Selected Issues

Abstract

Selected Issues

Impact of Monetary Policy on Luxembourg1

Accommodative monetary policy has contributed to the performance of the Luxembourg economy through some expansion of aggregate demand and through its impact on the financial system. Banks have remained profitable and interest margins stable, while fee and commission income from fund and other activity has been healthy. The investment fund industry has benefited from various factors such as portfolio rebalancing, search for yield, and other market developments leading to strong inflows into various classes of investment funds, and through strong valuation effects. Scenario analysis suggest that the fund industry could be adversely impacted by sharp interest rate increases and that, because of interconnections, the banking system would also be affected. Margins of some banks could also decline when interest rate normalize. Against this backdrop, it is important to implement all 2017 FSAP recommendations that will contribute to making the financial system more resilient to shocks, including those arising from faster-than-expected monetary policy normalization.

A. Introduction

1. Monetary policy has been in exceptional territory since the global financial crisis, thereby supporting aggregate demand. While short-term interest rates reached the lower bound, various unconventional monetary policies were implemented by the ECB and by the Federal Reserve System (Fed), including long-term refinancing operations, asset purchase programs, outright monetary transactions, negative interest rates on the deposit facility driving short-term rates into negative territory, and novel communication tools such as forward guidance.

2. Performance of the Luxembourg financial sector has been generally strong. The strong growth of the investments fund industry has generated income for the financial sector at large, including banks through higher fee and commission income, and stimulated ancillary service activities, in particular administrative and depository services. Meanwhile, there is no evidence that banks have compressed their lending margins as lower funding costs were passed on to retail borrowers and corporates. Meanwhile the private banking industry has attracted new high net worth clients to whom specific services, including mortgage loans, are offered.

3. Accommodative monetary policy and search for yield have contributed to the rapid expansion of the Luxembourg investment fund industry. In addition to a reduction in bank intermediation since the financial crisis, QE and low interest rate policy have fueled the expansion of the fund industry through:

  • Abundant liquidity and a global search for yield, creating demand for assets offering higher yields than traditional bank savings instruments;

  • Relative expansion of investment funds into riskier assets (high yield bonds, emerging market bonds, equity), including as a result of portfolio rebalancing; and:

  • Rich valuation effects, notably in stock markets.

4. This chapter is organized as follows. First, it provides an overview of the impact of monetary policy on Luxembourg’s macroeconomy. Second, it analyzes the impact on the banking system, including risks that could result from normalization. Third, it studies the impact of accommodative monetary policy on the investment fund industry. Fourth, it concludes.

B. Macroeconomic Impact of Monetary Policy in Luxembourg

5. Easing of monetary policy impacts the macroeconomy through various channels. An easing of monetary policy leads to a decrease in real interest rates, which lowers the cost of borrowing resulting in greater investment spending and consumption, thus affecting saving-investment decisions, and increasing aggregate demand. Asset prices such as stock market valuations tend to increase, which affects financial wealth and consumer spending; balance sheet effects through collateral valuations also tend to facilitate borrowing. Easing of monetary policy stimulates credit supply as banks’ cost of capital declines. By affecting asset prices (or through direct asset purchases), monetary policy may trigger portfolio rebalancing toward riskier assets. Last, easing of monetary policy can trigger more risk taking as banks loosen their lending standards, and investors step up search for yield. The effects of portfolio rebalancing (from safe financial assets to more risky investment funds) and search for yield are relevant for Luxembourg’s investment fund industry.

6. There is evidence that banks passed on the lower nominal interest rates to households and firms (Figure 1). Luxembourg banks have passed on lower interest rates to their retail and corporate borrowers. Lending rates have stabilized around or slightly below 2 percent. In the meantime, core inflation fell to 1 percent in 2016, but started to rise more recently broadly in line with euro area developments. Deposit rates also fell, particularly so for corporates.

Figure 1.
Figure 1.

Impact of Monetary Policy on Interest Rates in Luxembourg

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

7. Households may have benefitted from wealth effects, particularly home owners, as a result of continuously rising housing prices (see Chapter II). According to the 2014 Household Finance and Consumption Survey of the Banque centrale du Luxembourg (BcL), real estate amounts to about 80 percent of household wealth on average, and about 1/3 of total real estate wealth is for investment purposes. Bank deposits account for about 50 percent of financial wealth, suggesting that lower income households may have benefitted much less from potential wealth effects of accommodative monetary policy.

