After a prolonged period of zero interest rates, interest rates finally came back to life, as illustrated in the accompanying figure by the sharp rise in rates starting in 2021. However, the downward trend in interest rates can be traced roughly back to the financial crisis and continues into the 2020s. During this long episode, which put considerable strain on the banking sector, banks were nevertheless able to gradually improve their profitability, as indicated by the green, dashed line. In the years leading up to the COVID-19 crisis, the average profitability, measured as the pretax profit-to-total asset ratio (expressed as percentage), had already reached a solid level of 0.6-0.7 percent. Over the long term, there is a visible negative correlation between interest rates and bank profitability, as rates decline while profitability rises.
The sharp rise in inflation famously led to the tightening of monetary policy, reflected in the rapid increase in interest rates starting in 2021. After the COVID-19 crisis, the average profitability of banks grew alongside interest rates, breaking the decade-long negative correlation between these variables: now, there is a clear positive correlation between interest rates and bank profitability.
The figure also illustrates how the yield curve changes between 2009 and 2023. During the financial crisis, the yield curve was still steeply upward-sloping, but it gradually flattened year by year until the slope was close to zero at the beginning of the 2020s. Toward the very end of the period, the yield curve even inverted. This could indicate that markets were anticipating imminent interest rate cuts (which was indeed correct!).
But what does the data say? A simple fixed-effects regression model with two variables (and year dummies), where the dependent variable is ROA and the explanatory variable is the 10-year government bond interest rate, yields a negative coefficient (-0.32) for the entire period from 2009 to 2023. While this result is somewhat counterintuitive, the correlation was also evident in the figure. On average, long-term interest rates in the dataset were 2 percent during 2009–2023, and the average profitability during the same period was 0.47 percent. Relative to the average profitability, this coefficient is far from economically insignificant.
The result changes when the same two-variable regression is run for the years 2021–2023. The coefficient is nearly the same, but the sign reverses: a one-percentage-point change increases bank profitability by 0.28 percentage points. Interpreted differently, this implies that interest rate cuts begin to weaken the average profitability of banks. During 2024, the ECB has already lowered its deposit rate three times, so based on this result (correlation), it should also impact bank earnings. Additionally, the year dummy for 2023 is statistically insignificant, suggesting that the interest rate variable possesses substantial explanatory power.
And surely, there are many nuances to this relationship. Changes in interest rates do not affect all banks in the same way, as some banks’ business models may be more sensitive to changes in interest rates than other banks’ business models. However, we will leave that topic for another time.
