An important question in international macroeconomics is the degree to which advanced economy (AE) investors impact the economic policies of emerging market economy (EME) borrowers. The basic question about how EMEs can manage their economies has been discussed generally through the debate about the Mundell-Fleming ‘policy trilemma’ and its implications for open capital markets in papers such Calvo et al. (1996), Reinhart and Reinhart (2009), Reinhart and Rogoff (2009) and Rey (2015). Furthermore, the vulnerability of EMEs to AE investors has been highlighted by financial outflows from these countries following crisis periods from the late 1990s and up through the recent Covid period (e.g. McLoughlin et al. 2014, Korinek et al. 2020, Zakrajsek et al. 2022). Some have argued more specifically that increases in AE interest rates or other global risk factors have induced investors in these countries to reduce their holdings in EME securities, thereby generating negative spillover effects (Forbes and Warnock 2012, Kalemli-Özcan 2019).
In a recent paper (Faia et al. 2024), we revisit the question of whether interest rate increases in advanced economies trigger capital outflows from emerging market debt securities. Using novel euro area investor holdings data at a disaggregated level, we trace the portfolio responses to monetary policy shocks of individual investor groups, which in turn shape the aggregate spillovers.
Motivating trends in investor types
We consider three main investor categories that provide a good spectrum of the heterogeneous variation in asset holding. The variation comes from distinct institutional features and responses to shocks: Faia et al. (2022) uncover substantial heterogeneity in these investors’ portfolio responses to the ECB’s large-scale asset purchases. Based on the ECB naming conventions, our investor groups include: (1) insurance companies and pension funds (ICPFs) that tend to be more passive and invest longer term; (2) banks and money market funds, collectively called ‘monetary financial institutions’ (MFIs), that tend to rebalance more actively than insurance and pension funds, but are constrained by regulation; and (3) ‘other financial institutions’ (OFIs), a broad group including investment and mutual funds. As described below, this last investor group includes the most active participants in the EME debt market.
Trends in these investor groups can be seen by examining the portfolio shares of EME securities held by the main investor categories we consider. Figures 1 and 2 present information on the composition of foreign investors for EME long-term debt securities using two different data sets.
Specifically, Figure 1 reports the EME holdings by euro area investors, while Figure 2 graphs the holdings of EME country securities by all foreign investors.
Figure 1 Euro area investor holdings of emerging market debt by issuer sector
Source: ECB Securities Holding Statistics
Note: Figure 1 plots the aggregate market value of debt securities issued by emerging market economies and held by Euro area investors recorded in ECB’s securities holdings data. All observations are end-of-year values, and the units are in billions EUR. The“all other sectors” group includes non-financial companies, households and non-profit institutions serving households (NPISHs), and governments. The emerging market economies covered in the public SHS-ECB dataset include Argentina, Bulgaria, Brazil, Chile, China, Czech Republic, Croatia, Hungary, Indonesia, India, Latvia, Lithuania, Mexico, Poland, Romania, Russia, Slovenia, Slovakia, Turkey, South Africa.
Figure 2 EME exposure to heterogeneous foreign investors
(a) All emerging markets
(b) Country heterogeneity: an illustration
Source: IMF Coordinated Portfolio Investment Survey
Note: Figures 2 illustrate the foreign investor base of EME debt using the IMF Coordinated Portfolio Investment Survey, a country-by-country bilateral dataset covering cross-border portfolio holding on a sector-by-sector basis since 2013. Both panel (a) and (b) include issuers from all sectors. Panel (a) reports total cross-border holding with investor sector breakdown for all emerging markets Panel (b) focuses on the differences in investor base for countries with varying country fundamentals, using Chile and Argentina as case studies.
Both cases point to a similar picture: while investment funds hold the largest amount of EME debt for all issuer sectors, the size of holdings by long-term, stable investors such as ICPFs and MFIs is also on the rise. Figure 1 indicates that these two sets of investors account for about 40% of all of the euro area’s holding of EME government debt in 2021. Similarly, Figure 2 shows that for all foreign investments into EME debt, insurers and pension funds have been taking on a substantial proportion.
The shifts in the foreign investor profile are not uniform across EME issuers, however. Panel (b) of Figure 2 compares the foreign investor base of long-term debt issued by Argentina and Chile. For Chile, the share of holdings by insurance and pension funds exceeds those of investment funds, but the opposite is true for Argentina. Country fundamentals play a role in determining capital inflows from default-averse stable investors (see also Zhou 2024 and Fang et al. 2024). These patterns suggest a natural sorting of debt holdings across investor groups based upon these characteristics: slower-to-adjust investor groups, such as banks and insurance/pension funds, tend to hold securities from countries with more stable monetary regimes or inflation targeters, while investment funds (OFIs) are more willing to be exposed to issuers of riskier bonds with high returns.
