By Philipp Hartmann, Deputy Director General Research, ECB; and CEPR Research Fellow, Philippe Molitor, Principal Financial Stability Expert, ECB, Annachiara Tanzarella, Student Research Assistant, Directorate General Research, ECB, and Bruun de Jong, Student Research Assistant, Directorate General Research, ECB. Originally published at VoxEU.
Russia’s invasion of Ukraine added another one to a series of crises affecting Europe over the last decade and a half. This column finds that fragmentation of euro area financial markets has been relatively limited, as policymakers have benefitted from reforms after and experiences with past crises, European countries and authorities reacted with quite some unity, and direct financial exposures to Russia and Ukraine are relatively small. Cross-country divergences of euro area asset prices appear to have been particularly driven by bond markets and related to the countries with the greatest energy and financial exposures as well as to changes in economic uncertainty.
The European Economic and Monetary Union (EMU) has weathered several severe crises, notably during the last one and a half decades. The crises usually went hand in hand with reductions of financial integration, often referred to as financial fragmentation (see Adam et al. 2002, Baele et al. 2004, ECB 2007 and follow-up reports, European Parliament 2016, FSB 2019, and Claessens 2019 for broad discussions of financial integration/fragmentation and relevant literature). This can be seen in Figure 1, which shows two aggregate indicators of financial integration – one based on cross-border price differentials in key financial markets (price-based composite indicator, the blue line) and another based on cross-border asset holdings (quantity-based composite indicator, the yellow line). For example, the Great Financial Crisis and the European sovereign debt crises together brought measured financial integration down to levels similar to the times of the euro’s introduction. The reasons for these centrifugal forces are multifaceted, but two particularly important and interrelated ones are that EMU has been constructed bottom-up from single sovereign countries where the unification of monetary policy was not accompanied by a commensurate fiscal union. As member countries have differing fiscal and economic strengths, they do not have the same ability to stabilise during crises.
Figure 1 Low-frequency price-based and quantity-based composite indicators of euro area financial integration
(quarterly data, price-based indicator: Q1 1995 – Q1 2022, quantity-based indicator: Q1 1999 – Q1 2022)
Sources: ECB and ECB calculations
Notes: The price-based composite indicator aggregates ten indicators for money, bond, equity and retail banking markets, measuring cross-border asset price or yield convergence, while the quantity-based composite indicator aggregates five indicators for the same market segments (except retail banking), measuring cross-border asset holdings. The quarterly values of the price-based composite indicator are computed as the average of the monthly values of the respective quarter while the values of the quantity-based composite indicator refer to the end of the respective quarter. The indicators are bounded between zero (full fragmentation) and one (full integration). Increases in the indicators signal greater financial integration, i.e. greater cross-border asset price or yield convergence and greater cross-border asset investment within the euro area, respectively. From January 2018 onwards the behaviour of the price-based indicator may have changed due to the transition from EONIA to €STR interest rates in the money market component. Detailed descriptions of both indicators and their input data are in the Statistical Web Annex of ECB (2022) and Hoffmann et al. (2019). OMT stands for Outright Monetary Transactions.
As a consequence of each crisis, new institutional arrangements and policy tools have been introduced in the fiscal, monetary, prudential, and structural policy areas (see Hartmann and Smets 2018, Rancoita et al. 2020, Tordoir et al. 2020, Afman et al. 2021, Buti 2021, Freier et al. 2022, ECB 2022b, 2022c, as well as ESRB 2022 for broad overviews). They implied steps towards the completion of EMU and financial integration recovered, though more in prices than in quantities (see the right-hand side of Figure 1). When the COVID-19 crisis hit (penultimate red vertical line to the right), a sharp re-fragmentation set in again. But it turned out to be less severe and more short-lived than the previous episodes, notably as powerful innovations in policy were made again (e.g. Hartmann et al. 2021, ECB 2022a).
A ‘Clinical’ Analysis Of Euro Area Financial Integration During The Early Phase Of The Ukraine War
From the extreme right of Figure 1, the impact of the Ukraine crisis on euro area financial integration looks relatively contained as well (with quarterly data). Figure 2 provides a detailed picture of euro area financial integration developments with daily data (blue line) and relevant events (vertical lines with capital letters) during the first three months of the war. (This high-frequency indicator is only available for price data.)
