1. Introduction and Motivation
“International spillovers from the monetary policy of one country to other economies are a corollary of globalisation. This entails that we, as policymakers, have to rise to the challenge of conducting monetary policy in the presence of these unintended side-effects. […] We should not underestimate the challenges of living in the ever more closely interconnected global economy.”
Vítor Constâncio, Vice President of the ECB, Hong Kong, 15 October 20152
Over the past years, there has been a discussion in academia and among policymakers on the international effects (that is, spillovers) of advanced economies’ standard and in particular non-standard monetary policy in response to the global financial crisis and its legacy. Originally, the focus of interest was spillovers originating from the Federal Reserve’s monetary policy, but more recently the international effects of the European Central Bank’s (ECB) standard and non-standard monetary policy measures have gained increased attention. Since August 2007, and, in particular after October 2008, the ECB has introduced a number of non-standard monetary policy measures, which are unprecedented in nature, scope, and magnitude; and have included significant changes in the operational framework and large bond purchasing programs.
Spillovers from the ECB’s non-standard monetary policy on Southeastern Europe (SEE) are particularly relevant because the SEE countries are located in the direct neighborhood of the euro area and are already EU members (Bulgaria, Croatia, and Romania) or have the prospect of joining the EU in the future (Albania, Bosnia and Herzegovina, the Former Yugoslav Republic of Macedonia, Montenegro, and Serbia3). Moreover, the economies of the SEE countries are closely interlinked with the euro area through several channels that could potentially transmit euro area monetary policy shocks: first, the significant presence of euro-area-headquartered banks and a high degree of euroization constitute a financial or banking transmission channel of monetary policy shocks. Second, high trade integration and sizeable remittances flows could act as real channels for shock transmission.
Additionally, the heterogeneous monetary policy frameworks and exchange rate regimes of SEE countries provide an interesting case for cross-country comparisons with regard to the role of exchange rate regimes in shaping spillovers: in SEE they range from inflation targeters with (managed) floating exchange rates (Albania, Romania, and Serbia), via stabilized arrangements with the euro as a reference currency (Croatia and the former Yugoslav Republic of Macedonia) and euro-based currency boards (Bosnia and Herzegovina and Bulgaria) to the unilateral adoption of the euro as the sole legal tender (Montenegro).
2. A Framework for Estimating Spillovers
A major challenge for estimating spillovers of non-standard monetary policy is the measurement of the degree of monetary accommodation provided through non-standard monetary policy measures, since they are typically not captured by the key monetary policy rate. Therefore, alternative indicators have to be found.
For the purpose of estimating spillovers from the ECB’s non-standard monetary policy measures on SEE, I mostly resort to the size of the Eurosystem balance sheet, which mirrors the liquidity provision of the Eurosystem and, thus, reflects most non-standard monetary policy measures that were introduced by the ECB. The Eurosystem balance sheet assets started to increase notably with the introduction of non-standard monetary policy measures in October 2008, and, thereafter increased further significantly as a result of the various instruments used, with some variations over time owed to their nature and set up (Figure 1).
Key policy rate (in percent) and inverted Eurosystem balance sheet assets
(in EUR trillion)
Note: The key policy rate is the rate of main refinancing operations (MROs).Source: ECB.Key policy rate (in percent) and inverted Eurosystem balance sheet assets
(in EUR trillion)
Note: The key policy rate is the rate of main refinancing operations (MROs).Source: ECB.Key policy rate (in percent) and inverted Eurosystem balance sheet assets
(in EUR trillion)
Note: The key policy rate is the rate of main refinancing operations (MROs).Source: ECB.In order to be able to compare the spillovers across the SEE countries, a separate structural Bayesian Vector Autoregressive (BVAR) model is estimated for each country. Each model includes a number of euro area variables (output, price level, Eurosystem balance sheet assets, a measure of financial stress, the spread between the EONIA and the key monetary policy rate, and the key monetary policy rate as a separate variable) and SEE country variables (output, price level, exports to the euro area, trade partners’ output, interbank interest rate, key monetary policy rate [if applicable] and exchange rate vis-à-vis the euro [if applicable]). The model is estimated in a monthly frequency from January 2008 to December 2015 with four lags. It is assumed that SEE country variables do not influence euro area variables (so-called block exogeneity). The shocks are identified via sign and zero restrictions in order to distinguish a non-standard monetary policy shock from a conventional monetary policy shock and from endogenous balance sheet movements.
3. Results and Cross-country Comparison
As a response to the ECB non-standard monetary policy shock, output rises in approximately half of the countries, with the largest response being observed for Croatia, followed by Serbia and Bulgaria (Figure 2)4. The output of the Albanian, former Yugoslav Republic of Macedonian, and Romanian economies does not respond significantly to the shock. Compared to the transmission of the balance sheet shock within the euro area, the spillovers are sizeable, and in the case of Croatia and Serbia, even larger than in the euro area itself.
