Conditional dynamics of monetary policy shocks: the mitigating role of macroprudential policy in CESEE (OeNB Bulletin Q1/25)
Clara De Luigi, Markus Eller, Anna Stelzer 1
This study examines the economic effects of monetary policy (short-term interest rate shocks) and its interaction with macroprudential policy in 11 EU member states of Central, Eastern and Southeastern Europe (CESEE) over the period from 2000 to 2019. Employing a smooth transition vector autoregressive model, we assess how the impacts of interest rate shocks vary with the intensity of macroprudential policies and across different exchange rate regimes. We find that in countries with flexible exchange rates, monetary policy tightening tends to persist longer and is often offset by easing macroprudential measures, particularly when these policies are already stringent. This pattern is less evident in countries with fixed exchange rates, where interest rate shocks do not always represent independent monetary policy actions. Overall, muted macrofinancial responses across the sample suggest that macroprudential measures may counterbalance the effect of monetary policy (interest rate) shocks and that traditional monetary tools have been less effective in the latter half of our sample period. These results highlight the importance of incorporating macroprudential indicators into monetary policy analysis and contribute to discussions on policy coordination, offering insights to help optimize policy mixes to enhance economic resilience.
JEL classification: C32, E52, E61, G28
Keywords: monetary policy, macroprudential policy, smooth transition VAR, CESEE
Particularly in times of high uncertainty, understanding the interplay between policies is crucial to informing policy decisions aimed at balancing price and financial stability. In this context, the interaction between monetary policy (MP) and macroprudential policy (MPP) is complex and disputed. Both theoretical and empirical studies suggest that MP and MPPs are often complementary: MPPs tend to have positive long-term effects in terms of reducing macroeconomic volatility and appear to be better tailored to address risks in specific channels and market segments, while MP measures affect the broader economy. There is less consensus, however, on the short-term effects of the interaction between MP and MPP: Since MP and MPP often operate through the same channels (lending, balance sheet, risk-taking, etc.), their interaction may have undesired side effects (for a broader discussion, see Albertazzi et al., 2021).
The theoretical literature finds that the relationship between MP and MPP depends on the stage of the business and financial cycles and the nature of economic shocks. MP and MPP appear to be complementary when these cycles are synchronized and in the presence of financial shocks. 2 However, as financial cycles tend to outlast business cycles and, therefore, the two are not always synchronized, conflicts between policies aimed at stabilizing them can arise. Accordingly, the optimal policy mix is influenced by financial frictions and distortions within the economy. Due to the contractionary effects of tighter MPPs on credit and output, Du and Miles (2014) suggest that lowering the policy rate is optimal when capital requirements are tightened. Angeloni and Faia (2013) identify an optimal mix with mildly countercyclical capital requirements and an active monetary policy. The impact also varies depending on the MPP tool employed: Chen et al. (2020) note that tightening the loan-to-value ratio would be more than twice as contractionary as tightening the loan-to-income ratio when debt levels are high and monetary policy is constrained.
Regression studies report mixed results on the economic effects of the interaction between monetary and macroprudential policies. For instance, Nier and Kang (2016) observe that the effects of MPPs are unaffected by the MP stance. Takáts and Temesvary (2019) find that tight MPP mitigates the effect of MP on lending regardless of whether MP is tight or loose, whereas Zhang and Tressel (2017) report that loosening MP is more effective when MPP is also loose. Additionally, several studies indicate that more stringent capital requirements reduce the effectiveness of MP on the bank lending channel, which could hinder the achievement of inflation targeting objectives (see, for instance, Imbierowicz et al., 2019; Acharya et al., 2020; Buch et al., 2022).
Due to the fact that time series are usually not long enough to employ multivariate time series models, only few studies have employed a vector autoregressive (VAR) framework to analyze MP-MPP interactions. Kim and Mehrotra (2018) study the macroeconomic effects of MPP and MP within a structural panel VAR model for the Asia-Pacific region, finding that these policies generally complement each other in achieving price and financial stability. However, they note potential policy dilemmas during periods of low inflation coupled with high credit growth due to short-term trade-offs between these objectives. In a broader cross-country study using a panel VAR model including 32 advanced and emerging economies, Kim et al. (2019) observe that MPP loosens endogenously in response to contractionary MP shocks to stabilize credit. Moreover, the endogenous response of MPP to credit shocks is more pronounced in financially less developed countries and in those with less flexible exchange rates.
Countries in Central, Eastern and Southeastern Europe (CESEE) offer compelling case studies for jointly analyzing macroprudential and monetary policies. They implemented a diverse range of macroprudential tools earlier than many developed economies, therefore offering longer time series, and they have undergone shifts in monetary policy regimes, emphasizing the importance of employing nonlinear models for analysis. Eller et al. (2021) analyze the effectiveness of MPP in CESEE countries depending on the level of the policy rate, using a factor-augmented VAR model with regime switches. They find that tighter MPP would be effective in containing domestic private credit growth and the volumes of gross capital inflows, with the effectiveness of MPPs being greater and swifter in a low-interest rate environment.
