OeNB Report 2024/12: OeNB Euro Survey in Central, Eastern and Southeastern Europe: first results for 2023

Katharina Allinger, Elisabeth Beckmann, Söhnke Bergmann, Melanie Koch, Thomas Scheiber

This report presents results from the 2023 wave of the OeNB Euro Survey in Central, Eastern and Southeastern Europe (CESEE). This survey has been conducted since 2007 in selected CESEE countries, essentially the countries that have not yet (or only recently) adopted the euro. It serves to identify the trust of individuals in institutions, their monetary expectations and their financial decisions, especially with regard to de facto euroization, the use of financial assets denominated in euro.

Addressing the standard indicators of euroization, this report serves to show how determinants of demand for safe haven assets have developed over time and how indicators of demand for euro cash and deposits are correlated with socioeconomic characteristics and beliefs. We further present preliminary results on the perceived importance of cash in general, in the respective local currencies and in euro. Following Croatia’s accession to the euro area in 2023, we also present results on how the changeover to the euro influenced cash holdings and, more importantly, on how it affected inflation perception and expectations. Finally, we address ad hoc topics covered by the 2023 wave and conclude with descriptive results on the socioeconomic characteristics of respondents and references to the documentation on methodological aspects of the 2023 wave.

Findings presented here are primarily descriptive and should not be interpreted as causal evidence. For some assessments in this report, the number of observations is low, limiting the statistical power of the data analysis.

1 Euroization in CESEE

In this section, we address the following questions: How frequently do consumers in CESEE use euro cash for payments and what amounts do they typically hold? How frequently do consumers in CESEE hold euro deposits (deposit substitution)? Have there been significant changes in currency and deposit substitution since 2022? And how have demand-side determinants evolved since 2022, given that demand for euro cash and deposits tends to be driven by trust in banks, expectations regarding the stability of the local currency as well as network effects. Lastly, we look into how frequently cash is used for payments.

1.1 Background and motivation

De facto euroization has remained a widespread phenomenon in a number of CESEE countries beyond the euro area. Previous research has shown that, for individuals, de facto euroization is largely demand-driven. To test determinants of demand for (foreign currency) cash holdings, Stix (2013) develops a model which considers (i) transaction costs, (ii) trust in banks vs. social capital and (iii) network effects and expectations regarding inflation and the exchange rate. He shows that empirically, lack of trust in banks and memories of past banking crises are key determinants for cash preferences. In addition, he finds cash demand to have been driven by network effects ¹ of currency substitution and expectations regarding the stability of local currencies and foreign “safe haven” currencies. In contrast, he is able to rule out social capital, approximated by trust in other people, as a driver of cash preferences.

Regarding the prevalence of deposit substitution, theoretical research has highlighted the minimum variance portfolio (Ize and Levy-Yeyati, 2003), market failure (Broda and Levy-Yeyati, 2006), implicit bailout guarantees (Ranciere et al., 2010) as well as network effects (Craig and Waller, 2004) as significant factors driving deposit substitution and its persistence. Brown and Stix (2015) show that demand for deposit euroization is driven by distrust in the stability of the local currency. In addition, hysteresis effects, i.e., the experience of economic and financial crises during transition from planned to market economies in the 1990s, continues to affect expectations across cohorts. Moreover, deposit substitution is driven by network effects as well.

Following transition-related economic turbulence in the 1990s, policymakers in CESEE have built 30 years of “new” track records – partly amid global crises such as the financial crises and the coronavirus pandemic in the last decade. In 2022, a pronounced decline in trust and increase in foreign demand for euro cash were likely related to the initial shock of Russia’s invasion of Ukraine in February 2022. Although the war in Ukraine is ongoing, 2023 did not see any further shocks that might have directly affected households in CESEE in the form of, e.g., Sberbank closures. Furthermore, 2023 saw a renewed momentum in EU accession. Finally, 2023 brought continued rate hikes both in the euro area and in CESEE as well as a first decline from the high inflation rates of 2022. These changes may have affected (euro) cash demand and savings deposits denominated in euro.

