Business combination and earnings before amortization.

ABSTRACT

This study investigates the information content of earnings before amortization of goodwill and other intangible assets for high tech companies. Using 2001 Compustat Research Insight Database, we regress stock prices against traditional earnings per share and the new earnings per

INTRODUCTION

FASB has issued two new statements SFAS # 141 and SFAS # 142 eliminating the use of pooling method for business combinations and changing the rules for goodwill amortization. The objectives are to better reflect the investment made by an acquired entity, improve the comparability of reported financial information, and provide more complete information. The new standards will be effective beginning July 1, 2001. The elimination of the use of pooling method for business combinations has met with strong resistance from the business community, especially the high tech "new economy" companies such as SISCO. The major concern is the goodwill created by the adoption of purchase method and the subsequent amortization will depress their reported earnings, thus lead to the depression of their stock price, putting them in a competitive disadvantage against foreign counterparts.

Prior research indicates that accounting profit is only an artifact. The use of either pooling of interest or purchase methods for business combinations and the subsequent amortization of goodwill do not change the economic substance of the business transaction and should have no significant effect upon a company's stock price in a positive or negative manner, if the market is efficient as it is generally believed. This study investigates the information content of earnings per share before the amortization of goodwill and other intangible assets for high tech "new economy" companies. This new earnings per share, or often called "cash earnings per share" to distinguish from the traditional earnings per share, is constructed by adding back the amortization of goodwill and other intangible assets to the reported traditional earnings per share. We use the simple earnings capitalization model and regress stock prices against traditional earnings per share and the new or cash earnings per share respectively. We find that the estimated coefficient on the new earnings per share is highly significant and the [R.sub.2] is higher, indicating that the new earnings per share has information content to explain stock prices. Test results support the position of FASB in its two new statements to eliminate the use of pooling method for business combinations and change the rules for goodwill amortization.

Section 2 is a review of the literature of the information content of earnings including or before amortization of goodwill and intangibles. In section 3 we give a brief description about the working data sample drawn from 2001 Compustat Research Insight and provide descriptive statistics of the sample. Section 4 discusses the evidence obtained from our empirical test model. Section 5 summarizes and concludes.

THE INFORMATION CONTENT OF VARIOUS MEASURES OF EARNINGS

FASB's proposed elimination of the pooling of interest method for business combination has been a very controversial issue among the business community. High tech companies are among the most vocal opponents of FASB's proposal. Their dissatisfaction is based on the concern that investors and financial analysts value companies on their earnings per share and price/earnings ratios. Even when a fair price is paid, goodwill amortization from the purchase accounting method depresses earnings per share and thus penalizes firms' stock prices, as it is believed that market imposes a fairly rigid PE multiple [Lindenberg et al, 1999]. They even assert that the recognition of goodwill and subsequent amortization put U.S. firms in a competitive disadvantage in the international merger and acquisition arena [Vincent 1997]. Many high tech companies, such as Sisco Systems, Sun Microsystems, and America Online, use extensively the pooling of interest method to acquire smaller companies. They and their representatives fear that the elimination of the pooling of interest method will have severe adverse consequence for their growth and the unprecedented U.S. economy's expansion as a whole. They threaten to take the issue to congress if FASB insists on the elimination of the pooling method [Beresford 2001].

The academia is generally supportive of FASB's position. FASB's proposal to eliminate pooling method and high tech firms' strong opposition has aroused great research interest regarding the information content or value relevance of the various measures of earnings and the efficiency of the capital market. Issues examined include which earnings are more informative, the traditional earnings before extraordinary items, cash flow from operations, or earnings before the amortization of intangibles? How does goodwill amortization affect a company's profitability, and what is the reaction of the capital market to different earnings numbers?

Vincent [1997] collects a sample of fifty-seven purchase transactions and thirty-five pooling transactions during the 1979-1986 period. He examines whether the difference between the two accounting methods for business combination is reflected in stock prices for one method and not the other. His model is based on the theory that share price can be expressed as a weighted average of the firm's book value of equity and its accounting earnings. Findings of the study are inconsistent. There is evidence that pooling firms enjoy an equity valuation advantage over purchase firms, but there is no consistent evidence to link this advantage to the differences in financial reporting. Balsam and Lipka [1998] study the explanatory power of three currently available earnings measures, basic earnings per share, primary earnings per share, and fully diluted earnings per share and whether any one is the better than the other two. They run regression on the simple earnings capitalization model with the annual data from Standard and Poor's Compustat database of corporate annual report for the years 1975 through 1993, and use [R.sub.2] as an index for the usefulness of the measures to explain stock prices. Their findings are consistent with the current literature that all three measures provide investors with useful information in valuing stock prices. Specifically they find that reporting two measures is more informative than reporting one, and the new standard for basic earnings per share and diluted earnings per share is as informative as the old standard requiring primary earnings per share and fully diluted earnings per share.