Figure 2 High-frequency price-based composite indicator of euro area financial integration and relevant events during the Ukraine crisis
(daily data, 10 January 2022 – 23 May 2022)
Sources: ECB and ECB calculations
Notes: The high-frequency price-based composite indicator of financial integration is a variant of the low-frequency price-based indicator developed by Hoffmann et al. (2019), as shown in Figure 1, at daily frequency rather than monthly or quarterly frequency. While covering the same market segments its focus on high-frequency data implies that it can incorporate only a lower number of indicators overall. Its construction is described in Kochanska, Mulder and Zito (2020). Black vertical lines with capital letters mark selected relevant events:
A. U.S. and Russia accuse each other at U.N. Security Council meeting of stoking Ukraine crisis (31 January 2022); B. Russia and Belarus start large military drills close to the Ukrainian border (10 February 2022); C. Russian military announces pulling back some troops close to Ukraine (15 February 2022), D. Fighting escalates in separatist regions of eastern Ukraine (17 February 2022); E. U.S. President Biden and Secretary of State warn that Russia may be at the brink of invading Ukraine (17 February 2022); F. Russian invasion of Ukraine and agreement of EU leaders on a second sanctions package against Russia (24 February 2022); G. European countries (third package) and U.S. agree on further sanctions, including exclusion of some Russian banks from SWIFT and blocking the Russian central bank from access to its foreign reserves (26 February 2022); H. European Union said to consider joint bond sales to finance higher energy and defence spending emerging from Russia’s invasion of Ukraine as well as US and UK ban on Russian oil imports (8 March 2022); I. ECB accelerates reduction in net asset purchases against the background of continued high inflation and Ukraine not any longer pressing for NATO membership as well as signalling openness to neutrality with security guarantees (10 March 2022); J. Russia and Ukraine hint at progress in peace negotiations (16 March 2022); K. EU leaders announce that they will jointly buy gas, liquefied natural gas and hydrogen this year (21 March 2022); L. President Putin’s decree demanding foreign buyers to pay for Russian gas in roubles or have their supplies cut (1 April 2022); M. Euro area economic sentiment reaches an unexpected one-year low, German Finance Minister opens up to an oil embargo against Russia and euro area sovereign yields/spreads edge up in the run-up of US Fed and Bank of England monetary policy meetings (2 May 2022); N. Following a Bank of England rate increase and announcement to start (as the first major central bank) actively selling corporate bonds in September, following several ECB Governing Council Members’ comments on the ECB’s future path of rates and following sizeable falls in German industrial production and factory orders, the Italian sovereign spread and corporate credit risk in Europe reach two-year highs (5 May 2022).
First, the time at which integration began to be affected could be dated to end-January 2022, when a stand-off between Russia and the US happened in the United Nations security council as to whether Russia planned to attack Ukraine (event A in Figure 2), or on the first day of the invasion when EU leaders also agreed on a second package of sanctions against Russia (24 February 2022; event F). Second, right after the invasion and the first sanctions the price-based indicator signals sharp fragmentation, although not particularly large compared to the previous episodes shown in Figure 1. Third, after a series of news about a relatively united European response to the crisis and about signs of a moderation of the conflict (events H, I, J, and K in Figure 2), more than half of the initial fragmentation reversed (see also De Guindos 2022). Fourth, this re-integration trend stopped with President Putin’s decree requiring foreign buyers to pay for Russian gas in roubles (event L; see also Zagrandi et al. 2022). After that (until the end of our data period on 23 May 2022), measured financial integration across euro area countries hovered below the pre-crisis level but above the trough during the days following the invasion.
It is interesting to see that events pointing to a worsening of the conflict (public arguments of politicians, military escalations or sanctions/retaliation; e.g. events A, B, D, E, F, G, H, L and M) seem to drive the euro area integration indicator down, while moderation of the conflict (peace talks, softening of negotiation positions or military de-escalation; e.g. events C, I and J) and unified European approaches to deal with the economic consequences (potential joint debt issuance or joint energy purchases; events H and K) seem to drive it up again.
Figure 3 High-frequency comparison of price-based financial fragmentation in the euro area before and after the beginning of the Ukraine and COVID crises
(daily data, brown solid line: 13 January – 25 April 2022, brown dashed line: 20 December 2021 – 28 March 2022, blue line: 10 January – 21 April 2020)
Sources: ECB and ECB calculations
Notes: The high-frequency price-based composite indicator is described in the notes to Figure 2. On the horizontal axis 0 is the day of the beginning of the crisis, with negative numbers indicating days before and positive numbers days after it. The beginning of the COVID crisis is dated 21 February 2020 (blue line). For the Ukraine crisis two alternative starting dates are displayed, 31 January 2022 (dashed brown line) and 24 February 2022 (solid brown line).