1.2 Indicators of asset euroization over time

Euroization of assets held by private individuals has two main dimensions: the use of foreign currency cash (currency substitution) and the use of foreign currency deposits (deposit substitution). The former can only be elicited by survey evidence. Chart 1 (left-hand panel) shows the percentage of individuals holding euro cash in the nine CESEE countries covered by the OeNB Euro Survey in 2023. ² In Czechia, Hungary, Romania, Bosnia and Herzegovina, and North Macedonia this percentage significantly increased from 2022 to 2023. Chart 1 (right-hand panel) shows the median amount held in euro by individuals with euro cash. Hungary and Romania show an increase both in terms of the frequency of euro cash holdings and the amounts held. In Czechia and North Macedonia, the number of people holding euro cash increased, but the amounts held by those who have euro cash remained rather stable. ³ Given the recent changeover to the euro in Croatia, Croatian respondents were only asked about amounts of euro cash held at home. Evidently, the sharp increase in median amounts is partly due to conversion of kuna cash holdings into euro.

When we look at the two panels we see that in some countries (e.g., Czechia), euro cash is widespread, but the amounts held are low. In these countries, euro cash is likely held for transaction purposes abroad. For other countries, we observe both a high frequency of euro cash holdings and high amounts held (e.g., Serbia). For these countries, euro cash holdings likely serve the store of value function. Chart 2 shows the reasons respondents state for having euro cash. For North Macedonia and Serbia euro cash is, evidently, also an important means for domestic payments.

Combining the evidence from the two panels in chart 1 and projecting them onto the whole population following Scheiber and Stix (2009), we see that the average amounts held as euro cash per capita increased in five countries and decreased in three countries in 2023 (table 1). ⁴

To assess the overall importance of euro cash, the currency substitution index (based on Feige and Dean, 2004, and Scheiber and Stix, 2009) expresses euro cash in percent of total currency in circulation. More recently, we see an increase in currency substitution for five countries, and a decrease for three countries, by less than 3 percentage points at most (chart 3).

Deposit substitution (chart 4) decreased slightly in seven out of ten countries in 2023, with the highest decrease observed in Romania (6.2 percentage points) and Serbia (3.8 percentage points). In Czechia and Hungary, deposit substitution increased slightly in 2023.

Taken together, the currency substitution index and the deposit substitution index provide a picture of the overall extent of de facto euroization. The combined index is calculated as follows:

\[EI = (FCC + FCD)/(LCC + FCC + LCD + FCD)\]

where FCC is foreign currency cash, FCD is foreign currency deposits, LCC is local currency cash, and LCD is local currency deposits. FCC data are derived from the OeNB Euro Survey, the remaining information is based on aggregate statistics from national central banks. Since 2022, de facto euroization of assets has remained relatively stable with a slight decrease in six countries and a slight increase in two countries (chart 5). The largest change was observed in Romania with a decrease of 3.8 percentage points. Thus, the slight increase observed in seven countries in 2022 proved to be temporary and may reflect initial strong reactions to the war in Ukraine.

1.3 Demand-side indicators of euroization over time

The figures presented so far are affected by both supply (e.g., density of bank branch networks, availability of foreign currency deposits) and demand-side factors (e.g., saving preferences and trust in institutions). Understanding the demand-side determinants is particularly important for policymakers as they are less amenable to regulation than supply-side factors.

The OeNB Euro Survey elicits information about whether individuals prefer to hold their savings in cash rather than as deposits. Chart 6 shows that in Romania and Serbia, the majority of individuals prefer to save in cash. This does not mean that bank accounts would be held only by a minority of individuals. On the contrary, more than 80% have a current account in Romania, and even 89% do so in Serbia. The chart further shows a significant decline in the preference for cash savings in Serbia and North Macedonia since 2022. This is possibly related to recent high inflation rates. In Bosnia and Herzegovina there is a slight increase to the level of 2019-21. In Croatia, finally, preferences for cash savings have apparently not been affected by the changeover to the former safe-haven currency.

Turning to demand for euro deposits, chart 7 shows the percentage of individuals who would choose to deposit an amount equivalent to two average monthly salaries in euro rather than in local currency or another foreign currency. In Hungary, Poland, North Macedonia, and Serbia, we see a decrease in preferences for euro deposits. Bosnia and Herzegovina is the only country that shows a significant increase in preference for euro deposits. This might be related to the new momentum observed in EU accession procedures in 2023. Interestingly, in Bosnia and Herzegovina, this demand indicator for deposit substitution is moving in the opposite direction from what the macro data would imply (see chart 4). One reason for this could be that chart 7 shows a hypothetical scenario only, which is beyond the actual means of some households.

