Macroprudential Policy in the Irish Mortgage Market: Taking Stock - Deputy Governor Sharon Donnery

05 October 2021 Speech

Sharon Donnery

Remarks delivered virtually at Bank of Lithuania Macroprudential policy conference 2021

View presentation slides here

Ladies and gentlemen, it is a pleasure to be with you today to speak about the important topic of macroprudential regulation in the mortgage market, a topic that has grown hugely in importance to central banks in Europe since the global financial crisis, and one that is core to the mission of the Central Bank of Ireland. Today, I want to share with you our experience of the measures to date, including during the recent pandemic stress, and how we are now taking stock, listening to experiences and ensuring the framework is robust, balanced and effective in serving the public in an evolving real estate market.1

The measures – a recent history

One of our key goals in the Central Bank of Ireland is to safeguard financial stability – ensuring that the financial system can absorb, rather than amplify, adverse shocks, and that banks can continue to serve households and businesses through times of stress. Simply put, our approach to meet this goal is to build resilience when times are good, so that this resilience can be drawn upon when times are bad. The active use of macroprudential policies, including our mortgage measures, is key to meeting this goal.

Over the last half-decade we have had two main pillars of our macroprudential policies, the bank capital measures and the mortgage measures. More recently, we have also increased our analysis on macro-prudential policies for the non-bank sector, reflective of the relative size of the international industry in Ireland, as well as its increasing interconnectedness to parts of the domestic economy. On the bank capital side, we have a number of instruments at our disposal, including the countercyclical capital buffer (CCyB); the other systemically important institutions buffer (O-SII); and, the systemic risk buffer (SyRB). While our current assessment is that the macroprudential regime, both in banking and in the mortgage market, has generated substantial resilience, which could be drawn upon during the pandemic, we nonetheless recognise that the operating environment for our macroprudential regime is constantly evolving, and at certain junctures, a deeper review of our frameworks is necessary. That is why this year and into 2022 we are carrying out a framework review across its three pillars: the mortgage measures, the bank capital regime, and market-based finance.2 While all these policies of course interact, today I will focus on our mortgage measures. We are acutely aware that the mortgage measures directly affect individuals and families in Ireland, including their decisions about potentially purchasing a home. It is for this reason that the public interest is at the heart of our approach.

The Irish mortgage measures were introduced in February 2015, at a time when memories of the financial crisis were fresh in Irish minds. The effects of the crisis in Ireland were particularly severe, with the unemployment rate increasing by over ten percentage points to peak at about 16 per cent in 2011/2012, while property prices fell by over 50 per cent from their peak from 2007-2013. The financial crisis had left significant scars across all segments of the Irish economy and society. Losses on residential mortgages – following what had been a prolonged period of unsustainable lending standards – had contributed significantly to the banking crisis, which culminated in the Financial Measures Program of 2011 and associated support from the International Monetary Fund (IMF), the European Central Bank (ECB) and the European Commission.3

While unsustainable mortgage lending and the ensuing arrears crisis had direct effects on bank balance sheets and on households themselves, there were also wider effects: many Small and Medium Enterprises had seen their balance sheets impaired after misallocated borrowing to expose themselves to the property boom;4 consumption across the economy had been artificially inflated by housing wealth and collateral effects that had gone into sharp reverse:5 windfall revenues linked to the construction sector had given a misleading picture of how sustainable the Irish public finances were;6 and, negative equity was a reality for 40 per cent of mortgaged households, restricting borrowers with a desire to move home, to switch to avail of lower interest rates, or to borrow against their collateral for consumption purposes.7So the unsustainable mortgage lending, and subsequent crisis, entailed large and persistent costs for society as a whole.

Learning lessons from the past, since their inception in 2015, our measures regulate both loan to income and loan to value ratios at mortgage origination. They have had two objectives since 2015: increasing the resilience of banks and borrowers to negative economic and financial shocks, and dampening the pro-cyclicality of credit and house prices so a damaging credit-house price spiral does not re-emerge.