With the aim of providing further insights into the interaction of monetary and macroprudential policy in CESEE, we use an approach similar to Eller et al. (2021) and the intensity-adjusted MPP index developed by Eller et al. (2020), but we look at the opposite research question, analyzing the effectiveness of MP (or short-term interest rate) shocks across environments with different degrees of MPP tightness. We use the short-term interest rate as a proxy for the monetary policy stance and refer to MP and interest rate shocks interchangeably. However, in some of the countries studied, which operate under varying degrees of fixed exchange rate arrangements, monetary policy is constrained, and domestic interest rates may reflect shifts in the anchor country’s rates (see section 1.1). Despite limited scope for independent monetary policy in these countries, interest rate changes still capture shifts in monetary conditions that affect credit dynamics and economic activity, whether driven by domestic policy or external factors. We therefore believe it is a compelling research question to explore the interaction between interest rate shocks and MPP, regardless of whether these shocks can be strictly classified as traditional MP shocks. 3
We employ a smooth transition vector autoregression (ST-VAR) model to identify different MPP regimes based on the intensity-adjusted MPP index. ST-VAR models are particularly useful for analyzing different policy environments as they enable the identification of potential nonlinear policy effects conditional on a selected indicator variable, in our case, a macroprudential policy index. Our analysis reveals significant variations in the effects of MP (or short-term interest rate) shocks, shaped by the tightness of MPP and differences between flexible and fixed exchange rate regimes. In countries with flexible exchange rates, monetary policy tightening shocks tend to persist longer and are often countered by easing macroprudential measures, particularly when MPPs are already tight. This pattern is less pronounced in countries with fixed exchange rates, where MP is constrained, with notable exceptions such as Bulgaria and Croatia. Across our sample, macrofinancial responses are generally muted, likely due to macroprudential counterbalancing and the limited effectiveness of conventional monetary tools in boosting inflation after the global financial crisis (GFC). We also observe some peculiar cases of price and credit “puzzles,” especially in countries with fixed exchange rates where domestic monetary policy has lesser degrees of freedom.
The remainder of the paper is organized as follows: Section 1 reviews the evolution of policies across CESEE countries and provides the details of data compilation, section 2 outlines the econometric framework employed, section 3 summarizes our empirical results and section 4 discusses policy implications and concludes.
1 Key policy developments and data compilation
In order to study the differences in the response to MP shocks across countries, we conduct a country-by-country analysis. Due to data constraints, we narrow our analysis to the 11 EU member states in CESEE: Bulgaria, Croatia, Czechia, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia and Slovenia. Before delving into the empirical analysis, we review the main developments in these countries in the policy areas crucial for our analysis: monetary, exchange rate and macroprudential policies.
1.1 Monetary and exchange rate policies in CESEE
Following the transformational recession of the early 1990s, many countries in CESEE experienced severe hyperinflation and significant economic instability. This turmoil often necessitated measures aimed at stabilizing currencies and restoring monetary order. By 2000, most of these countries had adopted forms of fixed exchange rate regimes – including currency boards, official pegs and tightly managed floats – and were primarily using exchange rate-targeting monetary policy frameworks (see Belhocine et al., 2016). However, over the years distinct paths emerged, as outlined in table 1, which shows the evolution of these policies from 2000 to 2019 (our sample period).
Monetary policy frameworks and exchange rate regimes in CESEE in 2000 and 2019 | ||||
Monetary policy framework | Exchange rate regime | |||
2000 | 2019 | 2000 | 2019 | |
Bulgaria | CB | CB | CB | CB |
Croatia | ET | ET | FLOAT | IPEG (2006) |
Czechia | IT | IT | FLOAT | FLOAT |
Estonia | CB | MU (2011) | CB | EURO |
Hungary | ET | IT (2001) | PEG | FLOAT |
Latvia | ET | MU (2014) | PEG | EURO |
Lithuania | CB | MU (2015) | CB | EURO |
Poland | IT | IT | FLOAT | FLOAT |
Romania | MT | IT (2005) | FLOAT | FLOAT |
Slovakia | MT (IT) | MU (2009) | FLOAT | EURO |
Slovenia | ET (IT) | MU (2007) | FLOAT | EURO |
Source: Authors´ compilation based on IMF AREAER Database and national central banks. | ||||
Note: Monetary policy
frameworks: CB = currency board, ET = exchange rate targeting, MT =
monetary targeting, IT = inflation
targeting. Exchange rate regimes: CB = currency board, PEG = officially announced exchange rate peg, IPEG = unofficial exchange rate peg, FLOAT = managed or free float, MU = monetary union (euro area). Years in brackets indicate the year of change or an implicit monetary policy anchor in place. (IT) means that the monetary framework was at least implicitly targeting inflation. |
One group of countries maintained fixed exchange rate regimes and progressively integrated more closely with the euro area, eventually either joining it or entering its precursor, the European Exchange Rate Mechanism (ERM II). 4 This group includes the Baltic countries, Bulgaria, Croatia, Slovakia and Slovenia. Euro area membership milestones include Slovenia in 2007, Slovakia in 2009, Estonia in 2011, Latvia in 2014, Lithuania in 2015 and Croatia in 2023, with Bulgaria joining the ERM II in 2020 while continuing its currency board against the euro. 5 We consider these countries as having a fixed exchange rate.
Conversely, another group shifted from fixed to flexible exchange rate regimes and adopted inflation-targeting frameworks. Czechia led this transition in 1997, followed by Poland in 1998, Hungary in 2001 and Romania in 2005. Despite adopting flexible exchange rates, these countries occasionally employed managed exchange rate practices to address specific economic challenges. During the GFC, for instance, the central banks of Poland and Hungary resorted to foreign exchange swaps to alleviate liquidity pressures in their financial sectors. In 2016 and 2019, Romania managed the Romanian leu within a narrow 2 % depreciation band against the euro to stabilize its currency. Finally, Czechia frequently resorted to foreign exchange interventions to counteract appreciation pressures on the Czech koruna. The most significant instance was the “exchange rate commitment,” which was in place from 2013 to 2017. During this period, amid zero-lower bound constraints, the Czech National Bank implemented sizable foreign currency interventions (cumulatively totaling EUR 76 billion) to combat deflationary pressures and expedite the return to the 2 % inflation target (for an overview, see Franta et al., 2022).