Mean Household Wealth by Category

article image
Source: BCL household survey 2014

8. Private consumption has risen somewhat, in particular since 2011 (Figure 2). However, it remains volatile, making it difficult to identify a clear trend. While there is no clear evidence of an acceleration of new mortgage credit, debt servicing costs have declined for the stock of existing mortgages at floating interest rates. This should have contributed to stimulating consumption. However, for prospective home-owners, in particular younger households, a rising affordability problem may on the other hand have contributed to moderating consumption.

Figure 2.
Figure 2.

Accommodative Monetary Policy and Private Consumption

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

9. Bank credit to non-financial corporations appears to be rising, but statistical issues make interpretation difficult. Credit to resident non-financial corporations has been traditionally very volatile, perhaps due to the small number of large firms in Luxembourg. Some statistical issues, however, complicate the analysis. In 2015, a break in the series arose as a result of a reclassification of one firm from the category “other financial institutions” to “non-financial corporations”. In addition, the recent acceleration of lending to non-financial corporations by domestically oriented banks may partly be driven by borrowing by multinationals which could be used to fund activities of foreign subsidiaries.

A01ufig1

Credit to Resident Non-Financial Corporations

(Percent, Year-on-year growth rate)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

Source: BCL.

C. Impact of Monetary Policy on the Banking System

10. Luxembourg banks have withstood the low interest environment well and protected their lending spreads (Figure 3). The four large domestically oriented banks have somewhat better profit performance than European banks on average. They have passed on lower rates to domestic borrowers but lending spreads have remained broadly constant and nonperforming loans (NPLs) low, and so their net interest margins have remained stable since the global financial crisis, slightly above the EU average. Operating costs have remained below the European average.

Figure 3.
Figure 3.

Domestically Oriented Banks’ Performance

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

A01ufig2

Lending Spreads

(Percent)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

Sources: Haver Analytics.

11. In general, the banking system has performed relatively well since monetary policy became accommodative. Fee and commission income of the banking system have benefited from various activities such as those related to the investment fund industry. There is limited evidence that low interest rates have contributed to balance sheet expansion by lifting the demand for mortgages, as new mortgages have continued to grow along a broadly constant trend despite declining rates, perhaps because of housing supply constraints. Other activities such as private banking have been healthy.

A01ufig3

Banking System Net Income

(Percent of total assets)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

Sources: CSSF and IMF staff calculations.

12. Going forward, the normalization of interest rates could affect some domestic banks. The growing share of fixed rate mortgages in the new loan production relative to historical practices could expose the domestically oriented banks active in real estate lending to some compression of margins that would materialize when monetary policy normalizes (see Chapter 2). Credit risk could result from highly indebted households who financed real estate purchases with mortgages at floating interest rates (see Chapter 2). Last, balance sheet liquidity mismatches are limited given that the ratio of mortgage loans to retail (household) deposits remains below 100 percent.

A01ufig4

Ratio of Loans to Households / Household Deposits, in Domestically Oriented Banks

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

Source: BCL.

13. The Luxembourg banking system has become more interconnected domestically and globally, making it susceptible to external shocks that could result from sharp interest rate increases (see next section). According to sectoral financial accounts, in the three years to June 2017, liabilities of Luxembourg monetary and financial institutions (MFIs) to domestic sectors have grown by €125 billion, about 2.5 times 2016 GDP, reaching €395 billion. MFIs have become more connected among themselves, with local offices of multinational firms (“other financial institutions”) and with investment funds. During the same period, investment funds’ liabilities have increased by about €1 trillion, while total gross liabilities of the sector “other financial institutions” (which includes various entities conducting international Treasury operations) vis-à-vis all domestic sectors (including itself) and the rest of the world have increased by about €4 trillion. Linkages between local custodian banks and investment funds, both from deposits and derivative contracts, remain significant.

A01ufig5

Individual Luxembourg Bank Exposures to Investment Funds and Money Market Funds, 2017:Q3

(Share of total bank assets/liabilities)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

Source: BCL, statistical individual bank reporting framework.

Interlinkages Among Institutional Sectors in Luxembourg, June 2017

(In billions of euros)

article image
Sources: Total assets and liabilities are from BCL Table 05.08 and bilateral sectoral accounts are from dataset General FSAP questionnaire “10a_Luxembourg_General Data FSIs Table 10 LU-Rest_world”

Includes households and non-profit institutions serving households.

