BEGIN:VCALENDAR
VERSION:2.0
PRODID:talks.ox.ac.uk
BEGIN:VEVENT
SUMMARY:Estimating Very Large Demand Systems (New Insights) - Jeremy Large
(University of Oxford)
DTSTART;VALUE=DATE-TIME:20231012T140000
DTEND;VALUE=DATE-TIME:20231012T150000
UID:https://new.talks.ox.ac.uk/talks/id/2f17adc3-172f-4d57-8155-add49a998a
00/
DESCRIPTION:We present a discrete choice\, random utility model and a new
estimation technique for analyzing consumer demand for large numbers of pr
oducts. We allow the consumer to purchase multiple units of any product an
d to purchase multiple products at once (think of a consumer selecting a b
undle of goods in a supermarket). In our model each product has an associa
ted unobservable vector of attributes from which the consumer derives util
ity. Our model allows for heterogeneous utility functions across consumers
\, complex patterns of substitution and complementarity across products\,
and nonlinear price effects. The dimension of the attribute space is\, by
assumption\, much smaller than the number of products\, which effectively
reduces the size of the consumption space and simplifies estimation. Nonet
heless\, because the number of bundles available is massive\, a new estima
tion technique\, which is based on the practice of negative sampling in ma
chine learning\, is needed to sidestep an intractable likelihood function.
We prove consistency of our estimator\, validate the consistency result t
hrough simulation exercises\, and present the latest estimates from our mo
del using supermarket scanner data.\nSpeakers:\nJeremy Large (University o
f Oxford)
LOCATION:Manor Road Building (Seminar Room C & online via Zoom)\, Manor Ro
ad OX1 3UQ
TZID:Europe/London
URL:https://new.talks.ox.ac.uk/talks/id/2f17adc3-172f-4d57-8155-add49a998a
00/
BEGIN:VALARM
ACTION:display
DESCRIPTION:Talk:Estimating Very Large Demand Systems (New Insights) - Jer
emy Large (University of Oxford)
TRIGGER:-PT1H
END:VALARM
END:VEVENT
BEGIN:VEVENT
SUMMARY:Estimating Very Large Demand Systems - Jeremy Large (University of
Oxford)\, Emmet Hall-Hoffarth (University of Oxford)
DTSTART;VALUE=DATE-TIME:20220601T143000
DTEND;VALUE=DATE-TIME:20220601T153000
UID:https://new.talks.ox.ac.uk/talks/id/1ebf875e-b6a8-48ca-afe8-ec743f750d
64/
DESCRIPTION:We present a discrete choice\, random utility model and a new
estimation technique for analyzing consumer demand for large numbers of pr
oducts. In our model each product has an associated unobservable vector of
attributes from which the consumer derives utility. We allow the consumer
to purchase multiple products at once in a consumption bundle.\n\nBecause
the number of bundles available is massive\, a new estimation technique\,
which is based on the practice of negative sampling in NLP\, is needed to
sidestep an intractable likelihood function. We prove consistency of our
estimator\, validate the consistency result through simulation exercises\,
and estimate our model using supermarket scanner data.\nSpeakers:\nJeremy
Large (University of Oxford)\, Emmet Hall-Hoffarth (University of Oxford)
LOCATION:Manor Road Building (Seminar Room G and online via Zoom)\, Manor
Road OX1 3UQ
TZID:Europe/London
URL:https://new.talks.ox.ac.uk/talks/id/1ebf875e-b6a8-48ca-afe8-ec743f750d
64/
BEGIN:VALARM
ACTION:display
DESCRIPTION:Talk:Estimating Very Large Demand Systems - Jeremy Large (Univ
ersity of Oxford)\, Emmet Hall-Hoffarth (University of Oxford)
TRIGGER:-PT1H
END:VALARM
END:VEVENT
BEGIN:VEVENT
SUMMARY:Jeremy Large: A Random Utility Model for Choices over Largish Bask
ets of Goods: Complements\, Substitutes\, and Price Effects (joint with Jo
shua Lanier and John Quah) - Jeremy Large (University of Oxford)
DTSTART;VALUE=DATE-TIME:20211130T124500Z
DTEND;VALUE=DATE-TIME:20211130T140000Z
UID:https://new.talks.ox.ac.uk/talks/id/bdfc7a9d-bbe7-436a-81f8-2392436214
8d/
DESCRIPTION:Whether online\, in a supermarket\, or elsewhere\, people now
assemble consumption bundles from an extremely wide variety of goods. We m
odel this as a discrete choice between bundles\, to maximize a random util
ity depending on the attributes of the goods in the bundle. We do not obse
rve attributes directly\, but discern them in revealed preferences. We sug
gest attributes may be orders-of-magnitude fewer than goods - much reducin
g the effective consumption space. Under quasi-linear preferences\, our mo
del conforms to the intuition that price complements tend to appear togeth
er in consumption bundles\, while\, on the contrary\, substitutes tend not
to. We estimate consistently\, at scale\, by using techniques similar to
negative-sampling for embedding in machine learning. This involves estimat
ing from a big dataset every purchased good's latent attributes\, jointly
with every consumer's preferences over attributes.\nSpeakers:\nJeremy Larg
e (University of Oxford)
LOCATION:Online\, Join Zoom Meeting: https://zoom.us/j/99759486691?pwd=MnV
odzd0ZlFiWlRFMHRNT3FQa0dmUT09
TZID:Europe/London
URL:https://new.talks.ox.ac.uk/talks/id/bdfc7a9d-bbe7-436a-81f8-2392436214
8d/
BEGIN:VALARM
ACTION:display
DESCRIPTION:Talk:Jeremy Large: A Random Utility Model for Choices over Lar
gish Baskets of Goods: Complements\, Substitutes\, and Price Effects (join
t with Joshua Lanier and John Quah) - Jeremy Large (University of Oxford)
TRIGGER:-PT1H
END:VALARM
END:VEVENT
END:VCALENDAR