Despite some overlaps and specific adaptations within these two groups, we believe that this classification into fixed versus flexible exchange rate regimes provides a useful yardstick for presenting our findings throughout the remainder of this paper.
We use the countries’ short-term interest rate (STIR), typically the three-month money market rate, as an empirical proxy for the MP stance, as is commonly done in the literature (see Jarocinski, 2010, among many others). It is important to recognize that in countries with conventional fixed exchange rates, monetary policy is constrained compared to that in countries with freely floating regimes. Despite these constraints, there remains some temporary flexibility to deviate from the interest rates of the anchor country, primarily to defend a specific exchange rate. Under a currency board arrangement, the scope for MP action is even more limited and largely confined to adjustments in reserve requirements and banking regulation. Nonetheless, even within such a framework, shocks to the STIR can occur, particularly due to sudden, large-scale capital outflows triggered by external events. In both conventional fixed exchange rate regimes and currency boards, an MP shock in the anchor country can induce a corresponding shock in the domestic STIR, typically moving in tandem with the anchor country’s key rates. Such changes, however, should not be mistaken for independent domestic MP shocks but should rather be interpreted as repercussions of the commitment to the respective exchange rate policy. 6
1.2 Macroprudential policies in CESEE
While MPP became prominent in advanced European countries following the GFC, several CESEE countries had already begun implementing similar measures earlier. This preemptive adoption was largely in response to rapid domestic private sector credit growth prior to the GFC. Initially termed “administrative” measures, these early interventions focused primarily on curbing excessive credit expansion.
Eller et al. (2020) provide a detailed analysis of MPP decisions across 11 CESEE EU member states, tracking various macroprudential instruments over time. They developed an intensity-adjusted index – the macroprudential policy index (MPPI) – to quantify the overall stance of MPP. 7 According to their findings, there was a noticeable tightening in MPP across CESEE from the late 1990s up to the GFC, mostly driven by stricter capital and liquidity requirements. After the GFC, the intensity of these measures stabilized until 2010, then intensified again, with a greater emphasis on borrower-based measures and the introduction of capital buffers from 2014 onward. In terms of the instruments used and the timing of MPP activation, the CESEE countries can be grouped into three categories:
-
Bulgaria, Croatia and Romania, which implemented macroprudential policies in the late 1990s or early 2000s due to strong credit growth, strengthened these measures before the GFC.
-
Estonia, Poland and Slovenia also employed such policies before the GFC but to a lesser extent and with a less varied toolkit.
-
Czechia, Hungary, Lithuania, Slovakia and, to some degree, also Latvia, initially utilized MPP tools minimally but significantly tightened their policies just after the GFC, driven primarily by stricter capital-based measures.
Chart 1
Chart 1 illustrates how our two key policy variables by country developed over time, tracing the movement of the MPPI alongside the STIR. The early part of the sample shows high volatility in short-term rates among several CESEE economies, while the latter part, post-GFC, is characterized by generally lower interest rates – a trend common to both advanced and developing economies. Concurrently, the MPPI indicates a general upward trend across most countries after the GFC, reflecting the increased use of macroprudential measures discussed above. These opposing trends of monetary easing and macroprudential tightening, especially in the years following the GFC and prior to the COVID-19 pandemic, raise the question of how complementary both policy dimensions were in managing economic and financial stability – this issue will be further investigated in section 2. Chart 1 also validates our choice of model, a ST-VAR. While some trends can be identified in the development of the MPPI, we cannot detect any clearly distinct macroprudential regimes, which would have made a regime-switching model a natural alternative.
1.3 Dataset
Our sample captures both global data and country-specific macroeconomic and financial data for the 11 CESEE EU member states. We use quarterly data and chose to limit our sample to the period between 2000 and 2019 for reasons of data availability and because certain data series were highly volatile before 2000 and after the outbreak of the COVID-19 pandemic.
Dataset | |||
Variable | Description | Source | Transformation |
Financial variables | |||
STIR | Typically, three-month money markets rates | IMS IFS | |
LTIR | Ten-year Maastricht government bond yields | IMS IFS | |
REER | Real effective exchange rate (CPI-based) | BIS | 100*log(x) |
MPPI | Intensity-adjusted macroprudential policy index | Eller et al. (2020) | |
credit | Credit to domestic private sector, LC million | IMS IFS | 100*log(x), qoq |
Macroeconomic variables | |||
HICP | All-items HICP, 2015=100 | Eurostat | qoq |
GDP | GDP at market prices, LC million clv2015, swda | Eurostat | 100*log(x), qoq |
Global control variables | |||
WUI | World Uncertainty Index | Ahir et al. (2018) | 100*log(x) |
p_oil | APSP crude oil (USD/barrel) | IMF PCPS | 100*log(x), qoq |
Source: Authors’ compilation. | |||
Note: IFS = International
Financial Statistics dataset; LC = local currency; clv = chain-linked
values; swda = seasonally and working-day adjusted;
APSP = average petroleum spot price; PCPS = Primary Commodity Price System dataset. The last column indicates the transformation applied to the respective series. "qoq" refers to quarter-on-quarter differenced data. |
Our set of country-specific macroeconomic and financial data includes real GDP growth, HICP inflation, private domestic credit growth, the short-term interest rate (STIR), the long-term interest rate (LTIR), the real effective exchange rate (for the group of countries with flexible exchange rates) and the intensity-adjusted MPPI introduced in the previous subsection. Moreover, we include two global control variables, the World Uncertainty Index (WUI) and oil prices, to capture periods of high global macrofinancial and price volatility. This selection of variables is motivated by previous research that has examined MP or MPP shocks in the CESEE countries or economic spillovers of MP shocks from the euro area to CESEE (e.g., Feldkircher, 2015; Benecká et al., 2020; Eller et al., 2021). Table 2 gives an overview of the variables used, their respective transformations and the main sources they were obtained from.