D. Impact of Monetary Policy on the Investment Fund Industry

Event Study

14. Since the global financial crisis, the ECB and the Fed have announced many unconventional monetary policy decisions. The table below summarizes the month of main announcements by the ECB and the Fed regarding unconventional monetary decisions as well as, in the case of the Fed, communications regarding exit from QE or interest rate normalization (Table 1).

Table 1.

Luxembourg: Monetary Policy Announcements

article image
Sources: Federal Reserve and ECB.

15. Event study analysis suggests that monetary policy announcements impacted the patterns of aggregate net inflows into various types of Luxembourg investment funds.2 The impact of monetary policy announcements on the change in net flow patterns is visible for the entire sample of bond, equity and mixed UCITS funds.3 For example:

  • End of QE2 by the Federal Reserve in June 2011: Shortly after QE2 ended, net flows into investment funds turned negative until the end of 2011. There was a sharp volatility of stock market indices during this period.

  • ECB Long-term refinancing operation (LTRO): the December 2011 LTRO was followed by a reversal of net flows, from a monthly average of about -€10 billion in the last five months of 2011 to +€10 billion in the following 5 months.

  • ECB announcement of the PSPP: The March 2015 announcement which came after the earlier APP announcement was followed by a clear increase in aggregate net inflows into investment funds, which reached about €34 billion on average until July 2015, despite a moderate increase in market volatility (compared to a monthly average of €18 billion during the last 6 months of 2014).

  • Fed communications of interest rate normalization: Communication about interest rate normalization by the Feb started in July 2015. Average monthly net inflows dropped sharply to on average €4.8 billion until March 2016.

A01ufig6

Net Inflows in UCITS Luxembourg Investment Funds and Unconventional Monetary Policy

(Billions of euros)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

Source: Luxembourg authorities.

16. Different monetary policy announcements impacted investment funds differently according to their investment policy. While the ECB LTRO seemed to have mostly impacted net inflows into bond funds, equity funds and to a lesser extent mixed funds (which invest in bonds and equity) were more positively impacted by the London speech by Draghi and the QE3 announcement by the Fed. The PSPP announcement seems to have impacted all categories of funds, while the start of the Fed communication on interest rate normalization triggered net outflows mostly among bond funds. As shown in the left-hand chart of Figure 4 showing absolute net flows into different bond funds, the response for advanced economies corporate bond funds was generally stronger than for emerging market bond funds or for high yield bond funds. However, the relative expansion of emerging bond funds and of high yield bond funds was much stronger than of more standard corporate bond funds as unconventional monetary policies were implemented after the global financial crisis.

Figure 4.
Figure 4.

Evolution of Different Classes of Bond Funds

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

A01ufig7

Net Inflows in UCITS Luxembourg Investment Funds and Unconventional Monetary Policy

(Billions of euros)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A001

Source: Luxembourg authorities.

Econometric Analysis

17. We perform a regression analysis to assess investment funds’ response to monetary policy conditions (Box I). The time series regression analysis is performed on data at a monthly frequency over the period 2007:M1 to 2016:M6. The two dependent variables are investment funds’ net inflows and gross outflows (capturing aggregate redemptions by fund type), both expressed as a share of the previous month’s net investment fund asset value. The proxy for monetary conditions is the EONIA. We control for various indicators of market conditions, indicators of real economy performance, and for the previous period’s investment fund flows considered.

18. Net inflows into investment funds are significantly impacted by short-term interest rates (Table 2). An increase in the EONIA is significantly associated with a decline of net inflows into investment funds. The effect is significant for bond funds, the subcategory of corporate bond funds, and for equity funds. In decreasing order of magnitude, the effect is larger for corporate bond funds, equity funds and aggregate bond funds (see paragraph 20 for a quantification). The larger coefficients obtained for equity funds and corporate bond funds relative to the coefficient for aggregate bond funds offers some suggestive evidence of a search for yield among investors, whereby a more accommodative monetary policy is associated with a rebalancing of portfolios toward more risky investments. It is also quite notable that the coefficient remains statistically significant after we control for possible sluggishness of investment funds’ flows with a lagged dependent variable, for market conditions and for real economy performance.4

Table 2.

