Capital requirements and mortgage pricing: Evidence from Basel II☆,☆☆
Introduction
Capital requirements play a central role in financial regulation and have been at the forefront of academic and policy debates since the 2008 crisis. Capital requirements affect banks’ risk taking (Acharya and Steffen, 2015), and can smooth or reinforce the credit cycle (Behn, Haselmann, Wachtel, 2016, Jiménez, Ongena, Peydró, Saurina, 2017). Heterogeneity in regulatory treatment for similar assets can affect the competitive environment (Repullo, Suarez, 2004, Behn, Haselmann, Vig), and create incentives for regulatory arbitrage (Acharya et al., 2013). The financial crisis had its roots in the mortgage market,1 but mortgages have received limited attention in the otherwise extensive literature on the effects of capital regulation.
This paper provides new evidence on the relationship between capital requirements and mortgage lending, with particular relevance to lending behavior in crises. We develop a new strategy to identify the pass-through of capital requirements to mortgage rates by taking advantage of quasi-experimental variation arising from a regulatory regime change. Our novel, loan-level data on requirements and prices allow us to control for demand and lender-level effects by using within-lender variation. Our baseline estimates show that each 1pp decrease in risk-weighted capital requirements as a result of the implementation of Basel II led to a 10–16bp decline in mortgage rates on average, controlling for risk and alternative mechanisms.
The implementation of the Basel II regime in the UK resulted in a large, coordinated fall in mortgage capital requirements across most lenders. This implementation occurred in January 2008, although the reforms were originally agreed in 2004. From that date, UK banks were allowed to use internal risk models to calculate capital requirements. The objective was to make regulation more risk sensitive, but the reforms also had the effect of sharply reducing average risk weights for residential mortgages. By accident, these reforms came in the midst of the financial crisis, giving us some insight into the effect of lowering aggregate capital requirements in a crisis period. While the timing of Basel II’s implementation was coordinated across banks, not all lenders were equally affected. The reforms led to considerable heterogeneity in capital requirements both across and within lenders, only some of which can be explained by differences in underlying risk. The residual variation in capital requirements is what we use for identification.
To implement our strategy, we construct a unique dataset joining several regulatory databases. Our main dataset contains loan-level data on approximately 7 million mortgages originated between 2005 Q2 and 2015 Q4 in the UK from the Financial Conduct Authority’s Product Sales Database (PSD). We merged our main dataset with: (1) new granular data on mortgage risk weights that vary by lender, year and leverage (loan-to-value, or LTV ratio) for all lenders that adopted internal models; (2) additional information on lender-specific capital requirements and resources from the Prudential Regulation Authority’s Historical Regulatory Database (De Ramon et al., 2016); and (3) loan-level performance data for a subsample of 1.3 million mortgages, from 2010 and 2011 snapshots from the Financial Conduct Authority and the Council of Mortgage Lenders.
Unlike most of the literature on capital requirements, our focus will be on mortgage rates (prices) rather than quantities. Prices are important for two reasons. First, interest rates affect the size and quality of the pool of applications that lenders receive (Jaffee, Russell, 1976, Stiglitz, Weiss, 1981). Second, interest payments can affect borrowers’ consumption, particularly since many UK mortgage borrowers are liquidity-constrained with higher marginal propensities to consume out of liquidity shocks (Cloyne, Surico, 2016, Kaplan, Violante, 2014). While we do not explore these two issues in depth, providing evidence on the pass-through of capital requirements to interest rates is a necessary step toward a complete understanding of the effect of capital requirements. Our loan-level price data allow us to study this under-examined dimension, and together with estimates of pass-though can inform the debate about the cost of capital regulation (Kisin, Manela, 2016, Plosser, Santos).
Our identification strategy consists of two complementary steps. In the first step, we focus on the differential effects between lenders who adopted internal models and those who did not. Then we use our new, granular information on risk-weighted capital requirements to estimate pass-through on mortgage rates.
