Asset price inflation and monetary policy

Keynote speech by Yves Mersch, Member of the Executive Board of the ECB and Vice-Chair of the Supervisory Board of the ECB, at the celebration of INVESTAS’ 60th anniversary

Luxembourg, 27 January 2020

Introduction

The outlook for euro area economic activity and inflation at present warrants the highly accommodative monetary policy implemented by the ECB through our package of policy measures. Nonetheless, the Governing Council remains attentive to the potential side effects of those measures. Monitoring and analysing those side effects is part and parcel of the Governing Council’s ongoing monetary policy assessment and our recently announced strategy review.

Indeed, the prolonged period of substantial accommodation and the unconventional nature of our measures call for vigilance on the efficacy of the policy measures and might affect the strategic calibration and the appropriateness of the monetary policy stance. This vigilance is particularly warranted in the light of some signs that monetary policy is encouraging increased risk-taking and contributing to elevated asset price inflation and income inequality.

To my mind, this raises two concerns for the long-run efficacy of our measures, which I wish to discuss today.

The first is the contribution of stretched asset prices to vulnerabilities in the financial system, which may in turn trigger future crises. The second is the part they may play in creating a disparity between public perceptions of inflation and official measures. That disparity can undermine public support for our unconventional measures and eventually erode trust in the ECB.

Risk-taking and asset prices

There is substantial evidence that monetary accommodation incentivises financial firms to increase risk-taking behaviour. With looser policy, banks lend more to riskier firms,^[1] a tendency that is even more pronounced among banks with a high share of retail deposits in periods of negative rates.^[2] Lower interest rates also lead to higher leverage for investment banks.^[3]

Greater risk-taking extends beyond the traditional banking system. When non-bank financial institutions (for example pension funds and insurance companies) target a certain nominal rate of return, an environment of lower overall returns means they have to “search for yield” – in other words, acquire riskier assets to maintain nominal returns.^[4] We are already seeing some signs of increased risk-taking by non-banks such as pension and mutual funds.^[5]

This increased risk-taking channel is in part an intended aim of monetary policy. Greater lending by financial institutions supports consumption and investment, generating activity and inflationary pressures within the economy. The question is whether the increased risk-taking is excessive or not. At the moment, banking supervision is committed to de-risking of the banking system after the global financial crisis. And these unusual times call for heightened vigilance regarding the financial stability consequences of our monetary policy actions.

We have recently seen a marked increase in certain asset prices, which in some cases are at historically very elevated levels. For example, US equities are trading at unprecedented price-to-earnings ratios. Residential real estate prices in the euro area are also at historical highs relative to rents, notably in large cities.^[6] Residential property prices in the seven largest German cities have doubled since 2010, compared with a nationwide change of around 60%.^[7] In September 2019, the European Systemic Risk Board issued warnings to five countries including Luxembourg that medium-term vulnerabilities had been identified in residential real estate markets. It also issued stronger recommendations for remedial action to a further six countries.

To an extent, those elevated asset prices are a function of the long-running downward trend in interest rates experienced by advanced economies over the past three decades.^[8] That trend can be explained by a range of potential factors, including an ageing population, income distribution,^[9] rising saving in emerging markets^[10] and a general rise in risk aversion.^[11]

But concerns remain that the increase in asset and housing prices is excessive and results from the exceptionally long period of extremely accommodative monetary policy. As several economists have noted, monetary policy is a key determinant of the financial system’s ability to create money.^[12] In their view, money creation, credit creation and asset price determination are tightly interdependent.

Elevated asset prices relative to fundamentals increase the risk of a future correction in housing or equity markets. Such a correction would affect banks directly by reducing the value of collateral backing loans and indirectly by affecting confidence, leading to weaker overall economic activity. There is, therefore, no clean separation between the pursuit of monetary stability and that of financial stability in the medium term.

The limits of macroprudential policy

The risks to asset prices from loose monetary policy have brought macroprudential policies into sharper focus. These policies are exercised at the national level by the relevant competent authority where central banks are supposed to play a leading role. The idea here is that macroprudential policies can offset the build-up of risk, leading to an overall optimal policy mix. Certainly, macroprudential policies can be effective in restraining increases in residential property prices, but they are no panacea.

Lending by foreign-based banks dilutes the effectiveness of capital-based measures.^[13] Moreover, our current tools are bank-based and therefore exercise little control over the growing role of the non-bank financial sector in lending to households. For example, in the Netherlands around a third of mortgage lending to households is now provided by pension funds, insurers and mortgage funds.^[14]

Given these factors, it is worth considering whether the current suite of macroprudential tools remains appropriate. Capital measures, while effective at increasing bank resilience, are much less successful at leaning against excessive housing price inflation, in part due to the factors I just mentioned. Loan-to-value (LTV) and loan-to-income (LTI) constraints may be more effective since they act directly on borrowers and are more difficult to circumvent. Indeed, these measures have proven effective in reducing housing price inflation.^[15]

LTV limits increase the resilience of the household sector, reducing the risk of households finding themselves with negative equity if housing prices fall. LTI limits reduce the likelihood that some households will be forced to reduce non-housing expenditure if interest rates increase. This in turn helps mitigate the potential contagion from falling house prices to a more general economic downturn.

Yet LTV and LTI limits are themselves not without difficulties.

The sharp increases in house prices and rents witnessed in recent years have made it expensive to obtain housing. Insufficient housebuilding, owing to capacity constraints in the construction sector and regulatory restrictions, has exacerbated the shortage of housing. And the burden is felt most acutely by the poorest in society, who spend a larger share of their income on housing costs than richer households.