The measures – analytical support

Evidence-based policymaking is crucial for the Central Bank of Ireland and from their very inception, the measures have had economic research and data analysis at the heart of their design. Granular data collection had been central to our policy approach since a stress testing program in 2011 that accompanied the aforementioned recapitalisation package of the Irish banking system. During 2014, these data were used to provide an analytical support base for our decisions.

Firstly, the data showed clearly that originating loan to income and loan to value ratios had risen sharply in the years running up to the crisis, and that both were associated with higher default rates during the crisis.8

Secondly, the data showed that there was a substantial difference in default rates between first-time buyers and those moving home (referred to locally as second and subsequent buyers), with the latter much more likely to default, even when controlling for a wide range of observable factors.

Motivated by these research findings, our measures apply a loan to value limit of 90 per cent for first-time buyers while a loan to value limit of 80 per cent is in place for second and subsequent borrowers.9This approach has also been taken to address concerns regarding fairness, access to mortgage finance and homeownership for first-time buyers.

The performance of the measures against their objectives, 2015 - March 2020

Since their introduction, the measures have undoubtedly had an effect on the Irish mortgage and housing markets. Mortgages are among the longest maturity loans on banks’ balance sheets, but even after just six years, close to 40 per cent of loans on bank balance sheets have been issued under the measures.10

The measures have underpinned an improvement in credit quality. It is clearly very difficult to ascertain how borrowers and lenders would have reacted in the absence of our measures during a period characterised by weak housing supply and rapid growth in housing prices, following a more than 50 per cent decline from 2007 to 2013. However, we can look at the distribution of loan to income ratios among loans originated in recent years, and in the run-up to the previous crisis for an indication. The constraining role of the measures is clear, with a large bunching of borrowers at the loan to income limit of 3.5 in recent times, compared to a longer right tail of higher risk loans in previous times.

Aside from underpinning the resilience of borrowers and lenders, the measures have also guarded against the re-emergence of a credit-house price feedback loop. In the short run, this is evident both in survey data of house price expectations of market participants, which shifted downward immediately after the introduction of the measures. This indicated an understanding that previous cyclical dynamics in the housing market would be less likely with such restrictions in place. Similarly, the effects were evident in observed price data, with regions more acutely affected by the calibration of the loan to income and loan to value limits experiencing disproportionate slowdowns in price growth in 2015 and 2016.11Looking over the entire period since the measures have been introduced, counterfactual exercises suggest house prices may have been up to 25 per cent more expensive had the measures not been introduced to guard against looser lending standards.12

Performance through a crisis

The COVID-19 pandemic was the first major crisis since the introduction of the measures and our assessment is that, on the eve of an unprecedented shock, we arrived in March 2020 with a much more resilient banking system than would have been the case in the absence of our various macroprudential policy measures and the wider reforms to bank capital frameworks since the global financial crisis.13Borrowers were more resilient too, with originating loan to income and loan to value ratios at more modest levels than was the case during previous periods with similar economic conditions.

To mitigate the threat to public health, Ireland had implemented among the most stringent public health measures in Europe, leading to the shuttering of large sections of the economy. As a result, consumption fell by almost a fifth and 1.2 million people, or close to half the labour force were reliant on the State for either welfare payments or wage subsidies by Q2 2020.14This direct fiscal support was supported by immediate and accommodative monetary policy actions, while a series of macroprudential and microprudential actions allowed the banking sector to absorb the shock.15One of these specific actions, which resulted from extensive engagement between banks and their regulators, was the extension of payment moratoria to borrowers. By mid-summer, payment moratoria on more than €11.5 billion of debt had allowed households to postpone repayments. This provided critical liquidity relief which has thankfully only translated into longer-term difficulty for a very small share of the mortgage market, indicative of the success of the policy response and the strength of the recovery.

