Mergers & Acquisitions: The Good, The Bad, And The Ugly

Mergers & Acquisitions: The Good, The Bad, And The Ugly (And How To Tell Them Apart) by S&P Global Market Intelligence

by Richard Tortoriello; Temi Oyeniyi, CFA; David Pope, CFA; Paul Fruin, CFA & Ruben Falk

Year-to-date through July, over $800 billion of merger-and-acquisition (M&A) activity has been announced in the U.S. Should acquiring-company shareholders expect to benefit? In this study we show that, among Russell 3000 firms with acquisitions greater than 5% of acquirer enterprise value, post-M&A acquirer returns have underperformed peers in general. A number of deal-related and fundamental attributes can be used to separate the ‘good’ from the ‘bad’ (and, sometimes, the really ugly).

1. Introduction

The academic literature on M&A is vast but comes to few definitive conclusions. Although nearly all studies agree that M&A creates value for target-company shareholders, studies on post-M&A results for acquirers have no such unanimity.

This disparity is due to a simple truth: target company shareholders almost always receive a takeover premium. However, post-M&A acquirer returns depend on fundamental performance, which is affected by many factors, including deal size, due diligence adequacy, corporate culture, deal structure, valuation, funding sources, and management experience. Although post-close M&A research results are mixed, we’d cite a few relevant studies:

Our findings confirm many of the results cited above and, in addition, suggest that preacquisition growth rates in assets and shares outstanding are associated with post-M&A acquirer returns. We also show that poor post-acquisition stock performance directly reflects deteriorating post-acquisition fundamentals in terms of profitability, return on capital, and earnings growth.

Figure 2 shows U.S. and Canada completed M&A activity by gross transaction value over the past 18 years. Note that 2015 represented a new peak in terms of transaction values and that with the exception of 1998 to 2000 most transactions have been cash transactions, although stock transactions have picked up in value as of late.

2. What Causes Acquirer Post-Acquisition Underperformance?

Acquisitions may be pursued for a variety of reasons: to revive stagnant revenue growth, enter a new market, gain new products/technologies/talent, reduce competition, etc. However, a key premise of any acquisition is that “the whole will be greater than the sum of the parts.” 2 While acquiring-company management almost universally tout expected “synergies” and efficiency gains, our research shows that, on average, such synergies either do not exist or are only realized over an extended time horizon (i.e., well over three years).

Within the Russell 3000, M&A transactions can be summed up quite neatly: in aggregate, acquirers tend to underperform peers for an extended period following a significantlysized acquisition (section 2.1). We see the underlying cause for this underperformance in deterioration in post-acquisition fundamentals. Section 2.2 shows that a number of key fundamental ratios weaken following significant M&A activity.

2.1 Acquirer Pre- and Post-Acquisition Returns

Figure 3 shows M&A industry- and universe-relative3 acquirer returns for the Russell 3000 that closed between 2001 and 2013, measured from one year pre-close to three years postclose. Median returns are consistently negative, signaling universe/industry underperformance, and hit rates 4 are low and downward trending. (Although we present statistics going out three years, underperformance continues for at least five years.)

Note that by the third year following an acquisition only about 40% of acquirers have outperformed their industry- or universe-relative benchmarks. Pre-close market caps are in line with industry means and post-close market caps only modestly higher, so market cap effects (the so-called “size effect”) on performance are likely minimal. All returns and hit rates are significant at the one percent level.

2.2 Acquirer Pre- and Post-acquisition Fundamentals

Despite the oft-heard claim of potential M&A synergies, acquirers lag peers on a variety of metrics for an extended period following an acquisition. The graphs in this section display the industry-relative median values for the Russell 3000 M&A universe with close dates from 2001 to 2013.

Figure 4 shows that profit margins (left chart) fall below the industry median following an acquisition. Net margins deteriorate more than operating (EBIT) margins, due to an increase in below-the-operating-line items (interest expense and “special charges”). As a result, earnings per share growth (right chart) declines. From the perspective of the average acquisition, M&A tends to be dilutive to earnings growth over an extended period.

Figure 5 shows that return on equity and return on invested capital (ROIC) both decline relative to industry peers following a significant acquisition (left chart). This is partly a result of increased interest expense and other charges and, in the case of ROIC, partly due to a large rise in debt relative to peers (right chart).

