Data-driven optimization of peer-to-peer lending portfolios based on the expected value framework

In recent years, peer-to-peer (P2P) lending has been gaining popularity amongst borrowers and individual investors. This can mainly be attributed to the easy and quick access to loans and the higher possible returns. However, the risk involved in these investments is considerable, and for most investors, being nonprofessionals, this increases the complexity and the importance of investment decisions. In this study, we focus on generating optimal investment decisions to lenders for selecting loans. We treat the loan selection process in P2P lending as a portfolio optimization problem, with the aim being to select a set of loans that provide a required return while minimizing risk. In the process, we use internal rate of return as the measure of return. As the starting point of the model, we use machine-learning algorithms to predict the default probabilities and calculate expected values for the loans based on historical data. Afterwards, we calculate the distance between loans using (i) default probabilities and, as a novel step, (ii) expected value. In the calculations, we utilize kernel functions to obtain similarity weights of loans as the input of the optimization models. Two optimization models are tested and compared on data from the popular P2P platform Lending Club. The results show that using the expected-value framework yields higher return.     

A knowledge‐based framework for incorporating investor's preference into portfolio decision‐making

This paper describes a framework for the integration of a rule‐based system capable of identifying an investor's risk preference into a quantitative portfolio model based on risk and expected return. By inferring rules consisting of an investor's objective and subjective risk preferences, the integrated methodology provides the assets suitable for the preferences. Through investment in the portfolio composed of the assets, the investor is able to obtain the following bene?ts: reduction of costs and time spent to determine target assets, and alleviation of anxiety from ‘out‐of‐favor’ assets. The framework is applied to the development of a knowledge‐based portfolio system for constructing an investor's preference‐oriented portfolio. In the procedure of the system for ?nding an optimal portfolio, the system uses an arti?cial intelligence method of a case‐based reasoning to obtain preference thresholds for an investor when the investor's past investment records are available. Experimental results show that the framework contributes signi?cantly to the construction of a better portfolio from the perspective of an investor's bene?t/cost ratio than that produced by the existing portfolio models. Copyright © 2004 John Wiley & Sons, Ltd.     

A dynamic model of active portfolio management with benchmark orientation

This paper studies optimal dynamic portfolios for investors concerned with the performance of their portfolios relative to a benchmark. Assuming that asset returns follow a multi-linear factor model similar to the structure of Ross (1976) [Ross, S., 1976. The arbitrage theory of the capital asset pricing model. Journal of Economic Theory, 13, 342–360] and that portfolio managers adopt a mean tracking error analysis similar to that of Roll (1992) [Roll, R., 1992. A mean/variance analysis of tracking error. Journal of Portfolio Management, 18, 13–22], we develop a dynamic model of active portfolio management maximizing risk adjusted excess return over a selected benchmark. Unlike the case of constant proportional portfolios for standard utility maximization, our optimal portfolio policy is state dependent, being a function of time to investment horizon, the return on the benchmark portfolio, and the return on the investment portfolio. We define a dynamic performance measure which relates portfolio’s return to its risk sensitivity. Abnormal returns at each point in time are quantified as the difference between the realized and the model-fitted returns. Risk sensitivity is estimated through a dynamic matching that minimizes the total fitted error of portfolio returns. For illustration, we analyze eight representative mutual funds in the U.S. market and show how this model can be used in practice.     

Genetic Variation in Financial Decision‐Making

Individuals differ in how they construct their investment portfolios, yet empirical models of portfolio risk typically account only for a small portion of the cross‐sectional variance. This paper asks whether genetic variation can explain some of these individual differences. Following a major pension reform Swedish adults had to form a portfolio from a large menu of funds. We match data on these investment decisions with the Swedish Twin Registry and find that approximately 25% of individual variation in portfolio risk is due to genetic variation. We also find that these results extend to several other aspects of financial decision‐making.     

Endogenous Life‐Cycle Housing Investment and Portfolio Allocation

This paper develops a life‐cycle portfolio allocation model to address the effects of housing investment on the portfolio allocation of households. The model employs a comprehensive housing investment structure, Epstein–Zin recursive preferences, and a stock market entry cost. Furthermore, rather than resorting to calibration we estimate the value of the relative risk aversion and elasticity of intertemporal substitution. The model shows that housing investment has a strong crowding out effect on investment in risky assets throughout the life‐cycle. We further find that the effect of the presence of housing investment on households portfolio allocation is larger than the effect of having EZ recursive preferences.     

Survival or die: a survival analysis on peer‐to‐peer lending platforms in China

This study examines the probability of survival of online peer‐to‐peer (P2P) lending platforms in China. Our empirical findings show that shareholder's background, platforms’ risk management and institutional environment are three major factors that influence the probability of survival of P2P platforms in China. In addition, the effects of risk control and institutional environment are less pronounced in state‐owned P2P platforms. We also show that platforms that die from abscondence of owners tend to be more short‐lived than others, while platforms that die from liquidity problems are usually due to lack of high‐quality risk management techniques.     