Lindenberg et al. [1999] examine the claim that the market values firms based on a fairly rigid price/earnings multiple and that amortization due to the use of pooling method for business combination will have a negative effect upon a firm's valuation as earnings are depressed. They argue that purchase or pooling and goodwill amortization are simply accounting artifacts, and have no effect on a firm's future cash flows. According to modern finance theory, the choice of purchase or pooling and amortization, or not, should have no effect on market prices. Their data set includes more than 3,000 companies and over 1,000 mergers and acquisitions from 1991 to 1999. The results from their regressions show that earnings due to goodwill amortization do not affect stock prices and thus pooling should not be viewed as favorable because it avoids goodwill amortization. They also find that, contrary to common belief, price/earnings ratio expands by a sufficient amount in response to amortization, making amortization irrelevant to stock valuation.

Ayers et al. [2000] report that pooling pays a higher acquisition premium that purchase and estimate the effect of pooling method on the balance sheet and income statement. Based on their sample of firms using the pooling method over the period 1992 through 1997, they find that EPS would have been 18.5 percent lower and median return-on-equity and market-to-book ratios 22 percent lower, on average, had these firms used the purchase method. Hopkins et al. [2000] use experiment to investigate how pooling vs. purchase method affect equity analysts' valuation judgments. They find that analysts value stock higher when companies apply either pooling method or purchase method with immediate write-off of the acquisition premium as in-process research and development. They also find that analysts' stock price estimates are lowest when the acquisition happens three years ago, applies purchase method and rarely amortizes goodwill.

Moehrle el al [2001] investigate the information content of various earnings measures, traditional earnings before extraordinary items, the new FASB proposed earnings before amortization, and cash flow from operation, and the relative explanatory power of each measure for market-adjusted returns. Using all S&P 1,500 firm-year observations with non-zero amortization expense from 1988 to 1998, they regress market-adjusted returns against each of the three measures. They report that not surprisingly, the accrual accounting-based measures outperform the cash flow measure. They also find that the traditional earnings before extraordinary items and the new FASB proposed earnings before amortization, or the so-called "cash earnings" by First Call, are equally informative, even in cases where amounts are substantially different. They conclude that their findings support FASB's revised position to eliminate pooling and to change the way of goodwill amortization.

We study the information content of earnings per share before the amortization of goodwill and other intangible assets for high tech "new economy" companies. We, following Moehrle el al [2001], construct this new earnings per share, or often called "cash earnings per share" to distinguish from the traditional earnings per share, by adding back the amortization of goodwill and other intangible assets to the reported traditional earnings per share after taxes but before extraordinary items. Based on prior research, we use the simple earnings capitalization model and regress stock prices against traditional earnings per share and the new or cash earnings per share respectively. We focus on high tech companies since they are fiercely opposed to FASB's new position on eliminating the pooling method. If we find significant relation between firms' market price and the new earnings per share before the amortization of goodwill and other intangibles, the position of FASB to eliminate the pooling method and to change the rules for goodwill amortization will be supported.

DATA DESCRIPTION AND DESCRIPTIVE STATISTICS

This study focuses on the high tech "new economy" companies as they are among the most vocal opponents of FASB's proposal to eliminate the use of the pooling method for business combinations. Sample for the high tech companies include firms in the drug, computer, networking and telecommunicationindustries. Table 1 is a description of the industries in the sample in the three-digit SIC code. These industries are also believed to be among the most active in business combinations and have relatively bigger expense on amortization of intangibles.

Data used in this study are from Standard & Poor's 2001 Compustat Research Insight. All variables in this study are measured on a per share basis. Firm-year observations are eliminated of which (1) December is not the fiscal year end, (2) stock price three months after the fiscal year end is missing or negative, (3) earnings per share data is missing, and (4) amortization expense of intangibles is negative or zero. We further delete firms reporting negative earnings [Balsam and Lipka 1998]. Negative current earnings post a difficult theoretical dilemma since the simple earnings capitalization model indicates that investors pay a certain multiple for current earnings. Researchers argue that the model is misspecified for loss firms and suggest alternative value variables such as book value of equity [Hayn 1995, Collin et al 1997, Burgstahler and Dichev 1997]. The data set spans from 1990 to 2000, as this is the period that high tech companies are at their peak growth and the activities of business combinations are unprecedently hectic.