2 Modeling conditional monetary policy effects
To explore the country-specific responses to a monetary policy shock (or STIR shock, in the case of countries with fixed exchange rates), depending on whether the economy faces a comparatively tight or loose macroprudential policy environment, we utilize a ST-VAR framework, as will be elaborated in this section.
In particular, we estimate separate models for each CESEE country in our sample. 8 Our M time series of interest, the global control variables and country-specific quantities (which are discussed in the previous section and are presented in table 2) are collected in a vector yt for t=1,…,T.
The ST-VAR model for each country takes the form:
yt=(A11yt−1+…+A1Pyt−P+c1)×St(ut−1)+(A01yt−1+…+A0Pyt−P+c0)×(1−St(ut−1))+ϵt,
where Aip are M×M-coefficient matrices for state i∈{0,1} and lag p=1,…,P. ci is an M×1-vector of intercepts and St(ut−1)∈[0,1] denote state indicators, which are bounded between zero and one and transition smoothly between states. St(ut−1) depends on the signal variable ut−1, which is given by the MPPI, and allows us to analyze (potentially) nonlinear effects of shocks to the STIR, dependent on different macroprudential regimes. Lastly, ϵt∼N(0,Ω) is a Gaussian error term with zero mean and M×M-covariance matrix Ω, which is a standard choice in multivariate time series models.
Intuitively speaking, this models the economy as being in one of two states, a state of tighter or looser macroprudential regulation. Each state has its own coefficients matrices that describe how variables interact with each other, namely Aip for state i∈{0,1} and lags p=1,…,P. 9 A ST-VAR model does not switch abruptly between these two states but instead uses a smooth transition function which gradually changes the influence from one set of coefficients to the other, based on the current macroprudential stance. At any given time, the model therefore calculates a weighted average of the state-specific coefficients, with the weights being determined by the transition function. 10 This empirical framework offers some advantages compared to other regression methods. While it is possible to estimate effects for each regime separately in a ST-VAR, as it is the case in a regime-switching model, one regime might have relatively less observations than the other, which makes estimates unstable and imprecise in a traditional threshold model. The ST-VAR, on the contrary, leverages all the information available by using the varying degrees of a particular regime indicated by St (see also the respective discussion in Auerbach and Gorodnichenko, 2012). Therefore, inference of any regime is based on a larger set of observations. In addition, using a ST-VAR allows for modeling dynamic interactions between variables, capturing possible feedback loops or temporal dependencies.
2.1 Identification
Identification of an exogenous shock to the interest rate in this model is achieved via standard zero restrictions on contemporaneous responses, i.e. a Cholesky decomposition of Ω. The chosen identification scheme largely follows Eller et al. (2021) and is summarized in table 3.
Identification scheme defining zero impact restrictions | |||||
Global
variables |
MPPI |
Slow
macrofinancial variables |
STIR |
Fast
macrofinancial variables |
|
Global variables | x | 0 | 0 | 0 | 0 |
MPPI | x | x | 0 | 0 | 0 |
Slow macrofinancial
variables |
x | x | x | 0 | 0 |
STIR | x | x | x | x | 0 |
Fast macrofinancial
variables |
x | x | x | x | x |
Source: Authors’ compilation. | |||||
Note: Bold letters
indicate a vector of multiple variables. Global variables are the oil
price and the WUI. "MPPI" is the intensity-adjusted
macroprudential policy index. Slow macrofinancial variables include real GDP growth and inflation. "STIR" represents the short-term interest rate to account for the impact of monetary policy. Fast macrofinancial variables are the long-term interest rate (LTIR), the real-effective exchange rate (REER) and private sector credit growth. |
It implies that global variables (the oil price and the WUI) are exogenous, that the MPPI as a slow-moving policy variable does not react instantaneously to macroeconomic changes in the period of the STIR shock due to legislation and implementation lags (as also argued in Meeks, 2017) and that slower-moving macrofinancial variables (economic growth and inflation) do not react contemporaneously to monetary policy measures (i.e. changes in the STIR). Moreover, while macroprudential policy is contemporaneously exogenous to monetary policy (again motivated by longer legislation and implementation lags), MP may react on impact to MPP conditions (for a similar reasoning, see Kim and Mehrotra, 2018; and Kim et al., 2019). Finally, faster-moving macrofinancial variables like the real effective exchange rate (REER), the LTIR or credit growth react on impact to all other variables in the system (as motivated by Creel and Levasseur, 2005).
2.2 Prior setup
To manage a potentially large parameter space, we rely on shrinkage in the form of a Minnesota-style prior. It assumes that most of the parameters are close to zero but allows for deviations if the data indicate a necessity. This allows us to shrink the number of parameters we need to estimate in a data-driven way and makes the model more manageable and reliable. 11 Combining these priors with the likelihood of the model results in mostly standard conditional posterior distributions. Our sampling strategy follows Hauzenberger et al. (2021), to which we refer the interested reader for further technical details on the sampling algorithm.
3 Empirical results
In the following, we will describe our empirical findings and discuss in detail the interaction of MP (interest rates) and MPP in CESEE. Our first set of results concerns the state allocation resulting from our model. We then further describe how the effects of STIR shocks vary depending on the level of macroprudential regulation in an economy.