1.4 How have determinants of demand for euroization developed since 2023?

Demand for euro cash and deposits is heavily influenced by trust in the local currency and the euro as well as by expectations regarding inflation and exchange rate developments. In chart 8, dots above the 45-degree line show an increase of trust five years ahead for all countries covered by the survey compared with 2022, both for the local currency and for the euro. This constitutes a recovery to the levels of trust in 2021. The drop in trust in 2022 was likely partly due to the war in Ukraine.

As in previous waves, trust in the euro is on average higher than trust in the local currency. For euroization, relative trust plays an important role. Thus, while trust in the local currency increased since 2022, it remains lower than trust in the euro. Therefore, a reshuffling of portfolios into local currency is unlikely.

Chart 9 puts the increase in trust in the stability of currencies into perspective with trust in national and international institutions. It shows a similar picture to trust in currencies: Since 2022 trust in both national and European institutions has increased in all countries. This increase is reassuring as trust declined sharply from 2021 to 2022, which, again, was likely related to increased uncertainty amid Russia’s war of aggression in Ukraine.

In addition, the 2023 wave of the Euro Survey elicited whether respondents considered the euro to have had a positive or a negative effect on the euro area economies. In this respect, Bosnia and Herzegovina and Romania are the only two countries where the majority think that the euro had a positive effect on the economies of the euro area (chart 10). The share is lowest in Serbia with a positive assessment shared by 31% of respondents. At the same time, the share of “neutral” or “do not know” responses ranges from 21% (Bosnia and Herzegovina) to 38% (Serbia). In other words, the share of those who considered the euro to have had a negative effect on euro area economies was below one third of all respondents in all countries.

In all countries except North Macedonia a positive assessment of the euro for the euro area economies is positively and significantly correlated with trust in the ECB and trust in the EU. Probit regressions based on this survey evidence provide preliminary insights into how indicators of demand for euro assets and the frequency of euro cash holdings are related to socioeconomic characteristics, individual experiences and beliefs and network effects. Table 2 shows average marginal effects computed from probit regressions where the dependent variables are (i) euro cash holdings, (ii) preference for saving in cash, (iii) preference for saving deposits in euro. Control variables are based on Stix (2013) and Brown and Stix (2015). Results show that education and income are important correlates of demand for euroized assets. Self-employed individuals and those who receive remittances are more likely to be holding euro cash. Network effects (as measured by respondents’ perception of whether holding euro cash is common in the respective countries) are significantly and positively correlated with all three variables: euro cash holdings, preferences for saving in cash and preferences for having a saving deposit in euro. Those who lost money during previous crises are more likely to prefer having savings deposits in euro but less likely to hold euro cash. Somewhat counterintuitively, remembering high inflation is positively correlated with a preference for cash savings. However, cash savings are frequently denominated in euro. A similar reason likely drives the positive and significant effect of inflation expectations on preferences for cash savings.

The results for having euro cash and preferring cash savings differ for several control variables, e.g., education. Those who have euro cash are a mix of those who retain cash from travel abroad, those who make payments in euro at home and those who hold euro cash for saving purposes. Only the latter are similar in their socioeconomic characteristics, beliefs and expectations to those who prefer to save in cash. Therefore, the difference in results between columns 1 and 2 and columns 4 and 5 is not surprising. Finally, a positive assessment of the euro for the economies of the euro area is associated with a higher likelihood of having euro cash and a higher preference for holding saving deposits in euro, but significantly negatively related to cash savings. ⁵

1.5 Cash dependency? A first glimpse at the role of local currency and euro cash

In 2023, for the first time, the survey also enquired about the amounts of local currency cash people hold. Because this is an extremely sensitive question, the question was posed indirectly, asking respondents to share their thoughts about the average level of other people’s cash holdings in their local currency and in euro. To allow comparing the resulting median amounts (chart 11) across countries, the local currency amounts are converted to euro at current exchange rates. The chart shows that the “country ranking” in terms of local currency cash versus euro usage is similar to the currency substitution index ranking based on monetary aggregates (see chart 3). It also shows that the sum of local currency and euro cash holdings exceeds EUR 600 in all countries. Considering that we are dealing with median amounts, these amounts are substantial and in line with results shown in chart 6, which indicates a continued high preference for cash as a means of saving.

In five out of nine countries, cash is also used as a means of payment for everyday purchases in more than half of the cases (chart 12). In four countries, only every tenth purchase is cashless. This question was also asked for the first time in 2023, which is why we cannot tell whether this is a stable result or whether cashless payments have increased as they did in other countries. In summary, though, charts 11 and 12 show a high cash dependency for both savings and payments especially in the Western Balkan countries but also in Romania and Bulgaria.