Figure 6 shows the interplay of cash flow from financing, investing, and operating activities (left chart) following an acquisition. In the first year following an acquisition, while financerelated cash flow jumps, due to new debt and share issuance, investing cash flow drops, due to the cash cost of acquisitions. Operating cash flow is relatively unaffected, but dips slightly relative to peers.

The right chart shows that funds from operations (all operating cash flow activity, except changes in working capital) declines slightly relative to peers following an acquisition, while at the same time working capital needs rise modestly. In sum, post-acquisition operating cash flows also weaken, but not as badly as earnings.

3. Factor Identification

We reviewed a variety of deal-related factors using an event-study format, including target premium paid, deal valuation, post-announcement price drift, cross-border vs. domestic M&A, relative and absolute deal value, and consideration type. The only deal-related factors that showed significant predictive power were relative deal value and % stock/cash consideration. The larger the deal size relative to the acquirer and the more stock consideration paid, the greater the subsequent underperformance.

We also reviewed a variety of non-deal related fundamental factors based upon the event studies shown in Section 2. We found that the most significant non-deal factors with regard to post-M&A returns relate to growth (see Mortal and Schill, 2015), cash availability (see Lang et al, 1991), and profitability. We note that non-deal factor performance may relate more to the general factor performance than to performance specific to M&A.

3.1 Factor Identification – Regression Analysis

We applied regression analysis to isolate factors, both deal-related and generic, that have a strong statistical relationship to M&A returns. The seven factors that pass this test, on a univariate basis, are shown in Table 1. We regress one-year forward stock returns minus equal weighted one-year returns for the Russell 3000 against raw factor values for each acquirer measured as of the close of each acquisition.

Multiple regressions, shown at bottom of Table 1, reveal four factors (green shading) that are statistically related to M&A returns: % stock consideration, trailing 12-month (TTM) asset growth, one-year change in shares outstanding, and cash and equivalents to assets. In the next section we’ll include these factors in a simple multi-factor model.

3.2 Factor Identification – Event Study Returns

Table 2 applies an event-study format to the four factors that passed our regression analysis test, plus % deal value. (We add relative deal value because, despite its weakness in regression testing, we find that in portfolio tests it both adds alpha and helps to differentiate the % stock consideration values, which tend to be binary in nature – either 0% stock or 100%.) Each factor is sorted into quintiles, and industry-group relative returns are calculated for each quintile one-year and three-years post-acquisition close.

Factor returns that are significant at the one percent level are shaded green. Note that, with the exception of one-year change in shares, all of the alpha generated by these factors (at least using an event study format – portfolio study results differ) is on the short side. Also note the extended time horizon (three years) over which these factors generate negative alpha, based only on measurement of factor values at the time of the acquisition.

4. Five-Factor Model – Portfolio Study

In this section, we combine the four factors identified in section 3.1 above – % stock consideration, TTM asset growth, one-year change in shares outstanding, and cash and equivalents to total assets – along with % deal value, into a simple five-factor model. All factors in the model are equal-weighted, with one exception: as our strongest overall and deal-related factor, we double-weight % stock consideration.

The methodology used is described in detail in section five, Methodology and Database Characteristics. Put simply here, we use a 365 day lookback window, form portfolios monthly, and calculate one-month forward returns for each portfolio on an industry-relative basis.

Table 3 shows that hit rates for the top quintile (of the non-Fama-French-adjusted portfolios) are near 71% (significant at the 1% level), while those for the bottom quintiles are near 40% (significant at the 5% level). Turnover rates are low, averaging 40% for the top quintiles and 32% for the bottom quintiles. Company-level hit rates are 51% and 48% for the top and bottom quintiles, respectively.

Top and bottom quintile returns are statistically significant, even after adjusting for Fama French factors, with strong performance after accounting for risk on the short side. Annualized excess returns are 7.41% for the top quintile and -5.22% for the bottom quintile (12.63% long minus short).

Figure 7 shows the value of $1,000 invested over the past 15 ½ years for the top and bottom quintiles of our model versus the equal-weighted Russell 3000. Note that the top quintile portfolios had a large draw-down during the 2007-2008 financial crisis, but has recovered well since that period. The bottom quintile, however, has had minimal upward momentum over the entire time frame, reinforcing our view that M&A models are best implemented as “sell short or avoid” strategies.

