By Nicholas J. Mastracchio, Jr.

Business valuation is one of the fastest growing practice areas in the accounting profession. Many engagements require valuation of a closely held business; they may be performed for the purposes of a sale, estate and gift taxes, employee stock ownership plans (ESOPs), equitable distribution litigation, and minority shareholder actions. Because often no actual sale is taking place, the valuators’ estimate of value is not confirmed by the marketplace, preventing the valuators from assessing the accuracy of their approach.

Valuating Publicly Traded Stocks

Everyone involved in valuation will agree that in theory, the value of a stock is the present value of its future benefits. When elimination of the pooling of interests method was discussed last year, many were concerned about the possible impact. One author said that the most obvious benefit of pooling is the elimination of goodwill and the burden of amortization on future income statements; if pooling were eliminated, he predicted that accounting earnings would go down and adversely affect value (“Will Poolings Survive?” The CPA Journal, January 1999). Later that year, a Salomon Smith Barney survey concluded that eliminating pooling would have no effect because the market valued stock based on its future cash flow rather than accounting earnings (“Salomon Smith Barney Survey Says Losing Poolings of Interest Will Be OK,” The CPA Journal, September 1999). No one disagreed that a measure of these future performances was the driving force behind valuations.

On the other hand, many advisors acknowledge that investors are interested in historical earnings. Historical price/earnings ratios are reported each day, although some say that this number has become less relevant.

Historical Earnings in Closely Held Companies

Rather than follow the stock, investment analysts of small closely held companies work with management to estimate future earnings or cash flow. In most cases management has not made projections; however, future earnings must still be estimated in order to determine value.

In practice, historical earnings are used to estimatefuture earnings. A five-year average of normalized earnings is typically used to define earnings for closely held business valuations. (Normalization is the elimination of nonrecurring items, adjustments of related-party transactions to fair market value, and other adjustments to make the financial statements more predictive of the future.) Texts on valuating closely held businesses typically caution that a five-year average may not always be appropriate butgive little guidance on when not to use five years and what to use instead. A review of New York State Supreme Court cases involving valuations of closely held businesses indicates that the five-year average of earnings is almost always the basis of the valuation of a closely held company. In the rare exceptions, the valuator has explained the departure from five years.

Because of this conventional wisdom, five years is the predominant choice of valuators, four years being the next-best alternative, and lower numbers of years being less attractive. However, no empirical studies address the issue, leaving five years as an unsupported, arbitrary convention.

Testing the Five-Year Approach

The following test of the five-year rule of thumb involved the earnings capitalization method, requiring the division of earnings by a capitalization rate to determine value. (A price/earnings ratio is the inverse of the capitalization rate.)

Because the market values of closely held companies are rarely known, the test used only publicly traded firms. “Earnings” was defined as earnings before extraordinary items. The test period covered 10 years, with a five-year period of economic growth and a five-year period of decline. The five-year periods ending in 1995 and 1994 were tested as representative of years of growth and strong economic conditions. The five-year periods ending in 1989 and 1990 were chosen for the period of decline.

The database of all publicly traded companies was stratified by industry and analyzed in order to identify industries with at least 50 companies that had positive earnings over the 10-year period. The following six industries met the criteria and were included in the study:

From each industry, one group of 40 firms and one group of 10 firms were randomly selected. The data for the group of 40 were used to compute industry capitalization rates. Five different capitalization rates were determined: five-year, four-year, three-year, two-year, and one-year averages, each ending with the most recent year. The earnings and market value for each company within an industry were plotted and used to compute an average capitalization rate for each industry for the one-, two-, three-, four-, or five-year period.

The industry capitalization rates were used to calculate a value for the group of 10 firms in each of the industries. For each of the four different test periods (1989, ’90, ’94, and ’95) the values calculated through the capitalization rates were compared to published market values at the end of the period to evaluate which rate relative to historical earnings most closely predicted the firm’s actual value. Percentage differences between the estimated values and the actual values were calculated.

The calculations were made for 1989, ’90, ’94, and ’95 for the six industries. A total of 200 valuations per industry and 2,400 earnings capitalization valuations were made overall. The results contradict current thinking on the advisability of using a five-year average and support using a lower number of years.

Test Results

Comparisons were compiled by industry and for all four time periods to determine which average is a better indicator of market value than its longer-term alternatives. Using binomial evaluation, the results are expressed in terms of relative probability.

The Exhibit presents the overall results for the various earnings capitalization periods. The results of the industry calculations were similar to the overall results. Higher probability indicates greater confidence that the shorter period will produce a more accurate estimate. In no case did four or five years prove to be superior to shorter periods.

Specifically, the Exhibit shows that 47.43% of the time one-year average earnings will be a better estimate of value than a two-year estimate. In all other comparisons, a one-year average has an overwhelming probability of being better than an average computed for a longer period. One-year calculations did well in some cases, but overall the data favors using two years’ worth of data.

Editors:
James L. Craig, Jr., CPA

Thomas W. Morris The CPA Journal