Euro area investment in EME debt securities
As suggested by our stylised facts above, the heterogeneity in investor demand and the changing nature of EME risk may impact foreign monetary policy spillovers through differences in portfolio holding adjustments across these groups. We investigate this possibility by estimating the responses of investor portfolio shares of EME bond holdings. For this purpose, we use panel local projection estimations of these portfolio shares in response to identified high-frequency monetary policy shocks (Jordà 2005). By focusing on these shares, we follow a tradition in international finance that relates capital flows to investor portfolio adjustment, including for instance Hau and Rey (2005) and Camanho et al. (2022).
Figure 3 presents the impulse responses of EME debt security holdings to a 25 basis point tightening. These estimates show no significant decline of debt shares toward EMEs, with the important exception of OFIs (investment funds). Specifically, the shares held by OFIs for corporate bonds (panel c) and sovereign bonds (panel d) decline significantly. As noted earlier, these investors are the most active institutional investors because they are less constrained by prudential policies and often seek higher yields. By contrast, the shares held by ICPFs (insurance and pension funds) and MFIs (banks and money market funds) show no significant decline. Their passive behaviour may therefore contribute to the moderation in the response of the aggregate bond flows.
Figure 3 Impulse response of euro area emerging market debt allocation to surprise short-rate hikes (portfolio weight)
(a) Debt issued by all sectors
(b) Debt issued by financial corporations
(c) Debt issued by non-financial corporations
(d) Debt issued by governments
Source: ECB Securities Holdings Statistics.
Note: Figure 3 plots quarterly impulse responses of Euro area (EA) investors’ emerging market debt holding (market value) as share of total market value of securities portfolio to 25 bps monetary policy surprise reflected in the short-term interest rate (3-month OIS). The monetary policy surprise is identified via high-frequency movements in the asset prices around ECB monetary policy event windows. Impulse responses are estimated using local projections by investor sector (bank+MMF, ICPF, and other financial institutions), and by issuer sector (all sectors, financial corporations, non-financial corporations and government). The control variables include 4 lags of monetary policy shock and lagged changes (for 4 quarters) of the dependent variables. The unit of the y-axis is percentage point. 68% and 90% confidence interval with robust standard error are reported.
One concern about our finding could be that our data sample is too short to allow us sufficient power to find significant responses. To better gauge whether the sample size reduces the explanatory power for asset substitution estimates across other countries, we re-estimate our analysis focusing on debt holdings from AEs outside of the euro area. Strikingly, in contrast to the EME holdings, Figure 4 shows that portfolio shares of AE debt decline significantly, particularly for shares held by banks and investment funds. Overall, therefore, euro area investor holdings do appear to respond significantly to monetary policy shocks, but the reversal is from AE instead of EME debt securities.
Figure 4 Euro area investors’ non-euro area advanced economy debt allocation response to surprise short-rate hikes (portfolio weight)
(a) Debt issued by all sectors
(b) Debt issued by financial corporations
(c) Debt issued by non-financial corporations
(d) Debt issued by governments
Source: ECB Securities Holdings Statistics
Note: Figure 4 plots quarterly impulse responses of Euro area investors’ holding of debt securities (market value) issued by issuer outside Euro Area classified as Advanced Economies (AE) as share of total market value of securities portfolio to 25 bps monetary policy surprise reflected in the short-term interest rate (3-month OIS). The monetary policy surprise is identified via high-frequency movements in the asset prices around ECB monetary policy event windows. Impulse responses are estimated using the local projection by investor sector (bank+MMF, ICPF, and other financial institutions), and by issuer sector (all sectors, financial corporations, non-financial corporations and government). The control variables include 4 lags of monetary policy shock, lagged changes (for 4 quarters) of the dependent variables, and country-level variables (inflation, quarterly changes in unemployment rate and industrial production). The unit of the y-axis is percentage point. 68% and 90% confidence interval with robust standard error are reported.
Validation from security-level holding data
Our analysis above is based upon aggregate portfolio data for all euro area investors. We address potential shortcomings of our aggregate analysis through confidential security-level micro data for German investors.
These data contain information about the face value of bond holdings and therefore provide a separation between changes in portfolio holding and valuation effects. They are also available at a monthly frequency and contain a wider range of issuer countries.
Figure 5 shows the impulse responses of German investors’ EME portfolios using both debt face values (top panels) and portfolio weights (bottom panels). For this analysis, we focus on sovereign bonds, as these securities represent the largest share of EME securities held by euro area investors. Consistent with our results using the aggregate data, we observe no significant decline for long-term, passive investors such as insurance and pension funds (ICPFs), while investment funds (OFIs) demonstrate the strongest contraction of their EME portfolio.
Figure 5 Impulse response of German investors’ EME government bonds allocation to monetary surprises
Source: Research Data and Service Centre (RDSC) of the Deutsche Bundesbank, Securities Holdings Statistics (SHS-Base plus), 2012M12–2022M6, own calculations.