After accounting for risk (Fama French) factors, the strongest M&A returns are on the short-sale side of the model. Do the bottom-quintile portfolios have enough liquidity to be shorted? We use market cap as our proxy for liquidity. Table 4 shows the average and median market caps, measured one day prior to acquisition close, for each quintile of our model, over the January 2001 to May 2016 time frame. We note that the median market cap for the bottom quintile is over one billion dollars, so we believe there is some evidence that the strategy is shortable.

5. Methodology and Database Characteristics

We begin with the S&P Global Market Intelligence Transactions database in XpressfeedTM for the Russell 3000. Our U.S. M&A data is robust beginning in 2001, with over 15,000 transactions that have deal values within the January 2001 to April 2013 period. (There are also about 8,000 likely small deals for which terms were not disclosed.) The April 2013 end date allows a full three years of forward returns for all transactions, where available.

We then narrow the Russell 3000 M&A universe to transactions with a total cumulative acquirer-relative deal value over the past 365 days greater than 5%. This is calculated 1) as gross deal value as a percent of acquirer enterprise value (for non-financials that have a positive enterprise value [EV]) or 2) as total consideration to shareholders as a percent of acquirer market cap (for financials and companies without a positive EV). The 5% threshold results in a universe of over 9,000 unique transactions over the 2001 to 2013 period.

All factors and returns use an industry-relative approach, unless noted otherwise. Median factor values / returns for an acquiring company’s GICS Industry Group5 are subtracted from the acquirer’s factor values / returns to arrive at an industry-adjusted value.

We use three research formats, an event study format, a regression analysis, and a multi-factor portfolio backtest. For the event study, returns and fundamental characteristics (net profit margin, ROE, etc.) are aggregated and the median value taken, so that the end result is the universe-wide median of the industry-median-relative values. The time horizon used extends to three years following the close.

The regression analysis begins with a panel of data that has as its left column (or dependent variable range) the one-year forward return of each acquirer in our universe, beginning from the close date, minus the one-year forward return of the market, defined as the equalweighted Russell 3000. The right column (or independent variable range) contains the raw values for each investment factor (e.g., % deal value, asset growth), with all non-deal related factors measured on an industry-relative basis. In the case of multiple analyses the dependent variables are multiple columns of raw factor values.

For the portfolio backtest format, monthly portfolios of M&A acquirers are formed using a 365 day lookback window (e.g., all closed transactions with a cumulative deal value greater than 5% over the past year) and returns are calculated for the forward month.

All non-deal related factors are adjusted for the industry group median, and all factors for a given portfolio are then percentile-ranked, with the percentile values for each factor finally added together. If a stock is missing one or more factors, it is excluded from the portfolio. In order to use all available data, we calculate model performance over the January 2001 to April 2016 time frame, resulting in 184 separate portfolios.

Although we restrict ourselves to U.S. transactions in this study, the S&P Global Markets Intelligence Transactions database in Xpressfeed contains over 600,000 global M&A transactions going back to 1998 for the U.S., 2001 for Europe, Africa, Australia and New Zealand, and 2005 for Asia and LATAM. The database contains detailed data on a large variety of transaction- and consideration-related features, including tracking changes in consideration packages offered over time.


On the whole, significant merger and acquisitions activity results in long-term underperformance for acquirers. Empirical evidence shows that this underperformance is a rational response to weakening fundamentals. Post-acquisition profit margins, earnings growth, and return on capital all decline, on average, resulting from large increases in debt, interest expense, and “special charges,” without offsetting increases in cash flow or income generation. In other words, when it comes to M&A – management promises to the contrary – the whole is often less than the sum of the parts.

However, in the aggregate it is possible to differentiate between good and bad M&A transactions, based on a few deal- and fundamental-based characteristics. In our view, stock-based acquisitions underperform in part due to inflated stock being used as currency near stock market tops with relatively little discipline enforced by lenders. Large deals, or a string of moderately-sized deals within a short period of each other, likely underperform for many of the same reasons (stock deals tend to be large) and because large-scale merger integration can cause corporate indigestion.

Fundamental factors can also help separate the value-creating from the value-destroying. High pre-acquisition asset growth and growth in shares outstanding is a long-term negative for acquirers. High pre-acquisition cash balances, which may encourage ill-timed or otherwise poorly-thought-out deals, also bode negatively for post-deal stock performance.

Using these few simple factors together in a multi-factor model has historically worked well in separating the M&A acquirers to avoid (or sell) from those to keep (or go long).

See the full PDF below.