Valuations of private companies are performed for a number of reasons, including tax, transactions and disputes. Although there are lots of methods for valuing a private company, one of the most common is to apply a multiple to the subject company’s profit or turnover to estimate an "enterprise value" (EV) for the business.
For example: ABC Ltd EV = ABC Ltd EBITDA x EV/EBITDA multiple
Business valuation multiples are based on companies that, ideally, are comparable in terms of activities, size, geographical location and financial performance. The multiples are often derived from quoted companies, since they have valuations that are readily accessible and have been established by the market.
The companies are then compared by factors such as size, growth and profit margin, and a view is taken on what multiple should apply to the business being valued. Sometimes, return on assets, return on equity or other factors are also taken into account, depending on the sector in question. These factors are often used to justify a higher or lower multiple for a private business, but how strong is the relationship between valuation multiples and these factors in practice?
We provided Shucheng Hu, an MSc student from the University of Warwick's statistics department, with this question and a large dataset. We worked with Schucheng, Professor Simon French and the university to analyse market data for companies listed in the UK to see whether relationships can be observed in actual market data between business valuation multiples and factors such as size, growth and margin. The data was extracted as at 31 May 2019 for more than 5,500 companies with primary or secondary listings on London’s main market and Alternative Investment Market (AIM), before removing data points with invalid, missing, extreme or non-meaningful data. After cleaning, there were around 2,500 data points to which a number of scatter plots and regression models were fitted.
We found the relationship between multiples and the three areas as follows:
Comparing EV/EBITDA multiples to EBITDA growth forecasts (based on two-year compound annual growth rate), the relationship was surprisingly weak. This goes against the usual rules that apply in an M&A situation, whereby a buyer will often pay a higher multiple for a company on a steeper growth trajectory.
Size is commonly assumed to affect the level of EV/EBITDA multiple due to larger companies being less risky and able to cope with external shocks. However, it was again difficult to observe much of a relationship between EV/EBITDA multiples and size.
Profitability is often cited as a factor in considering a business valuation multiple. There was little relationship observed between EV/EBITDA multiples and EBITDA margin, but this may be logical when it is considered that a company improving only its EBITDA margin (all else being equal) would achieve a higher EBITDA and so may not expect to be rewarded with a valuation based on a higher multiple applied to this higher EBITDA (a "double whammy").
However, a relationship was possibly observed between EV/Revenue multiples and EBITDA margin. This could make sense when it is considered that (all else being equal) a higher margin business will generate more cash for every pound of turnover than a lower margin business, and when valuing a business, cash is king.
Despite transforming the data sets using natural logarithms, attempting to identify relationships through a statistical technique known as ‘clustering’, and considering differences between stock market listings and industries, no more meaningful relationships could be found than the one between EV/Revenue multiples and EBITDA margin (although the university noted that even this relationship was not strong, and further analysis would be required to establish it as being statistically significant).
In practice, EV/Revenue multiples tend to be used for valuing businesses that are currently loss-making, but expected to generate profit in future, or for certain businesses, whose value is driven primarily by revenue (eg SaaS companies), but otherwise are seldom used and not typical multiples for business valuation in M&A.
One conclusion that was drawn from the study is that considering such a large dataset may be masking underlying relationships that would be more apparent in smaller, more-focused datasets. In practice, valuation practitioners do tend to use groups of comparable businesses that are typically no larger than, say, the 10 most comparable, and this study would seem to support previous studies that have identified this pattern.
Business valuation multiples are driven by a number of factors, internal and external, specific and non-specific, which are difficult to capture in a study such as this. This perhaps supports the view that business valuation is as much an art as it is a science.
To continue the conversation, contact Mike Thornton.