Mortgage lending and credit risk: Micro-level data analysis
| Název práce v češtině: | Hypoteční úvěry a kreditní riziko: analýza s využitím mikro dat |
|---|---|
| Název v anglickém jazyce: | Mortgage lending and credit risk: Micro-level data analysis |
| Klíčová slova: | hypoteční úvěry, kreditní riziko, DSTI, LTV, mikrodata, český bankovní sektor |
| Klíčová slova anglicky: | mortgage loans, credit risk, DSTI, LTV, microdata, Czech banking sector |
| Akademický rok vypsání: | 2021/2022 |
| Typ práce: | diplomová práce |
| Jazyk práce: | angličtina |
| Ústav: | Institut ekonomických studií (23-IES) |
| Vedoucí / školitel: | doc. PhDr. Adam Geršl, Ph.D. |
| Řešitel: | skrytý - zadáno vedoucím/školitelem |
| Datum přihlášení: | 29.06.2022 |
| Datum zadání: | 29.06.2022 |
| Datum a čas obhajoby: | 19.06.2024 09:00 |
| Místo konání obhajoby: | Opletalova, O314, místnost. č. 314 |
| Datum odevzdání elektronické podoby: | 29.04.2024 |
| Datum proběhlé obhajoby: | 19.06.2024 |
| Oponenti: | doc. PhDr. Ing. et Ing. Petr Jakubík, Ph.D., Ph.D. |
| Zásady pro vypracování |
| Credit risk in mortgage loan portfolios typically depends on the combination of borrower-level characteristics (leverage and income developments, often captured by the debt-service-to-income DSTI or debt-to-income DTI ratios), loan characteristics (initial loan conditions including the level of collateralization as measured by the loan-to-value ratio, LTV), and macroeconomic developments (interest rates, GDP, exchange rate, inflation), which impact the micro-level characteristics. The relationship between credit defaults and LTV and DSTI ratios is investigated for example by de Haan and Mastrogiacomo (2020), using a logit regression, or Saha et al. (2022), where authors examine the factors that drive household mortgage defaults using a probit regression approach.
Given that such analyses require detailed micro-level data, typically not publicly available, there is a lack of studies exploring the links between the borrower-level and macro-level determinants of credit risk. For the Czech Republic, for example, we are not aware of any study of this type. Moreover, mortgage lending tends to be regulated these days by macroprudential policy instruments. This holds also for the Czech Republic, where the Czech National Bank (CNB) sets caps on LTV, DSTI and DTI to maintain the financial stability of the Czech economy acknowledging that extremely high levels of those credit loan conditions might lead to higher credit risk for banks. Many papers examine the link between macroprudential policy instruments and household debt in many ways; however, again, there are only a few studies that focus on the impact of macroprudential borrower-based measures using micro-level data. Gross and Población (2017) assesses the efficiency of LTV and DSTI caps in four European countries using microdata from the Eurosystem Household Finance and Consumption Survey (HFCS). This study is extended by Jurča et al. (2020), which is focused on the case of Slovakia. Both conclude that borrower-based measures improve resilience of borrowers to macroeconomic shocks. |
| Seznam odborné literatury |
| [1] Cussen, M., O’Brien, M., and Onorante, L. (2015) 'Assessing the Impact of Macroprudential Measures,' Central Bank of Ireland Economic Letter, Series 2015, Vol. 3.
[2] Gerlach-Kristen, P. and Lyons, S. (2018) 'Determinants of mortgage arrears in Europe: evidence from household microdata,' International Journal of Housing Policy, 18(4), pp. 545–567. [3] Gross, M. and Población, J. (2017) 'Assessing the efficacy of borrower-based macroprudential policy using an integrated micro-macro model for European households,’ Economic Modelling, 61, pp. 510–528. [4] Gross, M., Tressel, T., Ding, X., and Tereanu, E. (2022) ‘What Drives Mortgage Default Risk in Europe and the U.S.?' IMF Working Paper No. WP/2022/065. [5] de Haan, L. and Mastrogiacomo, M. (2020) 'Loan to Value Caps and Government-Backed Mortgage Insurance: Loan-Level Evidence from Dutch Residential Mortgages,’ De Economist, 168(4), pp. 453–473. [6] Herrala, R. and Kauko, K. (2007) 'Household Loan Loss Risk in Finland - Estimations and Simulations With Micro Data', SSRN Electronic Journal. [7] Jurča, P. et al. (2020) 'The Effectiveness of Borrower-Based Macroprudential Measures: A Quantitative Analysis for Slovakia', IMF Working Papers, 20(134). [8] Saha, A., Rooj, D. and Sengupta, R. (2022) 'Loan to value ratio and housing loan default – evidence from microdata in India,' International Journal of Emerging Markets. |
| Předběžná náplň práce |
| The thesis will contribute to the literature on determinants of mortgage credit risk by conducting a quantitative analysis of the link between household defaults, borrower-based and loan-based characteristics, and macro-financial determinants including house prices, whose effect on credit risk (defaults) is often unclear. It will also take into account the impact macroprudential instruments by the CNB In contrast to previous analyses on this topic, this paper will work with detailed individual-level data, and the relationships will be tested in unique subsamples not available before, such as on regional level. Thanks to this approach, new perspectives and findings can be explored.
Methodology: The thesis aims to study the link between residential mortgage loan defaults and examined ratios (DSTI, DTI, LTV etc.) on anonymized individual-level data. To determine the relationships, the thesis will use the binary logistic model, where the dependent variable will distinguish whether the loan defaulted or not. Independent variables will include borrower and loan characteristics as well as selected macro-financial determinants including house prices. The main data source is an anonymized dataset of mortgages received from a large commercial bank in the Czech Republic, which covers over 150 thousand households. The volume of residential loans in the examined portfolio is over 300 billion CZK. The data sample will contain data up to 2022. The extent of the data will allow for the testing of hypotheses on different subsamples; apart from subsamples on regional level, the work intends to compare the estimations using resulting from initial and latest data. Hypotheses: 1. Hypothesis #1: The level of DSTI has a significant effect on the household default. 2. Hypothesis #2: The level of LTV has no significant effect on the household default. 3. Hypothesis #3: The significance and size of the effects differ across Czech regions. Outline: 1. Introduction and motivation 2. Literature review 3. Data analysis 4. Estimating models 5. Discussion of results 6. Conclusion and suggestion of further improvements |