Conference Agenda

Recent revisions to capital regulation link mortgage capital requirements to loan-to-value (LTV) ratios. We show that applying the same LTV-based regime to all U.S. banks would not increase capital requirements for most institutions, and when internal capital policies already embed LTV risk, incentives to rebalance mortgage portfolios across LTV buckets significantly weaken. Exploiting plausibly exogenous variation in LTV ratios, we document that regulatory capital buffers respond to both LTV variation at mortgage origination and subsequent market-driven changes, even absent explicit LTV regulatory prescriptions. Banks’ internal capital management and restrictions governing collateral revaluation shape the effectiveness of LTV-based capital regulation.

Using a unique Bank of England dataset spanning the UK residential mortgage market from 2005 to 2023, this paper examines the relationship between securitisation and loan default risk. We find that securitised mortgages exhibit higher default risk than non-securitised loans, consistent with banks selectively securitising higher-risk assets under asymmetric information. Analysing key regulatory interventions - including the Internal Ratings-Based (IRB) approach, the Mortgage Credit Directive (2016), and the Simple, Transparent, and Standardised (STS) framework (2019) - we show that this association weakens following each reform, reflecting improvements in underwriting standards and incentive alignment. Our findings offer policymakers a practical basis for evaluating regulatory effectiveness and advancing transparency and stability in the UK mortgage-backed securities market.

We study how banks and nonbank mortgage companies (NMCs) adjust their

lending relationship following a large monetary policy intervention. Using administrative

regulatory data on bilateral credit-line relationships between 105 large US

banks and 405 NMCs, we exploit within-NMC variation across banks to isolate

credit supply from demand. Following the 2020 quantitative easing (QE) program,

yield-seeking banks with lean liquidity buffers increased their credit-line supply

to NMCs while reducing their direct holdings of real estate loans and mortgagebacked

securities. Several patterns support a relative-return interpretation: credit

line yields remain stable around QE while Treasury and MBS yields decline, and

banks expand credit lines toward higher-yield NMC relationships. This balancesheet

reallocation was not indicative of excessive risk-taking: banks that reallocated

most aggressively exhibited stronger ex-post performance, including fewer security

losses and lower deposit withdrawals during the subsequent period of quantitative

tightening.

Strong conflicts of interest may arise within modern multidivisional banks, when the bank lending division possesses substantial private information on borrowers, and the investment banking division seeks such information, e.g., when acting as underwriters or bookrunners. However, sharing of private information could also benefit investors, if it provides certification on the quality of the securities sold. We study this issue in the context of asset-backed securities (ABS) deals, which represents an interesting case of banks acquiring information on loan quality while investors rely on limited and opaque information. Our results show that ABS pay a higher spread (+27bps) when bookrunners can access private information on the assets backing the deal, suggesting that investors discount the potential conflict of interest more than they price the possible certification effect. Crucially, our results control for the endogeneity of the matching between lender-originators and bookrunners and of the choice of whether to disclose the conflict of interest. We further exploit the introduction of mandatory disclosure in the ABS market (Regulation AB) that led the share of deals revealed to have potential information sharing to rise from 6% to 34%. We show that ABS with potential information sharing previously hidden request an additional spread. The impact is stronger

for more opaque deals, confirming that investors’ response is exacerbated by limited information.