This dissertation is an empirical investigation of security excess returns associated with the announcement of corporate spinoffs. Spinoff excess returns exist, but the sources of these returns are not clear. Varying levels of disaggregated accounting information result from spinoffs. The purpose of this dissertation is to relate these information levels to spinoff excess returns.

A sample of 79 voluntary spinoffs from 1980 to 1987 is categorized according to levels of disaggregated accounting information. Analyses of the security returns for the entire sample and for the subsamples formed by information levels are performed. Additionally, the importance of the size of the spinoff and the combined impact of spinoff size and information levels are investigated. Daily excess returns from the CRSP Excess Returns File are the source of the dependent variable measure. Although the market reaction to spinoffs has been studied previously I the reaction to spinoffs from 1980 to 1987 has not been studied in detail.

The key findings and implications of the empirical investigation include:

1. Spinoff announcements do result in significant excess returns. This is a confirmation of similar findings from earlier investigations.
2. The size of the spinoff has a significant relationship to the announcement return. This is also a confirmation of results from earlier spinoff studies.
3. Levels of disaggregated accounting information are not significantly related to spinoff announcement returns. Another possible source of spinoff excess returns has been investigated.
4. Levels of disaggregated accounting information are related to the dispersion of spinoff announcement returns. The value of accounting information in a new setting is known.
5. Interaction between spinoff size and information levels is related to spinoff announcement returns. A qualification of the effect of size on spinoff excess returns is demonstrated.
6. Levels of disaggregated accounting information are related to spinoff postannouncement returns. This finding suggests postannouncement drift and a topic for further research.