3.1 Resulting state allocation
In chart 2 , we illustrate the regime indicator variable, the MPPI (represented by solid black lines on the left-hand axis), alongside the corresponding posterior median of the state allocation (St, depicted with blue solid lines on the right-hand axis). This representation is shown separately for countries with fixed and flexible exchange rate regimes, in the upper and lower panels, respectively. The lower and upper horizontal dashed lines allow us to illustrate which observations of the MPPI are associated with a comparatively loose or tight macroprudential policy regime, respectively. 12
We observe that the main dynamics of the MPPI are closely mirrored in the St indicator. In most countries, both the MPPI and the St exhibit limited volatility throughout the sample. Notably, a continuous upward trend is apparent, for instance, in Poland, Hungary, Czechia and Slovakia (starting from 2015 in the latter two cases), which reflects developments in the MPPI that were pointed out above. Additionally, there are periodic shifts toward the extremes (St values close to zero or one) in countries such as Estonia and Lithuania. Overall, however, we can observe that a majority of countries is characterized by extended periods where the St values range between zero and one. This substantiates our choice of the ST-VAR model. Unlike traditional threshold models, which require distinctly separated regimes of low or high state allocation values, the ST-VAR approach also benefits from the information gained during periods of moderate indicator values and allows for the identification of a continuum of regimes.
Chart 2
3.2 Monetary policy effects under different macroprudential conditions
Charts 3 to 6 display the country-specific responses of selected domestic variables in our system to a one standard deviation tightening shock in STIRs, measured one and four quarters after impact, respectively. 13 The impulse responses are summarized in boxplots, with the solid line indicating the median response, the shaded box representing the middle 50% of the posterior distribution, and the whiskers marking the 68% posterior credible set. Responses are reported separately for a tight (T) and loose (L) macroprudential policy stance. In addition, we show the difference in responses between the tight and loose MPP regime (D).
We start in charts 3 and 4 with the results for the countries with flexible exchange rates (Poland, Hungary, Czechia and Romania). First, it becomes apparent that within a relatively loose MPP environment, the impact of monetary policy tightening tends to linger longer, as indicated by positive median responses of the STIR. This suggests that the effectiveness of monetary policy tightening might be partially mitigated by the more lenient MPP stance, potentially necessitating a prolonged period of tighter monetary policy. In Czechia, persistence in the monetary policy tightening shock can also be observed within a tight MPP environment.
Chart 3
Turning to the MPP response, we observe predominantly negative median responses, particularly in cases of a tight MPP regime. This suggests that monetary policy tightening might be partially offset by subsequent macroprudential easing, corroborating the findings of Kim et al. (2019). Tightening monetary policy within an already tight MPP environment could worsen financing conditions and exacerbate loan repayment pressures. Therefore, some macroprudential easing could alleviate these pressures and allow monetary policy to remain tight for an extended period if needed.
Chart 4
Interestingly, macrofinancial variables most of the time do not exhibit statistically significant responses to monetary policy tightening. This lack of response could potentially be attributed to the counterbalancing effect of MPP. However, there are a few notable exceptions. For instance, under loose MPP conditions, monetary policy tightening leads to a decline in inflation and credit growth (responses of credit growth can be found in the annex) in Czechia, while in Poland, inflation, GDP growth and credit growth increase, and the REER appreciates (the latter only one year after impact).
Shifting the focus to countries with fixed exchange rate (Bulgaria, Estonia, Latvia, Slovenia, Slovakia, Lithuania and Croatia) in charts 5 and 6, the results reveal a more distinct pattern than those for the countries with flexible exchange rates discussed above. We can no longer observe that there is more persistence in the STIR shock in a loose MPP environment. On the contrary, the persistence of the STIR shock is now more strongly pronounced within a tight MPP environment.
Given the constrained flexibility of MP and the limited role of the exchange rate channel for the transmission of interest rate changes in countries with fixed exchange rates, other economic policy areas, such as MPP, may need to align in the same direction to achieve the desired effects. However, responses of the MPPI vary significantly across countries with fixed exchange rates, with both positive and negative responses observed under different MPP environments, which highlights the complexity of policy interactions. Notably, while an easing macroprudential response to positive interest rate shocks was quite robust for countries with flexible exchange rates, it can be observed for countries with fixed exchange rates only in Bulgaria (in both loose and tight MPP regimes), Croatia (in the loose MPP regime) and Lithuania (in the tight MPP regime one year after the shock).
Turning to the response of macrofinancial variables, we again observe a mixed pattern. The responses of GDP growth and inflation are in most cases statistically insignificant, with occasional exceptions such as positive GDP growth responses in Bulgaria and Latvia under specific MPP environments and positive inflation responses in Bulgaria, Latvia and Slovenia. In countries with fixed exchange rates, positive credit growth responses can be more frequently observed than in countries with flexible exchange rates, namely in Bulgaria, Latvia, Slovenia, Lithuania and Croatia. Notably, in Bulgaria and Croatia, the positive response of credit growth following an interest rate tightening shock is accompanied – and possibly triggered – by contemporary macroprudential easing. The muted response of macrofinancial variables in countries with fixed exchange rates is not surprising since, in these countries, interest rate shocks can propagate through fewer transmission channels (Mishra et al., 2012, or Eklou, 2023, for example, argue that the exchange rate channel can be undermined in economies with fixed exchange rates).
Chart 5
To summarize, our analysis uncovers notable variations in the effects of MP (interest rate) shocks, influenced by the tightness of MPPs and differences between flexible and fixed exchange rate regimes. In countries with flexible exchange rates, MP shocks tend to persist longer and MPPs often respond with easing measures, suggesting a counterbalancing effect that mitigates some macrofinancial impacts of monetary policy tightening. This pattern is less evident in countries with fixed exchange rates, except if they actively use MPPs, like Bulgaria and Croatia. Across our sample, responses of macrofinancial variables to interest rate shocks are often insignificant, possibly due to the counterbalancing effect of MPPs (as also documented in Imbierowicz et al., 2019; Kim et al. 2019; Acharya et al., 2020; Buch et al., 2022) and the challenges central banks faced in elevating inflation rates with conventional tools in the wake of the GFC – a phenomenon not unique to CESEE. However, there are also some exceptions including counterintuitive responses. “Inflation puzzles” emerge in a few cases, namely in Poland, Bulgaria, Latvia and Slovenia, where we observe positive inflation responses to interest rate tightening. 14 Moreover, “credit puzzles” are evident in some instances such as Poland, Bulgaria, Latvia, Slovenia, Lithuania and Croatia, where positive credit responses are observed. It can be noted that these patterns are more pronounced in countries with fixed exchange rates, which do not have an independent monetary policy and tend to be more susceptible to external shocks, thus often exhibiting greater macroeconomic volatility (as highlighted by Khan, 2017).