Our first step uses a triple-difference regression model to exploit the quasi-experimental variation associated with the switch to Basel II to identify the causal impact of differences in risk-weighted capital requirements on interest rates. Under Basel I, risk weights were uniform within the same asset class: every mortgage from every lender received a 50% risk weight. Under Basel II, lenders had to self-select into using one of two methods for calculating risk weights: internal-ratings based (IRB) models or the standardized approach (SA). Selection was based largely on lender size, due to economies of scale in developing and maintaining risk models (Competition and Markets Authority, 2015). Firms stood to benefit from lower risk weights under IRB versus SA, but the fixed costs of adoptation were large and the choice practically irreversible.2 All the largest UK lenders adopted IRB models, while the smaller ones, with few exceptions, opted for the standardized approach.
Average risk weights, shown in Fig. 1, fell for both groups, but considerably more so for the IRB group, giving IRB lenders an advantage (the IRB-SA gap), on average, over SA lenders in terms of (lower) risk-weighted capital requirements. The size of this IRB-SA gap was larger for lower LTV ratios.
Fig. 2 shows changes in mortgage rates. There is an obvious downward trend in average rates after the financial crisis, following from the Bank of England’s cuts to the policy rate. But relative to this trend, the pattern of variation between IRB versus SA, and high versus low LTV ratios, is similar to the pattern for risk-weighted capital requirements. The lower panel shows the difference between IRB and SA prices, which removes the macro trend. IRB lenders price lower than SA lenders on average. Under Basel I, this gap is little different between high and low LTV ratios, and evolves in parallel over time. But under Basel II, a differential opens up and persists between high and low LTV ratios, consistent with IRB lenders exploiting their comparative advantage at low LTV ratios by pricing more aggressively.3
We estimate a triple difference model leveraging only the joint variation across lenders (IRB/SA), over time (pre-/post-Basel II), and across LTVs (low/high). This strategy allows us to control for bank-specific factors (e.g. funding costs), time-invariant specialization in low/high LTV, and any different effect of the crisis at low/high LTV. We add a rich set of borrower, loan, and property characteristics that capture most of the information available to the lender at origination. Soft information does not play an important role in UK mortgage pricing, but we test the quality of our controls with an analysis of ex-post arrears.
Finally, we address other possible causal interpretations of the triple difference estimates. Differential exposure of IRB and SA lenders to the financial crisis could also have led to differences in the relative pricing of low- versus high-LTV mortgages. To test alternative channels, we horse-race the adoption of IRB models against measures of lenders’ exposure to the financial crisis: access to securitization, foreign asset holdings, and capital buffers. Overall, we find that as a result of the risk weight ‘discount’, IRB lenders decreased prices by 32bp and increased their low-LTV portfolio shares by 12pp relative to SA lenders. These effects do not seem to be driven by borrower selection or by differential exposure to the crisis.
In the second step of our identification strategy, we follow the same intuition, but quantify the pass-through of risk-weighted capital requirements on rates. We do so by exploiting the substantial heterogeneity in risk weights as captured by our survey data.4 We combine the quasi-experimental variation coming from the switch to Basel II with our granular information on risk weights. In particular, we construct an intensity-of-treatment measure (which varies in cross sections of lenders and LTV ratios) using observed granular variation in capital requirements and risk weights. We then interact this measure with a dummy for the Basel II period in order to quantify the effect of differences in capital requirements on mortgage rates. Thus, in this second model we can exploit the differential effect of the switch to Basel II not only between the two groups of lenders that adopted IRB and SA, but also within the group of IRB lenders. For the latter, variation in risk weights is driven by differences in their specific internal models.
We find that changes in capital requirements can have a strong effect on mortgage lending: a 1pp point lower risk-weighted capital requirement led to a reduction in mortgages rates by 10–16bp. This effect is predominantly driven by capital-constrained lenders with low buffers above minimum regulatory requirements. We also find stronger pass-through for the largest lenders. These lenders originate about 90% of all mortgage lending and any change in their pricing is likely to have implications for aggregate mortgage supply. Thus, our results speak to the policy debate about countercycal capital requirements as a possible tool to sustain mortgage lending during a crisis.