Research on the take-up of payment breaks by mortgage borrowers is instructive on the role of the mortgage measures during this period of stress.16Data show loans granted after the introduction of the measures had substantially lower levels of payment break take-up compared to those issued before the Global Financial Crisis. This is likely indicative of stronger underlying borrower resilience and more prudent levels of credit assessment by banks since the crisis, which have been copper-fastened by the measures.
 
Data on the relationship between payment breaks and the underlying level of loan to income and loan to value ratios provide further supportive evidence. Take for example, a borrower with a loan to income ratio of 2, who was half as likely to request a payment break compared to one with a loan to income ratio of 4. These comparisons reassure us that the measures played a role in preserving borrower resilience in the face of this extraordinary shock.
 
The resilience benefits are more directly visible in metrics such as default rates and payment break take-up rates. However, an equally important but less obvious benefit of the mortgage measures related to the resilience of the housing market to adverse shocks that could have resulted from overvaluation leading into the pandemic. As mentioned earlier, counterfactual estimates from Bank staff show that housing could have been much less affordable in March 2020 in the absence of the measures. Had we entered the pandemic with credit-fuelled overvaluation, the market would have been more vulnerable to a sharp house price correction. This would have come with ensuing elevated risks for borrowers and banks during the pandemic stress, and heightened risks of the banking sector acting as an amplifier of stress.

Our framework – taking a step back

I will now take a step back to discuss our framework. Each year since 2016, we have reviewed the mortgage measures against their stated objectives. In the past, changes implemented during these reviews have included an increase in the loan to value limit for some first-time buyers, and changes to the size and composition of allowances above the limits. While we will do this again in 2021, we are currently in the process of doing something significantly different in parallel: an overarching review of the entire framework around the mortgage measures. In this framework review, we are assessing deeper, longer-term issues to ensure our policy framework remains fit for purpose, not just now, but into the future.

At its core, this review will reflect the Bank’s goal of serving the public interest, with the areas of focus determined by: listening to our stakeholders, lessons learned from experiences across the globe; an assessment of key changes in the housing market and wider economic environment since 2015; and, focused evidence from our extensive analysis.

The macro-financial environment in Ireland

The first step to inform the review is to ensure we understand how the environment in which our measures operate may have changed since 2015. Research published today looks at how the housing and mortgage markets have evolved in Ireland in recent years, and also assess how Irish mortgage and housing affordability compare in a European context.

Domestically, much has changed in the housing market since 2015. The historical relationship between credit and house price dynamics is shifting, indeed weakening as the post-crisis increase in non-mortgaged actors in the housing market has maintained. The price elasticity of housing supply has remained very low in Ireland, perhaps beyond any of our expectations in 2014, meaning that recent price growth has simply not, up until now, been accompanied by the type of supply response needed to keep up with demographic demand. The cost of rents, for those looking to access new tenancies, has undergone significant growth, increasing affordability issues in the face of a continued imbalance between the demand and supply of housing services. These factors provide important context for any decisions we make with respect to the measures.17

Housing affordability pressures appear to be a challenge facing a number of societies globally. Some of the research published today, for example, suggests that a number of countries have seen house prices growing faster than incomes in recent years. And while, of course, it is difficult to have a consistent comparison of the level of house prices relative to incomes across countries, the same research suggests that estimates of the house price to income ratio in Ireland has remained below the median for a selection of Organisation for Economic Co-operation and Development (OECD) countries during the period of operation of the measures. This, of course, will not provide solace to an individual or couple experiencing difficulties in attempting to purchase a home. However, it does suggest the challenges being faced here are part of a wider set of global trends in housing markets.

Looking across countries at the experience of individual households in the mortgage market, it is clear that the era of the mortgage measures has been associated with a decrease in mortgage repayment burdens among those drawing down new mortgage finance.18This in part reflects the coincidence of the era of macroprudential mortgage policies with that of low global interest rates. This has had an influence in lowering mortgage borrowing costs across the continent.

However, the particularly large reductions in repayment burdens in Ireland have also been influenced by the implementation of the mortgage measures in Ireland.