Chart 6
4 Conclusions
This paper examined the interplay between monetary and macroprudential policies in 11 CESEE countries, using a ST-VAR model to assess how the economic effects of exogeneous shocks to the nominal short-term interest rate are influenced by the stringency of the macroprudential policy environment.
We find that macroprudential policies often moderate the impacts of interest rate adjustments, particularly in countries with flexible exchange rates. In these settings, macroprudential easing frequently offsets the effects of monetary policy tightening, especially in an already tight macroprudential environment, alleviating loan repayment pressures and enabling monetary policy to remain stringent for an extended period if necessary. The observation that macroprudential easing – in response to monetary policy tightening – is more feasible when macroprudential policies have already reached a relatively tight stance can be attributed to the political costs associated with macroprudential tightening, which makes it more viable to ease restrictions after a sufficiently stringent framework has been established.
Conversely, in countries with fixed exchange rates, the interaction between interest rate shocks and macroprudential policies is less straightforward. The constraints imposed by fixed exchange rates limit the scope of monetary policy actions, necessitating more nuanced macroprudential interventions. In such environments, we observe that macroprudential policies need to be particularly well calibrated to ensure they complement any changes in interest rates effectively.
The heterogeneity of varying impacts across different regimes suggests that a one-size-fits-all policy approach may not be appropriate. Our results suggest, however, that policymakers can learn about the speed and strength of monetary policy transmission when taking the macroprudential policy stance into account. Even outside CESEE, monetary policymakers should consider the specific macroprudential environment and exchange rate arrangement when designing and implementing policy measures, as these factors can significantly influence the success and repercussions of such policies. Possible complementarities and conflicts between monetary and macroprudential policies highlight the importance of considering the state of business and financial cycles when making policy decisions. Furthermore, effective communication between monetary and macroprudential authorities are essential to align policies and manage potential conflicts, which is particularly important in the euro area with its diverse economic landscape.
Our findings contribute to the broader discourse on optimal policy mixes, particularly in regions with more volatile macrofinancial environments such as CESEE. As CESEE countries continue to develop and integrate further into the global financial system, the insights from this analysis can be valuable not only for local policymakers but also for those in similar emerging markets. Future research could build on this foundation by exploring the long-term effects of policy interactions or assessing how the response to monetary policy tightening varies depending on the stringency of different macroprudential tools (e.g. capital-based versus borrower-based measures). Such studies would provide deeper insights into the mechanics of monetary policy effectiveness across diverse macroprudential policy contexts.
5 References
Acharya, V. V., B. Imbierowicz, S. Steffen and D. Teichmann. 2020. Does the Lack of Financial Stability Impair the Transmission of Monetary Policy? Journal of Financial Economics 138 (2). 342–365.
Ahir, H., N. Bloom and D. Furceri. 2022. World Uncertainty Index. NBER Working Paper No. w29763. https://ssrn.com/abstract=4039482
Albertazzi, U., A. Martin, E. Assouan, O. Tristani, G. Galati, and T. Vlassopoulos. 2021. The role of financial stability considerations in monetary policy and the interaction with macroprudential policy in the euro area. Occasional Paper Series 272. European Central Bank. https://ideas.repec.org/p/ecb/ecbops/2021272.html
Alessandri, P. and H. Mumtaz. 2019. Financial Regimes and Uncertainty Shocks. Journal of Monetary Economics 101. 31–46.
Angeloni, I. and E. Faia. 2013. Capital Regulation and Monetary Policy with Fragile Banks. Journal of Monetary Economics 60 (3). 311–324. https://doi.org/10.1016/j.jmoneco.2013.01.003
Auerbach, A. J. and Y. Gorodnichenko. 2012. Measuring the Output Responses to Fiscal Policy. American Economic Journal. Economic Policy 4 (2). 1–27.
Backé, P. and S. Dvorsky. 2018. Enlargement of the euro area toward CESEE: progress and perspectives. In: OeNB. Focus on European Economic Integration Q3/18. 43–56. https://ideas.repec.org/a/onb/oenbfi/y2018iq3-18b4.html
Belhocine, N., E. Crivelli, N. Geng, T. Scutaru, J. Wiegand and Z. Zhan. 2016. Taking Stock of Monetary and Exchange Rate Regimes in Emerging Europe. IMF European Department, Departmental Paper No. 2016/012. https://www.imf.org/external/pubs/ft/dp/2016/eur1603.pdf
Benecká, S., L. Fadejeva and M. Feldkircher. 2020. The Impact of Euro Area Monetary Policy on Central and Eastern Europe. Journal of Policy Modeling 42 (6). 1310–1333. https://doi.org/ https://doi.org/10.1016/j.jpolmod.2020.05.004
Buch, C. M., S. Eickmeier and E. Prieto. 2022. Banking Deregulation, Macroeconomic Dynamics and Monetary Policy. Journal of Financial Stability 63. 101057. https://doi.org/ https://doi.org/10.1016/j.jfs.2022.101057
Caggiano, G., E. Castelnuovo and N. Groshenny. 2014. Uncertainty Shocks and Unemployment Dynamics in US Recessions. Journal of Monetary Economics 67. 78–92.
