Liquidity linkages in European sovereign bond markets can amplify fundamental economic shocks
Liquidity linkages in European sovereign bond markets can amplify fundamental economic shocks
The designers of the Maastricht Treaty envisaged that a combination of fiscal rules and market discipline would ensure the safety of euro area national government bonds. The European sovereign debt crisis revealed serious flaws in this approach, as the abrupt repricing of sovereign bonds forced some countries to seek financial assistance from the official sector and raised fears over the survival of the euro. A prominent literature (e.g. DeGrauwe and Ji 2013, Aizenman et al. 2013, Dewachter et al. 2015) questions whether the repricing that occurred during the crisis was larger than implied by developments in economic fundamentals. Stable sovereign bond markets are crucial to a well-functioning economy and financial system. The destructive power of the bank-sovereign ‘doom loop’ demonstrates the importance of understanding the sources of such mispricing.
The misalignment of yields with observable economic fundamentals implies a role for either expectations of self-fulfilling defaults, as described by Bocola and Dovis (2019), or for amplifications of sovereign bond market tensions related to flights-to-safety and sudden liquidity contractions. Despite the undoubted importance of the latter (and the fact that fears of imminent flights and liquidity contractions are likely to be triggers that shift expectations of self-fulfilling defaults), there is little direct empirical evidence of the transmission channels through which such catalysts for amplification operate. In a recent contribution (Clancy et al. 2019) we address this gap in the literature by documenting significant own- and cross-market interdependencies between liquidity and tail risks that amplify shocks likely attributable to economic fundamentals. We do so using intra-day data on the German, Italian and Spanish sovereign bond markets from the MTS trading platform and sophisticated risk-modelling techniques, such as the VAR-for-VaR methodology of White et al. (2015), complemented with the marginal expected shortfall approach of Brownlees and Engle (2014).
Our analysis delivers a compelling set of results. We provide strong evidence of a two-way relationship between the perceived likelihood of liquidity contractions and the risks faced by individual liquidity providers in euro area sovereign bond markets. This validates the widely accepted view that “liquidity begets liquidity”, as described in Foucault et al. (2013), and confirms the fact that liquidity dry-ups and the associated risk of price drops can become self-fulfilling in a way that reduces market quality and amplifies turbulence. We provide evidence that such self-reinforcing effects can spread to other assets consistent with well-understood informational externalities (Cespa and Foucault 2014) and hedging relations (Dunne 2019). An important aspect of our analysis indicates that these effects are further augmented by the ‘safe-haven’ status of the German Bund.
Having established the existence of market-based sources of amplification, we consider (1) whether policy responses, such as the ECB’s commitment to do “whatever it takes” to ensure the survival of the euro, have beneficial calming effects on the examined sovereign bond markets; and (2) whether amplification effects experienced across our sample are dependent on background macroeconomic conditions
We find evidence of a post-crisis dampening of cross-market effects following crisis-era announcements of changes to euro area policies and institutional architecture, proxied by the famous “do whatever it takes” speech by ECB President Draghi. Figure 1 shows how risks associated with the provision of liquidity (measured in terms of the expected size of the lowest quantile of returns – also called the 1% value-at-risk) are related across pairs of markets after a shock to absolute returns in one of the markets. The response of the 1% value-at-risk facing dealers in the Italian and German sovereign bond markets following a unitary shock to absolute returns in the Italian market after Draghi’s speech in July 2012 practically disappears (becoming statistically insignificant for both markets). This implies that the speech produced conditions in which dealers feel safe when they provide liquidity and one less prone to sudden liquidity contractions.
Figure 1 Quantile-IRFs: Before and after “whatever it takes”
Notes: This figure displays Value-at-Risk (VaR) impulse response functions for the German (DE) and Italian (IT) markets in response to a unitary shock to the Italian absolute returns (measured over 15 minute intervals) for sub-sample periods before and after Draghi’s famous ‘do whatever it takes’ speech in July 2012. Standard error bounds (two above and below) are depicted as the grey areas and these are computed as in White et al. (2015). The “before” sample covers the period 01/01/2012 – 25/07/2012. The “after” sample covers the period 27/07/2012 – 31/12/2012. The vertical axis shows the VaR response in percentage terms. The intervals over which the response is shown are also 15 minutes in length, implying that effects are shown for roughly a two-day period following a shock.