The results presented for 2023 in this report are preliminary and descriptive. In summary, they show that currency substitution increased slightly in five out of eight countries whereas deposit substitution decreased in seven countries. Overall euroization slightly decreased in six out of eight countries. Trust in the stability of the local currency and the euro recovered to the levels observed in 2021 after the sharp decline observed in 2022. Trust in national and international institutions also recovered to levels observed prior to Russia’s invasion of Ukraine in February 2022. After the first 25 years of the euro, between 31% (Serbia) and 72% (Bosnia and Herzegovina) of CESEE citizens perceive the euro as having been positive for the euro area economies. Preliminary results, based on survey questions that were posed for the first time in 2023, indicate that cash amounts held in both local currency and euro cash are high at above EUR 600 median amounts. Payments in cash for everyday purchases predominate in five out of nine countries.

2 The impact of euro adoption on prices in Croatia – public perceptions versus economic estimates

On January 1, 2023, Croatia adopted the euro as its legal tender. This chapter addresses three questions related to euro adoption. First, did the changeover to the euro have an inflationary impact? Second, can we explain the change in public opinion and the discrepancy with official estimates? Third, should we be worried about negative spillover effects to trust in institutions or the euro?

2.1 Did euro adoption have an inflationary impact?

In Croatia, annual HICP inflation had peaked at 13% in November 2022, followed by a gradual deceleration to 5.4% in December 2023. Developments were thus broadly in line with those in the rest of the euro area. Monthly HICP inflation rates remained positive until August 2023, with the highest values observed in March (0.9%) and April (1.2%).

According to preliminary estimations by economists, the effect of euro adoption seems to have been mild. Falagiarda et al. (2023) show that while the month-on-month growth rate of service prices was relatively high in January 2023, other prices developed in line with historical patterns. The January increases in service prices seem to be partially due to underlying inflationary pressures beyond euro adoption. Using a hypothetical month-on-month growth rate based on historical averages, Falagiarda et al. (2023) conclude that the upper bound for the impact of euro adoption on prices was around 0.4 percentage points above annual HICP inflation. This is in line with estimations and analyses in the context of previous euro changeovers. The conclusion of a mild impact is also confirmed by a cross-check with micro-data from Croatian retailers.

However, this assessment is in stark contrast to public perceptions in Croatia. In October 2023, hardly any respondents considered euro adoption to have decreased prices or to have remained without effect. Instead, 18% believed prices to have increased somewhat, while 79% felt that prices had increased strongly due to euro adoption (chart 13).

We had asked the same question in a forward-looking way in October 2022, to be able to assess a possible change in perceptions following the euro changeover. Our data show that Croatians appear to have held more positive views on the inflationary impact of euro adoption in October 2022 than they did in October 2023. A few months before adoption, 57% of respondents had anticipated a slight increase in prices and only 31% had expected a strong increase in prices.

2.2 Can we explain the change in public opinion and the discrepancy with official estimates?

It is ultimately impossible to give a definite answer to this question. Nonetheless, we offer some possible explanations.

Regarding the change in public opinion, one factor that led respondents to change their views regarding the inflationary impact of euro adoption might have been changes in their overall views on inflation. OeNB Euro Survey data on qualitative inflation expectations and inflation perceptions over a 12-month horizon show that both perceptions and expectations have declined from 2022 to 2023 (see charts A1 (perceptions) and A2 (expectations) in the annex). However, if we compare the inflation levels expected in October 2022 with the inflation levels perceived in October 2023, the data show that there is a large share of people who considered inflation to be higher than initially expected (chart A3 in the annex). Perceiving stronger increases in prices than expected might have led respondents to attribute some of the unexpected increase to the euro adoption.

This may also partially explain the discrepancy with economic estimates. Respondents may have misattributed underlying inflation dynamics to the euro adoption. Moreover, according to anecdotal evidence, the news media in Croatia have been rather critical, reporting frequently on price increases due to euro adoption. This might have added to worsening public sentiment about the changeover. In general, eliciting inflation expectations and perceptions from individuals is a difficult task. Moreover, we know from the existing literature that the formation of such beliefs is a complex process and can lead to biased views.