Overview. The rest of the paper is organized as follows. Section 2 describes the setting and the data. Section 3 sets out our relationship to the literature and develops our main hypothesis. Section 4 explains the identification strategy. Section 5 presents our results and robustness checks. Section 6 concludes and discusses policy implications.
Section snippets
Background
Under the Basel Accords (implemented in the EU under the Capital Requirement Regulations) banks have to meet capital adequacy requirements which are expressed as a percentage of risk-weighted assets (RWAs).5 Currently, banks are required to hold capital resources of at least 8%
Related literature
Our paper is related to the empirical literature on the effects of capital requirements. We highlight our contribution to three strands of literature. First, we provide evidence that the two-tiered system of capital requirements introduced by Basel II had an effect on competition. Repullo and Suarez (2004) theorised that lenders would specialize in the segment of the loan market for which their chosen methodology to calculate risk weights (and thus capital requirements) gave them a comparative
Identification strategy
This section explains how we identify the causal effect of risk-weighted capital requirements on mortgage rates. For the reasons outlined in the previous section, we expect risk-weighted capital requirements to have a positive effect on prices.
Risk weights and prices are indeed strongly correlated (the Pearson correlation coefficient is around 0.6), and the graphical analysis discussed in the introduction (see Figs. 1 and 2) shows strikingly similar patterns of variation in prices and risk
Results
In this section we describe our results. First, in Section 5.1 (triple difference model) we exploit the change in regulatory regime from Basel I to Basel II and discuss the effect of the IRB-SA risk weight gap on prices, portfolio shares, and lending growth. Then, in Section 5.2 (pass-through model), we leverage our granular data on variation in risk-weighted capital requirements to estimate their marginal effect on prices.
Conclusions
In this paper, we exploit within-lender variation to identify the pass through of capital requirements to mortgage rates, and find that each 1pp point lower risk-weighted capital requirement led to a reduction in mortgages rates by 10-16bp. Our variation comes from risk-weighted capital requirements for UK mortgage lending that vary by lender, year and leverage (i.e. the loan-to-value ratio, or LTV). The 2008 implementation of Basel II’s new rules to calculate risk weights provides us with a
CRediT authorship contribution statement
Matteo Benetton: Methodology, Formal analysis, Writing - original draft. Peter Eckley: Methodology. Nicola Garbarino: Methodology, Formal analysis, Writing - original draft. Liam Kirwin: Formal analysis, Writing - original draft. Georgia Latsi: Formal analysis.
References (54)
- et al.
Securitization without risk transfer
J. Financ. Econ.
(2013) - et al.
The greatest carry trade ever? Understanding eurozone bank risks
J. Financ. Econ.
(2015) - et al.
Capital requirements and bank behavior in the UK: are there lessons for international capital standards?
J. Bank. Financ.
(2012) How do banks set interest rates?
Eur. Econ. Rev.
(2008)- et al.
Bank size and risk-taking under basel II
J. Bank. Financ.
(2011) - et al.
Bank lending during the financial crisis of 2008
J. Financ. econ.
(2010) - et al.
The failure of models that predict failure: Distance, incentives, and defaults
J. Financ. Econ.
(2015) - et al.
Loan pricing under Basel capital requirements
J. Financ. Intermed.
(2004) - et al.
Loan pricing under Basel II in an imperfectly competitive banking market
J. Bank. Financ.
(2008) From financial crisis to great recession: the role of globalized banks
Am. Econ. Rev.
(2012)
Using loopholes to reveal the marginal cost of regulation: the case of fuel-economy standards
Am. Econ. Rev.