Note: Figure 5 provides additional evidence on German investors’ portfolio allocation to EME government bond using monthly security-level holding data. In the top three panels, the dependent variable is the absolute month to month changes in the face value of holding. The bottom three panels use portfolio weights as the dependent variable. The category “Fund” corresponds to other financial institutions in the baseline ECB data. 68% and 90% confidence interval with robust standard error are reported.
As these investment funds are the most responsive, a natural question is: who are the investors responsible for these variations? To answer this question, we exploit the more refined breakdown of investor responses to monetary policy surprises by (a) bond funds, (b) mixed (allocation) funds, (c) retail funds, and (d) institutional funds.
Figure 6 shows the results using face values in the top and portfolio shares in the bottom panels. In both the face values and the shares, the mixed funds and the retail investors demonstrate the most significant decline in EME holdings. These results are consistent with the view that these institutions tend to service clients that require active rebalancing or are more flexible in their investment mandates.
Figure 6 Impulse response of German mutual funds’ EME government bonds allocation to monetary surprises, by type of funds
Source: Research Data and Service Centre (RDSC) of the Deutsche Bundesbank, Investment Funds Statistics Base, 2012M12–2022M6, own calculations.
Note: Figure 6 plots the impulse responses of German investors’ EME debt holding (face value) as share of total market value of securities portfolio, in response to surprise monetary tightening. Investors include all mutual funds and are broken down in Bond Funds, Market Funds, Retail, Institutional. The monetary policy surprise is identified via high-frequency movements in the asset prices around ECB’s monetary policy event windows (Altavilla et al., 2019). Impulse responses are estimated using the local projection. The control variables include 3 lags of monetary policy shock and lagged changes (for 3 months) of the dependent variables, as well as issuer country-level controls. 68% and 90% confidence interval with robust standard error are reported.
Concluding remarks
The analysis described here studies the monetary spillovers onto foreign securities, with a focus on EME debt, using securities holdings held by different investor types in the euro area. We find no consistent evidence of spillovers to capital flows through EME bond holdings, except for those linked to the changes in the shares held by investment funds (OFIs). As a growing literature is showing, these investment funds appear to be the most active investors. We also examine whether the strength and nature of the spillovers change with the stance of monetary policy. Specifically, we distinguish the effects of conventional versus unconventional policies and the impact of synchronised tightening between the US and the euro area. Moreover, we find that information shocks induce very different responses in investment fund behaviour than traditional monetary shocks. Overall, our results provide new evidence for the connection between monetary shocks and investment funds. They also suggest a rich array of future research issues.
Authors’ note: This column represents the personal opinions of the authors and does not necessarily reflect the views of the Deutsche Bundesbank or the Eurosystem.
References
Calvo, G A, L Leiderman, and C M Reinhart (1996), “In- flows of Capital to Developing Countries in the 1990s”, Journal of Economic Perspectives 10(2): 123–139.
Camanho, N, H Hau, and H Rey (2022), “Global portfolio rebalancing and exchange rates”, The Review of Financial Studies 35(11): 5228–5274.
Faia, E, J Salomao, and A Ventula Veghazy (2022), “Granular investors and international bond prices: Scarcity-induced safety”, available at SSRN.
Faia, E, K K Lewis, and H Zhou (2024), “Do Investor Differences Impact Monetary Policy Spillovers to Emerging Markets?”, CEPR Discussion Paper 19389.
Fang, X, B Hardy, and K K Lewis (2024), “Who holds sovereign debt and why it matters”, CEPR Discussion Paper 17338.
Forbes, K J and F E Warnock (2012), “Capital flow waves: Surges, stops, flight, and retrenchment”, Journal of International Economics 88(2): 235–251.
Hau, H and H Rey (2005), “Exchange Rates, Equity Prices, and Capital Flows”, The Review of Financial Studies 19(1): 273–317.
Jordà, O (2005), “Estimation and inference of impulse responses by local projections”, American Economic Review 95(1): 161–182.
Kalemli-Özcan, S (2019), “U.S. Monetary Policy and International Risk Spillovers”, Jackson Hole Proceedings.
Korinek, A, J A Ocampo, and B Erten (2020), “Managing capital flows to emerging markets”, VoxEU.org, 11 August.
McLoughlin, C, D Reinhardt, and L Gauvin (2014), “Policy uncertainty spillovers to emerging markets: Evidence from capital flows”, VoxEU.org, 5 August.
Reinhart, C M and K S Rogoff (2009), This time is different: Eight centuries of financial folly, Princeton University Press.
Reinhart, C M and V R Reinhart (2009), Capital flow bonanzas: An encompassing view of the past and present, Vol. 5, 9–62, University of Chicago Press.
Rey, H (2015), “Dilemma not trilemma: the global financial cycle and monetary policy independence”, NBER Working Paper 21162.
Zakrajsek, E, B Wei, V Yue, and S Gilchrist (2022), “Global financial risk factors and sovereign risk”, VoxEU.org, 26 February.
Zhou, H (2024), “The Fickle and the Stable: Global Financial Cycle Transmission via Heterogeneous Investors”, available at SSRN.
