The relatively weak pass-through of deposit rates in Croatian banks to ECB policy rates has met considerable backlash from the Croatian public. Raising deposit rates can benefit an individual bank. However, if all banks increase their deposit rates at the same time, this would have an adverse effect on their profitability. On the back of such considerations banks sometimes procrastinate and restrain the magnitude of pass through from policy rates to deposit rates. Interest rates in Croatia increased considerably following a substantial increase in deposit rates by one systemically important bank in October last year. This prompted depositors to shift their funds and boosted competition on the deposit market. However, deposit migration quickly moderated and clients increasingly started to opt for time deposits in their own banks, despite significant and even increasing interest rate differentials. This suggests a considerable “stickiness” of deposits or depositor inertia, which is one of the sources of banks’ market power. Depositor inertia, coupled with market characteristics, such as the relatively high concentration of market for deposits on the national and local level, the limited potential for adjusting interest rates on the stock of loans due to the widespread practice of interest rate fixation, the strong capital position of banks and considerable central bank reserves, supported persistent differentials in deposit rates. Although the analysis primarily focuses on factors relevant for each individual bank, the main conclusions can be generalised, in the sense that the structural features of the banking system, coupled with sizeable client inertia, have dampened the magnitude of interest rate pass-through.

Interest rates on household time deposits in Croatia recorded a relatively moderate growth in the past two years, even after accounting for the considerable jump in deposit rates since October last year. From mid-2022 to September 2023, the ECB lifted its key interest rates by 450 basis points in total, which has been the strongest monetary policy tightening cycle so far. However, the pass-through of these increases to interest rates on household time deposits has been relatively weak, which has initiated a lot of public criticism in Croatia. While slower growth in deposit rates compared to key central bank rates during the episodes of monetary policy tightening is an empirical regularity well documented in the literature (Dreehsler et al. (2017), Breyer et al. (2024)), the intensity of pass-through has been even weaker in the current cycle of monetary policy tightening (Messer and Niepmann, 2023). The pass-through of key interest rates in Croatia was among the weakest in the euro area, despite the momentum gathered since October 2023. The average interest rate on household time deposits in Croatia rose by 230 basis points by the end of 2023, while the average growth in the euro area stood at around 300 basis points (Figure 1). Interest rates in Croatia moved along the lower bound of the euro area range until October 2023, when they rose markedly on account of one systemically important bank^[1] having upped its rate from the very low 0.5% to 3%. Thus, the average time deposit rate in Croatia rose from 1.5% in September 2023 to a peak of 2.5% in October, which was followed by its very slow gradual decline.

Figure 1 Average interest rate on household time deposits

Sources: ECB and CNB.

Banks were relatively synchronised in increasing their time deposit rates, with a slight uptick in the difference between deposit rates across banks. By raising interest rates, banks seek to attract deposits from competing banks in order to increase liquidity surpluses held with the central bank. This increases the income they earn on the account of such deposits and consequently boosts their profit, while also making their operations more resilient to different disruptions (Liquidity stress testing of credit institutions). However, banks may be reluctant to engage in such an increase because this also gradually raises the costs of the existing deposits, given the repricing of maturing time deposits and higher inflows of funds from transaction accounts into time deposits (Rosan and Beriša, 2024). Also, data show that competing banks tend to react to such actions in a matter of days in an attempt to curb the risk of deposit outflows, which dampens the effect of the initial interest rate hike on deposit inflows (Figure 2). However, the intensity of the increase in interest rates varied considerably between different banks, so the range of deposit interest rates remained very wide. Large and even growing price differentials characterised by transparent information^[2] as well as quick and easy transfer of funds between banks are an interesting phenomenon, especially in an environment of elevated public attention and strong resentment with the relatively low level of interest rates on time deposits in Croatia.

Figure 2 Daily interest rates on household time deposits

Notes: Bank quintile groups are shown in red and grey. Bank* refers to the interest rate of a systemically important bank derived from the publicly available Excerpts from decisions on interest rate levels (12-month time deposits in euro). As a consequence, this series does not exhibit usual daily oscillations that can be caused by different characteristics of deposits and special arrangements with individual clients. The shown group of systemically important banks excludes the Bank*. The figure also shows that some systemically important banks also increased interest rates at the time of T-bills subscription from 13 to 20 November (two solid vertical black lines) in order to stem deposit outflows.
Sources: Database with standards on lending to consumers; excerpts from decisions on interest rate levels for the Bank*.

The size, that is, the systemic importance of a bank^[3], is one of the key determinants of the level of interest rates. Systemically important banks (Figure 2, red colour), in the second half of 2023, offered interest rates on time deposits that were, on average, 0.7 percentage points (p.p.) lower (Figure 3) than the rates offered by other banks (Figure 2, grey colour). This estimate takes into account differences in maturity and amounts of individual deposits as well as the exact time and location of contracting the deposit in order to control for the compositional effects. Depositors of large banks, which have an extensive network of branches and ATMs and provide a wide range of services (Avernas et al., 2023), are less sensitive to interest rate differentials due to reluctance to shift their deposits to other banks. This means that large banks are less affected by competition, that is, they are able to attract deposits even with below-market interest rates. This holds both for local (county) markets and the national market. Thus, even a small bank at national level can have a significant market share at local level, enabling it to maintain relatively low deposit rates (Figure 3).