Risk‐Adjusted Long‐Term Contrarian Profits: Evidence from Non‐S&P 500 High‐Volume Stocks

Can trading volume help unravel the long‐term overreaction puzzle? With portfolios of non‐S&P 500 NYSE stocks, we show that (1) both the high‐ and low‐volume (abnormal volume) contrarian portfolios earn a much higher market‐adjusted excess return than the normal‐volume contrarian portfolio, (2) however, when leverage‐induced risk is factored in, excess returns from contrarian portfolios with normal‐ and low‐volume stocks are insignificant, (3) only excess returns from high‐volume contrarian stocks are significant and cannot be explained by the time‐varying risk and return framework, and (4) such high‐volume, risk‐adjusted excess returns arise mainly from winner (glamour) stocks.     

Lock‐In Effects in Relationship Lending: Evidence from DIP Loans

Do prior lending relationships result in pass‐through savings (lower interest rates) for borrowers, or do they lock in higher costs for borrowers? Theoretical models suggest that when borrowers experience greater information asymmetry, higher switching costs, and limited access to capital markets, they become locked into higher costs from their existing lenders. Firms in Chapter 11 seeking debtor‐in‐possession (DIP) financing often fit this profile. We investigate the presence of lock‐in effects using a sample of 348 DIP loans. We account for endogeneity using the instrument variable (IV) approach and the Heckman selection model and find consistent evidence that prior lending relationship is associated with higher interest costs and the effect is more severe for stronger existing relationships. Our study provides direct evidence that prior lending relationships do create a lock‐in effect under certain circumstances, such as DIP financing.     

Time‐Varying Risk–Return Trade‐off in the Stock Market

We uncover a strong comovement of the stock market risk–return trade‐off with the consumption–wealth ratio (CAY). The finding reflects time‐varying investment opportunities rather than countercyclical aggregate relative risk aversion. Specifically, the partial risk–return trade‐off is positive and constant when we control for CAY as a proxy for investment opportunities. Moreover, conditional market variance scaled by CAY is negatively priced in the cross‐section of stock returns. Our results are consistent with a limited stock market participation model, in which shareholders require an illiquidity premium that increases with CAY, in addition to the risk premium that is proportional to conditional market variance.     

Where Is the Risk in Value? Evidence from a Market‐to‐Book Decomposition

We study the value premium using the multiples‐based market‐to‐book decomposition of Rhodes‐Kropf, Robinson, and Viswanathan (2005). The market‐to‐value component drives all of the value strategy return, while the value‐to‐book component exhibits no return predictability in either portfolio sorts or firm‐level regressions. Existing results linking market‐to‐book to operating leverage, duration, exposure to investment‐specific technology shocks, and analysts’ risk ratings derive from the unpriced value‐to‐book component. In contrast, results on expectation errors, limits to arbitrage, and certain types of cash flow risk and consumption risk exposure are due to the market‐to‐value component. Overall, our evidence casts doubt on several value premium theories.     

NEURO‐GENETIC PREDICTIONS OF CURRENCY CRISES

We create a neuro‐genetic (NG) model for predicting currency crises by using a genetic algorithm for specifying (1) the combination of inputs, (2) the network configuration and (3) the training parameters for a back‐propagation artificial neural network (ANN). The performance of the NG model is evaluated by comparing it with standalone probit and ANN models in terms of utility for a policy decision‐maker. We show that the NG model provides better in‐sample and out‐of‐sample performance, as well as provides an automatic and more objective calibration of a predictive ANN model. We show that using a genetic algorithm for finding an optimal model specification for an ANN is not only less laborious for the analyst, but also more accurate in terms of classifying in‐sample and predicting out‐of‐sample crises. For a sufficiently parsimonious, but still nonlinear, model for generalized processing of out‐of‐sample data, the creation and evaluation of models is performed objectively using only in‐sample information as well as an early stopping procedure. Copyright © 2011 John Wiley & Sons, Ltd.     

Economic capital for nonperforming loans

A portfolio of nonperforming loans requires economic capital. We present two models for forecasting the portfolio loss and its probability distribution. In the first model, the loss for each nonperforming loan entails a change in provision over the risk horizon. The risk determinants are the single-name concentration, measured by the Herfindahl–Hirschmann index, as well as a systematic factor and the idiosyncratic risk. Our second model allows for interportfolio diversification with a portfolio of performing loans because banks typically own both performing and nonperforming loans. In this model, the nonperforming loan is identified with its systematic risk. Both models allow for closed-form expressions of economic capital and for the capital charge of the single loan. We calibrate the macroeconomic model parameters statistically with a loss panel; the microeconomic parameters depend on the portfolio. The portfolio risk for nonperforming loans mainly depends on the volatility of the systematic economic factor. The dependence becomes more pronounced when interportfolio diversification is taken into account. The magnitude of interportfolio diversification is also marked. Finally, we calculate regulatory capital charges according to Basel II for past-due loans. The regulatory charges are on average smaller than our economic charges and, additionally, take the volatility of economic activity into account only implicitly.     

