Mortgage Portfolio Research Articles

Optimizing Ensemble Learning to Reduce Misclassification Costs in Credit Risk Scorecards

Credit risk scorecard models are utilized by lending institutions to optimize decisions on credit approvals. In recent years, ensemble learning has often been deployed to reduce misclassification costs in credit risk scorecards. In this paper, we compared the risk estimation of 26 widely used machine learning algorithms based on commonly used statistical metrics. The best-performing algorithms were then used for model selection in ensemble learning. For the first time, we proposed financial criteria that assess the impact of losses associated with both false positive and false negative predictions to identify optimal ensemble learning. The German Credit Dataset (GCD) is augmented with simulated financial information according to a hypothetical mortgage portfolio observed in UK, European and Australian banks to enable the assessment of losses arising from misclassification costs. The experimental results using the simulated GCD show that the best predictive individual algorithm with the accuracy of 0.87, Gini of 0.88 and Area Under the Receiver Operating Curve of 0.94 was the Generalized Additive Model (GAM). The ensemble learning method with the lowest misclassification cost was the combination of Random Forest (RF) and K-Nearest Neighbors (KNN), totaling USD 417 million in costs (USD 230 for default costs and USD 187 for opportunity costs) compared to the costs of the GAM (USD 487, USD 287 and USD 200). Implementing the proposed financial criteria has led to a significant USD 70 million reduction in misclassification costs derived from a small sample. Thus, the lending institutions’ profit would considerably rise as the number of submitted credit applications for approval increases.

Climate Risks in the Commercial Mortgage Portfolios of Life Insurers: A Focus on Sea Level Rise and Flood Risks

In this paper, we explore the climate risk exposure of U.S. life insurers' commercial mortgage loan portfolios, focusing on sea level rise (SLR) and flood risks. While vast literature and industry reports document the potential impact of climate change on the insurance industry, most focus on the property insurance market. Life insurers hold a significant number of commercial mortgages, representing approximately 15% of their assets by the end of 2021. These investments are susceptible to both physical and transitional risks of climate change and are much less researched or understood. From 2012 to 2019, we find that the size of newly issued commercial mortgages by life insurers in the U.S. grew from $37 billion to $56 billion (by 51%). We find heterogeneity across life insurers in terms of the average loan size, the loan origination frequency, as well as loan locations. In terms of locations, life insurers tend to originate loans in metropolitan areas, many of which are on the coast. Some life insurers focus on specific geographical areas that happen to be exposed to high levels of SLR or flood risks (e.g., Florida), while some diversify across metropolitan areas (coastal or noncoastal) in the U.S. Our findings have important policy implications for regulators and researchers.

Applying Machine Learning Techniques to Evaluate Climate-Related Risks in Real Estate Mortgage Valuations

Facing the escalating effects of climate change, the real estate industry faces risks from physical perils and shifting towards a low-carbon economic model. These risks have substantial consequences for the assessment and effectiveness of real estate mortgage portfolios. Conventional approaches to evaluating mortgage risk frequently fail to capture the intricate, non-linear connections between climate variables and loan results. In this paper, we present a new machine-learning framework that aims to quantify climate-related risks in real estate finance. We utilize neural networks and gradient-boosting algorithms to forecast the likelihood of mortgage defaults and the potential loss resulting from defaults in different climate stress scenarios. A robust and forwardlooking risk assessment is developed by integrating property-level exposure data, loan characteristics, and macroeconomic indicators. The empirical findings prove that our models perform superior to conventional econometric methods regarding predictive precision and computational effectiveness. The framework offers a robust instrument for investors, lenders, and regulators who aim to effectively address climate risks and enhance their ability to withstand and adapt to an unpredictable future.