While LTV and LTI limits are relatively effective in slowing housing booms, they work by crowding out marginal borrowers, who are precisely those most affected by the housing shortage. LTV limits crowd out those without a large enough deposit, usually young and/or poor households, while LTI limits exclude low-income households.

This distributional impact of LTV and LTI limits can render their use politically controversial, and may lead to inaction bias, limiting the effective use of policy measures to counter housing price inflation. Dealing with distributional issues lies beyond the remit of central banks and prudential authorities, but we must remain mindful of such consequences arising from our actions.

Furthermore, there remains a need to clarify the range of available macroprudential tools and to calibrate the balance between rules and discretion. Greater clarity is required on governance arrangements, both nationally and internationally, as well as on the potential interaction with other policy areas. These are areas I have discussed on previous occasions.^[16]

A role for monetary policy in supporting financial stability

So there are good reasons to believe that macroprudential policy in the euro area is currently constrained in its effectiveness. Jeremy Stein has argued that when macroprudential policy creates leakages to foreign banks or the shadow banking sector, monetary policy should be used for financial stability purposes because “it gets in all the cracks”.^[17] That would mean using monetary policy at the European level to mitigate the build-up of risks in the financial sector.

Using monetary policy for these purposes – often termed “leaning against the wind” – brings benefits in terms of reducing the likelihood and severity of financial crises, but comes at the price of below-target inflation during booms.^[18] That is a trade-off that needs to be balanced carefully, taking into account the risks on both sides.^[19]

At the ECB, such considerations are incorporated into our monetary policy strategy via our two-pillar approach, which explicitly includes a role for monetary developments in our policy assessment. Strong bank credit growth will be reflected in faster-growing monetary aggregates, which would warrant monetary policy tightening in response.

House prices and trust in monetary policy

The risks arising from strong housing price inflation extend beyond financial stability.

At present, owner-occupied housing costs are not included in the Harmonised Index of Consumer Prices (HICP) that is used to formulate our inflation aim of below, but close to, 2% over the medium term. There are a number of technical explanations for this exclusion, but it is clear that households view the cost of housing as an important part of their lifetime expenditure. Rents represent around 6.5% of the basket used for measuring inflation. For many, rents alone or mortgage payments easily exceed a third of their take-home pay.

So there may be a significant gap between what households perceive to be the increase in their cost of living and what is measured by the HICP. Research shows that perceptions play an important role in determining economic behaviour.^[20]

Incorporating owner-occupied housing in the reference rate of inflation for monetary policy would provide a clearer signal for monetary policy to lean against housing price booms. Indeed, the United States, Japan, Sweden and Norway already integrate owner-occupied housing into their reference inflation indices. If it were to be included in the HICP, it could raise measured inflation rates in the euro area by around 0.2 to 0.5 percentage points in some periods.^[21] Taking that into consideration, core inflation would lift from its current 1.3% to its long-run trend, or even higher, thereby having a bearing on the monetary policy stance.

The gap between perceptions and official measures of inflation can complicate the communication of policy decisions. If households believe that inflation is rampant then they will see little justification for unconventional measures, in particular negative interest rates.

Trust in the ECB fell markedly following the onset of the crisis. According to the Eurobarometer poll, net trust averaged around 25 percentage points in the years before the crisis, but fell to a low of -23 in spring 2014.^[22] This fall was in line with that experienced by other EU institutions such as the Commission and the Parliament. But while net trust in the ECB has recovered somewhat, and now stands at -2 percentage points, trust in those other institutions has recovered more rapidly. Whereas levels of net trust in the ECB and the Commission were historically similar, net trust in the Commission now stands 10 percentage points higher.

Even more noteworthy is the now quite marked divergence between support for the euro and trust in the ECB. Prior to the crisis, net support for the euro and net trust in the ECB generally moved in line with one another, with the currency enjoying a level of support around 20 percentage points higher. Support for the single currency weathered the crisis fairly well, and now stands at its highest ever level. But the gap between the two measures now stands at 60 percentage points, and has persisted ever since the introduction of unconventional measures, although other country-specific factors might also have played a role.

This reduced trust can influence expectations and blunt the effectiveness of policy.^[23] Indeed, the low-for-long policy seems to have had little impact on the aggregate saving rate, although the increase in consumption since the beginning of 2019 has lagged behind real income growth. The so-called reversal rate may kick in at different points across sectors, with households more sensitive to the imposition – or even the fear of the imposition – of negative rates and more likely than market participants to behave in a fashion that counteracts the intended aim of the policy. The difference in savings cultures across European countries also plays a role.

A prolonged loss of trust in the ECB risks undermining the broad public support that is necessary for central bank independence. This is of particular concern when the range of non-conventional measures brings monetary policy closer to the realm of fiscal policy and the institutional effects of these policies are becoming more pronounced. It is vital, then, that our policy assessment incorporates insights from behavioural economics and political economy, rather than relying solely on linear models of the aggregate economy.

Conclusion

Let me conclude.

Asset prices are currently at very elevated levels. In part this is a consequence of long-running fundamental trends in interest rates. Yet there is substantial empirical evidence that monetary policy encourages risk-taking in the financial system, and the risks of an asset price correction are increasing.

Dealing with such risks solely through macroprudential policy is challenging. The effectiveness of macroprudential policies is curtailed by the presence of foreign banks and non-bank financial firms lending to households. Moreover, borrower-based measures can exacerbate the impact on inequality arising from housing shortages and housing price appreciation.

In such a situation, it is preferable for monetary policy to incorporate financial stability concerns into its policy deliberations. This is something that has long been recognised in our two-pillar strategy, which we would be well advised to maintain if not enhance.