In May 2018 political uncertainty increased sharply in Italy, and this had the potential to spill over to other euro area member states. For this period, we identify a structural break in cross-market conditional correlation that produces a persistent reduction in Italian sovereign bond market liquidity. Figure 2 shows how market depth and the correlation between the Italian and German bond markets declined progressively as the crisis developed. We regard the decline in liquidity as being due to fewer hedging opportunities and weaker informational externalities. This is because the Italian market become increasingly dominated by idiosyncratic movements.
Figure 2 Conditional correlations and quoted depth, 2018
Notes: This figure displays intra-day median of quoted depth at best bid and ask quotes within 15 minute intervals (green line) and the daily moving average of the intra-day depth observations (black line) for the Italian 10 year benchmark bond in Millions of euro (values shown on the left vertical axis). The GARCH-implied dynamic conditional correlation of the Italian benchmark bond returns with German Bund returns is shown as a magenta line (values shown on the right axis where conditional correlation is denoted as Rho). Sample: 01/01/2018 – 31/12/2018 (7159 obs.).
A comparison across shorter sub-samples suggests that the 2018 tensions in Italy could have been much more disruptive had background economic circumstances been less accommodative. Figure 3 shows a scatter plot of the bi-monthly average daily yield (y-axis) and the initial own-response of the 1% value-at-risk faced by dealers in the Italian market (x-axis). There is substantial variability in the magnitude of the tail quantile responses. The most negative (i.e. those furthest to the left) occur during the height of the euro area crisis (April-May 2012) and political turmoil in Italy (May-June 2018). Apart from the (small) tail-quantile responses in the months immediately following Draghi’s speech, the absolute size of the initial value-at-risk response rises with the magnitude of the within-sample average yield. This correlation is consistent with tail-quantile responses amplifying (and being amplified by) the economic fundamentals that largely determine sovereign bond yields. This suggests that the rise in market tensions during this period could have had a much larger effect if yield levels were nearer their historical levels.
Figure 3 VAR-for-VaR estimated responses at diering yield levels
Notes: This figure plots bi-monthly estimated (impact) responses of the 1% Value-at-Risk of Italian 10-year benchmark bond returns (horizontal axis) versus the average daily yield level during those months (vertical axis). The circles relate to months in the 2011/12 sample, while the crosses are months in the 2018 sample.
Overall, our findings demonstrate the potential for the provision of liquidity across sovereign markets to be vulnerable to sudden fractures, with possible implications for euro area economic and financial stability. In spite of substantial improvements in the euro area policy and institutional architecture, the 2018 disruptions show that national sovereign bond markets remain susceptible to market-driven amplifications of fundamental shocks that can spread beyond the local market.
Aizenman, J, M Hutchison, and Y Jinjarak (2013), “What is the risk of European sovereign debt defaults? Fiscal space, CDS spreads and market pricing of risk”, Journal of International Money and Finance 34: 37-59.
Bocola, L and A Dovis (2019), “Self-Fulfilling Debt Crises: A Quantitative Analysis”, American Economic Review, forthcoming.
Brownlees, C and R Engle (2017), “SRISK: A Conditional Capital Shortfall Measure of Systemic Risk”, The Review of Financial Studies 30: 48-79.
Cespa, G and T Foucault (2014), “Illiquidity contagion and liquidity crashes”, The Review of Financial Studies 27(6): 1615-1660.
Clancy, D, P Dunne and P Filiani (2019), “Liquidity and tail-risk interdependencies in the euro area sovereign bond market”, Central Bank of Ireland RTP No 11/RT/19.
De Grauwe, P and Y Ji (2013), “Self-fulfilling crises in the Eurozone: An empirical test”, Journal of International Money and Finance 34: 15-36.
Dewachter, H, L Iania, M Lyrio, and M de Sola Perea (2015), “A macro-financial analysis of the euro area sovereign bond market”, Journal of Banking & Finance 50: 308-325.
Dunne, P G (2019), “Positive Liquidity Spillovers from Sovereign Bond-Backed Securities”, Journal of Risk and Financial Management 12(2).
Foucault, T, M Pagano, and A Roell (2013), Market liquidity: Theory, evidence and policy, Oxford University Press.
White, H, T Kim, and S Manganelli (2015), “VAR-for-VaR: Measuring tail dependence using multivariate regression quantiles”, Journal of Econometrics 187: 169-188.