Our research published today also tries to assess housing costs across the population in the round, using a range of survey sources, rather than focussing solely on house price and rental indices, which measure new transaction activity only. This research suggests that loan to income ratios across the mortgaged population are similar in Ireland to elsewhere in Europe, and have fallen more than most since 2013. Looking at renters, OECD estimates suggest that the share of lower-income households with particularly high rental burdens is below OECD median levels. The research released today notes that the particularly high share of households with subsidised rental costs in Ireland is likely to form part of that explanation. However, this doesn’t take away from the very real challenges for individuals that I mentioned above, particularly for new tenants, for whom rental costs have increased significantly.

The objectives of macroprudential mortgage policies

Aside from analysing how the Irish housing market has changed in recent years, a key aim of our overarching framework review is to reassess whether the objectives of our policies remain appropriate. As I’ve said, the current objectives of our policies are two-fold. Firstly, increasing the resilience of banks and borrowers to negative economic and financial shocks. Secondly, dampening the pro-cyclicality of credit and house prices so a damaging credit-house price spiral does not re-emerge. Collectively, these can be thought of as “benefits” of mortgage measures. Research published today by my colleagues with Professor David Aikman of King’s College London considers the range of benefits that accrue to an economy when such measures are in place.19In looking across the research literature in the area, one way to describe the benefits of such policies is that they accrue to all citizens, rather than new mortgage borrowers only, operating through the reduction of both the likelihood, as well as the severity, of a recession stemming from spiralling housing-credit dynamics.

Macroprudential mortgage policies achieve the above benefits through a range of channels. They restrain the effective demand for owner-occupied housing by limiting the contribution of mortgage finance to growth in house purchases relative to incomes. They also limit the role that exuberant expectations can have in amplifying house price growth, with a number of knock-on benefits, for example in lowering the risk of unsustainable growth in borrowing backed by property to fund consumption. Our own experience in Ireland up to 2008 shows clearly that housing booms infiltrate the entire economy when they become unsustainable, influencing the labour market, consumption, government spending and tax policy, as well as the obvious bank lending, borrower indebtedness, and housing overvaluation channels.

Our research with Professor Aikman also highlights the range of potential costs that macroprudential mortgage policies can impose on the economy, which are in many cases experienced more directly and immediately by the population. These macroeconomic costs include reduced activity in the housing market, reduced capacity to borrow financed by home equity, reduced housing-related consumption of items such as “white goods”, and potentially reduced housing supply. The literature is only beginning to emerge on some of these more immediate costs.20However, there is also the potential for broader, more difficult to measure longer-run costs that we must consider, such as the impact of housing equity on lifecycle consumption and the cost of housing servicing costs in retirement in the event that longer-term homeownership rates remain lower than was previously the case.

Overall, while there are potential downsides to macroprudential mortgage policies, our view is that the benefits heavily outweigh the costs. The benefits may at times not be as visible as the direct effects on individuals or households, but as we saw recently in a time of great crisis, the resilience built for both borrowers and lenders through the measures was crucial to avoid the financial stresses and economic hardships that could have emerged. In addition, for supply-constrained housing markets, it is important to remember there are many other policy levers that can be used to influence housing construction such as planning levies, building regulations and the tax system. As our research outlines, the economy is likely better served by a policy mix that stimulates additional housing supply through reductions in construction costs, rather than through increased price levels resulting from higher borrower indebtedness.21

Distributional effects

One topic I would like to discuss before I conclude relates to the distributional effects, or the effects that are particular to specific groups of the borrower population, of macroprudential mortgage measures. Looking at data on borrowers, it is clear that the market in Ireland has been changing during the era of our macroprudential mortgage policies. The age profile of borrowers has changed, with the median first-time buyer age increasing from 30 to 35 since 2008.22The income profile has also changed, with average incomes of mortgage entrants continuing to increase since 2015. Taken together, these findings of increasing age and income among mortgage borrowers at first glance may suggest the measures have played a role in limiting access to the market among younger, lower-income borrowers.

