For more than 60 years, Bob Dylan’s “Don’t Think Twice, It’s All Right” has endured as a brilliant song and a terrible principle for valuation—especially when it comes to multiples. Managers and finance practitioners should always think twice about multiples. When multiples are used properly and the correct peer groups are selected, they can provide a quick estimation, serve as a reality check against a traditional discounted cash flow model, and slot into common shorthand. But investment or financial decisions should never be made based primarily, let alone solely, on multiples.1
In some cases, multiples can be particularly misleading. In this article, we’ll look at five of the most common situations where multiples can provide an incomplete or distorted picture.
1. Shocks to an industry or the broader market
It’s easiest to compare multiples when conditions hold steady over time: each 12- or 24-month earnings forecast is an iterative prediction of a company’s earnings over a very long horizon. But during short periods of turbulent change, the current or 12-month forward multiples can lead to nonsense results; the short term (informed by current, highly atypical performance) tells investors or managers little about what cash flow will be after broad turbulence has passed. Likewise, any future comparison between how a company performed during a crisis and how its consensus earnings appear after a crisis can be wildly inapt. We’d expect, for example, that some retailers would sell more at times when people were “just stocking up before the hoarders get here,”2 or that physical fitness chains would have lower earnings during pandemic lockdowns. But unless these occasions are regular and not one-offs, they tell us little to nothing about a company’s prospects to create value over the longer term.
Market reactions during the initial phase of the COVID-19 pandemic provide a case in point. Consider the multiples of one biotech company. In the months leading up to the global pandemic, the company’s forward EBITDA expectations were negative. When the worldwide health crisis became evident, however, the company shifted toward developing COVID-19 vaccines. Investors applauded, sending the company’s two-year3 enterprise value (EV) multiple to a dizzying 33-times EBITDA. Over the following months, however, as the analysts caught up and adjusted their forecasts, the company’s EV/EBITDA multiple fell and then flattened out to between four and seven times consensus earnings.
The COVID-19 pandemic’s economic effects, of course, were not limited to single companies; the crisis affected entire industries. For example, airlines’ earnings declined significantly (Exhibit 1). As a result, earnings multiples for all major airlines expanded. Why? Because EV incorporates cash flow during and well after a crisis has passed; when the pandemic first hit, the airlines’ earnings denominator shrank much more than its EV numerator. When travel resumed a little over a year after the pandemic, multiples quickly corrected to historical levels. Now, if one were to look back and consider an airline’s historical multiples during the first year or so of the pandemic, that period would stand out as an aberration. And conducting a multiples analysis during those fraught times invited massive error; minor differences in small earnings estimates led to a wide distribution of one-off multiples.
2. Periods of heavy investment
Multiples also get distorted when companies incur very large capital expenditures within a very short span. That may seem counterintuitive, since capital expenditures are not explicitly included in a typical multiple. However, consider the following example: Company A, a refinery, has just completed a major technological upgrade. Its competitor, Company B, will begin investing tomorrow and continue investing over the next two years for the same technology. Since the cost of the upgrade is known, all else equal, Company B’s observed net enterprise value (enterprise value before excess cash) will be lower by the present value of the upgrade cost. Consequently, the observed EV/EBITDA multiple will shrink, even though the two refineries will have the same capabilities to generate cash flows in the long term. The lower multiple doesn’t mean lower growth. But it would be impossible to know that unless one were to take a closer look.