Group of Governors and Heads of Supervision Announces Higher Global Minimum Capital Standards
Outcome of the Consultative Process on Proposals for International Convergence of Capital Measurement and Capital Standards
A New Capital Adequacy Framework
International Convergence of Capital Measurement and Capital Standards
An Assessment of the Long-term Economic Impact of Stronger Capital and Liquidity Requirements
Bank for International Settlements
Regulatory Consistency Assessment Programme (RCAP): Analysis of Risk-weighted Assets for CreditRisk in the Banking Book
Higher bank capital requirements and mortgage pricing: Evidence from the counter-cyclical capital buffer
Rev. Financ.
Procyclical capital regulation and lending
J. Financ.
Risk heterogeneity and credit supply: evidence from the mortgage market
NBER Macroeconomics Annual 2012, Volume 27
Estimating the elasticity of intertemporal substitution using mortgage notches
Technical Report
Regulatory capital arbitrage and the potential competitive impact of basel ii in the market for residential mortgages
J. Real Estate Financ. Econ.
Cited by (24)
Financial ratios, credit risk and business strategy: Application to the PDO Parma ham sector in single production and non-single production firms
2024, Journal of Agriculture and Food ResearchMacroprudential Policy and mortgage leverage decisions—Evidence from micro data
2023, Economic Analysis and PolicyComplexity and the default risk of mortgage-backed securities
2023, Journal of Banking and FinanceEuro area banking and monetary policy shocks in the QE era
2022, Journal of Financial StabilityBank credit risk and macro-prudential policies: Role of counter-cyclical capital buffer
2022, Journal of Financial StabilityCitation Excerpt :Auer et al. (2022) find that additional capital requirements resulting from the activation of the CCyB are associated with higher growth in banks’ commercial lending. Therefore, in addition to the minimum common equity requirement and the Capital Conservation Buffer (CCB), banks are also required to satisfy sufficient counter-cyclical capital buffers to sustain lending during a crisis and to maintain financial system resilience (Benetton et al., 2021; Bui et al., 2017; Bennani et al., 2014). The amount that banks are required to hold will ultimately depend on whether systemic risks are rising or declining (Jahn and Pirovano, 2019).
Capital requirements, mortgage rates and house prices
2022, Journal of Banking and FinanceCitation Excerpt :Tzur-Ilan (2019) study the effect of an increase in capital requirements for a very specific segment of the market: mortgages with variable interest-rate portions of 25 percent or more, high LTVs and high loan amounts. Benetton et al. (2021) use a triple difference-in-differences identification strategy to exploit variation in which IRB lenders benefitted from an additional reduction relative to STA lenders of 8.9 percentage points at low LTVs. In their benchmark model this results in a 32 basis points decrease in mortgage rates or 3.55 basis points for each percentage point change in risk weights.
- ☆
The authors contributions to this paper were completed while employed by the Bank of England and do not represent the views of their current employers. This paper was previously circulated as “Specialization in Mortgage Risk Under Basel II”. The authors appreciate comments and suggestions from Paul Calem, Klaus Düllmann, Alessandro Gavazza, Charles Grant, Paul Grout, Leonardo Gambacorta, Benjamin Guin, Tommaso Oliviero, Daniel Paravisini, José-Luis Peydró, Tarun Ramadorai, Victoria Saporta, Ieva Sakalauskaite, Glenn Schepens, Enrico Sette, Matthew Willison, members of the Banking Inquiry Panel of the UK Competition & Markets Authority; participants at conferences organised by Columbia University, the London School of Economics, the Bank for International Settlements, the European Banking Authority, the Federal Reserve Bank of Cleveland and the Office of Financial Research, the European Economic Association, the European Finance Association and the Bank of England. We would like to thank Paolo Siciliani for his advice throughout this research project and Marco Schneebalg and Peter McIntyre for valuable assistance.
- ☆☆
The views in this paper are those of the authors and do not necessarily reflect the views of the Bank of England, the Monetary Policy Committee, the Financial Policy Committee, or the Prudential Regulation Authority.