Model Complexity and Out-of-Sample Performance: Evidence from S&P 500 Index Returns

Andreas Kaeck, Paulo Rodrigues, Norman J. Seeger

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

We apply a range of out-of-sample specification tests to more than forty competing stochastic volatility models to address how model complexity affects out-of-sample performance. Using daily S&P 500 index returns, model confidence set estimations provide strong evidence that the most important model feature is the non-affinity of the variance process. Despite testing alternative specifications during the turbulent market regime of the global financial crisis of 2008, we find no evidence that either finite- or infinite-activity jump models or other previously proposed model extensions improve the out-of-sample performance further. Applications to Value-at-Risk demonstrate the economic significance of our results. Furthermore, the out-of-sample results suggest that standard jump diffusion models are misspecified.

Original languageEnglish
Pages (from-to)1-29
Number of pages29
JournalJournal of Economic Dynamics and Control
Volume90
Issue numberMay
DOIs
Publication statusPublished - May 2018

Fingerprint

Model Complexity
Jump-diffusion Model
Specification Test
Confidence Set
Financial Crisis
Value at Risk
Stochastic Volatility Model
Feature Model
Specifications
Standard Model
Jump
Model
Economics
Stochastic models
Specification
Testing
Evidence
Alternatives
Range of data
Demonstrate

Keywords

  • Forecasting
  • Jump-diffusion models
  • Lévy-jump models
  • Non-affine variance models
  • Out-of-sample specification tests

Cite this

@article{0ac234bd9fa148d9aac2199a0337f1b9,
title = "Model Complexity and Out-of-Sample Performance: Evidence from S&P 500 Index Returns",
abstract = "We apply a range of out-of-sample specification tests to more than forty competing stochastic volatility models to address how model complexity affects out-of-sample performance. Using daily S&P 500 index returns, model confidence set estimations provide strong evidence that the most important model feature is the non-affinity of the variance process. Despite testing alternative specifications during the turbulent market regime of the global financial crisis of 2008, we find no evidence that either finite- or infinite-activity jump models or other previously proposed model extensions improve the out-of-sample performance further. Applications to Value-at-Risk demonstrate the economic significance of our results. Furthermore, the out-of-sample results suggest that standard jump diffusion models are misspecified.",
keywords = "Forecasting, Jump-diffusion models, L{\'e}vy-jump models, Non-affine variance models, Out-of-sample specification tests",
author = "Andreas Kaeck and Paulo Rodrigues and Seeger, {Norman J.}",
year = "2018",
month = "5",
doi = "10.1016/j.jedc.2018.01.040",
language = "English",
volume = "90",
pages = "1--29",
journal = "Journal of Economic Dynamics and Control",
issn = "0165-1889",
publisher = "Elsevier",
number = "May",

}

Model Complexity and Out-of-Sample Performance: Evidence from S&P 500 Index Returns. / Kaeck, Andreas; Rodrigues, Paulo; Seeger, Norman J.

In: Journal of Economic Dynamics and Control, Vol. 90, No. May, 05.2018, p. 1-29.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Model Complexity and Out-of-Sample Performance: Evidence from S&P 500 Index Returns

AU - Kaeck, Andreas

AU - Rodrigues, Paulo

AU - Seeger, Norman J.

PY - 2018/5

Y1 - 2018/5

N2 - We apply a range of out-of-sample specification tests to more than forty competing stochastic volatility models to address how model complexity affects out-of-sample performance. Using daily S&P 500 index returns, model confidence set estimations provide strong evidence that the most important model feature is the non-affinity of the variance process. Despite testing alternative specifications during the turbulent market regime of the global financial crisis of 2008, we find no evidence that either finite- or infinite-activity jump models or other previously proposed model extensions improve the out-of-sample performance further. Applications to Value-at-Risk demonstrate the economic significance of our results. Furthermore, the out-of-sample results suggest that standard jump diffusion models are misspecified.

AB - We apply a range of out-of-sample specification tests to more than forty competing stochastic volatility models to address how model complexity affects out-of-sample performance. Using daily S&P 500 index returns, model confidence set estimations provide strong evidence that the most important model feature is the non-affinity of the variance process. Despite testing alternative specifications during the turbulent market regime of the global financial crisis of 2008, we find no evidence that either finite- or infinite-activity jump models or other previously proposed model extensions improve the out-of-sample performance further. Applications to Value-at-Risk demonstrate the economic significance of our results. Furthermore, the out-of-sample results suggest that standard jump diffusion models are misspecified.

KW - Forecasting

KW - Jump-diffusion models

KW - Lévy-jump models

KW - Non-affine variance models

KW - Out-of-sample specification tests

UR - http://www.scopus.com/inward/record.url?scp=85044728606&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85044728606&partnerID=8YFLogxK

U2 - 10.1016/j.jedc.2018.01.040

DO - 10.1016/j.jedc.2018.01.040

M3 - Article

VL - 90

SP - 1

EP - 29

JO - Journal of Economic Dynamics and Control

JF - Journal of Economic Dynamics and Control

SN - 0165-1889

IS - May

ER -