Banks with more capital and abundant central bank reserves and banks with a larger share of loans with fixed interest rates also tend to offer lower interest rates than the rest of the market. Banks with CET1 ratio and above median share of central bank reserves in total assets offered interest rates that were, on average, by 0.6 p.p. lower than interest rates offered by banks with CET1 ratio and the share of central bank reserves below median. Such a market position is probably supported by the perception of security and greater headroom to face potential drain of central bank reserves in case of deposit outflows. Loan portfolio maturity is also potentially an important determinant of deposit rates (Polo (2021), Kho (2023)). Banks with long-term assets may find it more difficult to adjust interest rates on loans and consequently try to avoid the pass-through from higher policy rates to deposit rates. Loan maturity was not a statistically significant determinant of domestic banks’ interest rates in the observed period. However, a bank’s exposure to interest rate risk depends not only on loan maturity, but also on the possibility to reprice existing loans. Accordingly, banks with an above-median share of loans with variable interest rate in total loans offered time deposit rates that were, on average, by 1 p.p. higher than those offered by banks with an above-median share of loans with interest rates fixed for at least a year. Banks with predominance of loans linked to EURIBOR rather than to the national reference rate (NRR) in loan portfolio also offered higher deposit rates, reflecting a stronger and swifter adjustment of the EURIBOR to market conditions. Capital position and liquidity surpluses, as well as the share of loans with fixed interest rates, are highly correlated with the size of the bank, so there is a possibility that part of the impact on interest rates stems from factors related to market concentration and is not exclusively associated with the amount of capital and liquidity reserves and the share of loans with fixed interest rates. Also, the number of banks in Croatia is relatively small for the purpose of statistical analysis. All these factors reduce the reliability of estimates and make it hard to isolate and precisely quantify the impact of each individual factor on the deposit rate level.

Figure 3 Determinants of interest rate levels by bank

Notes: A 5% confidence interval is shown. The red colour indicates statistically significant variables at the level of 5%, while the grey colour denotes statistically insignificant variables. Given that the variables are highly correlated, the estimates are based on separate linear regressions with fixed effects for deposit maturity, amount of deposit (converted into a categorical variable), total amount of all counterparty deposits (indicates a person's wealth; converted into a categorical variable), month when deposit was made and county. For ease of interpretation, continuous variables were converted into binary/categorical variables in the following manner: the share in the county was converted into a binary variable by applying the threshold for market share of 10%; all other variables were converted into binary variables by applying the criterion of below/above median. Observed period: June 2023 – December 2023.
Sources: Database with standards on lending to consumers and CNB calculations.

Higher deposit rates initially prompted depositors to shift significant funds from overnight to time deposits, with the intensity of flows abating with time. Time deposits increased significantly after banks raised interest rates in October (Figure 4.a). In the last quarter of 2023, the inflow of new time deposits tripled from the previous quarter and stood at around 13% of all deposits, representing 58% of the stock of time deposits in September. Deposit inflows were at their peak in October and went down noticeably towards the end of the year, before jumping again at the very end of December. This can be attributed to the announced decrease in interest rates on time deposits from the beginning of 2024.

Figure 4 Daily inflows of household time deposits

Notes: The dotted line marks the moment when a systemically important bank raised interest rates on time deposits. “New clients” denote persons holding new deposits in a bank in which they previously held no transaction or savings accounts.
Source: Database with standards on lending to consumers.

The rise in interest rates against the backdrop of pronounced interest rate differentials prompted clients to shift their deposits between the banks, with the magnitude of flows quickly declining. Deposits of new clients, which can be used as a proxy for deposit migration between banks, increased ten-fold in the last quarter of 2023 compared to the previous quarter (Figure 4.b). In the last quarter of 2023, these deposits reached 15% of total new time deposits on average, while in the quarter before, they stood only at around 5% of new time deposits. Deposit flows followed interest rate differentials, with new time deposits shifting to banks offering higher interest rates (Figure 5). However, the inflow of deposits of new clients started to decline very quickly after the initial jump in October. This suggests that only a small group of depositors was sensitive to interest rate differentials and reacted to increasing dispersion by shifting their deposits between banks. Only a small number of depositors sensitive to differences in interest rates enabled some banks to maintain interest rates at a below-market level. These banks also saw substantial inflows of funds into time deposits, with the relative importance of flows to banks with below-market deposit rates gradually increasing towards the end of the last quarter.

Figure 5 Relation between the share of time deposits of new clients and interest rate levels

Note: “New clients” denote persons holding new deposits in a bank in which they previously held no transaction or savings accounts. Observed period: June – December 2023.
Source: Database with standards on lending to consumers.

In conclusion, only a moderate increase in interest rates and persistent interest rate differentials between banks can be explained by market structure as well as a relatively strong client inertia with respect to shifting (time) deposits between banks. Interest rates on household time deposits increased only moderately in Croatia. The increase in interest rates was particularly pronounced in October last year, which was followed by a inflow of funds into time deposits and in particular flows from banks offering lower interest rates to banks offering higher interest rates. However, the magnitude of deposit flows between banks dropped very quickly, which enabled some banks to maintain deposit rates at below-market levels. Deposit rates increased the least in banks with higher market shares, high levels of capital and liquidity and a large share of interest rates with long fixation periods in their loan portfolios. The factors in addition to enabling some banks to maintain low interest rates on time deposits also had an impact on overall market dynamics. The results of this analysis suggest that abundant capital and liquidity, a high degree of concentration and a large share of fixed-rate loans, as well as considerable client inertia, are key factors that have dampened the pass-through from higher policy rates to deposit rates.

1. The box focuses on this event, although it is important to bear in mind that banks’ deposit rate decisions might have been influenced by other factors as well. For example, the conclusions reached at the meeting of banks with the representatives of the Government and the Ministry of Finance in September 2023 might have influenced banks’ decisions on deposit rates. In addition, in late October 2023, the CNB published its first information list with banks’ offer of deposit rates in order to make it easier for consumers to compare different deposit rates offered by banks and to stimulate bank competition. ↑

2. The CNB's information list with the offer of deposits can be found here. ↑

3. A definition of systemically important banks can be found here. ↑