However, these findings must be set in a wider context. Firstly, the Irish population at large is ageing, and a range of societal and economic trends have been contributing to the rising age of mortgage market participants. Secondly, incomes at large have been rising during the economic recovery, as evidenced by the similar growth rates among new first-time borrowers and the population under 50 since 2015. Thirdly, there has been a continual and, in a cumulative sense, expanding gap between demand and supply in the housing market. This has exerted continued upward pressure on house prices and has caused an increase in the number of borrowers, especially at lower incomes, who are bound by the measures.

While a loosening of mortgage borrowing limits may seem attractive as a way to address such distributional outcomes by improving access to the mortgage market among those most constrained, there is much uncertainty as to how this would work in practice. In a supply-constrained market, the increased purchasing power of all borrowers could result in similar levels of access to the mortgage market, coupled with higher levels of prices and, as a result, more highly-indebted households. Simply put, more money chasing the same number of homes resulting in higher debt for households and individuals.

As things stand, there are two features of our framework that address issues relating to borrowing by specific cohorts. Firstly, the difference in the loan to value limit on first-time buyers relative to second and subsequent borrowers acknowledges both that the evidence suggests first-time buyers are lower-risk, but also that accumulation of down-payments can be more challenging for this group, for whom access to a mortgage implies access to homeownership.23

Secondly, the existence of allowances for lending above each macroprudential limit allows for flexibility and the consideration of the circumstances of individual borrowers. This means that the allowances act as a separate channel through which effects on certain borrower groups can be considered. Of course, lenders make their own commercial decisions on the borrowers receiving these allowances but the research up to now suggests that borrower groups constrained by the measures do indeed receive allowances.24However, the data also suggest that higher-value properties are more likely to receive an allowance, with this effect driven by the predominance of borrowers in urban centres among those accessing allowances. Taken in the round, the evidence base suggests that allowances play a role with respect to the distributional consequences of macroprudential policy, but cannot be a substitute for government policy to address wider issues in the housing market.

Conclusion

I have outlined a wide, but non-exhaustive, list of issues currently being considered by the Central Bank as part of our overall framework review. Our regular annual review will conclude as normal later this year while into 2022, we will continue our work to ensure our strategy and toolkit remain fit for purpose in view of the evolution of our financial system and economy.

A challenge for all macroprudential policy makers is to continue to make the case that the benefits of our policies justify the existence of the measures. These benefits are of course invisible to citizens day-to-day as they relate to the mitigations of risks relating to economic recessions and financial crises. This is a challenge the Central Bank has worked hard to meet, regularly explaining the decisions we have taken and sharing the evidence we consider through publishing our research and analysis. We will continue this approach in communicating the outcome of the framework review next year.

As the Central Bank of Ireland has stated on numerous occasions, these measures are a permanent feature of the Irish mortgage market. They play a key role in increasing the resilience of banks and borrowers to negative economic and financial shocks, and dampening the pro-cyclicality of credit and house prices so a damaging credit-house price spiral does not re-emerge. We have recently seen the policies contribute to meeting their objectives as the resilience built up since the introduction of the mortgage measures showed its worth during the pandemic.

The Central Bank of Ireland’s mission is to serve the public interest by maintaining monetary and financial stability while ensuring that the financial system operates in the best interests of consumers and the wider economy. Our mission will continue to guide our decisions as we conclude our regular annual review this year, and as we conduct our wider framework review over the next year.

I thank you all for your attention.


 

1I would like to thank Fergal McCann, Robert Kelly and Paul Reddan for their contributions to my remarks.

2Donnery, Sharon. Macroprudential Policy – Lessons in the Pandemic Era. Address to Virtual Workshop at the Central Bank of Ireland (19 February 2021).

3 Donnery, S., Fitzpatrick, T., Greaney, D., McCann, F., and O’Keeffe, M., 2018. Resolving Non-Performing Loans in Ireland: 2010-2018. Central Bank of Ireland, Quarterly Bulletin Signed Article, QB2 2018.