Optimal asset allocation for a large number of investment opportunities

This paper introduces a stock‐picking algorithm that can be used to perform an optimal asset allocation for a large number of investment opportunities. The allocation scheme is based upon the idea of causal risk. Instead of referring to the volatility of the assets time series, the stock‐picking algorithm determines the risk exposure of the portfolio by concerning the non‐forecastability of the assets. The underlying expected return forecasts are based on time‐delay recurrent error correction neural networks, which utilize the last model error as an auxiliary input to evaluate their own misspecification. We demonstrate the profitability of our stock‐picking approach by constructing portfolios from 68 different assets of the German stock market. It turns out that our approach is superior to a preset benchmark portfolio. Copyright © 2005 John Wiley & Sons, Ltd.     

On the Origins of Risk‐Taking in Financial Markets

Financial investment behavior is highly correlated between parents and their children. Using Swedish data, we find that the decision of adoptees to hold equities is associated with the behavior of both biological and adoptive parents, implying a role for both genetic and environmental influences. However, we find that nurture has a stronger influence on the share of financial assets invested in equities and on portfolio volatility, suggesting that financial risk‐taking is substantially environmentally determined. The parental investment variables substantially increase the explanatory power of cross‐sectional regressions and so may play an important role in understanding cross‐sectional heterogeneity in investment behavior.     

Some consequences of banks' LDC loans: A note

The realized returns on LDC loans made prior to 1980 and hypothetically sold at (below par) secondary market prices in either 1986, when a secondary market first began, or, say, in 1989 are moderately positive. The realized risk-adjusted returns to investors on a portfolio of large LDC lending bank stocks are unmistakenly lower than the return on a broad market index, though they are not significantly so. Similarly, although the difference is not statistically significant, the average realized return on LDC lending banks' stocks is well below that of a portfolio of non-LDC lending banks. The article concludes that while shareholders suffered economic losses as a result of LDC lending activity, these losses were not quantitatively large.     

Predatory mortgage lending

Regulators express growing concern over predatory loans, which we take to mean loans that borrowers should decline. Using a model of consumer credit in which such lending is possible, we identify the circumstances in which it arises both with and without competition. We find that predatory lending is associated with highly collateralized loans, inefficient refinancing of subprime loans, lending without due regard to ability to pay, prepayment penalties, balloon payments, and poorly informed borrowers. Under most circumstances competition among lenders attenuates predatory lending. We use our model to analyze the effects of legislative interventions.     

THE ROBUSTNESS OF ABNORMAL RETURNS FROM THE EARNINGS YIELD CONTRARIAN INVESTMENT STRATEGY

We examine the sensitivity of the abnormal profitability of the earnings' yield (E/P)‐based contrarian investment strategy to the following two risk measurement issues: (a) return‐measurement interval over which systematic risk is estimated and (b) time variation in systematic risk. We conduct our analysis using the capital asset pricing model to parameterize risk. We find that the estimates of systematic risk of E/P‐ranked portfolios are not sensitive to the return‐measurement interval. Consequently the abnormal profits to the E/P‐based contrarian investment strategy observed in prior studies are not artifacts of the return‐measurement interval. Furthermore, although both the raw and abnormal returns to E/P‐ranked portfolios exhibit mean reversion, time variation in systematic risk ensuing from this mean‐reverting behavior does not substantially affect abnormal profits to E/P‐ranked portfolios. JEL classification: G11, G12, G14     

The Over‐the‐Counter Theory of the Fed Funds Market: A Primer

We present a dynamic over‐the‐counter model of the fed funds market and use it to study the determination of the fed funds rate, the volume of loans traded, and the intraday evolution of the distribution of reserve balances across banks. We also investigate the implications of changes in the market structure, as well as the effects of central bank policy instruments such as open market operations, the discount window lending rate, and the interest rate on bank reserves.     

CEO Risk‐taking Incentives and Bank Loan Syndicate Structure

This paper investigates the effects of a borrowing firm's CEO risk‐taking incentives on the structure of the firm's syndicated loans. When CEO risk‐taking incentives are high, syndicates are structured to facilitate better due diligence and monitoring efforts. These syndicates have a smaller number of total lenders and are more concentrated, and lead arrangers will retain a greater portion of the loan. Moreover, CEO risk‐taking incentives have a lesser effect on the syndicate structure when lead arrangers have a good reputation and a prior lending relationship with a borrowing firm, while they have a greater effect on the syndicate structure when borrowing firms have low information transparency, are financially distressed or have low growth prospects.     

A Tale of Two Anomalies: Higher Returns of Low‐Risk Stocks and Return Seasonality

Prior studies have shown that low beta and low volatility stocks earn higher average returns than high beta and high volatility stocks, contradicting the prediction of the capital asset pricing model and the fundamental relationship between risk and return. In this paper, we demonstrate that this phenomenon is driven by the seasonality of stock returns. We show that the risk‐return tradeoff does hold in the nonsummer months, and that switching to a portfolio of low‐risk stocks in summer outperforms—both in terms of absolute and in risk‐adjusted returns—buy and hold strategies as well as the Sell in May strategy of switching to treasury bills in summer.