Mergers and acquisitions waves in the UK: a Markov-switching approach

Abstract

This article further investigated wave behaviours for mergers and acquisitions – M&A in the UK during the 1969Q1/2004Q1 period by means of Markov-Switching models. Previous analysis had focussed on traditional models that incorporate the potentially limiting assumption of constant transition probabilities across regimes. The consideration of more general models with time-varying transition probabilities across regimes along the lines of Diebold et al. (1994) provide a useful route for assessing to what extent M&A waves are driven by economic variables usually considered in the related literature. The empirical implementation considered lagged conditioning variables referring to real output growth, real growth in money supply and real stock market returns. The evidence indicated that one should reject the constant transition probability model in favour of the time-varying transition probability model and therefore the usual aggregate variables considered in the empirical literature on M&A indeed appear to play some role in determining the wave behaviour of M&A in the UK, though the effects are asymmetric across the different regimes.

Acknowledgements

The author acknowledges CNPq for the financial support and an anonymous referee for comments on a previous version of the article but the usual caveats apply.

Notes

¹ Useful introductions to Markov-switching models appear in Hamilton (1994, 1999) and Kim and Nelson (1999).

² The approach considered by Diebold et al. (1994) does not introduce autoregressive dynamics and therefore can be thought to some extent as an extension of Hamilton (1990) to the time varying transition probabilities case.

³ The reader is referred to Diebold et al.(1994) for a step by step exposition on the necessary recursive calculations. The Matlab code for the implementation of the EM algorithm was kindly provided by G. Weinbach.

⁴ It is interesting that despite the over-optimism effect, SY explicitly mention the role of expectations on synergetic effects.

⁵ Interest rate could in principle exert a negative effect on mergers by a similar argument that motivated the relevance of money supply growth. The evidence as indicated by the study of Benzing (1993) for the US failed to capture any such influence. As noted by a referee it would be important to distinguish between cash and stock deals in order to have a deeper understanding of the role of interest rates. Unfortunately, the available aggregate data does not allow to pursue that type of analysis.

⁶ In the constant transition probability case, one can obtain ρ  =  (1 − p ⁰⁰)/((1 − p ¹¹) + (1 − p ⁰⁰)). See for example, Hamilton (1994), chapter 22.

⁷ Interest rate in terms of long term yield for government bonds was also considered as an alternative monetary conditioning variable. In that case, however, no convergence was achieved for the EM algorithm even with different choices of starting values for the parameters.