2.3 Should we be worried about these results creating negative spillover effects to trust in institutions or the euro?

Regardless of whether the perceptions documented in chart 13 are rational or not, one might wonder if they had negative implications for other preferences, beliefs, or behaviors. Chart 14 shows the share of respondents that had confirmed having trust in the ECB, the HNB and the euro since 2019. Compared to previous waves, trust in the ECB and trust in the HNB was higher in 2023 – but the difference compared to the previous year is not statistically significant. Looking at the period 2019 to 2023, trust in the stability of the euro was lowest in 2022 – possibly due to Russia’s invasion in Ukraine and resulting economic and geopolitical developments and elevated risks. Both in 2019/2020 and in 2023, the trust result was almost 10 percentage points higher than in 2022.

In short, 2023 did not see a worsening of these trust indicators, despite the fact that people reported a strong inflationary impact of euro adoption.

In summary, we find strong discrepancies between first estimations of the inflationary impact of euro adoption in Croatia and public perceptions thereof. The public’s perceptions of the inflationary impact seem to have worsened when comparing the pre-adoption and post-adoption results. We provide some possible explanations, but ultimately it is impossible to say why the views have worsened over time or differ so much from economic estimates. On the positive side, measures of trust in the ECB, the HNB and the euro have not worsened over the same period.

3 Background information on the 2023 wave

The questionnaire of the OeNB Euro Survey consists of core questions that are asked each year and ad hoc questions on topics of current policy relevance and/or more in-depth research ideas. A new interactive dashboard allows interested readers to explore the core results on euroization and its determinants. Furthermore, the OeNB Euro Survey regularly includes three basic questions on financial literacy regarding interest rates, inflation and risk diversification. Results, including a comparison with results for Austria, can also be explored using an interactive dashboard .

The non-core questions in 2023 covered the following topics:

1. The euro at 25: the view from CESEE

2. Financial inclusion: does the digital transformation help or hinder?

3. Euroization and cash preference

4. Euroization and euro introduction in Croatia: did euro introduction affect cash savings?

5. Eliciting household income and dealing with nonresponse

6. Investment portfolio including crypto assets

While some of these ad hoc questions are intended for in-depth research projects that will be available from the OeNB Euro Survey website in due course, others have already been used for policy debates, e.g., the ECB’s Economic Reform Programme Assessments, and presentations, e.g., for the ECB’s International Relations Committee.

Finally, the questionnaire regularly includes questions on the socioeconomic characteristics of respondents. Table 3 presents descriptive statistics for these data by country. It shows that the countries covered by the OeNB Euro Survey are very diverse not only in terms of euroization as shown in section 1, but also in terms of demography and socioeconomic characteristics.

The survey was in the field during October 2023 in nine CESEE countries. Albania had to be excluded from the survey for this wave because of an overlap with census fieldwork. The 2023 survey was the first survey wave to collect data on unit nonresponse. For further details on sampling, fieldwork, response rates and weighting as well as editing and consistency checks, see the methodology section of the OeNB Euro Survey website .

4 References

Broda, C. and E. Levy-Yeyati. 2006. Endogenous deposit dollarization, in: Journal of Money, Credit and Banking 38. 963–988.

Brown, M. and H. Stix. 2015. The Euroization of Bank Deposits in Eastern Europe, in: Economic Policy 30(81). 95–139.

Craig, B. and C. J. Waller. 2004. Dollarization and currency exchange, in: Journal of Monetary Economics 51. 671–689.

Falagiarda, M., C. Gartner, I Mužić, and A. Pufnik. 2023. Has the euro changeover really caused extra inflation in Croatia? SUERF Policy Brief 571.

Feige, E. L. and J. W. Dean. 2004. Dollarization and Euroization in Transition Countries: Currency Substitution, Asset Substitution, Network Externalities, and Irreversibility, in: V. Alexander, Melitz J. and G. M. von Furstenberg (eds), Monetary Unions and Hard Pegs: Effects on Trade, Financial Development, and Stability. Oxford University Press. 303–319.

Ize, A. and E. Levy-Yeyati. 2003. Financial dollarization, in: Journal of International Economics, 59, 323–347.

OeNB Report 2023/6 : Economic Trends in CESEE.

Ranciere, R., A. Tornell and A. Vamvakidis 2010. Currency mismatch and systemic risk in emerging Europe, in: Economic Policy 25. 597–658.

Scheiber, T. and H. Stix. 2009. Euroization in Central, Eastern and Southeastern Europe – New evidence on its extent and some evidence on its causes. OeNB Working Paper 159.

Stix, H. 2013. Why do people save in cash? Distrust, memories of banking crises, weak institutions and dollarization, in: Journal of Banking & Finance 37. 4087–4106.

Annex I