Casiraghi, M., K. Habermeier and T. Harjes. 2022 . Monetary Policy Frameworks – Choice of Exchange Rate Arrangement. IMF Monetary and Capital Markets Department Technical Assistance Handbook by the Monetary and Macroprudential Policies Division (MCMMP). https://www.imf.org/-/media/Files/Publications/Miscellaneous/English/2022/mcm-technical-assistance-handbook/choice-of-exchange-rate-arrangement.ashx
Chan, J. C. C. 2021. Minnesota-Type Adaptive Hierarchical Priors for Large Bayesian VARs. International Journal of Forecasting 37 (3). 1212–1226.
Chen, J., D. Finocchiaro, J. Lindé and K. Walentin. 2020. The Costs of Macroprudential Deleveraging in a Liquidity Trap. CEPR Discussion Papers 14564. https://EconPapers.repec.org/RePEc:cpr:ceprdp:14564
Creel, J. and S. Levasseur. 2005. Monetary policy transmission mechanisms in the CEECs: How important are the differences with the euro area? Documents de Travail de l’OFCE 2005-02. Observatoire Français des Conjonctures Économiques (OFCE). https://ideas.repec.org/p/fce/doctra/0502.html
Darvas, Z. 2009. Monetary Transmission in Three Central European Economies: Evidence from Time-Varying Coefficient Vector Autoregressions. Working Papers 0903. Department of Mathematical Economics. Economic Analysis. Corvinus University of Budapest. https://ideas.repec.org/p/mkg/wpaper/0903.html
De Jonghe, O., H. Dewachter and S. Ongena. 2020. Bank Capital (Requirements) and Credit Supply: Evidence from Pillar 2 Decisions. Journal of Corporate Finance 60. 101518. https://doi.org/ https://doi.org/10.1016/j.jcorpfin.2019.101518
Du, C. and D. Miles. 2014. Interaction Between Monetary Policy and Regulatory Capital Requirements. CEPR Discussion Paper DP10200.
Eklou, K. M. 2023. The Anatomy of Monetary Policy Transmission in an Emerging Market. IMF Working Papers WP/23/146.
Eller, M., N. Hauzenberger, F. Huber, H. Schuberth and L. Vashold. 2021. The Impact of Macroprudential Policies on Capital Flows in CESEE. Journal of International Money and Finance 119. 102495. https://doi.org/10.1016/j.jimonfin.2021.102495
Eller, M., R. Martin, H. Schuberth and L. Vashold. 2020. Macroprudential Policies in CESEE–an Intensity-Adjusted Approach. In: OeNB. Focus on European Economic Integration Q2/20. 65–81.
Feldkircher, M. 2015. A Global Macro Model for Emerging Europe. Journal of Comparative Economics 43 (3). 706–726. https://doi.org/10.1016/j.jce.2014.09.002
Franta, M., T. Holub and B. Saxa. 2022. Exiting from an Exchange Rate Floor in a Small Open Economy: Balance Sheet Implications of the Czech National Bank’s Exchange Rate Commitment. International Journal of Central Banking 73. June.51–105. https://www.ijcb.org/journal/ijcb22q2a2.htm
Gefang, D. and R. Strachan. 2009. Nonlinear Impacts of International Business Cycles on the UK – a Bayesian Smooth Transition VAR Approach. Studies in Nonlinear Dynamics & Econometrics 14 (1).
Hauzenberger, N., M. Pfarrhofer and A. Stelzer. 2021. On the Effectiveness of the European Central Bank’s Conventional and Unconventional Policies Under Uncertainty. Journal of Economic Behavior & Organization 191. 822–845. https://doi.org/10.1016/j.jebo.2021.09.041
Holcner, V. and J. Neubauer. 2015. Modelling Expected Defense Specific Inflation. 1648. March. https://doi.org/10.1063/1.4912947
Huber, F. and T. O. Zörner. 2019. Threshold Cointegration in International Exchange Rates: A Bayesian Approach. International Journal of Forecasting 35 (2). 458–473.
Imbierowicz, B., A. Löffler and U. Vogel. 2019. The transmission of bank capital requirements and monetary policy to bank lending. Discussion Papers 49/2019. Deutsche Bundesbank. https://ideas.repec.org/p/zbw/bubdps/492019.html
Jarocinski, M. 2010. Responses to Monetary Policy Shocks in the East and West of Europe: A Comparison. Journal of Applied Econometrics 25. 833–868. https://doi.org/10.1002/jae.1082
Josifidis, K., J. P. Allegret and E. Beker Pucar. 2009. Monetary and Exchange Rate Regimes Changes: The Cases of Poland, Czech Republic, Slovakia and Republic of Serbia. Panoeconomicus 56 (2). 199–226. https://ideas.repec.org/a/voj/journl/v56y2009i2p199-226id215.html
Khan, N. 2017. Monetary Regimes and External Shocks Reaction: Empirical Investigations on Eastern European Economies. The Romanian Economic Journal 20 (66). 63–81.
Kim, J., S. Kim and A. Mehrotra. 2019. Macroprudential Policy in Asia. Journal of Asian Economics 65. 101149.
Kim, S. and A. Mehrotra. 2018. Effects of Monetary and Macroprudential Policies – Evidence from Four Inflation Targeting Economies. Journal of Money, Credit and Banking 50 (5). 967–992.
Meeks, R. 2017. Capital Regulation and the Macroeconomy: Empirical Evidence and Macroprudential Policy. European Economic Review 95. 125–141.
Mishra, P. and P. Montiel. 2012. How Effective is Monetary Transmission in Low-Income Countries? A Survey of the Empirical Evidence. IMF Working Papers WP/12/143.