Addressing the economic and demographic complexity via a neural network approach: risk measures for reverse mortgages

The study deals with the application of a neural network algorithm for fronting and solving problems connected with the riskiness in financial contexts. We consider a specific contract whose characteristics make it a paradigm of a complex financial transaction, that is the Reverse Mortgage. Reverse Mortgages allow elderly homeowners to get a credit line that will be repaid through the selling of their homes after their deaths, letting them continue to live there. In accordance with regulatory guidelines that direct prudent assessments of future losses to ensure solvency, within the perspective of the risk assessment of Reverse Mortgage portfolios, the paper deals with the estimation of the Conditional Value at Risk. Since the riskiness is affected by nonlinear relationships between risk factors, the Conditional Value at Risk is estimated using Neural Networks, as they are a suitable method for fitting nonlinear functions. The Conditional Value at Risk estimated by means of Neural Network approach is compared with the traditional Value at Risk in a numerical application.

Impacts of extreme weather events on mortgage risks and their evolution under climate change: A case study on Florida

We develop an additive Cox proportional hazard model with time-varying covariates, including spatio-temporal characteristics of weather events, to study the impact of weather extremes (heavy rains and tropical cyclones) on the probability of mortgage default and prepayment. We compare the survival model with a flexible logistic model and an extreme gradient boosting algorithm. We estimate the models on a portfolio of mortgages in Florida, consisting of 69,046 loans and 3,707,831 loan-month observations with localization data at the five-digit ZIP code level. We find a statistically significant and non-linear impact of tropical cyclone intensity on default as well as a significant impact of heavy rains in areas with large exposure to flood risks. These findings confirm existing results in the literature and also provide estimates of the impact of the extreme event characteristics on mortgage risk, e.g. the impact of tropical cyclones on default more than doubles in magnitude when moving from a hurricane of category two to a hurricane of category three or more. We build on the identified effect of exposure to flood risk (in interaction with heavy rainfall) on mortgage default to perform a scenario analysis of the future impacts of climate change using the First Street flood model, which provides projections of exposure to floods in 2050 under RCP 4.5. We find a systematic increase in risk under climate change that can vary based on the scenario of extreme events considered. Climate-adjusted credit risk allows risk managers to better evaluate the impact of climate-related risks on mortgage portfolios.

A Framework for Integrating Extreme Weather Risk, Probability of Default, and Loss Given Default for Residential Mortgage Loans

This paper considers a hypothetical case in which a bank wants to build a routine climate stress test exercise on residential mortgage loans. The bank has regularly updated the probability of default (PD) and loss given default (LGD) on each residential mortgage loan under the internal-rating-based (IRB) approach of Basel II/III. Additionally, the bank estimates the stressed PD and stressed LGD associated with a predetermined extreme weather event. Using simulation techniques, this paper shows that the loss of the bank’s residential mortgage portfolio can reach a median of around 36% of the portfolio value. This remarkable loss comes from the effects of default correlation and property damage. Banks should pay more attention to such impacts of extreme weather events.

Proposal for calculating regulatory capital requirements for reverse mortgages

In the current context in which many people worry about the sustainability of pension systems, reverse mortgages are gaining popularity because they are a way to supplement elderly people's incomes. However, it is necessary to provide banks with an adequate risk measurement and management procedure for reverse mortgages to increase the commercialization of these products, which will result in greater well-being for the retirement age population. In this paper, we propose a method to measure risk and estimate the regulatory capital requirements for a portfolio of reverse mortgages owned by a financial institution according to Basel II and III. The method considers house price risk, mortality risk and interest rate risk; consequently, regulatory capital requirements need to be computed using a Monte Carlo simulation procedure. The proposed method is general and can accommodate several scenarios for reverse mortgage specifications, including fixed or variable mortgage rates and different income stream schemes (with the lump sum as a particular case). The results for the U.K. show that reverse mortgage providers face higher risk when the lender initially advances a higher amount, with the lump-sum case indicating the highest risk, for relatively younger borrowers, the female population, higher interest rates and floating mortgage rates.

The Nexus of climate crisis and asset liability management in frontier economies

Climate change is a pressing issue that is disrupting the economic systems through financial and monetary impairment. The extent is more severe in the frontiers markets and compliance with the COP26 agenda is the only recourse for low-income economies. The banks play a critical role in such countries to finance business operations and incorporating sustainability practices in asset liability management can ensure financial resilience. In this paper, we assess if climate change has any impact on the mortgage portfolio and deposit mobilization of the banking sector. Using a comprehensive sample of 175 banks across twenty-three frontier markets and employing data between 2011 and 2021, we found that climate risk is negatively associated with mortgage exposure and deposit growth. We argue that the impact on mortgages is via physical risk while for deposit mobilization the pressure emanates from transitional factors. The results remained robust even after differentiating the banks into big and small categories suggesting that climate change is bringing in risks that will impact the financial sector without any discrimination. The paper has important implications for both transitional and physical risks related to climate change.

