Characteristics are covariances: A unified model of risk and return

Bryan T. Kelly, Seth Pruitt, Yinan Su

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We propose a new modeling approach for the cross section of returns. Our method, Instrumented Principal Component Analysis (IPCA), allows for latent factors and time-varying loadings by introducing observable characteristics that instrument for the unobservable dynamic loadings. If the characteristics/expected return relationship is driven by compensation for exposure to latent risk factors, IPCA will identify the corresponding latent factors. If no such factors exist, IPCA infers that the characteristic effect is compensation without risk and allocates it to an “anomaly” intercept. Studying returns and characteristics at the stock-level, we find that five IPCA factors explain the cross section of average returns significantly more accurately than existing factor models and produce characteristic-associated anomaly intercepts that are small and statistically insignificant. Furthermore, among a large collection of characteristics explored in the literature, only ten are statistically significant at the 1% level in the IPCA specification and are responsible for nearly 100% of the model's accuracy.

Original languageEnglish (US)
JournalJournal of Financial Economics
DOIs
StatePublished - Jan 1 2019
Externally publishedYes

Fingerprint

Risk and return
Principal component analysis
Latent factors
Factors
Anomaly
Cross section
Time-varying
Risk factors
Expected returns
Modeling

Keywords

  • Anomaly
  • BARRA
  • Conditional betas
  • Cross section of returns
  • Factor model
  • Latent factors
  • PCA

ASJC Scopus subject areas

  • Accounting
  • Finance
  • Economics and Econometrics
  • Strategy and Management

Cite this

Characteristics are covariances : A unified model of risk and return. / Kelly, Bryan T.; Pruitt, Seth; Su, Yinan.

In: Journal of Financial Economics, 01.01.2019.

Research output: Contribution to journalArticle

@article{7c0d1135742944c3ab9747b4c7d99114,
title = "Characteristics are covariances: A unified model of risk and return",
abstract = "We propose a new modeling approach for the cross section of returns. Our method, Instrumented Principal Component Analysis (IPCA), allows for latent factors and time-varying loadings by introducing observable characteristics that instrument for the unobservable dynamic loadings. If the characteristics/expected return relationship is driven by compensation for exposure to latent risk factors, IPCA will identify the corresponding latent factors. If no such factors exist, IPCA infers that the characteristic effect is compensation without risk and allocates it to an “anomaly” intercept. Studying returns and characteristics at the stock-level, we find that five IPCA factors explain the cross section of average returns significantly more accurately than existing factor models and produce characteristic-associated anomaly intercepts that are small and statistically insignificant. Furthermore, among a large collection of characteristics explored in the literature, only ten are statistically significant at the 1{\%} level in the IPCA specification and are responsible for nearly 100{\%} of the model's accuracy.",
keywords = "Anomaly, BARRA, Conditional betas, Cross section of returns, Factor model, Latent factors, PCA",
author = "Kelly, {Bryan T.} and Seth Pruitt and Yinan Su",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.jfineco.2019.05.001",
language = "English (US)",
journal = "Journal of Financial Economics",
issn = "0304-405X",
publisher = "Elsevier",

}

TY - JOUR

T1 - Characteristics are covariances

T2 - A unified model of risk and return

AU - Kelly, Bryan T.

AU - Pruitt, Seth

AU - Su, Yinan

PY - 2019/1/1

Y1 - 2019/1/1

N2 - We propose a new modeling approach for the cross section of returns. Our method, Instrumented Principal Component Analysis (IPCA), allows for latent factors and time-varying loadings by introducing observable characteristics that instrument for the unobservable dynamic loadings. If the characteristics/expected return relationship is driven by compensation for exposure to latent risk factors, IPCA will identify the corresponding latent factors. If no such factors exist, IPCA infers that the characteristic effect is compensation without risk and allocates it to an “anomaly” intercept. Studying returns and characteristics at the stock-level, we find that five IPCA factors explain the cross section of average returns significantly more accurately than existing factor models and produce characteristic-associated anomaly intercepts that are small and statistically insignificant. Furthermore, among a large collection of characteristics explored in the literature, only ten are statistically significant at the 1% level in the IPCA specification and are responsible for nearly 100% of the model's accuracy.

AB - We propose a new modeling approach for the cross section of returns. Our method, Instrumented Principal Component Analysis (IPCA), allows for latent factors and time-varying loadings by introducing observable characteristics that instrument for the unobservable dynamic loadings. If the characteristics/expected return relationship is driven by compensation for exposure to latent risk factors, IPCA will identify the corresponding latent factors. If no such factors exist, IPCA infers that the characteristic effect is compensation without risk and allocates it to an “anomaly” intercept. Studying returns and characteristics at the stock-level, we find that five IPCA factors explain the cross section of average returns significantly more accurately than existing factor models and produce characteristic-associated anomaly intercepts that are small and statistically insignificant. Furthermore, among a large collection of characteristics explored in the literature, only ten are statistically significant at the 1% level in the IPCA specification and are responsible for nearly 100% of the model's accuracy.

KW - Anomaly

KW - BARRA

KW - Conditional betas

KW - Cross section of returns

KW - Factor model

KW - Latent factors

KW - PCA

UR - http://www.scopus.com/inward/record.url?scp=85066097292&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066097292&partnerID=8YFLogxK

U2 - 10.1016/j.jfineco.2019.05.001

DO - 10.1016/j.jfineco.2019.05.001

M3 - Article

AN - SCOPUS:85066097292

JO - Journal of Financial Economics

JF - Journal of Financial Economics

SN - 0304-405X

ER -