Operating performance of European bank mergers

Abstract

With very few exceptions the accepted viewpoint established by (predominantly) US research is that bank operating performance is not improved after merger. In this article we concentrate on European banks and investigate post-merger operating performance for 35 publicly listed bank mergers that were completed between 1992 and 1997. We find that industry-adjusted mean cash flow return did not significantly change after merger but stayed positive. We also find that the merger led to a significant decrease in profitability and capitalisation. Our key finding, in contrast to the US evidence, is that cost-efficiency ratios improved, although the improvement was not large enough to offset the profitability decrease. We also find that low profitability levels, conservative credit policies and good cost-efficiency status before merger are the main determinants of industry-adjusted cash flow returns and provide the source for improving these returns after merger.

Keywords:

• bank merger
• operating performance
• cash flow

Notes

Thomson Financial SDC database as reported in Pasiouras et al. (2007).

When we were collecting our sample Norway was still not a member state in the EU, we included these two countries as they have been considered very important in other papers such as Vander Vennet (1996, 1999) and Cybo-Ottone and Murgia (2000).

Cornett and Tehranian (1992) investigated the post-merger operating performance for 30 bank mergers where the purchase price was more than USD 100 million. Their refinement strategy was based on eliminating acquisitions from the sample ‘if the bidder was involved in more than one acquisition of USD 100 million or greater, and/or if the bidder undertook more than five acquisitions (of any size) over the event period’ (Cornett & Tehranian, 1992, p. 215). Vander Vennet (1996, 1999) studied a large sample of EU bank M&A without imposing any restriction on the time gap between acquisitions announced by the same acquirer.

The relative size is calculated as the target to acquirer market value of equity in year −1 relative to the announcement year of the deal.

Tier I capital is defined as Common Shareholders’ Equity plus Perpetual Preferred Stock plus minority interest in equity accounts of consolidated subsidiaries less goodwill. Total Capital includes Tier I Capital plus a bank's loan loss reserves up to maximum of 1.25% or risk-adjusted assets plus various convertible and subordinated debt instruments. Risk-adjusted assets figure is comprised of two components: risk adjusted on balance sheet assets and risk adjusted off balance sheet assets.

Non-performing loans are those loans that do not generate interest or commission income because of customers’ default payment. These loans are usually classified as very high credit risk loans and they are at only one higher grade from being classified as bad debts.

We also computed the Durbin–Watson d-statistics and found that the independent variables in our models do not suffer from serial correlation.

The use of a 2-year period before the merger announcement year and 3 years post-merger was mainly driven by data availability as for many deals data was not available on Bankscope for longer pre-merger period. Moreover, if we use a longer post-merger period we end up with a much smaller sample.

Cornett and Tehranian (1992) and Healy et al. (1992) deduct the change in the equity value from 5 days before the merger announcement to the day the target is delisted from the stock exchange. We believe that this could be a long period. For instance, generally, the merger in the financial services industry takes time to be approved by the regulatory bodies, and therefore the change in equity value during this long period might be attributable to other factors that are not related to the merger event.

In cases where the target bank and the acquiring bank are from two different countries, when we calculate the industry-adjusted performance measures we use the weighted-average industry figures for the countries of both parties weighted by the total book value of assets of each bank.

In all the deals that we study the purchase method of consolidation was used, therefore our sample does not suffer from any problem related to the use of differing accounting methods of merger.

Practitioners pretend that within the first year after merger, the expected cost saving is about 50% while full savings can be achieved after 3 years (Rhoades, 1998). Moreover, others found on average savings of 35% of the acquired bank's data processing and back-office operations expense within 6–9 months after the mergers (Keefe, Bruyette & Woods, 1990).

Cornett and Tehranian (1992) found that the ROA ratio did not improve but the ROE improved after merger, for the combined entity.

Rose (1992) used the ROA and ROE ratios as well and found that the ROE ratio improved, whereas the ROA did not.

Linder and Crane (1992) used the operating income to assets ratio, but also found that this ratio improved for mergers within the BHC.

We used two weighting procedures that gave similar results, these are the Market Value of Equity and Market Value of Assets for acquirers and targets.

Another opposite explanation is that because high pre-merger personnel expenses to average assets could be a source of inefficiency, therefore this creates a scope to improve the efficiency status after merger by less dependence on human capital.