Although firm managers devote resources to tax minimization in order to increase shareholder value, there is no empirical evidence that investors value such tax management. The policy issue is that if firms devote resources to tax management which is undervalued by markets, firm resources may be better off spent elsewhere. This study finds that investors do value such tax minimization in terms of increased returns on common stock. What is surprising is that such valuation occurs in light of the well-known extreme measurement error inherent in publicly-available data on tax rates. The results are robust to a number of specific assumptions.

1.0 Introduction
Numerous studies have documented variations in firms’ effective tax rates (ETRs). Since taxes have a direct cash flow effect, such differences should be perceived by analysts and investors as having an effect on firm values. The major purpose of this paper is to test whether there is such an association between ETRs and firm value, the latter of which is measured by returns on common stock. No prior studies have specificly addressed this question.

The answer to the above research question is not obvious for at least two reasons. First, there may not be sufficient variation in firms’ ETRs to allow investors to discern major cash flow differences among firms, or within firms over time. This is particularly true for Federal ETRs; since 1986, many Federal tax preferences have been removed, making Federal tax planning more difficult. On the other hand, state/local and international statutory tax rates vary widely across jurisdictions, and varying rules also allow tax planning opportunities. For this reason, our analysis separates ETRs into three components: Federal, state/local, and foreign.

A second major reason is that the only source of publicly-available information on ETRs is from the firms’ financial statements. Such rates are detailed (by SEC rules) in a footnote to the financial statements. Alternatively, investors can calculate the overall effective rate directly from the firm’s income statement by dividing income tax expense by pretax income. As discussed in the next section, there are numerous measurement errors in such financial statement data which may make it difficult for investors to discern true ETRs.

The research question is important for at least two reasons. From a practice/professional perspective, if firm performance does not increase with lower ETRs, then resources spent on tax planning may be better spent elsewhere. The policy implication which follows is that legislation aimed at providing planning opportunities (e.g., tax incentives) may, on overall basis, be perceived as of low-order importance by investors and thus by managers.

Briefly, the findings are that higher Federal, state/local, and foreign tax rates are all associated with lower returns. The results are robust to a number of model specifications. The next section provides a Literature Review, which is followed by a discussion of the Research Design, Results, and Conclusions.

2. Literature Review

2.1 Effective Tax Rates
To measure the impact of taxes on a firm, simply examining total taxes paid is misleading because of scale effects; i.e., larger firms pay more in total taxes. Instead, an effective rate is used, which divides some measure of taxes paid by some scale, usually net income. Since tax returns are not publicly available, the only way to construct an effective tax rate measure is to use publicly-available financial statement data. As discussed below, there are numerous reasons why such measures are proxies for the true (cash flow) impact of taxes.

There are a number of differences between financial (“book”) and tax accounting which could cause the proxy problem. One is timing; many items of revenue and expense are recognized in different years depending on whether financial or tax rules are used. Depreciation is good example; financial accounting uses straight line depreciation, whereas tax uses accelerated methods. A second reason is permanent differences, whereby revenue is realized under one system but not the other, or expense is recognized under one system but not the other. A prevalent example is income from overseas subsidiaries; tax rules recognize such income only when the cash is repatriated, whereas such income may never be recognized for book purposes. Indeed, such differences may result in enormous differences between true and reported ETRs. As reported in Plesko (2001), there is a low correlation between reported book ETRs and ETRs computed from actual tax return data.

Numerous studies (Gupta and Newberry,1992;Wilkie,1988) have documented variations in book-based effective tax rates between firms and across time. Rates were found to vary due to statutory rate changes and tax preferences. These studies examined overall tax rates, and did not break ETRs into Federal, state/local, and foreign components. This study goes beyond these papers by breaking taxes out into these three components.

2.2 Tax Rates and Performance
To our knowledge, no prior studies have specifically examined the association between ETRs ( or tax expense) and firm performance as defined by returns on common stock. The reason for the absence of such studies may be due to expectations that no surprising results should be found. That is, much like any other component of net income, taxes should have a mechanical affect on performance as simply part of net income. Numerous earnings response coefficient (ERC) studies have documented fairly regular stock price relationships with various components of earnings (although not taxes).

However, such a mechanistic view implicitly assumes that analysts/investors do not specifically examine tax expense for, say, a signal that the firm is managing its taxes. Further, we do not expect a mechanical stock price relationship with tax rates, since rates (as opposed to raw tax expense) are performance measures, and not simply a component of income. Casual empiricism suggest that analysts have recently begun to focus on variations in ETRs, instead of capping them at the top statutory tax bracket. For example, most financial analysts’ texts devote entire chapters to the analysis of income taxes. Some of these texts specifically recommend the examination of firm effective tax rates, over time and vis-à-vis the firm’s industry (White, et al, 1997, pp. 445-446). Another reason we may fail to observe a stock price/ETR relation is because of the noisiness in reported ETRs (as reported above) which may cause analysts and investors to disbelieve( or at least heavily discount) ETRs.

