the larger multiple and look deceivingly dearer
than firms with less debt?
ii)
the multiples are
uniformly
measured, especially for the EPS and BVPS variables. Take EPS for example
For a sample of firms with different financial year end dates – While trailing PE multiple is uniform
with respect to the
period in which earnings are measured
, current PE isn’t and has this undesired impact: a high growth firm with an earlier
YE date could have a larger current PE and look deceivingly dearer than another high growth firm with a later YE date.
All expense items should be classified uniformly
according to the taxation rulings
into operational or CAPEX to eliminate
the difference in EPS caused by reporting flexibility. Otherwise, firms that expense items will have higher PE and look
dearer than firms that capitalize and amortize the same items.

** Subscribe** to view the full document.

FINS3641 SAV
Week 8: Fundamental Principles and Earnings Multiples
6
Usage of Relative Valuation – 2
nd
Quality Control Test – Descriptive Test
It is also critical that we know (or have a sense of)
for each sector, the cross sectional and time series distribution of multiples among the firms
the cross sectional and time series distribution of the average multiples in one sector relative to those of
the other sector and the entire market
in order to
pass judgment on whether the multiple of a given firm or a sector is too high or low
Hence we need to, for each and every sector and the entire universe of stocks,
i)
For a
normally distributed
sample of multiples, compute the mean and standard deviation.
ii)
For a sample of multiples that is
skewed
by the outliers (i.e., excessively large values), compute the median
and percentile values in the sample
iii)
Do not rely on any
biased
sample of multiples that have
firms removed / constrained due to excessively large multiples
firms removed due to negative multiples that do not make sense
Ensure that all firms are used in the set of (unbiased) sample descriptive statistics. For firms with
meaningless multiples
, take a firm with negative earnings or a negative PE multiple, instead of dropping
the firm and having biased sample descriptive statistics, we may
Divide the aggregate market value by the aggregate net income/loss of the sample
Compute the earnings yield (the reciprocal of PE ratio) for each and every firm instead