4 McCann, F. and McIndoe-Calder, T., 2014. Property debt overhang: the case of Irish SMEs. Central Bank of Ireland, Research Technical Paper Vol. 2014, No.14.

5LeBlanc, J. and Lydon, R., 2019. Indebtedness and spending: What happens when the music stops? Central Bank of Ireland, Research Technical Paper Vol. 2019, No.14.

6Addison-Smyth, D. and McQuinn, K., 2010. Quantifying Revenue Windfalls from the
Irish Housing Market. The Economic and Social Review, Vol. 41, No. 2, Summer, 2010, pp. 201–223.

7 Kennedy, G. and Lyons, P., 2017. Negative equity in the Irish housing market: recent developments. Central Bank of Ireland, Macro-Financial Review 2017:II.

8Hallissey, N., Kelly, R., and O’Malley, T., 2014. Macro-prudential Tools and Credit Risk of Property Lending at Irish banks. Central Bank of Ireland, Economic Letter, Vol. 2014, No.10.

9Buy-to-let buyers need to have a minimum deposit of 30%, further details on the mortgage measures can be found here.

10Mortgages issued under the mortgage measures framework are those mortgage loans approved and drawn down since 9 February 2015. See Central Bank of Ireland Financial Stability Review 2021:I for further details.

11Acharya, V.V., Bergant, K., Crosignani, M., Eisert, T. and McCann, F.J., 2020. The anatomy of the transmission of macroprudential policies (No. w27292). National Bureau of Economic Research.

12The range of estimates in the local literature is reflective of the uncertainty inherent in such exercises, with some estimates closer to 10 per cent.

13For example, the Basel III reforms.

14Supports included the pandemic unemployment payment (PUP), the temporary wage subsidy scheme (TWSS), the employment wage subsidy (EWSS) and support via the live register. See Central Bank of Ireland Quarterly Bulletin No.3, 2020 for further detail.

15See Central Bank of Ireland press release, “Focused on protecting consumers and supporting individuals in financial difficulty”, March 2020.

16Gaffney, E., and Greaney, D., 2020. COVID-19 payment breaks on residential mortgages. Central Bank of Ireland, Financial Stability Note, Vol. 2020, No.5.

17Donnery, Sharon. Perspectives on the Irish Housing Market – the past five years. Address to Annual ESRI/Department of Housing Conference on the Irish housing and mortgage market (13 November 2019).

18Kelly, J., Kennedy, G., and Lambert, D., 2021. The cost of housing and indebtedness across European and OECD households. Central Bank of Ireland, Financial Stability Note, Vol. 2021, No.10.

19Aikman, D., Kelly, R., McCann, F., and Yao, F., 2021. The macroeconomic channels of macroprudential mortgage policies. Central Bank of Ireland, Financial Stability Note, Vol. 2021, No.11.

20See for example Richter et al. (2019), who show that LTV restrictions do indeed reduce economic output, but that a relatively large reduction of 10 points in the LTV ratio is required to reduce output by as much as a 25 bps hike in monetary policy rates.

21Aikman, D., Kelly, R., McCann, F., and Yao, F., 2021. The macroeconomic channels of macroprudential mortgage policies. Central Bank of Ireland, Financial Stability Note, Vol. 2021, No.11.

22Gaffney, D., and Kinghan, C., 2021. Mortgage lending in Ireland during the 2010s. Central Bank of Ireland, Financial Stability Note, Vol. 2021, No.9.

23Kelly, R., O’Malley, T., and O’Toole, C., 2015. Designing Macro-prudential Policy in Mortgage Lending: Do First Time Buyers Default Less? Central Bank of Ireland, Research Technical Paper, Vol. 2015, No.2.

24Kinghan, C., and McCann, F., 2019. Lending above macroprudential mortgage limits: The Irish experience since 2015. Central Bank of Ireland, Financial Stability Note, Vol. 2019, No.8.