Review of Quantitative Finance and Accounting, 13 (1999): 63±82
# 1999 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands.
Resolution of Uncertainty and Asset Prices:
Why the Timing of Information Release
Might be Relevant After All
University of Haifa
University of Utah, Eccles School of Business, Salt Lake City, Utah 84112
Abstract. The objective of this paper is to reexamine the effects of the timing of information releases on
security prices. We extend Ross (1989) by allowing the timing of information releases to affect the martingale
probabilities. We show that if the early release of information is expected to resolve part of the uncertainty about
the economy wide shock, it will positively affect asset prices in general and, under some conditions, the price of
the information generating ®rm will rise more than the price of other ®rms. Our results are consistent with
puzzling empirical observations documented in both the accounting and ®nancial economics literatures.
1. Introduction and summary
In a recent paper Ross (1989) describes the price behavior of one New York City bond
issue during the ®scal crisis in the late 1970s. In April, it was announced that the results of
a city revenue collection audit scheduled to begin in July would be reported in August,
rather than in November, as previously believed. The price of the bond rallied as a
consequence of this announcement, even though the audit itself had not yet started. That is,
bond prices rose in response to the news that information would be released earlier than
expected, even though the earlier release merely re¯ected implementation of a new
administrative procedure. This puzzling behavior motivated Ross (1989) to study the
question of the timing of announcements.
He argues that the timing, per se,of
information release is irrelevant as long as future cash ¯ows are unchanged. He assumes
that in a no-arbitrage economy the information released does not affect the martingale
probabilities. This assumption is critical for the timing irrelevancy result but does not
explain the puzzle of the bond price rally.
Here we reexamine Ross's assumption through an investigation of the conditions under
which the release of information affects individuals' consumption choices and thereby the
martingale probabilities. Under these conditions early announcements are relevant for
asset pricing and will be associated with price reactions. With the bond issue Ross cites, for