ТЕНДЕНЦИИ РЫНКА ИПОТЕЧНОГО КРЕДИТОВАНИЯ РОССИИ

In the current conditions in the financial market, there is an active growth of mortgage portfolios, the development of mortgage programs, and interest from financial institutions in this area of investment. The purpose of the study is to analyze the current state and development trends of the mortgage lending market in the Russian economy. In the course of the study, problems and limitations in the development of the mortgage products market were identified, development prospects were identified, and innovative mortgage lending instruments in the Russian market were considered.

Climate Risks in the Commercial Mortgage Portfolios of Life Insurers: A Focus on Sea Level Rise and Flood Risks

Climate Risks in the Commercial Mortgage Portfolios of Life Insurers: A Focus on Sea Level Rise and Flood Risks

Geographic Dependence and Diversification in House Price Returns: The Role of Leverage

Abstract We analyze the time variation in the average dependence within a set of regional monthly house price index returns in a regime-switching multivariate copula model with a high and a low dependence regime. Using equidependent Gaussian copulas, we show that the dependence of house price returns varies across time with changes in credit market conditions, which reduces the gains from the geographic diversification of real estate and mortgage portfolios. More specifically, we show that a decrease in leverage, measured by the loan-to-value ratio, and to a lesser extent an increase in mortgage rates, are associated with a higher probability of moving to and staying in the high dependence regime.

The impact of macroeconomic scenarios on recurrent delinquency: A stress testing framework of multi-state models for mortgages

Transition probabilities between delinquency states play a key role in determining the risk profile of a lending portfolio. Stress testing and IFRS9 are topics widely discussed by academics and practitioners. In this paper, we combine dynamic multi-state models and macroeconomic scenarios to estimate a stress testing model that forecasts delinquency states and transition probabilities at the borrower level for a mortgage portfolio. For the first time, a delinquency multi-state model is estimated for residential mortgages. We explicitly analyse and control for repeated events, an aspect previously not considered in credit risk multi-state models. Furthermore, we enhance the existing methodology by estimating scenario-specific forecasts beyond the lag of time-dependent covariates. We find that the number of previous transitions have a significant impact on the level of the transition probabilities, that severe economic conditions affect younger vintages the most, and that the relative impact of the stress scenario differs by attributes observed at origination.

Estimation of Flood Risk on a residential mortgages portfolio

In the context of the rapid changes that have occurred in recent years, characterized by veritable 'black swans' such as the COVID-19 pandemic and extreme weather events that are occurring with increasing frequency, the issue of climate change has come into the focus of banking regulators and supervisors. Therefore banking institutions, if they are subject to the Single Supervisory Mechanism, have been called upon to develop (and, subsequently, to integrate into their business practices) methodologies for the identification, quantification and management of such risks, mainly under the profiles of: - Transition Risk, associated with policies undertaken by governments to foster climate change mitigation and adaptation; - Physical Risk, associated with the occurrence of extreme climatic events and its impact on the bank's assets. This paper analyzes one of the most significant hazards within the Physical Risk domain, which is Flood Risk. The measurement is focused on the prospective evolution of the flood events on a portfolio of mortgages secured by residential properties. The impact of this risk driver is subsequently reflected through the movement of appropriate transmission mechanisms on the LGD and PD parameters relating to the exposures in the scope. Finally, the effect on loan adjustments is provided, by recalculating the expected losses that result from the stressed projections. The flood risk projection is executed on a long-term timeframe, developing over 3 climate scenarios up to 2050. The choice of this hazard is due to its relevance in terms of frequency of events and harmfulness, a relevance that is confirmed by its inclusion in both the top-down climate stress testing exercises carried out by the ECB and in the bottom-up climate stress testing exercise promoted by the ECB itself in 2022 and carried out by the SSM Banks. A comprehensive simulation framework, structured as follows, is then presented: - a macro-climate scenario simulation engine; - the downscaling of these scenarios to obtain localized climate effects on individual properties; - the transmission of these effects into a depreciation formula for the individual property; - the LGD stress associated with the devaluation of the collateral property, and the PD stress that goes along with it, obtained by correlation.

