Spillovers From ECB QE – Global Yet Varied

Central bank bond buying has accelerated at an unprecedented pace during the COVID crisis. The impact of large asset purchases on home markets is generally well understood, but the global spillovers are less clear. New research by the ECB, which for the first time uses data on an individual security level, provides a detailed insight into the impact of the first two years of the bank’s Public Sector Purchase Programme (PSPP) during 2015-16. Given the €6bn now purchased daily under the ECB’s new €750bn Pandemic Emergency Purchase Programme (PEPP) and the resumption of PSPP on an opened-ended basis late last year, the results are highly relevant for current capital flows and sectoral portfolio shifts.

Moving Past BoP Flows

Balance of payments data for the Euro-area very clearly show significant capital outflows as a result of the PSPP, first initiated in March 2015 (Chart 1). These flows peaked at 5% of Euro-area GDP in mid-2016, and the stock of assets now stands at €2.2tn.

Chart 1: Breakdown of Euro Area Net Portfolio Investment Flows

Source: ECB

Earlier studies showed that quantitative easing (QE) reduced long-term yields of eligible assets, as well as other debt securities, while simultaneously supporting equity markets both across the EU and globally.

By using detailed data by geography and security, ECB staff add to these findings to show that net selling of PSPP assets was uneven across countries and sectors. For example, insurance companies and pension funds showed larger shifts into longer-term assets, and stressed Euro-area economies were larger net sellers given greater scope for capital appreciation. The research also backed up earlier findings that ECB QE did not benefit emerging markets, with other advanced economies being the main beneficiaries.

Methodology: Model & Evidence-Based Approach

Key to the authors’ finding is the very granular data used. All 19 Euro-area countries are included, and the data covers individual securities (debt, equity, investment funds) and is split into six sectors, namely: monetary and financial institutions (MFIs), insurance companies and pension funds (ICPF), other financial institutions (OFI), non-financial corporations (NFC), general government, and households.

Six different hypotheses are tested to determine the exact nature of capital flows (excluding valuation impacts), using both descriptive evidence as well as the model-based results.

1) Euro-area investors are expected to be net sellers of assets targeted under the asset purchase programme and to rebalance into ‘closest substitutes’.

Results: The model confirms the relevant variable as significant. The authors note that more than €250bn was sold in eligible securities during the period, and foreign debt was purchased in similar size, with around 40% of this defined as closest substitutes. By sector, MFIs, households and OFIs were the largest sellers. However, other Euro-area debt securities saw larger net sales, with the largest chunk coming via banks.

2) Investors are expected to rebalance into longer-term securities (more duration risk).

Results: Again, the model confirmed the relevant variable as significant, and data shows most of the net debt securities purchased had an original maturity >10 years and driven by ICPFs and OFIs. Net sales were concentrated in original maturity of 2-5 years and mostly by MFIs and households (the same as in 1, above).

3) Euro-denominated bias in debt securities weakened.

Results: This hypothesis was also confirmed, with the variable showing as insignificant. Data show significant sales of euro-denominated assets and an increase in USD and GBP denominated debt securities, while equity purchases remained euro-denominated. Unlike in 1 and 2 above, all sectors were net buyers of USD-denominated debt and more than half through OFIs.

4) Euro-area portfolios shifted towards foreign debt securities (other than closest substitutes) including EM with increased preference for risk-taking.

Result: All sectors increased investment into debt securities issued by advanced economy banks and OFIs. This confirms the rebalancing towards closest substitutes as well as broad-based purchases of debt securities in advanced economies. Euro-area investors were not, however, net buyers of EM debt in aggregate, unlike US investors during earlier Fed QE. Euro-area MFIs were, however, net buyers of EM sovereign bonds.

5) Heterogeneity across sectors (households likely to be sellers of eligible assets, ICPFs skewed towards longer-term assets).

Results: NFCs and households were net sellers of eligible securities as well as broader euro-denominated securities, reflecting investments channelled through mutual funds rather than directly into foreign sovereign bonds.

6) Heterogeneity across countries (former crisis countries selling more eligible assets to rebalance towards more higher-yielding assets at home, non-crisis more towards longer-term).

Results: Different coefficients in the model results confirm the hypothesis, and with the coefficient almost three times larger in former crisis countries, this shows more significant portfolio rebalancing here (given a greater incentive to realise capital gains). And only in non-stressed countries did the authors see a shift into longer-term debt reflecting lower yields at home.

Chart 2: Euro Area Net Debt Transactions by Country

Source: ECB

Summary

Putting all this together, the study confirms key trends in QE capital flows discussed in earlier work, and a broad shift into foreign debt securities. But it also adds on valuable detail. Euro-area investors were net sellers in favour of foreign debt securities as expected, but with investors from former crisis countries more significant sellers. Investors also shifted towards longer-term securities, as expected, but this was not uniform across sectors. Neither was the preference to reduce Euro-denominated debt securities. And, importantly, EM countries were not net beneficiaries.

The study also showed expectations ahead of the announcement as significant, with Euro-area investors net buyers of eligible securities in H2 2014. Another interesting finding was that splitting out active capital flows and passive (valuation effects) shows active net sales not fully compensated with positive capital gains.

Translating these findings into the current environment would suggest we should again expect significant capital outflows but with important differences between countries and sectors with perhaps Italy as one of the largest net sellers. The global environment is of course very different than in 2015-16 with the very significant QE from the Fed swamping that of the ECB. Unprecedented capital outflows from EM also change the international spillovers more broadly. One interesting question: will the Euro-denominated bias again be weakened across sectors?