Luxembourg Impact of Short-Term Interest Rates on Net Inflows into Investment Funds

article image
Notes: Robust standard errors. ***, ** and * denote significant at 1%, 5% and 10% respectively.

Econometric Model

We estimate a econometric model linking investment fund flows to monetary conditions, global market and real economy developments as follows:

Δinflowit=αi+δiEONIAt+ΣνγiνYνtl+βiΔinflowit1+ɛt(1)

Where Δinflowit is either net inflows or gross outflows into/out of an aggregate category of investment funds i during month t, expressed as a ratio to the previous month’s net asset value of the fund category i. We consider three categories of funds: (i) UCITS bond funds, (ii) UCITS equity funds and (iii) UCTIS bond funds with a strategy centered on corporate bonds.

The explanatory variables are: (i) the monthly average of the euro overnight market rate (EONIA), (ii) a vector of control variables Yvtl at various lags l which include the growth rate of the monthly average of the eurostoxx index, the monthly average of the stoxx50 volatility index, the growth rate of the industrial production index for the euro area, and the growth rate of the industrial production index of the U.S. We also include a lagged dependent variable Δinflowit–1.

The econometric model is estimated with Newey West standard errors computed by allowing an autocorrelated structure of the residual εt with up to four lags.

19. The impact of monetary conditions on net flows is to a significant extent realized through a correlated response of aggregate redemptions to interest rates (Table 3). We find that higher short-term interest rates are associated with larger aggregate redemptions from all three categories of investment funds, and that the effects are of broadly the same order of magnitude, though they are somewhat larger for corporate bond funds. Aggregate redemptions are also positively associated with a decline in market performance and by higher financial market volatility.5

Table 3.

Luxembourg: Impact of Short-Term Interest Rates on Redemptions from Investment Funds

article image
Notes: Robust standard errors. ***, ** and * denote significant at 1%, 5% and 10% respectively.

20. The size of the effect of changes in the interest rate environment on investment funds’ flows is economically significant. We compute the standard deviations of the dependent variables (top of Table 4), and compare them with the predicted impact over one month of one standard deviation increase in each of the explanatory variables that are significant. We find that the impact of interest rates on net inflows (respectively redemptions) are particularly large as they explain 15–25 percent (respectively 60–70 percent) of the monthly standard deviations in the dependent variables. The impact of market returns and market volatility are also particularly large.

Table 4.

Luxembourg: Quantification

article image

21. These findings suggest that investors’ aggregate portfolio allocation behaviors among investment funds, and between investment funds and other financial assets are very dependent on the monetary and financial market environments. The empirical analysis suggests that the decision to allocate savings between financial assets offering a safe nominal return (such as deposits) and more risky types of investments—such as equity and corporate bonds—are correlated among investors when it comes to the response to short-term interest rate movements or general market conditions. It is also noticeable that bond UCTIS can be impacted at the aggregate level, despite the fact that they include a significant share of government bonds.

22. Short-term interest rates and market conditions also impact redemptions from funds invested in government bonds. We also run the regressions of Tables 2 and 3 on a sample in which we remove corporate bond funds from the aggregate UCTIS bond fund category.6 On net inflows we find that the effect becomes statistically insignificant. On redemptions, we find a positive significant impact of short-term interest rates, which is halved relative to the impact on the entire sample of UCTIS bond funds. Moreover, the effect of short-term interest rates becomes insignificant when we control for the market environment (which has a significant impact), suggesting that the shock to government bond funds is transmitted through the effect of interest rates on market performance and volatility. These findings are also consistent with the hypothesis that, while government bond funds may suffer from aggregate redemptions in the event of a negative shock, these redemptions could be somewhat compensated by gross inflows resulting from a portfolio rebalancing from riskier fund strategies.

Scenario Analysis

23. Scenario analysis shows that, while a smooth interest rate adjustment would be well absorbed by investment funds, a severe interest rate rise would have significant adverse effects on the fund industry:

  • The first scenario considered encompasses a gradual return of EONIA to its 2015 average level over a period of one year. In this scenario, we estimate that the net outflows out of bond and equity funds would be in the range 1.4–1.8 percentage points while redemptions would be 3–4 percentage points.