Hazlitt, Hayek, and How the Fed Made Itself into the World's Biggest Savings and Loan

The Federal Reserve is the biggest savings and loan in the world, in addition to being the most dangerous financial institution in the world. It does not and cannot know what the results of its own actions will be. But its mortgage portfolio growth would not be surprising to students of Hazlitt, Hayek, and Mises, who expect governments and central banks always to try to expand their power.

Residential Mortgage Portfolio, Product Innovation and Performance of Commercial Banks in Kenya

This study sought to investigate the relationship between residential mortgage portfolios, product innovation, and performance of commercial banks in Kenya. The study was anchored on the Modern Portfolio Theory, Agency Theory, and Asymmetric Information Theory. The study adopted a correlational descriptive research design and data collected from the annual residential mortgage surveys conducted by the central bank of Kenya (CBK) on commercial banks covering a 13-year period from 2006 to 2018. Further, the financial statements of commercial banks and Kenya Bankers Association database were used as a source of secondary data. Data were analyzed via panel data approaches. The Baron and Kenny (1986) approach was used to test the hypothesis. The results revealed that residential mortgage portfolio attributes, namely: portfolio quality and mortgage interest return significantly influence bank performance. The effect of mortgage terms on the relationship between mortgage portfolio quality and performance was negative and statistically significant. Loan to value (LTV) ratio was however found not to significantly intervene in the relationship between residential mortgage portfolios and the performance of banks operating in Kenya. From the findings, the study suggested that bank managers pay attention to the institutional environment and product characteristics in designing their mortgage loan portfolios.

Credit Card Defaulters Using Machine Learning

Abstract: As an increasing number of purchasers depend upon the credit score card to pay their regular purchases in on line and bodily retail store, the quantity of issued credit score playing cards and the overpowering quantity of credit score card debt via way of means of the cardholders have hastily accelerated. Therefore, maximum monetary establishments need to address the problems of credit score card default similarly to the credit score card fraud together with credit score card dump. Both the credit score card verification implemented to the cardholders and the default threat control after card issued are essential to the destiny achievement of maximum monetary establishments According to the Federal Reserve financial statistics, the default charge on credit score loans throughout all industrial banks is at an all-time excessive for the beyond sixty-six months, and it's miles possibly to hold to climb in the course of 2022. The delinquency charge suggests the share of beyond-due loans in the borrower’s whole mortgage portfolio. The mountain climbing delinquencies will bring about a good sized amount of cash lose from the lending establishments, together with industrial banks. Therefore, banks ought to have a threat prediction version and be capable of classify the maximum relative traits which can be indicative of humans who've a better possibility of default on credit score. In 2013, patron spending encompassed about 69% of USA gross home product. Of the $3.098 trillion of tremendous patron credit score within the United States within the closing region of 2013, they have been revolving credit score card for over 25% of it ($857.6 billion). A small boom within the accuracy of figuring out excessive-threat loans may want to save you losses of over $eight billion. Because of the dangers inherent in one of these huge part of the economy, constructing fashions for patron spending behaviors to restrict threat exposures on this region is turning into extra critical. For this to be a feasible option, the predictions want to be moderately correct. A strong version isn't most effective a beneficial device for the lending establishments to determine on credit score applications, however it could additionally assist the customers to be privy to the behaviors that could harm their credit score scores. The number one motivation at the back of threat prediction is to make use of monetary statistics, for instance, enterprise transactional statistics, change statistics and patron transactions, and so on to foresee the patron’s enterprise overall performance or character credit score card statistics and to lower bathrooms and vulnerability. Several threat prediction fashions are primarily based totally on statistical methods, inclusive of nearest neighbor, discriminant evaluation, and logistic regression. The aim of credit score default prediction is to assist monetary establishments determine whether or not or now no longer to lend to a patron. The ensuing check is mostly a threshold price that lets in the choice-makers to make the lending choice. The well-known version relies upon at the monetary ratios, earnings account, and statistics at the stability sheet

