetary policy on household debt is such that it could exacerbate credit cost inequality across the income distribution -- a result that is particularly striking in a country with pre-existing high levels of inequality. What implications might these results hold for policymaking? For central banks, our results contribute to our understanding of how monetary policy propagates through household liabilities. It also highlights a trade-off between distributional considerations and the efficacy of monetary policy – the ability of monetary policy to affect aggregate demand is greatest when it is borne by higher-MPC individuals.