Truth and Bias in M&A Target Fairness Valuations: Appraising the Appraisals

Matthew Shaffer is a PhD candidate at Harvard Business School and an assistant professor at the University of Southern California starting summer 2019. This post is based on his recent paper, and is part of the Delaware law series; links to other posts in the series are available here. Related research from the Program on Corporate Governance includes Using the Deal Price for Determining “Fair Value” in Appraisal Proceedings (discussed on the Forum here) and Appraisal After Dell, both by Guhan Subramanian.

Since the Trans Union case (Smith v. Van Gorkom, 1985), it has been considered effectively mandatory for target directors to seek and consider a third-party “fairness opinion” with supporting financial analyses and valuations (henceforth “fairness valuations”) before accepting a takeover offer. Despite their ubiquity in M&A, their role is, to put it mildly, controversial.

First, critics allege that the providers of these fairness opinions—which are often affiliated investment banks—have conflicts of interest, and incentives to rubber stamp or rationalize the negotiated deal price ex post, instead of providing a truly independent valuation “audit.” The providers claim they have implemented sufficient controls to protect the independence of their work.

Second, more fundamentally, critics argue that even unbiased, independent valuations would not be useful to directors when they have a market signal of the value of their company. For a public target, they argue, the pre-announcement stock price is a summary statistic for its value, and so the market premium is a sufficient statistic for a well-priced offer. What value can a subjective valuation—which must rely on uncertain forecasts and other discretionary inputs—add to this “market test”? (This basic question—on the relevance of valuation in checking a deal price when a market signal is available—is also at the center of the live debate over appraisal rights, as in the Aruba Networks case.)

But proponents can point to four factors that could, in principle, open the door for third-party valuation of public targets. First, there is evidence of fundamental mispricing in financial markets. Second, mispricing may be especially pronounced around M&A deals, where managers may have incentives to manipulate the information environment, or time the market. Third, if the target has been the subject of M&A rumors or expectations, it can be impossible to disentangle what portion of its stock price represents its standalone value vs. the expectation of an acquisition premium. Fourth and finally, the idiosyncratic synergies that the target will bring to its acquirer will not be impounded in its pre-deal stock price, and these may be relevant to directors.

In my dissertation, Truth and Bias in M&A Target Fairness Valuations: Appraising the Appraisals, I examine this controversy empirically, using a comprehensive sample of DCF fairness valuations of public targets from 2006-2016. I provide new evidence that suggests that these valuations can provide relevant information about the value of the target that is not impounded in its pre-deal stock price, which could render them useful even in the public-company case. But I also find evidence that the providers are, as alleged, partially ex post rationalizing deal prices, rather than giving a fully independent valuation audit. I describe these as my “validation tests” and my “bias tests” below.

Validation tests

Testing for whether fairness valuations can provide a valid signal about the value of the target firm poses an immediate research-design challenge. If we wished to, for example, validate a sell-side equity analyst’s valuations we could test for whether the stocks she rates a “buy” outperform those she rates a “sell.” But when the target is acquired, we cannot observe its ex post performance. I have two major ways of getting around this fundamental problem.

First, I find a small, but extremely useful, subsample of transactions that were unexpectedly terminated (usually due to regulatory scrutiny) after the fairness valuations had been disclosed and distributed to shareholders. In this subsample, over the medium- to long-term horizon, the erstwhile targets with higher fairness valuations have higher abnormal returns than those with lower fairness valuations. This holds even after conditioning on the foregone premium (which is equivalent to conditioning on the pre-deal market price). I interpret this to mean that the valuations can provide a signal about relative ex ante fundamental mispricing in targets. There are several possible, non-exclusive explanations for this result: Fairness opinion providers have advanced access to non-public information. Another possibility is that, put simply, fundamental valuation works. If there is random mispricing in financial markets, then a constrained fundamental valuation may be able to detect it, without requiring non-public information.

Second, for the transactions in my sample that are not terminated, I use the ex post performance of the acquirer to, essentially, back out the implied value of the target to its acquirer. Once again, I find that fairness valuations can forecast this value, incrementally to the target’s pre-deal stock price. Surprisingly, fairness valuations appear to on average provide a better overall forecast of the market’s appraisal of the value of the target to its acquirer on the acquisition announcement date. Likely the major reason for this is that these valuations can sometimes, explicitly or implicitly, include expected deal-level synergies, which are excluded from the pre-deal stock price.

Bias tests

I next explore whether the providers do, as alleged, rationalize transaction prices ex post, rather than providing an independent ‘true’ valuation. This once again poses an immediate research-design challenge: There is, I argue, no external normative benchmark for the true value of the target that can be credibly measured by the researcher. So instead, I examine the discount-rate assumptions that the providers use in their DCF valuations. One way in which providers could rationalize deal terms ex post would be to toggle their discount-rate assumptions upwards (downwards) to match relatively poorly (richly) priced deals. (Consider, as a memorable example, Morgan Stanley’s revisions to its analyst coverage of Snap shortly after its IPO.) Economic and financial theory—and the conventions of industry practice—provide more thorough guidance on the determinants of discount rates, which allows me to estimate a credible normative benchmark, and, thus, measure how providers deviate from this benchmark to rationalize deal prices.

The estimate from my most robust specification implies that in response to a 5 percentage point decrease in the negotiated deal price in terms of the implied premium, providers deviate their discount-rate assumptions upwards on average by .1 percentage point. While that may seem small, DCF valuations are highly sensitive to discount-rate assumptions. In a typical valuation in my sample, increasing the WACC by .1% would decrease the aggregate equity valuation by about 1.5 percentage points in terms of the implied premium—in other words, this is about 30% as large as “full rationalization.” However, I also find that this form of bias has decreased in recent years, especially for investment banks—likely in response to the Delaware Chancery’s criticism of Goldman Sachs’ post-hoc fairness opinion valuations in its In Re: Southern Peru decision.

Conclusion

This new evidence has implications for boards, regulators, courts, and M&A and valuation professionals. It suggests that third-party appraisal has potential relevance even in the controversial case of a public-company transaction, and it also suggests avenues for reform: The more that fairness opinion providers are simply rationalizing deal prices ex post, the less valuable they can be. But I caution against over-interpreting my findings: My evidence suggests that, directionally and on average, these valuations can provide an incrementally relevant signal. But that does not imply that they can be taken at face value in any particular case, that they are free of systematic bias, or that they should be given total deference or even very large weight when market signals are available. It only implies that, given the diverse complexities and subtle conflicts of corporate-control transactions, market signals may not provide directors and shareholders with all they need to know, and financial expertise and fundamental valuation may still help bridge the gap. If directors plan to continue to give weight to such financial analysis, they should take further steps to ensure the credibility and independence of the experts.

Both comments and trackbacks are currently closed.