Those studies which have linked taxes and firm performance have either looked at the LIFO-FIFO tax/financial accounting tradeoff, or more recently, whether investors value balance sheet aspects of deferred taxes. Givoly and Hayn (1992) and Tse and Holtzman (1995) suggest that investors view deferred taxes as liabilities. More recently, Ayers (1998) found that investors new information required by the enactment of SFAS No. 109 provided substantial value-relevant information over the information disclosure required by APB 11.

3. Research Design

3.1 Data
Our sample is drawn from all firms available on Compustat and CRSP from 1994 through 1999. To avoid outliers, firms with ETRs above 100% were capped at this value, which is the theoretical maximum in any one year. Also, observations for which variable values fell in the outer one percentiles were eliminated. Firms which did not have all data on Compustat, and for which insufficient data was available to calculate daily returns on CRSP, were also eliminated. Public utilities, financials, and insurance companies were also eliminated. This resulted in a final sample of 17,072 firm-years.

The variables used are as follows (where subscripts i and t refer to firm i and year t, respectively):

R_i,t = Raw daily percentage returns on common stock compounded over the
fiscal year (return is the change in daily price divided by price),

PTNI_i,t = Firm pre-tax net income, excluding extraordinary items,

OETR_i,t = Firm overall effective tax rate,

USETR_i,t = Firm Federal tax rate

SETR_i,t = Firm state/local tax rate, and

FETR_i,t = Firm foreign tax rate

Effective tax rates are computed as the sum of current plus deferred tax expense, divided by pre-tax net income. We also measure each of the ETRs in terms of tax cost/savings. We measure such savings three ways. The first savings measure is done at the industry level. First, we determine the (four digit SIC code) industry median effective tax rate for each category of explicit tax rate. These variables are denoted by an “S” on the end, e.g., USETR_i,t is the difference by which the firm’s effective tax rate differs from that of its industry median, multiplied by pretax income. Since these by-firm amounts are subtracted from their industry medians, a positive value indicates tax savings; a negative value, tax costs above its industry counterparts. By controlling for industry effects, we may be able to isolate firms’ management of taxes, or lack thereof. The second way we measure tax savings/cost is to subtract the observed ETR from the maximum statutory rate. The maximum statutory rates are 35% (US Federal); 12% (highest state income tax rate--Iowa); and 46% (highest foreign rate--Germany). The third way we measure tax savings/cost is via annual tax rate changes. That is, we subtract each firm’s current year ETR form that firm’s previous year ETR. Any positive (negative) value is considered tax savings (cost).

Descriptive data is reported in Table 1. As can be seen, there is a wide variation in ETRs, which adds credence to our mis-measurement arguments. Some rates are obviously inflated; many ETRs exceed 80%, and before truncation of our sample, some exceeded 100%!

Table 1 - Descriptive Data for the Variables^a

4. Models
To test whether various components of effective tax rates describe firm performance as reflected in stock returns we use the following model:

R_i,t = a₀ + B₁PTNI_i,t + B_j ETR_j + e_i,t - (1),

Where ETR_j are the various forms of firm tax rates, and other terms are defined previously. That is, we perform a series of regressions using various forms of explicit tax rates. The model, except for the addition of tax effects, is consistent with that of Dhaliwal et al (1999), and studies cited therein. In conventional financial economics theory, a firm is valued based on its cash flows, which are proxied by accounting information. Under the assumption of informationally-efficient capital markets, accounting information becomes known throughout the year through a variety of sources (quarterly reports, company earnings announcements, media articles, etc). Changes in the value of a firm’s stock (return) is an on-going process throughout the year in response to knowledge of the firm’s earnings. As noted in Dhaliwal, Subramanyam, and Trezevant (1999), and citations therein, the use of earnings levels as a proxy for unexpected earnings in a regression of returns and earnings has theoretical and empirical support.

5. Results

5.1 Tax Rates
Results of regressions of returns on effective tax rates for the numerous models are shown in Table 2. The first regression is a check of the veracity of our model and data; the independent variable is only pretax income . Results for this are shown in the first line in the Table; coefficient estimates, adjusted R², and overall model F statistics are nearly identical to those compared to those of previous studies (e.g., Dhaliwal, Subramanyam, and Trezevant, 1999). Since the data are pooled cross-sectional and time series, 4-digit SIC industry fixed effects dummy variables and year dummy variables are also included; the overall F statistic for the resultant fixed effects coefficients is shown in the table.

Table 2 - Regression Results: Association of Common Stock Returns with Effective Tax Rates (t statistics in parentheses)

The second model reports the effects of overall effective tax rates. As can be seen, each 1% increase in overall effective tax rates reduces stock returns by .3%. The third model bifurcates ETRs into Federal and state/local components; as can be seen, the higher the effective state and Federal tax rate, the lower the stock market return.