Table 2 reports the descriptive statistics for the selected sample. [P.sub.t] is cum-dividend price of the firm's stock price three months after the end of the fiscal year t plus its dividend per share for year t. [X.sub.t] is the reported net income after taxes but before extraordinary items. [X.sub.t] is Compustat actual data item 18 (IB). C[X.sub.t] is the reported net income after taxes but before extraordinary items before amortization of intangibles. C[X.sub.t] is the sum of Compustat actual data item 18 (IB) and 65 (AM). After data treatment, we have a total of 1,170 usable firm-year observations. The eleven-year mean of the reported net income (per share) after taxes but before extraordinary items is $0.822, while that of the new net income (per share) before amortization of intangibles is $0.929, indicating that the mean amortization of intangibles is $0.107, or a substantial 13 percent of the reported traditional net income.

TEST RESULT AND DISCUSSIONS

Table 3 reports the results of regressing stock price against the traditional earnings per share each year during the eleven years from 1990 to 2000. The estimated coefficient on the traditional earnings per share after taxes but before extraordinary items is both positive with a mean of 12.33916 and significant at the 1 percent level every year. The average coefficient estimate indicates that every $1 of earnings per share corresponds to $12.34 of market price. The adjusted [R.sub.2] has an average of 42.52 percent meaning that the traditional earnings per share explains 42.52 percent of the variation in equity market values, very much in line with prior reported results of researches in this area. For instance, Collins et al [1999] report an adjusted [R.sub.2] of 38 percent for all firms, and 54 percent for profit firms, during the 1975 to 1992 period using the simple earnings capitalization model.

In Table 4 we report the results of regressing stock price against the new or cash earnings per share each year during the eleven years from 1990 to 2000. The estimated coefficient on the new or cash earnings per share after taxes but before extraordinary items before amortization of intangibles is both positive with an average of 11.71027 and significant at the 1 percent level every year.

The average coefficient estimate indicates that every $1 of earnings per share corresponds to $11.71 of market price. The adjusted [R.sub.2] has an average of 45.53 percent meaning that the traditional earnings per share explains 45.53 percent of the variation in equity market values.

By comparing the results of Table 3 and Table 4, we find that the estimated coefficient of 11.71027 on the new EPS in Table 4 is lower than that of 12.33916 on the traditional EPS in Table 3, indicating that the market adjusts the coefficient down as the new EPS reports higher earnings due to the exclusion of amortization of intangibles. Both regressions are highly significant and the adjusted [R.sub.2] for the regression of price against the new EPS is actually higher than that of the traditional EPS, demonstrating strongly that the new or cash earnings per share after taxes but before extraordinary items before amortization of intangibles has information content in explaining equity market price. Test results do not support the claim of the high tech "new economy" companies that the market pays a fixed multiple for earnings numbers and will not adjust up or down for non-cash expenses such as amortization. Evidence from this study demonstrates the information content of the new proposed earnings before amortization of intangibles, supporting FASB's position to eliminate the pooling method for business combinations and changing the rule for the treatment of goodwill. Figure 1 illustrates the market adjustment regarding the two EPS numbers.

[FIGURE 1 OMITTED]

SUMMARY AND CONCLUDING REMARKS

FASB's proposed elimination of the pooling of interest method for business combination has met with strong resistance from the business community, of which the high tech companies are the most vocal spokesmen. They argue that investors and financial analysts value companies on fairly rigid price/earnings ratios. Even when a fair price is paid, goodwill amortization from the purchase accounting method depresses earnings per share and thus penalizes firms' stock prices. They even claim that the goodwill amortization puts U.S. firms in a competitive disadvantage in the international merger and acquisition arena.

This study tests the information content of earnings before the amortization of goodwill and other intangible assets for high tech "new economy" companies on the data of 2001 Compustat Research Insight Database. We regress stock prices against traditional earnings per share and the new earnings per share respectively which excludes amortization of goodwill. Test results show that the market adjusts price/earnings multiple higher or lower in regard to the inclusion or exclusion of amortization of intangibles, contrary to the fixed multiple for earnings numbers claim of the high tech "new economy" companies. Evidence from this study demonstrates the information content of the new proposed earnings before amortization of intangibles, supporting FASB's position to eliminate the pooling method for business combinations and changing the rule for the treatment of goodwill.

