EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
982. Forecasting M&A shareholder wealth effects to prevent value-destroying deals: Can it be done?
Invited abstract in session WD-59: OR in Accounting: Wealth and Risk, stream OR in Financial and Management Accounting.
Wednesday, 14:30-16:00Room: S08 (building: 101)
Authors (first author is the speaker)
1. | Joao Quariguasi Frota Neto
|
Alliance Manchester Business School, University of Manchester | |
2. | Konstantinos Bozos
|
Accounting & Finance, University of Leeds | |
3. | Marie Dutordoir
|
Alliance Manchester Business School, University of Manchester | |
4. | Konstantinos Nikolopoulos
|
Marketing and Management, Durham University |
Abstract
M&A announcements can result in substantial positive or negative ab-
normal acquiring-firm stock returns and sizeable associated dollar value
gains or losses. Despite the extensive existing M&A literature, whether such
loses and gains are predictable remains unknown. Similarly explored are the
choices of model for such forecasting. This paper fills this gap. We employ
acquirer, target, deal and macroeconomic features commonly used in the
literature and test the accuracy of parametric and non-parametric models.
As expected, given the high noise-to-ratio inherent of financial forecasting,
predictability is low. However, non-parametric models are able to consis-
tently forecast abnormal returns associated with M&As, sometimes sur-
passing their parametric counterparts. Our analyses of feature importance
shows that a handful consistently outweighs the relevance of the remaining
subset. We further construct two portfolios of M&As, one containing ex-
pected value-generating M&As and another expected value-destroying MA,
and show that they are significantly different.
Keywords
- Accounting
- Forecasting
- Machine Learning
Status: accepted
Back to the list of papers