Working papers
Asset Pricing
Changing Expected Returns can Induce Spurious Serial Correlation with Richard Roll and Avanidhar Subrahmanyam
Abstract: Changing expected returns can induce spurious autocorrelation in returns and possible bias in other financial models. We show why this happens with some simple examples and then investigate its prevalence in actual equity data. Assets may undergo changes in expected returns for a variety of reasons. Hence, autocorrelations computed from extended records can be subject to spurious bias. The bias might be difficult to measure because of noise but the re-sampling method discloses its rather ubiquitous presence in US equity data. Turning the phenomenon on its head, return serial correlation in an efficient market is evidence of changing expected returns. Estimated risk measures such as “beta” might also be subject to bias induced by non-stationary mean returns, but the direction of the bias is more ambiguous.
Machine Learning Classification Methods and Portfolio Allocation: An Examination of Market Efficiency with Yang Bai
Abstract: Richard Thaler proposes two questions to test efficient market hypothesis: (1) can we beat the market, and (2) are the prices correct? We introduce a novel methodology, machine learning classification methods, and apply them to answer these questions. First, our classification portfolios beat the market in time series and cross-sections. Our models capture the information about return state transition and have time invariant applicability. Second, the performance is attributed to the prediction accuracy both in sample (IS) and out of sample (OOS). The accuracy implies that our classification models can generate information about future return states with historical observations. Our study suggests the current prices may not fully reflect all historical information.
Asset Prices and Partisanship: Evidence from Daily Shopper Data with Jialu Shen and Ruixiang Wang
Are stock market anomalies anomalous after all with George Chalamandaris, Spyros Pagratis and Nikolaos Topaloglou
Abstract: We propose a stochastic spanning to evaluate whether anomalies are genuine under factor-model framework. Our approach is nonparametric and does not rely on any assumption of return distribution and investor risk preferences. It depends on the whole distribution of returns, rather than only on the first two moments. Of the anomalies considered, only a few expand the opportunity set of the risk-averter and have real economic content. Our approach is consistent in identifying genuine anomalies in and out of samples. This is in contrast to mean-variance (MV) spanning tests where anomalies identified in-sample, not out-of-sample.
Agnostic Tests of Stochastic Discount Factor with Richard Roll and Junbo Wang
The original version with RIchard Roll, "An Agnostic and Practically Useful Estimator of the Stochastic Discount Factor"
Abstract: We propose an estimator for the stochastic discount factor (SDF) that does not require macroeconomic proxies or preference assumptions. It depends only on observed asset returns yet is immune to the form of the multivariate return distribution. It does not depend on factor structure, thus making our SDF ideal for asset pricing test. We find that SDF is correlated with a large number of the factor/firm characteristics, which confirms Kozak, Nagel, and Santosh (2018)’s finding. Moreover, assets within the equity market are not fully integrated as SDFs constructed from different asset groups are uncorrelated.
Testing Asset Pricing Model with Non-Traded Factors: A New Method to Resolve (Measurement/Econometric) Issues in Factor-Mimicking Portfolio with Richard Roll, Junbo Wang, and Tengfei Wang (Internet Appendix)
Abstract: We relate factor-mimicking portfolios (FMPs) to the beta-pricing model and propose that each FMP should minimize the mispricing component of its underlying factor. We also examine FMP construction when the underlying factor contains noise and offer a new method to resolve this issue. For both classical and our newly proposed methods, we recommend enhanced necessary criteria. FMPs of several macroeconomic factors constructed by our method satisfy these criteria. We find that equities are priced by consumption growth, inflation, and the unemployment rate; and corporate bonds are priced by consumption growth, industrial production, and the default spread.
ASSET PRICING AND MACHINE LEARNING
A Generalized Machine Learning Framework for Linear Factor Model Test with Christopher Jones, Jinchi Lv and Junbo Wang
Abstract: We introduce a generalized statistical learning method, sparse orthogonal factor regression (SOFAR), in testing linear factor models with both large numbers of factors and testing assets. Our approach encompasses most of the existing methods in the literature and can be used in many other scenarios with large data sets. Applying SOFAR, we can select the PC factors from the whole swath of 219 candidate factors proposed by the literature simultaneously, identify test assets associated with the selected PC factors, and interpret them. We can also select the PC factors and correlated characteristics in the IPCA framework without bootstrapping. Without firm characteristics instrumenting, we find that four PC factors (market, investment, intangible, and frictions) are relevant to the covariance of asset returns and three types of factors (profitability, asset liquidity, and liquidity bets) price assets in cross-section. With characteristics as instruments, we only identify one factor, and the correlated characteristics are beta, size, momentum, and liquidity.