TABLE 1
Industries in the High Tech "New Economy" Sample *

283   Drugs
357   Computer and Office Equipment
360   Electrical Machinery and Equipment, Excluding Computers
361   Electrical Transmissions and Distribution and Equipment
362   Electrical Industrial Apparatus
363   Household Appliances
364   Electrical Lighting and Wiring Equipment
365   Household Audio, Video Equipment, Audio Receiving
366   Communication Equipment
367   Electronic Components, Semiconductors
368   Computer Hardware (Including Mini, Micro, Mainframes,
        Terminals, Discs, Tape Drives, Scanners, Graphics Systems,
        Peripherals, and Equipment)
481   Telephone Communications
737   Computer Programming, Software, Data Processing
873   Research, Development, Testing Services

* The three digit SIC codes and names of the industries are reported.
Industries are selected based on, among other reasons, whether firms
in the industry are likely to have significant intangible assets,
reported or unreported (Jennifer Francis and Katherine Schipper, 1999)

TABLE 2
Comparative Descriptive Statistics for Price,
Amortization Expense, and EPS

                       [P.sub.t] (PRICE)

Year   Obs.     Mean    StdDev   Minimum   Maximum

90      55     12.302   16.470     0.250   101.005
91      65     15.268   15.218     0.219    72.243
92      74     15.770   14.076     0.219    63.000
93      102    14.989   12.058     0.375    62.065
94      93     15.101   10.968     0.937    55.970
95      95     15.921   12.501     1.250    60.791
96      129    16.249   11.384     1.937    49.810
97      144    23.275   14.541     0.437    65.120
98      167    23.582   21.275     0.090   122.507
99      170    36.608   33.365     1.250   212.875
00      76     25.919   19.815     0.060    94.000

Mean   1,170   19.544   16.515     0.639    87.217

                    AMORTIZATION EXPENSE

Year   Obs.    Mean    StdDev   Minimum   Maximum

90      55     0.070    0.115     0.001     0.685
91      65     0.072    0.118     0.001     0.738
92      74     0.089    0.129     0.001     0.819
93      102    0.082    0.131     0.001     0.883
94      93     0.089    0.169     0.001     1.318
95      95     0.082    0.133     0.001     1.071
96      129    0.101    0.206     0.001     1.919
97      144    0.098    0.123     0.001     0.674
98      167    0.120    0.184     0.001     1.368
99      170    0.155    0.336     0.001     4.030
00      76     0.219    0.259     0.002     1.346

Mean   1,170   0.107    0.173     0.001     1.350

                      [X.sub.t] (EPSPX)

Year    Obs    Mean    StdDev   Minimum   Maximum

90      55     0.665    0.902     0.000     5.430
91      65     0.568    0.640     0.000     3.040
92      74     0.742    0.711     0.000     2.900
93      102    0.701    0.669     0.000     3.190
94      93     0.764    0.695     0.040     3.552
95      95     0.689    0.637     0.000     2.800
96      129    0.768    0.740     0.000     4.730
97      144    0.835    0.619     0.000     2.920
98      167    0.871    0.983     0.000     8.700
99      170    1.010    1.101     0.010     5.800
00      76     1.430    1.461     0.000     6.860

Mean   1,170   0.822    0.832     0.005     4.538

                      C[X.sub.t] (EPSAM)

Year    Obs    Mean    StdDev   Minimum   Maximum

90      55     0.735    0.952     0.006     5.523
91      65     0.640    0.711     0.002     3.778
92      74     0.831    0.774     0.007     3.529
93      102    0.783    0.730     0.010     3.633
94      93     0.853    0.727     0.063     3.776
95      95     0.771    0.694     0.015     3.218
96      129    0.869    0.793     0.020     4.755
97      144    0.934    0.665     0.002     3.146
98      167    0.991    1.116     0.001    10.068
99      170    1.166    1.296     0.036     9.150
00      76     1.649    1.595     0.002     7.706

Mean   1,170   0.929    0.914     0.015     5.298

TABLE 3
Coefficient Estimates from Regressing Rice on EPS

(1) [P.sub.t] = a + b[X.sub.t] + [e.sub.t]

Year   Obs.        A            b        Adj [R.sup.2]