Output responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. * denotes not strictly credible responses. ** For Montenegro, the response of industrial production is depicted.Output responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. * denotes not strictly credible responses. ** For Montenegro, the response of industrial production is depicted.Output responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. * denotes not strictly credible responses. ** For Montenegro, the response of industrial production is depicted.Following the euro area shock, all countries exhibit a distinct price level response (Figure 3), where the peak response is by far the strongest for Serbia, followed by the former Yugoslav Republic of Macedonia, Romania, and Bulgaria. The relatively strong price responses, which are more pronounced for around half of the countries than the euro area response itself, are in line with the high share of imports from the euro area that ranges from around one third to over 50 percent of all imports in SEE countries.
Price level responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. * denotes not strictly credible responses.Price level responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. * denotes not strictly credible responses.Price level responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. * denotes not strictly credible responses.Comparing the peak responses across countries, the results suggest that the exchange rate regime does not play a significant role in shaping the size and magnitude of output and price spillovers, as countries with flexible exchange rates lie on opposite ends of the range. This result is not surprising, as for the three countries operating under a (in principle) flexible exchange rate regime, none of the domestic exchange rates shows a pronounced response to the euro area balance sheet shock in the model output5. This result is in line with relatively stable exchange rates (especially for Albania and Romania) in the sample period, reflecting the general reluctance to tolerate exchange rate fluctuations also under the flexible exchange rate regimes adopted, in principle, in Southeastern Europe.
Regarding potential transmission channels, the results suggest that the impact of a nonstandard monetary policy shock on SEE exports to the euro area is the largest for the former Yugoslav Republic of Macedonia and Montenegro, followed by Bulgaria, Serbia, and Albania (Figure 4). On the other hand, the real transmission channel does not seem to be considerable in Bulgaria, where no significant increase of exports can be observed, and in Croatia and Romania, where exports react negatively to the balance sheet shock.
Export responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. *denotes not strictly credible responses.Export responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. *denotes not strictly credible responses.Export responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 12 months in percent. *denotes not strictly credible responses.Taking into account both the financial and the real transmission channels, it seems that for a majority of countries, spillovers are transmitted via exports, while the interest rate channel plays a significant role only in two countries, as suggested by the response of the short-term interbank market rate (Figure 5). This result might reflect the significant bank cross-border deleveraging many SEE countries experienced after the crisis. However, both indicators suffer from drawbacks due to data availability. Exports here denote only merchandise exports, since exports of services are not available, which might underestimate the export channel in countries with a high share of service exports, such as Croatia. In the case of financial flows, changes in the interbank market rate might not capture foreign direct or portfolio inflows, and, therefore, underestimate the role of the financial channel in transmitting shocks.
Interbank interest rate responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 6 months in percentage points.Interbank interest rate responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 6 months in percentage points.Interbank interest rate responses to ECB non-standard monetary policy shock
Source: Author’s calculations.Notes: Peak responses to a 1 standard deviation expansionary balance sheet shock (approx. +0.7%) within the first 6 months in percentage points.4. Conclusions
The results suggest that the ECB’s non-standard monetary policy measures have had sizeable spillovers to the SEE countries: while output responds positively to a Eurosystem balance sheet shock in approximately half of the SEE countries, pronounced price level effects can be found for all countries, potentially driven by import prices given the high euro area import share. With regard to possible transmission channels, the export channel seems to be more important than the financial channel (represented by interbank interest rates). Moreover, the results suggest that the exchange rate regime does not influence the magnitude and sign of the shock response. Whether spillovers are beneficial to the economy largely depends on the business cycle position of the respective SEE country and, more generally, on business cycle synchronization with the euro area. The current work could be extended into various directions, for example comparing nonstandard with conventional monetary spillovers, add more potential transmission channels, or compare the effects of a US monetary policy shock with a euro area monetary policy shock.
This contribution is based on Moder, I. (2017), “Spillovers of the ECB’s non-standard monetary policy measures on Southeastern Europe,” ECB Working Paper No 2095, Frankfurt am Main. The views expressed are those of the author and do not necessarily reflect those of the ECB or of the ESCB.
The full speech can be accessed under: http://www.ecb.europa.eu/press/key/date/2015/html/sp151015.en.html
While the potential EU candidate Kosovo* is also located in SEE, it cannot be included in the analysis due to limited data availability. *This designation is without prejudice to positions on status and is in line with UNSCR 1244 and with the ICJ Opinion on the Kosovo Declaration of Independence.
Montenegro’s output also responds positively, but is measured in industrial production, thus the effect on GDP is unknown.
The peak exchange rate responses are not shown here due to the weak and noncredible responses. Impulse response functions of the exchange rate reaction can be found in the Working Paper version.