HUMAN CAPITAL VALUATION
On Valuing Human Capital And Relating it to Macro-Economic Conditions with Dave Berger and Richard Roll
Abstract: Human capital is the largest component of aggregate wealth, but its relation to other macroeconomic variables is murky due to the lack of market prices. Valuing human capital using historical costs or expected income is characterized by substantial measurement error. We develop a human capital index using slave prices and relate its dynamics to that of other indicators including equities, GDP, real estate and interest rates. The human capital values are extrapolated from the 19th Century to the modern era. Their evolution has substantial implications for our understanding of the human capital dynamics, with applications to growth and portfolio allocation.
ECONOMICS
Interest Rates and Real Estate Values: The Divergence Effects of Real Rates and Expected Inflation with Richard Roll
Abstract: A common belief is that higher interest rates reduce real estate values, but corroborating evidence is underwhelming. A sensible explanation, one that suggests heterogeneity across countries, is that mortgage interest deductibility for the U.S. personal income tax makes much of the real value of the mortgage principal tax deductible when inflation is high. This implies that nominal mortgage rates could be positively related to U.S. house prices. We find evidence consistent with this possibility; the inflation component of nominal interest is associated positively with U.S. real estate prices but negatively with those in Canada, a country without mortgage interest deductibility.
Media Tone Goes Viral: Global Evidence from the Currency Market with Kari Heimonen and Heikki Lehkonen
Abstract: Using several million news and social media articles related to currencies, we examine the role of media tone in predicting the exchange rate returns of 12 developed and 24 emerging markets from 1998 to 2016. The text-based currency Media tone is a strong positive predictor of currency excess returns beyond fundamentals of one to three months ahead and six months cumulatively, with the average in-sample and out-of-sample R2s of 4.45% and 9.03% in the US. The one-month predictability is observed in four other developed markets and 18 emerging market currencies, with the latter showing a stronger pattern. This predictability encompasses previous month currency returns, currency factors, macro fundamentals, and market sentiment and is stronger for currencies that are freely floating, less liquid, difficult to value, costly to arbitrage, and which have high interest rates. The effect is channeled through expectation errors and driven by the forecasting component rather than risk. The predictability of Media tone is driven by non-mainstream financial media.
Abstract: Changing expected returns can induce spurious autocorrelation in returns and possible bias in other financial models. We show why this happens with some simple examples and then investigate its prevalence in actual equity data. Assets may undergo changes in expected returns for a variety of reasons. Hence, autocorrelations computed from extended records can be subject to spurious bias. The bias might be difficult to measure because of noise but the re-sampling method discloses its rather ubiquitous presence in US equity data. Turning the phenomenon on its head, return serial correlation in an efficient market is evidence of changing expected returns. Estimated risk measures such as “beta” might also be subject to bias induced by non-stationary mean returns, but the direction of the bias is more ambiguous.
Machine Learning Classification Methods and Portfolio Allocation: An Examination of Market Efficiency with Yang Bai
- The 3rd winner of Crowell Prize 2021, Panagora Asset Management
- AFA PhD poster 2021; World Finance Conference 2021
- Mentioned here
Abstract: Richard Thaler proposes two questions to test efficient market hypothesis: (1) can we beat the market, and (2) are the prices correct? We introduce a novel methodology, machine learning classification methods, and apply them to answer these questions. First, our classification portfolios beat the market in time series and cross-sections. Our models capture the information about return state transition and have time invariant applicability. Second, the performance is attributed to the prediction accuracy both in sample (IS) and out of sample (OOS). The accuracy implies that our classification models can generate information about future return states with historical observations. Our study suggests the current prices may not fully reflect all historical information.
Asset Prices and Partisanship: Evidence from Daily Shopper Data with Jialu Shen and Ruixiang Wang
- Midwest Finance Association meeting 2021
- FutFinInfo 2021; World Finance Conference 2021
Are stock market anomalies anomalous after all with George Chalamandaris, Spyros Pagratis and Nikolaos Topaloglou
Abstract: We propose a stochastic spanning to evaluate whether anomalies are genuine under factor-model framework. Our approach is nonparametric and does not rely on any assumption of return distribution and investor risk preferences. It depends on the whole distribution of returns, rather than only on the first two moments. Of the anomalies considered, only a few expand the opportunity set of the risk-averter and have real economic content. Our approach is consistent in identifying genuine anomalies in and out of samples. This is in contrast to mean-variance (MV) spanning tests where anomalies identified in-sample, not out-of-sample.