 90     55      0.67882      17.48362       0.9146
                 (0.84)     (24.07) **
 91     65      7.58638      13.5198        0.3125
               (3.61) **    (5.49) **
 92     74      3.80447      16.12519       0.6583
               (2.74) **    (11.90) **
 93     102     5.48007      13.57364        0.562
               (4.78) **    (11.43) **
 94     93      7.9791        9.328         0.3422
               (5.80) **    (6.99) **
 95     95      5.79899      14.69026       0.5556
               (4.59) **    (10.89) **
 96     129     8.59635       9.95866       0.4146
               (7.76) **    (9.57) **
 97     144    12.30181      13.13464       0.3081
               (7.25) **    (8.04) **
 98     167    14.80349      10.07614       0.2119
               (7.57) **    (6.75) **
 99     170    25.97909      10.51961       0.1153
               (7.94) **    (4.80) **
 00     76     15.4493        7.32125       0.2819
               (5.71) **    (5.52) **

Mean   1,170    9.85981      12.33916       0.4252
               (5.326) **   (9.586) **

TABLE 4
Coefficient Estimates from Regressing Price on Cash EPS

(2) [P.sub.t] = a + cC[X.sub.t] + [e.sub.t]

Year   Obs.        A             b        Adj [R.sup.2]

 90     55      0.20251      16.46342        0.9029
                 (0.23)     (22.43) **
 91     65      6.84242      13.15693        0.3677
               (3.38) **     (6.18) **
 92     74      3.35694      14.93705        0.6709
               (2.43) **    (12.24) **
 93     102     5.25755      12.43491        0.5622
               (4.53) **    (11.43) **
 94     93      6.70845       9.83944        0.4195
               (5.01) **     (8.21) **
 95     95      4.90475      14.28031        0.6239
               (4.16) **    (12.53) **
 96     129     7.35565      10.23124         0.504
               (7.01) **    (11.45) **
 97     144     11.61279     12.48805        0.3214
               (6.75) **     (8.29) **
 98     167     14.67055      8.99459        0.2178
               (7.52) **     (6.87) **
 99     170     26.00425      9.09761        0.1198
               (8.04) **     (4.90) **
 00     76      14.55878      6.88946        0.2982
               (5.30) **     (5.73) **

Mean   1,170    9.22497      11.71027        0.4553
               (4.942) **   (10.024) **

REFERENCES

Ayers, B.C., Lefanowicz, C.E., & Robinson, J.R.. (2000, March). The Financial Statement Effects eliminating the pooling-of-interests method of acquisition accounting. Accounting Horizons, 1-19.

Balsam, S., & R. Lipka. (1998). Share prices and alternative measures of Earnings Per Share. Accounting Horizon, 12, 234-249.

Beresford, D. R. (2001). Congress looks at accounting for business combinations. Accounting Horizon, 15, 73-86.

Burgstahler, D., & I. Dichev. (1997). Earnings, adaptation and equity value. The Accounting Review, 72, 187-215.

Collins, D. W., M. Pincus, & H. Xie. (1999). Equity valuation and negative earnings: The role of Book Value of Equity. The Accounting Review, 74, 29-61

Francis, J., & K. Schipper. (1999). Have financial statements lost their relevance? Journal of Accounting Research, 37, 319-352.

Hayn, C. (1995). The information content of losses. Journal of Accounting and Economics, 20, 125-153.

Hopkins, P.E., R.W. Houston, & M.F. Peters. (2000). Purchase, pooling, and Equity analysts' valuation judgments. The Accounting Review, 75(3), 257-281

Jennings, R., J. Robinson, R.B. Thompson II, & L. Duvall. (1996). The Relation between Accounting Goodwill Numbers and Equity Values. Journal of Business Finance and Accounting, 23(4), 513-533.

Jennings, R., M. LeClere, & R.B. Thompson II. (2001). Goodwill amortization and the usefulness of earnings. Financial Analysts Journal, 57(5), 20-28.

Lindenberg, E., M. P. Ross, & S. S. Barney. (1999). To purchase or to pool: Does it matter? Journal of Applied Corporate Finance, 12, 32-47.

Moehrle, S. R., J. A. Reynolds-Moehrle, & J. S. Wallce. (2001). How informative are earnings numbers that exclude goodwill amortization? Accounting Horizon, 15, 243-255.

Vincent, L. (1997). Equity valuation implications of purchase versus pooling Accounting. The Journal of Financial Statement Analysis, 5-19.