Hot and cold merger markets

Abstract

We study mergers and acquisition during the period from 1988 to 2005 and examine the impact of merger market intensity, i.e., merger waves, on the means of payment and the returns to target and acquirer shareholders. We use two proxies to measure the intensity of the merger market—the number of mergers in the trailing 12-month period prior to a merger and the total dollar volume of mergers in the trailing 12-month period prior to a merger—and use these measures to define hot and cold merger markets. We find that stock financing is more common after a stock price run-up for the acquiring firm and in hot merger markets. We also find that the acquisition premium is larger in hot merger markets. Returns to acquiring company shareholders are lower for stock financed mergers and are lower when merger markets are intense. Our results are consistent with the predictions of the behavioral theory for merger waves.

Appendix

We use a two-step Heckman (1979) selectivity bias adjustment as described below. Our approach parallels the approach used by Puri (1996).

The model for the returns to the target and acquirer returns (separate regressions) is:

$$ Y = X^{\prime } \beta + \gamma D + u $$

(1)

where Y is the return (the target or the acquirer), X is the vector of independent variables and D is a dummy variable that is 1 if the acquisition is stock financed and 0 otherwise.

The acquirer uses stock to finance the acquisition based on the following model:

$$ h = z^{\prime } b + \eta $$

(2)

where z is the vector of observable factors in determining the acquirer’s decision to use stock financing and b is the vector of coefficients. η is the vector of unobservable factors. Assuming that η~Ν(0, σ²),

$$ E\left[ {u|{\text{Stock}}} \right] = \sigma \pi \,{\text{Lambda}}_{\text{Stock}} [z^{\prime } b|\sigma ] $$

(3)

$$ E\left[ {u|{\text{Non-Stock}}} \right] = \sigma \pi \,{\text{Lambda}}_{\text{Non-Stock}} [z^{\prime } b|\sigma ] $$

(4)

where Lambda_Stock = n(.)/N(.), and Lambda_Non-Stock = −n(.)/l − N(.), and n(.) and N(.) are the standard normal density and the standard cumulative distribution function, respectively.

We can re-write Eq. (1) as follows,

$$ Y = X^{\prime } \beta + w\,{\text{Lambda}}_{\text{Stock}} D + w\,{\text{Lambda}}_{\text{Non-Stock}} (1 - D) + v $$

(5)

As Puri (1996) shows, consistent estimates can be obtained by a two-step process. First, we estimate the PROBIT likelihood function for stock financed mergers. This allows us to estimate Lambda for stock and non-stock mergers. We can then estimate w by OLS regressions. We correct for the OLS standard errors by using a robust regression model.

The inference on w is as follows. If w < 0, returns are lower in a stock merger. If w > 0, returns are higher in a stock merger. If w = 0, returns are independent of the means of payment.