Return to search

A macroeconomic factor test of the arbitrage pricing theory

An Iterated-Nonlinear-Seemingly-Unrelated-Regression Model with derived-macroeconomic factors was used to test the Arbitrage Pricing Theory. These derived macroeconomic factors (exogenous variables) were estimated by applying Principal-Components Analysis to a set of ten macroeconomic time-series residuals and five sets of residual portfolio returns (constructed to be orthogonal to the macroeconomic residuals). This specification reduces errors-in-variables that result when macroeconomic factors are estimated through the factor structure of security returns or defined by macroeconomic variables. / The Arbitrage Pricing Theory was tested over the three periods 1972-1987, 1972-1979, and 1980-1987. The results of these tests were consistent with the Arbitrage Pricing Theory over the subperiods 1972-1979 and 1980-1987, and less conclusive over the period 1972-1987, suggesting that the macroeconomic factor structure may be unstable over long time periods. In all three time periods, the sensitivity of a security to the macroeconomic factors was a significant determinant of security returns; in addition, security returns were not significantly affected by nonlinear factor risk. The implication of the Arbitrage Pricing Theory, that security returns are a linear combination of a security's sensitivity to a small set of macroeconomic factors, was strongly supported by the results of this study. / Source: Dissertation Abstracts International, Volume: 51-07, Section: A, page: 2442. / Major Professor: Elton Scott. / Thesis (Ph.D.)--The Florida State University, 1990.

Identiferoai:union.ndltd.org:fsu.edu/oai:fsu.digital.flvc.org:fsu_78273
ContributorsDukas, Stephen Peter., Florida State University
Source SetsFlorida State University
LanguageEnglish
Detected LanguageEnglish
TypeText
Format133 p.
RightsOn campus use only.
RelationDissertation Abstracts International

Page generated in 0.0078 seconds