The fourth model reports the effects of Federal and foreign effective tax rates on returns. To make the analysis meaningful, only firms reporting pretax foreign income are included in the analysis (this item is required to be reported by firms, if material). The sample size is reduced to 5161 firm-years. As can be seen, for the most part coefficients remain the same relative to those of prior models. In addition, the foreign tax rate coefficient is significant and negative.

5.2 Tax Savings
Table 3 reports regression results based on overall tax savings/dissavings, as measured by the variance of firm ‘s ETR from the firm’s industry median ETR, multiplied by the firm’s pretax income. If this variable can be thought of as tax management, then the results indicate the degree to which stock returns are associated with tax planning. Model 5 reports the equivalent of model 2 except using overall tax savings in lieu of overall ETR. The results show that reducing overall effective taxes (relative to that of the industry) is associated with increased common stock returns. Model 6 bifurcates tax savings into Federal and state; both are positively associated with returns. Finally, model 7 adds foreign tax savings, which is also significantly associated with higher returns.

Table 3 - Regression Results: Association of Common Stock Returns with Tax Savings Tax Savings=Industry Median Taxes Minus Total Firm Taxes
(t statistics in parentheses)

Table 4 reports results where tax savings is measured as the difference between the top statutory tax rates and the firm’s reported ETRs. Model 5A reports the equivalent of model 5 except using the different measure of overall tax savings. The results show that reducing overall effective taxes (relative to that of the top statutory tax rate) is associated with increased common stock returns. Model 6A is the equivalent of model 6, except using the different measure of tax savings. Here, Federal and state tax rate savings both are positively associated with returns. Finally, model 7A adds foreign tax rate savings, which is also significantly associated with returns.

Table 4 - Regression Results: Association of Common Stock Returns with Tax Savings Tax Savings=Statutory Tax Rates Minus Firm ETRs
(t statistics in parentheses)

Finally, Table 5 reports results where tax savings is measured as the difference between a firm’s current and previous year ETRs. Model 5B reports the equivalent of model 5 except using this different measure of tax savings. The results show that reducing overall effective taxes (relative to that of the prior year) is associated with increased common stock returns. Model 6B is the equivalent of model 6, except using the different measure of tax savings. Here, Federal and state tax rate savings (relative to the prior year) both are positively associated with returns. Finally, model 7B adds foreign tax rate savings relative to the prior year), which is also significantly associated with returns.

Table 5 - Regression Results: Association of Common Stock Returns with Tax Savings Tax Savings = Firm ETR_t-1 – Firm ETR_t
(t statistics in parentheses)

6. Conclusion
Results, using a variety of effective tax rate measurements, indicate that the stock market values low tax rates. What is surprising is that this effect occurs despite serious measurement issues relating to effective tax rates as reported in publicly-available financial statements. Further research, using actual tax return-derived ETRs as a benchmark, is necessary to see if the market’s valuation of financial statement-based ETRs is inaccurate, or if the market is efficient enough to see through the measurement errors and place an appropriate value on taxes.

References
Ayers, B. (1998). Deferrred tax accounting under SFAS No. 109: an empirical investigation of its incremental value-relevance relative to APB No. 11. The Accounting Review 73 (April): 195-212.

Cottle, S., Murray, R., and F. Block (1992) Graham and Dodd’s Security Analysis. New York: Mcgraw-Hill.

Dhaliwal,D., Subramanyam, K.R., and R. Trezevant (1999). Is comprehensive income superior to net income as a measure of firm performance? Journal of Accounting and Economics 26: 43-67.

Givoly,. D. and C. Hayn (1992). The valuation of the deferred tax liability: evidence from the stock market. The Accounting Review (April): 394-410.

Gupta, S. and K. Newberry (1992). Corporate average effective tax rates after the Tax Reform Act of 1986. Tax Notes 55 (May 4): 689-702.

Plesko, G. (2001). An Evaluation of Alternative Measures of Corporate Tax Rates. Working paper, Sloan School of Management, Massachussetts Institute of Technology.

Swenson, C. (1992). Some Tests of the Incentive Effects of the R&E Credit. Journal of Public Economics (November).

Swenson, C. (1999). Increasing firm value by reducing effective tax rates. Tax Notes 62

Tse, S. and M. Holtzman (1995). The cross-sectional valuation of deferred income taxes: effects of industry-specific policies. Working paper, University of Texas-Austin.

White,G., Sandhi, A., and D. Fried (1997). The analysis of financial statements. New York: John Wiley and Sons.

Authors
Joe Lua, Assistant Professor, Faculty of Business and Economics, University of Hong Kong, Hong Kong.

Charles W. Swenson, Professor and Leventhal Research Fellow, Marshall School of Business, University of Southern California, Los Angeles, CA, USA.