BEGIN:VCALENDAR
VERSION:2.0
PRODID:talks.ox.ac.uk
BEGIN:VEVENT
SUMMARY:Comparing Predictive Accuracy in the Presence of a Loss Function S
hape Parameter - Sander Barendse (University of Oxford)
DTSTART;VALUE=DATE-TIME:20190502T130000
DTEND;VALUE=DATE-TIME:20190502T140000
UID:https://talks.ox.ac.uk/talks/id/1593d11f-832d-4fc6-9877-0c1335c2c5a5/
DESCRIPTION:We develop joint out-of-sample tests for multiple testing prob
lems that arise when comparing predictive accuracy using loss or utility f
unctions that contain shape parameters. Our tests cover forecast compariso
n scenarios in which the shape parameter (vector) takes values in some sub
set of Euclidean space. We apply our tests to three such forecast evaluati
on problems. First\, we consider hypotheses of equal (superior) expected u
tility between two portfolio strategies\, defined over an interval of risk
aversion parameter values. Second\, we consider hypotheses of equal (supe
rior) predictive ability between two conditional quantile forecast models
using Murphy diagrams. Finally\, we consider hypotheses of equal (superior
) predictive ability of univariate quantile forecasts of portfolio returns
–as generated by multivariate models of the portfolio assets–by examin
ing all portfolios with positive weights summing to one. In empirical appl
ications we show that the new tests reject at least as often as benchmarks
tests\, such as the standard Wald test or Bonferonni multiple correction\
, and are better behaved than the benchmarks in practice\, in that p-value
s remain stable as we test at more elements of the multiple hypothesis. Mo
nte Carlo experiments verify that our tests have good size and power prope
rties in small sample.\nSpeakers:\nSander Barendse (University of Oxford)
LOCATION:Manor Road Building (Seminar Room C)\, Manor Road OX1 3UQ
URL:https://talks.ox.ac.uk/talks/id/1593d11f-832d-4fc6-9877-0c1335c2c5a5/
BEGIN:VALARM
ACTION:display
DESCRIPTION:Talk:Comparing Predictive Accuracy in the Presence of a Loss F
unction Shape Parameter - Sander Barendse (University of Oxford)
TRIGGER:-PT1H
END:VALARM
END:VEVENT
END:VCALENDAR