Agnostic Tests of Stochastic Discount Factor with Richard Roll and Junbo Wang
The original version with RIchard Roll, "An Agnostic and Practically Useful Estimator of the Stochastic Discount Factor"
Abstract: We propose an estimator for the stochastic discount factor (SDF) that does not require macroeconomic proxies or preference assumptions. It depends only on observed asset returns yet is immune to the form of the multivariate return distribution. It does not depend on factor structure, thus making our SDF ideal for asset pricing test. We find that SDF is correlated with a large number of the factor/firm characteristics, which confirms Kozak, Nagel, and Santosh (2018)’s finding. Moreover, assets within the equity market are not fully integrated as SDFs constructed from different asset groups are uncorrelated.
Testing Asset Pricing Model with Non-Traded Factors: A New Method to Resolve (Measurement/Econometric) Issues in Factor-Mimicking Portfolio with Richard Roll, Junbo Wang, and Tengfei Wang (Internet Appendix)
- Finalist of Crowell Prize 2019, PanAgora Asset Management
Abstract: We relate factor-mimicking portfolios (FMPs) to the beta-pricing model and propose that each FMP should minimize the mispricing component of its underlying factor. We also examine FMP construction when the underlying factor contains noise and offer a new method to resolve this issue. For both classical and our newly proposed methods, we recommend enhanced necessary criteria. FMPs of several macroeconomic factors constructed by our method satisfy these criteria. We find that equities are priced by consumption growth, inflation, and the unemployment rate; and corporate bonds are priced by consumption growth, industrial production, and the default spread.
ASSET PRICING AND MACHINE LEARNING
A Generalized Machine Learning Framework for Linear Factor Model Test with Christopher Jones, Jinchi Lv and Junbo Wang
Abstract: We introduce a generalized statistical learning method, sparse orthogonal factor regression (SOFAR), in testing linear factor models with both large numbers of factors and testing assets. Our approach encompasses most of the existing methods in the literature and can be used in many other scenarios with large data sets. Applying SOFAR, we can select the PC factors from the whole swath of 219 candidate factors proposed by the literature simultaneously, identify test assets associated with the selected PC factors, and interpret them. We can also select the PC factors and correlated characteristics in the IPCA framework without bootstrapping. Without firm characteristics instrumenting, we find that four PC factors (market, investment, intangible, and frictions) are relevant to the covariance of asset returns and three types of factors (profitability, asset liquidity, and liquidity bets) price assets in cross-section. With characteristics as instruments, we only identify one factor, and the correlated characteristics are beta, size, momentum, and liquidity.
HUMAN CAPITAL VALUATION
On Valuing Human Capital And Relating it to Macro-Economic Conditions with Dave Berger and Richard Roll
- Semi-Finalist of Best Paper Award 2016, Financial Management Association Meeting
Abstract: Human capital is the largest component of aggregate wealth, but its relation to other macroeconomic variables is murky due to the lack of market prices. Valuing human capital using historical costs or expected income is characterized by substantial measurement error. We develop a human capital index using slave prices and relate its dynamics to that of other indicators including equities, GDP, real estate and interest rates. The human capital values are extrapolated from the 19th Century to the modern era. Their evolution has substantial implications for our understanding of the human capital dynamics, with applications to growth and portfolio allocation.
ECONOMICS
Interest Rates and Real Estate Values: The Divergence Effects of Real Rates and Expected Inflation with Richard Roll
Abstract: A common belief is that higher interest rates reduce real estate values, but corroborating evidence is underwhelming. A sensible explanation, one that suggests heterogeneity across countries, is that mortgage interest deductibility for the U.S. personal income tax makes much of the real value of the mortgage principal tax deductible when inflation is high. This implies that nominal mortgage rates could be positively related to U.S. house prices. We find evidence consistent with this possibility; the inflation component of nominal interest is associated positively with U.S. real estate prices but negatively with those in Canada, a country without mortgage interest deductibility.
Media Tone Goes Viral: Global Evidence from the Currency Market with Kari Heimonen and Heikki Lehkonen
Abstract: Using several million news and social media articles related to currencies, we examine the role of media tone in predicting the exchange rate returns of 12 developed and 24 emerging markets from 1998 to 2016. The text-based currency Media tone is a strong positive predictor of currency excess returns beyond fundamentals of one to three months ahead and six months cumulatively, with the average in-sample and out-of-sample R2s of 4.45% and 9.03% in the US. The one-month predictability is observed in four other developed markets and 18 emerging market currencies, with the latter showing a stronger pattern. This predictability encompasses previous month currency returns, currency factors, macro fundamentals, and market sentiment and is stronger for currencies that are freely floating, less liquid, difficult to value, costly to arbitrage, and which have high interest rates. The effect is channeled through expectation errors and driven by the forecasting component rather than risk. The predictability of Media tone is driven by non-mainstream financial media.
