In a classic study, Beaver (1968) examined the trading volume of firms’ securities around the time of their earnings announcements. Specifically, he examined 506 annual earnings announcements of 143 non- December 31 year end NYSE firms over the years 1961– 1965 inclusive (261 weeks).
For each earnings announcement, Beaver calculated the average daily trading volume (of the shares of the firm making that announcement) for each week of a 17- week window surrounding week 0 (the week in which the earnings announcement was made). For each firm in the sample, he also calculated the average daily trading volume outside its 17- week window. This latter calculation was taken as the normal trading volume for that firm’s shares.
For each week in the 17- week window, Beaver averaged the trading volumes over the 506 earnings announcements in the sample. The results are shown in Figure 5.4 below. The dotted line in the figure shows the average normal trading volume outside the 17- week window.
As can be seen from the figure, there was a dramatic increase in trading volume, relative to normal, in week 0. Also, volume is below normal during most of the weeks leading up to week 0.
Subsequent research investigates factors affecting the week 0 increase in trading volume, based on the decision theory model of Section 3.3. A key driver of this volume is the extent to which prior beliefs about future firm performance differ across investors. If investors are primarily small, such as Bill Cautious in Example 3.1, their prior probabilities of a firm’s future performance will tend to be similar, since small investors are exposed to basically the same public information. Consequently, for a given information system their posterior probabilities will also be similar. Lacking investors with different opinions, there is little incentive for investors to trade among themselves, and trading volume will be relatively low.
If investors are primarily institutions, with more resources than small investors and larger share holdings, they will invest more heavily in developing private information about future firm performance. Consequently, their prior probabilities about future performance will differ from those of small investors. Since the institutions are sophisticated, they will be confident in their prior beliefs, so that the earnings announcement will have a relatively low impact. That is, an institution’s prior and posterior beliefs about future firm performance will be similar. If we further assume that the various institutions are equally sophisticated, their posterior probabilities will tend to be similar across institutions. Again, there is little incentive for institutions to trade among themselves, and trading volume will also be relatively low.
It follows that trading volume will be highest when the market for a firm’s shares consists of both small and institutional investors. Then, differences in investor beliefs (i. e., small versus large investors) about future firm performance are highest, in which case there is a relatively high incentive for trading following an earnings announcement.
In sum, theory predicts that trading volume is an inverted U- shaped function of the proportion of a firm’s shares held by institutions. Empirical evidence consistent with this prediction is presented by Ali, Klasa, and Li (2008).

Required
a. Why do you think Beaver found that trading volume increased in week 0?
b. Why do you think Beaver found that trading volume was below normal in the weeks leading up to week 0?
c. Do the findings of Beaver and Ali, Klasa, and Li support the decision usefulness of earnings information? Explain.
d. When trading volume is low surrounding an earnings announcement, does this mean that the change in share price surrounding that announcement will necessarily be low? Explain. Use the degree of decision usefulness of net income in your answer.