These outliers will result in averages that are not representative of the sample. In most
cases, services that compute and report average values for multiples either throw out these
outliers when computing the averages or constrain the multiples to be less than or equal to
a fixed number. For instance, any firm that has a price earnings ratio greater than 500 may
be given a price earnings ratio of 500.
When using averages obtained from a service, it is important that you know how
the service dealt with outliers in computing the averages. In fact, the sensitivity of the
estimated average to outliers is another reason for looking at the median values for
multiples.
Biases in Estimating Multiples
With every multiple, there are firms for which the multiple cannot be computed.
Consider again the price-earnings ratio. When the earnings per share are negative, the
price earnings ratio for a firm is not meaningful and is usually not reported. When looking
at the average price earnings ratio across a group of firms, the firms with negative earnings
will all drop out of the sample because the price earnings ratio cannot be computed. Why
should this matter when the sample is large? The fact that the firms that are taken out of
the sample are the firms losing money creates a bias in the selection process. In fact, the
average PE ratio for the group will be biased upwards because of the elimination of these
firms.
There are three solutions to this problem. The first is to be aware of the bias and
build it into the analysis. In practical terms, this will mean adjusting the average PE down
to reflect the elimination of the money-losing firms. The second is to aggregate the market
value of equity and net income (or loss) for all of the firms in the group, including the