Mojon, B. and G. Peersman. 2001. A VAR description of the effects of monetary policy in the individual countries of the euro area. Working Paper Series 92. European Central Bank. https://ideas.repec.org/p/ecb/ecbwps/200192.html
Nier, E. W. and H. Kang. 2016. Monetary and Macroprudential Policies – Exploring Interactions. BIS Paper 86e.
Ross, K. 2000. Post-stabilization inflation dynamics in Slovenia. Applied Economics 32 (2). 135–149. https://doi.org/10.1080/000368400322831
Shambaugh, J. C. 2004. The Effect of Fixed Exchange Rates on Monetary Policy. The Quarterly Journal of Economics 119 (1). February. 301–352. https://doi.org/10.1162/003355304772839605
Shevchuk, V. 2020. Monetary Policy Transmission Mechanisms in the Central and Eastern European Countries. Argumenta Oeconomica Cracoviensia. January. 51–68. https://doi.org/10.15678/AOC.2020.2303
Takáts, E. and J. Temesvary. 2019. Can Macroprudential Measures Make Cross-Border Lending More Resilient? Lessons from the Taper Tantrum. International Journal of Central Banking 15 (1). 61–105. https://ideas.repec.org/a/ijc/ijcjou/y2019q1a2.html
Wlodarczyk, P. 2017. Monetary Policy Transmission and the Labour Market in the Non-Eurozone Visegrad Group Countries in 2000-2014: Evidence from a SVAR Analysis. Comparative Economic Research. Central and Eastern Europe 20. 23–43. https://doi.org/doi:10.1515/cer-2017-0026
Zhang, Y. and T. Tressel. 2017. Effectiveness and channels of macroprudential policies: lessons from the Euro area. Journal of Financial Regulation and Compliance 25 (3). 271–306. https://ideas.repec.org/a/eme/jfrcpp/jfrc-10-2016-0094.html
6 Annex
The annex can be found in the pdf version of this study.
-
Oesterreichische Nationalbank, International Economics Section, clara.deluigi@oenb.at (corresponding author); Monetary Policy Section, anna.stelzer@oenb.at ; Joint Vienna Institute (JVI), meller@jvi.org . Opinions expressed by the authors of studies do not necessarily reflect the official viewpoint of the OeNB, the Eurosystem or the JVI. The authors would like to thank an anonymous referee and Katharina Allinger, Thomas Reininger, Fabio Rumler, María Teresa Valderrama and Julia Wörz (all OeNB) for helpful comments and valuable suggestions. We further thank Oskar Baackmann (Deutsche Bundesbank University of Applied Sciences) and Zoltan Walko (OeNB) for research assistance and data support. ↩︎
-
In such situations, MP and MPP can reinforce each other or be used to offset each other’s side effects; MP can mitigate the negative effects of MPP tightening on output, and MPP can counteract the side effects of MP on financial stability. ↩︎
-
Note that Eller et al. (2021) also refer to “policy rates” for the same set of CESEE countries, including countries which operate under a peg or currency board. Similarly, numerous studies analyze the monetary policies of countries before they adopted the euro, reflecting the importance of interest rate dynamics as a transmission mechanism also in a transitioning monetary policies and in fixed exchange rate regimes (see Mojon and Peersman, 2001, and Jarockinski, 2010, among others). Literature on emerging markets in Latin America and Asia also discuss how interest rates act as conduits for monetary shocks, even under fixed exchange rate systems (see Shambaugh, 2004, and Casiraghi et al., 2022). ↩︎
-
For a review of the enlargement of the euro area toward CESEE during 2010–2018, see Backé and Dvorsky (2018). ↩︎
-
While Croatia’s official exchange rate arrangement was a managed float, the Croatian kuna stabilized within a 2% band against the euro for most of 2006–2019. Prior to joining the ERM II, Slovakia and Slovenia were implicitly already practicing inflation targeting, as discussed in Josifidis et al. (2009) and Ross (2000). ↩︎
-
We consider countries that joined the euro area countries with fixed exchange rates as they adopted the euro in the middle of our sample and mostly transitioned from exchange rate targeting or currency boards to monetary union. ↩︎
-
The MPPI integrates capital-, liquidity- and borrower-based measures, along with minimum capital and reserve requirements. It weighs and aggregates these measures based on their implementation impact and legal status, differentiating between recommendations and binding regulations and between implementation and announcement dates. For more details, see Eller et al. (2020). ↩︎
-
Note that in order to simplify notation, we omit a country-specific indexation in the model, but besides the two global quantities (the oil price and the WUI), all variables, coefficients and other model ingredients are specific to one country. Further technical details and notation are provided in the annex. ↩︎
-
In our model, we chose p=3 to control for dynamics in past values of the variables. Our results are qualitatively robust to lag choices between one and four, which are standard choices for models using quarterly data. ↩︎
-
Further modeling choices such as the transition function or prior choices for the Bayesian estimation techniques are elaborated upon in the annex. ↩︎
-
Further details can be found in the annex. ↩︎
-
It should be noted that the MPPI was constructed to start with a value of zero in each country in the mid-1990s, a period when MPP measures were not yet implemented. Therefore, we can reasonably assume that the initial value of the MPPI represents a neutral MPP stance. Under this assumption and considering that the same weighting rules were applied to subsequent changes across all countries, we can compare the levels of policy tightness also across different countries. ↩︎
-
As a lot of responses of macrofinancial variables are statistically insignificant, we focus here on significant responses and provide responses of all variables in the annex. ↩︎
-
Similar “price puzzles” in response to MP shocks in CESEE countries have been documented in the literature (see, for instance, Creel and Levasseur, 2005; Darvas, 2009; Holcner and Neubauer, 2015; Wlodarczyk, 2017; Shevchuk, 2020). ↩︎