Figure 3 compares the initial fragmentation at the beginning of the COVID crisis with the initial fragmentation at the beginning of the Ukraine crisis. Irrespective of whether one dates the latter to the time of the escalation in the ‘war of words’ or the time of the effective invasion, its financial integration implications for the euro area remained clearly more contained than for the COVID crisis. In sum, while one has to be cautious that the Great Financial Crisis, the European sovereign debt crisis, the COVID crisis, and now the Ukraine crisis are of very different nature and magnitude, our integration measures suggest that the reforms making steps towards EMU completion and the adoption of new policy tools by various authorities have been increasing resilience of euro area financial integration over time.
An Exploratory Analysis Of Further Factors Accounting For Financial Fragmentation During The Ukraine Crisis
Cross-country asset price or yield differentials, as captured in the previous three figures (various blue and brown lines for the price-based indicators), can be explained by differences in the economic fundamentals that influence the valuation of assets, differences in market expectations about the future development of fundamentals and associated risk premiums or non-fundamental market dynamics/mispricing (e.g. from the dissemination of inaccurate information, herding behaviour, self-fulfilling expectations, market panics or illiquidity). The following figures illustrate some of the factors that seem to have driven the divergence of asset prices across euro area countries during the first three months of the Ukraine crisis. All relate to asset prices included in or affecting the price-based indicator of financial integration displayed in Figures 1, 2 and 3.
One of the major factors how the Ukraine war influences other countries is through energy imports (e.g. Verwey et al. 2022). Particularly European countries import a lot of gas and oil from Russia, albeit to differing degrees. When Russia limited deliveries in response to sanctions, prices skyrocketed. This accelerated inflation, constrained growth, and triggered fiscal expenditures to contain the impact on firms and households in euro area countries. Countries with higher energy dependence on Russia were affected more than countries who use other energy sources.
Figure 4 illustrates whether differing energy imports from Russia were reflected in euro area countries’ sovereign spreads during the main financial fragmentation period at the start of the war (23 February to 1 March 2022; recall the sharp decline in Figure 2 during this period). Two channels through which this could be the case is through higher public expenditures (as mentioned above) and lower tax revenues from lower growth, both of which would increase fiscal deficits and thereby tend to lower the creditworthiness of countries.
There seems to be one group of countries with low energy exposure to Russia (blue dots in Figure 4) for which no clear relationship is visible. But for the most-exposed countries (above a level of 1% of total non-euro area imports; red dots), greater dependence tended to imply a greater increase of their sovereign spread against Germany (also illustrated by a dotted red regression-fitted line for these countries). Interestingly, in the subsequent period of re-integration (see the increase of the price-based integration indicator in Figure 2 from event H to event L) the relationship flips sign, meaning that the highly exposed countries benefited from larger spread declines compared to the less-exposed countries (not shown in Figure 4). (Similar relationships hold for sovereign yields rather than spreads, where Germany is in the blue group of countries.) Obviously, given the low number of observations, all these correlations should be interpreted cautiously.
Figure 4 Euro area countries’ energy dependence from Russia and their sovereign spreads during the initial fragmentation period at the beginning of the Ukraine war
(vertical axis: change between 23 February and 1 March 2022 in basis points; horizontal axis: 2021)
Sources: Refinitiv, Eurostat and ECB calculations
Notes: Sovereign spreads are for 10-year government bond yields relative to Germany. Countries with energy exposure to Russia below 1 per cent are marked in blue and countries above 1 per cent in red. Dotted lines are fitted with simple bivariate regressions including an intercept. The black dotted line incorporates all 15 euro area countries for which reliable data were available (blue and red dots). The red dotted line only incorporates the countries with the largest energy exposures to Russia. Country abbreviations are as follows: AT=Austria; BE=Belgium; CY=Cyprus; ES=Spain; GR=Greece; FI=Finland; FR=France; IE=Ireland; IT=Italy; LT=Lithuania; MT=Malta; NL=Netherlands; PT=Portugal; SI=Slovenia; SK=Slovakia.