  • The second scenario illustrates a severe stress situation in which EONIA returns to its 2007 average level over a period of one year, meaning an interest rate rise of around 4 percent. In such a scenario of a very abrupt adjustment, the impact on the investment funds’ industry would be severe: net outflows would be in the range 11.8–14.6 percentage points while aggregate redemptions would be 24.7–22.6 percentage points. This scenario considers only the direct impact of interest rates on investment funds’ flows, holding market conditions constant. In the event of a large interest rate shock, financial market performance would fall and volatility would sharply increase. These market developments would in turn cause additional net outflows and redemptions, as well as negative valuation effects.

Table 5.

Luxembourg: Interest Rate Scenario Analysis

article image

E. Conclusions

24. Accommodative monetary policy has benefitted Luxembourg’s economy. Private consumption (and possibly investment) has risen, possibly as a result of wealth effects of higher house prices, and lower borrowing costs. However, intergenerational effects may be at play as younger households have faced growing affordability issues (see Chapter 2). The banking system appears to have passed on the monetary stimulus, while protecting its lending margins, and has become more interconnected in recent years, including with a strongly expanding investment fund industry, but also with cross-border “other financial institutions” operating treasury offices from Luxembourg. The investment fund industry has been a significant beneficiary of monetary policy, as a result of portfolio rebalancing effects, search for yield, and strong valuation effects in particular in stock markets.

25. Luxembourg faces risks related to the normalization of monetary policy going forward. Normalization of interest rates could trigger aggregate redemptions from the investment fund industry, in particular if the shock is very large and sudden. This could amplify existing liquidity mismatches between the underlying assets and redemption terms. Net asset value would also decline as a result of valuation effects. Banks could be impacted possibly through aggregate deposit outflows, and declining revenues. Net interest margins on fixed rate mortgages would also be compressed. Some households who borrowed at floating rates would also see their mortgage debt service increase.

26. To build resilience against such risks, the authorities should continue enhancing regulation and supervision in line with the 2017 FSAP recommendations, and in particular:

  • Better risk prevention and resilience through strong risk monitoring and supervision of the banking system and its cross-border linkages, and of the investment fund industry, and by further developing internal capacity and methodologies for system-wide stress testing of investment funds, and provision of guidance to the industry on the design of liquidity stress tests;

  • Stronger macroprudential oversight that will provide adequate macroprudential instruments such as borrower-based limits, and that will reinforce the willingness to act;

  • Better risk management in particular through provision of guidance to the investment fund industry on liquidity management tools.

1

Prepared by Thierry Tressel (EUR).

2

Monetary policy announcements are not exogenous as they sometimes were a reaction to market events. It is however notable that a number of these announcements were followed by reversals in the patterns of net investment flows.

3

During 2017, flows into Luxembourg investment funds were substantial while the ECB continued QE and the Feb very gradually started to normalize monetary policy.

4

In unreported regressions, we control only for the lagged dependent variables. We find that the coefficients on the EONIA are 50–100 percent larger in absolute value without market indicators, which suggests that part of the effect of monetary policy may be transmitted through market conditions.

5

In unreported regressions with subscription flows as the dependent variable we find smaller positive or non-significant effects of a decline in interest rates, confirming that the impact of a low interest rate environment is to a large extent driven by the response of redemptions which have declined significantly at an aggregate level as interest rates fell.

6

These regressions are available upon request.

Luxembourg: Selected Issues
Author: International Monetary Fund. European Dept.
  • View in gallery

    Impact of Monetary Policy on Interest Rates in Luxembourg

  • View in gallery

    Accommodative Monetary Policy and Private Consumption

  • View in gallery

    Credit to Resident Non-Financial Corporations

    (Percent, Year-on-year growth rate)

  • View in gallery

    Domestically Oriented Banks’ Performance

  • View in gallery

    Lending Spreads

    (Percent)

  • View in gallery

    Banking System Net Income

    (Percent of total assets)

  • View in gallery

    Ratio of Loans to Households / Household Deposits, in Domestically Oriented Banks

  • View in gallery

    Individual Luxembourg Bank Exposures to Investment Funds and Money Market Funds, 2017:Q3

    (Share of total bank assets/liabilities)

  • View in gallery

    Net Inflows in UCITS Luxembourg Investment Funds and Unconventional Monetary Policy

    (Billions of euros)

  • View in gallery

    Evolution of Different Classes of Bond Funds

  • View in gallery

    Net Inflows in UCITS Luxembourg Investment Funds and Unconventional Monetary Policy

    (Billions of euros)