The differences between current state and future state are particularly acute for green start-ups or high-tech companies. They can have very high growth expectations and mostly negative near-term earnings. Because of the challenge and expense of scaling up, multiples won’t be a useful way to compare early-stage companies with peers that are only a few years older. The best approach for assessing high-growth companies is to go back to discounted cash flow basics—“back from the potential future”—using probability-weighted scenarios to arrive at value today.4
3. Cyclical companies
A typical commodity company, precisely because it is subject to cycles, will not have a stable EBITDA. Yet its enterprise value changes much less over time. That’s because sophisticated investors price in the cycles. As a result, we typically see marked shifts in forward multiples at different stages during the commodity cycle. For example, forward multiples at the top of a cycle tend to be high because of an expected downturn, even though little to no growth is expected. Consequently, a point-in-time multiple for cyclical companies can be very deceptive (Exhibit 2). To gain greater insight, we recommend using a through-cycle multiple or a two- to three-year forward multiple. Practitioners can also use multiples such as EV over installed capacity, EV per barrel (in the case of an oil and gas company), or EV over mine reserves (in the case of a mining company).
When a company announces an acquisition, its market value should reflect its own value plus any synergies, net of any premium being paid to the target—after pricing in the possibility that the deal will not close. However, even when a deal closes and uncertainty is effectively reduced to zero,5 forward-multiples will likely still not be adequately reflected in analysts’ forecasts. Many analysts do not even adjust numbers reported to typical data aggregators until after closing, which leads to a distorted “mechanical” multiple. This is especially the case for deals with a long time lag between announcement and closing.
Consider one large acquisition in the consumer-packaged-goods industry. At closing, its EV/EBITDA multiple sharply increased to approximately 20 times EBITDA. In the year before closing, the EV multiple had been 15 times EBITDA, and two years before closing, it had been 13 times. Soon after the deal closed, though, the 20-times multiple began to regress to prior levels. Analysts were now including both the initial earnings impact of the acquisition and a more informed perspective about expected synergies.
For deals in which the target’s earnings and historical financials are publicly available, we recommend always double-checking the multiple. This is relatively easy to do: combine the acquirer’s expected earnings with those of the target and make a rough forecast to start. You can then refine your analysis from there, based on changes to EBITDA that one could reasonably expect.
5. Changes in strategy or business model
Companies that announce a new growth or portfolio strategy that shifts their business mix typically see some change in market expectations. However, given the disconnect between current profitability and short-term expectations compared with where the company is headed over a longer time, enterprise value multiples may not fully represent the company’s long-term trajectory (Exhibit 3).
For example, from 2003 to 2005, one industrial company traded at an EV multiple of ten to 12 times two-year forward EBITDA. In 2006, the company began to expand into medtech and life sciences. Yet it took a full decade for the company to fully transition and its underlying multiples (16 to 20 times earnings) to reflect the profitability of its new business model. Since the market hadn’t fully grasped how the company was changing, it also failed to appreciate which companies were its peers. Indeed, any peer comparison based on its 2006 model would have been a poor predictor indeed. That’s why it is always best to use different multiples for each business segment—particularly when a corporation is changing its businesses.
A current manifestation of this challenge is in the energy sector. One traditional energy company, for example, has been aggressively transitioning its portfolio to include renewables. For the first several years of its transition, the company’s multiple was consistently in line with those of traditional energy companies, even though it determinedly pushed into alternative energy. Only after a decade of making the transition—when its renewable portfolio exceeded more than 20 percent of its business—did the market adjust the company’s multiples.
Market misses are particularly likely when a company moves more sharply away from its long-time core businesses. Because a new business model has different fundamentals, a point-in-time model (for example, EV to one-year forward earnings) won’t give a full picture of the corporation’s underlying economics for a longer term. The more a company changes its business mix, the weirder its multiples may appear—at least before disaggregating into segments. Even in the age of large databases and AI, practitioners need to look at company details, economic circumstances, and facts on the ground to correctly apply multiples. Common sense and a little legwork go a long way.
Used properly, multiples can be an effective supplemental tool. But traditional EV multiples can provide an incomplete picture or inapposite results during industry- or market-wide shocks, periods of heavy investment, single points in time over a longer commodity cycle, after a merger or acquisition has just closed, and when a company is making profound changes to its business portfolio. In those cases, it’s always best to think twice about multiples—and sometimes, more than that.