On a recent television talk show, a self-appointed pop pundit pronounced this verdict on bank mergers: good for shareholders, bad for customers.

That kind of statement has a hollow ring. In the increasingly competitive financial services sector, shareholders cannot long prosper pursuing policies inconsistent with the interests of customers.

But of course the pundit's pronouncement is not at all accurate. Bank mergers have not been good for the shareholders of acquiring banks either! While bank stocks have certainly gone up, shareholders of acquiring banks have been impoverished relative to those of peer institutions.

Most merger deals have had the potential to perform well for all shareholders (and customers) but have failed in the implementation. One reason may be related to the widespread use by acquiring banks of pooling- of-interest accounting, a technique that permits the acquirer to avoid deducting from earnings the premium paid over the assessed value of acquired assets and liabilities.

Such a practice may, in our view, be taking the sting out of post-deal performance incentive. The suggestion is ironic, in that a method designed to make earnings look good in the short run may work to undermine them in the long run, in the process condemning the associated merger to failure.

Mitchell Madison Group measures how well large mergers (more than $1 billion) perform for shareholders by comparing the total returns (dividends plus stock-price appreciation) received by the acquiring bank's owners with those received by the shareholders of a group of large regional and superregional peer institutions for a three-year period (three months before the announcement).

If shareholder returns exceed those recorded by peers, the merger is a success; if they fall short, the merger is not. This is the true shareholder test of any merger: whether, given time to implement and demonstrate results, the shareholders would have been better or worse off than if they had simply purchased an industry mutual fund.

Based on this criterion, 14 of the 17 large U.S. bank deals with a mid- 1998 three-year post-deal track record were failures. In the aggregate, the group of bank acquirers, achieved a return 13% lower than that of peer banks. By way of comparison, midyear 1998 performance data showed that all large U.S. acquirers-nonbank as well as bank-underperformed their peers by only 3%, though acquirers in manufacturing lagged by 6%. Thus, the total- return shortfall in banking was four times that in the overall economy and twice that in manufacturing-the next-worst-performing sector.

What accounts for this less-than-stellar showing? As noted above, U.S. banks have a penchant for pooling-of-interest accounting. Indeed, about 90% of the bank acquisitions have used this methodology, as opposed to 26% of the acquisitions in all business sectors. We would not suggest that accounting treatment causes failures, but it may be a contributing factor.

Mergers fail either because the acquisition price is too high or the ensuing earnings gain is too low. In a sense, of course, this distinction is artificial. When a merger fails, it is always because the acquirer cannot boost earnings sufficiently to recoup the inevitable acquisition premium. Nevertheless, it is useful to distinguish between a failure caused by the payment of a price so high that it immediately creates an unrealistic improvement bogie and a failure that results from the inability to achieve gains that might be eminently attainable. In the first instance, the buyer overpays; in the second, he underimplements.

Since the data show that acquirers employing pooling-of-interest accounting tend to pay about the same acquisition premiums as those using a purchase-accounting approach, there is no systematic evidence of overpayment. The proximate cause of merger failure therefore would seem to be flawed implementation.

The ultimate causes are varied. Acquirers implement poorly because of:

Lack of follow-through on the deal logic.

Failure to act quickly and decisively.

Inability to align organization and culture.

The acquirer's choice of accounting methodologies appears to fall under the second point. Our analysis of 115 recent economywide deals reveals that three-quarters of those that employed the pooling-of-interest approach failed to do as well as those of peer institutions. By comparison, the proportion of failed mergers among those using purchase accounting came to less than half. Thus, to the extent that pooling helps to boost reportable earnings, it may be facilitating complacency.

The notion that pooling accounting slows down the implementation process is readily comprehensible. Pooling has been called the Diet Coke of business-combination accounting. It tastes good and reminds you of the original without the unpleasant calories. That is, it combines all the benefits of a merger without such weighty unpleasantnesses as goodwill and its attendant amortization.

But merger implementers must never divert their gaze from the premium, which reflects the indisputable reality of having "overpaid" for assets and liabilities in the expectation of greatly improving their future productivity. If managers are not ceaselessly confronted with the consequences of this premium-the need to penalize reported earnings-their sense of urgency in making the improvements in asset productivity that will offset the penalties is, quite naturally, lessened.

Though, absent the need to amortize goodwill, managerial resolve and celerity may be blunted, that of the marketplace remains unaffected. The marketplace looks at cash flows, not book values. It sees that the acquirer has spent a considerable amount of currency in the belief that the future cash flows of the combined entity, suitably discounted, will be much larger than previously anticipated-large enough, in fact, to exceed the price of the deal.

If it does not get ongoing evidence that implementation is successful and the cash flow is increasing on schedule, the marketplace will act with implacable logic to bid down the acquirer's stock price relative to its peers. The resulting decline-absolute or relative-in shareholder wealth is in part the price paid for excessive managerial preoccupation with accounting fictions rather than the harsh exigencies of cash generation.

Our point is not to condemn pooling accounting. It has its uses. We merely emphasize that it can increase the risk of merger implementation. Its history is suggestive, not conclusive.

Yet conjure with one additional comment. Our midyear-1998 data reveal that, unlike their U.S. counterparts, Canadian acquirers in the banking sector have outperformed peer institutions by 7%. Most Canadian bank acquisitions to date have not used pooling accounting (although the two deals recently announced-CIBC-TD and Royal Bank of Canada-Bank of Montreal- will, if they are approved).

Once again, that is not a QED-there are other factors. It merely reinforces the caution to acquirers, especially those that embark on mergers whose success depends not so much on cost savings as on revenue enhancements. The former are relatively easy to obtain; the latter are not. They require the kind of hard work that one would like to defer and that a misguided and cosmetic view of current earnings may encourage one to defer.

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.