One would expect that also euro area companies or corporate sectors would be affected in differential ways, depending on their relative reliance on oil and gas. Different industrial structures – with some countries having a greater share of energy-dependent sectors than others – could then imply the divergence of equity or corporate bond markets across internal euro area borders. Figure 5 displays the evolution of euro area stock market indices for two particularly energy- or oil-dependent corporate sectors – transport (grey line) and chemicals (violet line) – compared to a general market index (dashed black line). As one would expect, the transport index underperforms relative to the overall market during the initial fragmentation period (23 February to 1 March 2022; light orange shaded area) and overperforms relative to the market during the temporary re-integration period (7 March to 1 April 2022; light green shaded area). But the chemicals sector index behaves hardly different from the overall market index, except perhaps for the last week of the re-integration period. Even for the transport index, deviations from the general market index do not seem to be persistently large.
More generally, with the data available to us we could not find systematic results for the relationship between equity market returns and energy dependence, neither in cross-country correlations, nor in cross-sector correlations. Three considerations could explain this somewhat surprising result. First, stock market indices for corporate sectors may not be sufficiently granular, often mixing companies with high and low energy or oil consumption. Second, some energy-intensive companies or sectors may benefit from market structures in which they have pricing power. This market power would allow companies to pass the higher input costs on to their customers without reduced profits materialising in their equity market valuations. Third, the general macroeconomic implications of the Ukraine crisis for the euro area (lower growth and higher inflation) may dominate sector-specific implications, so that equity investors do not discriminate as much between sectors as one might have expected. In line with these considerations, if one decomposes the price-based composite indicator of financial integration in Figure 2 into its market components (not shown in the figures of this column), then its changes during the Ukraine crisis are particularly driven by bond markets and only to a limited extent by equity markets.
Figure 5 Euro area stock market indices during the beginning of the Ukraine war
(vertical axis: all indices normalised to 100 on 23 February 2022 and sector indices re-normalised to the level of the overall market index on 7 March 2022; horizontal axis: daily data from 22 February to 1 April 2022)
Sources: Bloomberg and ECB calculations
Notes: Overall market index is the Morgan Stanley Capital International (MSCI) Europe Index. Transport sector index is the MSCI Europe Transportation Index. Chemicals sector index is the STOXX Europe Chemicals Index.
Let us now turn from energy to financial exposures. Figure 6, which is constructed in a similar way as Figure 4, illustrates whether euro area banks’ exposures to Russia and Ukraine were reflected in their bond yields. Here we do not look at the fragmentation period during the first days of the war but at the temporary re-integration period (7 March to 1 April 2022). Therefore, the hypothesis reverses: as some signs of a relaxation of the conflict emerge and as European countries respond in a unified way, do the banks with the largest exposures to Russia and Ukraine experience a greater reduction in their bond yields compared to less exposed banks, as risk premiums decline?
The first observation from Figure 6 is that euro area banks’ direct exposures are overall quite low (see also ECB 2022d). While with limited country-level data the relationship is unavoidably weak and the correlations have to be interpreted with caution, the red dotted line is broadly consistent with the hypothesis. Note that several countries (dots) had to be removed from the figure for confidentiality reasons, but their data are included in the regression-fitted lines. Moreover, the corresponding figure for the fragmentation period shows similar results (not displayed in Figure 6), just with positive slopes of the fitted lines (as one would expect and also in line with Figure 4).
Figure 6 Euro area countries’ bank exposures to Russia and Ukraine and their bond yields during the temporary re-integration period during the Ukraine war
(vertical axis: change between 7 March and 1 April 2022 in basis points; horizontal axis: Q4 2021)
Sources: Bloomberg, ECB and ECB calculations
Notes: Bank exposures are to households, non-financial corporations and government institutions, covering cash deposits, debt, loans, securities and derivatives holdings. Countries with bank exposures below 0.25% are marked in blue and countries above this value in red. Dotted lines are fitted with simple bivariate regressions including an intercept. The black dotted regression-line incorporates all 11 euro area countries for which reliable data were available (blue and red dots). The red dotted line only incorporates the countries with the largest bank exposures to Russia and Ukraine. The dots of three countries had to be suppressed for data confidentiality reasons but their data are included in the regression-fitted lines. Irish data were a large outlier and therefore excluded from this figure. Country abbreviations: DE=Germany; LU=Luxemburg; rest see notes to Figure 4.
In sum, the evidence so far suggests that the direction of fragmentation and re-integration trends during the early months of the conflict seem to be broadly in line with the directions one would expect from some fundamental factors. But we cannot infer from this simple analysis that non-fundamental market dynamics or procyclical risk premiums have not played any role.