Assessment of the current state of the housing mortgage lending market in the Russian Federation

This article analyzes the mortgage lending market in the Russian Federation. The dynamics of such important indicators of the mortgage market as the dynamics of the volume of mortgage loans issued, the average size of mortgage loans, the weighted average mortgage rate, indicators of commissioning of housing stock in Russia, the share of overdue debt in the mortgage portfolio is presented.

House price prediction with gradient boosted trees under different loss functions

ABSTRACT Many banks and credit institutions are required to assess the value of dwellings in their mortgage portfolio. This valuation often relies on an Automated Valuation Model (AVM). Moreover, these institutions often report the models accuracy by two numbers: The fraction of predictions within and range from the true values. Until recently, AVMs tended to be hedonic regression models, but lately machine learning approaches like random forest and gradient boosted trees have been increasingly applied. Both the traditional approaches and the machine learning approaches rely on minimising mean squared prediction error, and not the number of predictions in the and range. We investigate whether introducing a loss function closer to the AVMs actual loss measure improves performance in machine learning approaches, specifically for a gradient boosted tree approach. This loss function yields an improvement from to of predictions within of the true value on a data set of transactions from the Norwegian housing market between 2013 and 2015, with the biggest improvements in performance coming from the lower price segments. We also find that a weighted average of models with different loss functions improves performance further, yielding of the observations within of the true value.

Mortgage lending market in Russia in terms of exit out of the pandemic

Aim. The work aimed to summarize the results of the development of the mortgage lending market during the acute phase of the COVID-19 pandemic and predict its development in the near future.Tasks. The work was performed to investigate the volume of issuance and growth rates of mortgage lending in Russia; to analyze the dynamics of the total mortgage portfolio of the banking system of the Russian Federation (RF), as well as the average cost of 1 square meter of housing in the primary and secondary markets in a pandemic, combined with government support; demonstrate possible changes in the Russian mortgage market during the exit out of the COVID-19 pandemic and consequences of the implementation of state market support programs.Methods. The methodological foundation of the study was the general scientific methods of cognition, namely analysis, synthesis, generalization, visualization, and forecasting. In particular, the authors analyzed the indicators of the mortgage lending and real estate markets, identified causal relationships between the state support for mortgage lending and changes in the main parameters of the real estate market, as well as forecasting the dynamics of its individual indicators.Results. The work presents the main indicators characterizing mortgage lending in Russia from 2014 to the present time, including the aggregate mortgage portfolio of Russian banks. The study describes the average cost of 1 square meter of housing in the primary and secondary markets. The authors of the article predicted changes in the mortgage lending market, the state of the quality of bank loan portfolios and trends in the residential real estate market as a whole.Conclusion. Under conditions of the COVID-19 pandemic, mortgage lending has shown its incredible growth in association with government support. However, according to the authors, there will be significant changes in the future, which will be the consequences of such a rapid development, namely an increase in overdue indebtedness on mortgage loans, a shortage of demand for built housing which has sharply become expensive. Banks will be forced to tighten monitoring of mortgage programs and select carefully new borrowers, ensuring their highest quality.

OPTIMAL CROSS-CURRENCY MORTGAGE DECISIONS

In this paper, we study optimal mortgage decisions in a cross-currency setting. In particular, we address the question on how a household should optimally split its mortgage portfolio in a fixed rate mortgage in the domestic currency and an adjustable rate mortgage denominated in a foreign currency subject to some risk constraints. We propose an affine factor model which allows to jointly investigate the impact of variations in interest rates as well as of exchange rate fluctuations on mortgage decisions. As a case study, we apply our model to real data on Swiss and German mortgage markets and we estimate parameters using a quasi-Kalman filter approach. We then study the impact of different income splits, risk attitudes, and mortgage specifications on the household’s portfolio choice in a mean–variance optimization approach.