Let us conclude with a look at factors that may disturb the accuracy of pricing assets across euro area countries. In the remainder of this column, we focus on economic uncertainty (see also Anayi et al. 2022 on the uncertainty effects of the Ukraine crisis). If uncertainty about the economic outlook of a country increases (declines), risk premiums can be expected to rise (decline) and financial market expectations may become more (less) prone to herding or more (less) vulnerable to self-fulfilling equilibriums.
Figure 7 Uncertainty about euro area countries’ outlook and their sovereign yields during the beginning of the Ukraine war
(vertical axis: change from February to March 2022 in basis points; horizontal axis: change from March to April 2022 in percentage points)
Sources: Refinitiv, European Commission and ECB calculations
Notes: Changes in economic uncertainty are measured with month-to-month differences of the Economic Uncertainty Indicator of the European Commission (2022a, section 3.6.8). This indicator is derived from surveys of consumers and company managers (in the industry, services, construction and retail trade sectors). Respondents are asked during the first three weeks of a month how difficult they find it to predict their future financial situation (for households/consumers) or their future business situation (for companies/managers). They can choose one of five possible answers: easy to predict, moderately easy to predict, moderately difficult to predict, difficult to predict or don’t know. A monthly value of the indicator is the aggregate share of respondents that answer difficult to predict or moderately difficult to predict. Sovereign yields are for 10-year government bonds. To match them with the timing of the Commission survey field work, changes are calculated as the difference between the average daily yield level during the first three weeks of a month. Dotted lines are fitted with simple bivariate regressions including an intercept. The green dotted line, covering changes from February (before the war) to March, incorporates all 16 euro area countries for which reliable data were available (green dots). The light blue dotted line, covering changes from March to April, incorporates 15 euro area countries, as Greece had to be excluded as an extreme outlier. Country abbreviations are the same as in Figures 4 and 5. The signs of the correlations reflected in the dotted lines do not change when Malta is removed from the February-March period or Slovakia from the March-April period.
Figure 7 shows how euro area countries’ sovereign yields were correlated with changes in uncertainty, as measured with the European Commission’s broad survey-based Economic Uncertainty Indicator (EUI; see European Commission 2022a), during the first months of the war. This indicator combines survey answers of consumers and company managers in different EU countries about how difficult they find it to predict their future financial or business situation.
Increasing uncertainty about the economic and financial outlook seems to be associated with higher sovereign yields (green dots and green dotted regression-fitted line for changes from February – before the war – to March) and declining uncertainty with lower sovereign yields (light blue dots and light blue fitted line for changes from March to April). The same relationship is found for uncertainty and sovereign spreads relative to Germany (not reported in Figure 7).
In other words, while further research is needed on the relative balance between fundamental and non-fundamental factors in driving the financial integration developments during the Ukraine crisis, it seems likely that in an environment of changing economic uncertainty both components will have played a role.
Authors’ note: All views expressed are those of the authors and should not be regarded as the views of the ECB or the Eurosystem.
Just noticed the Dedication in Glazyev’s “The Last World War“:
To my dear wife Oksana,
who loves her homeland Ukraine
and believes in its deliverance
from the neofascist occupation
Well then, that certainly clarifies the situation…
Yes, high-flying sentiments for a dastardly crew. I can only hope against hope that Zelensky and Zelenskaya deign to regale us with a Vogue supersplash part 2!
Uh, all this is written in a language I have a hard time parsing and understanding. But so far several things have stuck out for me:
As I recall it, Russia was and is willing to deliver to anyone who is willing to pay. Russia was even willing to stabilize the prices with long-term contracts.
Now, why would that be, since Russian energy is still cheaper than energy from other sources. For example India and Saudi-Arabia are making huge profits with cheap Russian energy.
Also, in figure 4. Lithuania seems to be an outlier that can alone explain the observed phenomena. Should it be removed from the data, the bivariate regressions would be much more alike, except the Russian energy dependency one with more negative slope.
This brings to mind Blazing Saddles:
Making money off the mentally ill, in grafs and opacity…
All those words, and a conclusion “Needs mores research.”
Put the conclusion first please. The One can decide if reading the argument is worthwhile.
Do any of them slam their heads on their desks and say, “Why didn’t we do this research before jumping when the USA snapped its fingers?”