The study examines the annual financial reports of a group of large U.S. corporations and purports to show “the federal income taxes paid or not paid by 275 of America’s largest corporations in 2001, 2002, and 2003.” Interestingly, the paper fails to mention that financial reports can be used to derive only crude estimates of corporations’ actual tax payments because tax returns, which are not publicly disclosed, are needed to reconcile the differences between tax law and accounting rules.
The report also claims that congressional leaders and the Bush Administration changed the tax laws to benefit “loophole-seeking corporations,” thus lowering many corporate tax bills. Such a serious charge should, at a minimum, be supported by citations of tax law changes and sufficient information to allow independent researchers to evaluate the allegations. Regrettably, the paper does not name any legislation, and its methodology section lacks essential details.
The study omits other key details that seriously weaken its conclusions. The following list summarizes the report’s major flaws and omissions:
The differences between tax laws and financial accounting rules create legitimate differences between values reported on corporate financial statements and tax returns. The study does not discuss reasons for these valid differences. It merely mischaracterizes them as tax avoidance schemes.
CTJ estimates average corporate tax rates using figures that almost certainly differ from their true values. The legitimate discrepancies between values reported on financial statements and those reported on tax returns cannot be accurately reconciled without access to private tax return data. The CTJ report does not mention this serious limitation to its analysis.
Previous attempts by CTJ to estimate corporate tax payments using only financial statement data have been proven grossly inaccurate. In January 2002, CTJ claimed that Enron received a net refund
of $278 million on its 2000 federal income taxes.
However, the Joint Committee on Taxation reported in 2003 that Enron paid
$63.2 million in 2000 federal income taxes.
Massachusetts Institute of Technology professor George Plesko estimates that the aggregate difference between taxable income reported on financial statements and reported on tax returns declined in 2000 and was negative in 2001.
This negative difference indicates that the amount of income reported to the Internal Revenue Service (IRS) was greater
than the amount of income reported on public financial statements, thus casting doubt on the claim that the book-tax difference reflects increased tax shelter activity. The ideal tax shelter reduces the amount of income reported to the IRS without affecting the amount of income reported on financial statements.
The report fails to address the difference between marginal and average tax rates. However, it presents its average
rate as the key tax rate faced by companies. Marginal tax rates are the tax rates applied to the next dollar that the firm earns, while average tax rates represent the percentage of tax
paid on total income. Marginal tax rates are important because they dictate the after-tax income earned from producing another good or investing in another project. CTJ’s own estimates suggest that companies pay marginal tax rates twice as high as their average rates.
The study fails to recognize the difference between the timing and magnitude of tax deductions. Consequently, it misrepresents the tax treatment of depreciation and net operating losses (NOLs) as illegitimate tax subsidies. When corporate managers use accelerated depreciation and/or NOLs in a given year, they trade away the use of these deductions in subsequent years.
The paper does not provide the proper statistical context for its estimated average corporate tax rate of 18.4 percent. Although it purports to show the average tax rate for the “typical” large U.S. company, its sample suggests that an average tax rate between 6.72 percent and 30.14 percent could be considered “typical.” CTJ does not report this range, yet this much variation, combined with CTJ’s imprecise methodology, sharply limits any usefulness of the 18.4 percent average.
The report’s most disturbing aspect is that it purports to show the actual tax payments of 275 large corporations, but never mentions the fact that this information is not publicly disclosed.
The following sections of this report detail the major shortcomings and omissions of the CTJ corporate tax study.
Corporate annual reports are prepared using Generally Accepted Accounting Principles (GAAP)—standards that are meant to provide uniform reports that investors, shareholders, and regulators can use to evaluate financial performance. Because most businesses are unique, GAAP allows corporate accountants substantial discretion over key decisions, such as choosing a method for revenue and expense recognition. Because tax laws do not offer this type of discretion, legitimate differences arise between tax payments disclosed on financial reports and the actual tax payments made to the IRS.
In fact, disparate financial reporting rules and tax laws typically lead to differences in the amount of pre-tax profits (taxable income) reported to shareholders and to the IRS. The divergence in the values on financial reports and tax returns are referred to as “book-tax differences” and are typically classified as either permanent or temporary.
Permanent book-tax differences arise when income and/or expenses are recognized under one set of rules but not under the other. Examples of permanent book-tax differences are the treatment of stock option compensation and tax-free interest from municipal bonds. Stock option compensation is recognized as an expense for tax reporting purposes but not under GAAP. Tax-free municipal bond interest is included in GAAP income but not in the pre-tax income reported to the IRS.
Temporary book-tax differences arise when the same total income and/or expenses are reported under both systems but in different time periods. For example, the IRS allows companies to use “accelerated” depreciation, whereby a greater portion of an asset’s cost is expensed at the beginning of its usable life and a correspondingly smaller share is deducted from income in future years. Although GAAP does not require companies to report accelerated depreciation expense, the total amount of depreciation taken remains the same under both GAAP and tax rules. In other words, the difference between the two sets of rules is the annual amount of depreciation expense, not the total depreciation expense.
Book-Tax Differences Compared to Tax Shelter Activity
Book-tax differences are difficult to reconcile using the information in publicly available annual reports—even when supplemented with confidential tax return data. Because of these difficulties, a group of academics and policy experts recently undertook an effort to make book-tax differences more transparent. As part of this ongoing effort, the Department of the Treasury recently introduced the form M-3, an expanded book-tax reconciliation report. Starting in tax years ending on or after December 31, 2004, the M-3 must accompany the U.S. tax returns of all corporations with at least $10 million in total assets.
The form M-3 includes 75 separate reconciliation items and is designed to “make differences between financial accounting net income and taxable income more transparent.” Some policymakers hold that large book-tax differences reflect increased abuse of tax shelter activity. However, as Professor Plesko points out, “Until the reporting requirements change, it is only possible to speculate on the magnitude of specific factors affecting the difference and the potential effects certain types of transaction[s] may have.”
Plesko also points out that many legitimate accounting factors and abusive tax shelter activities affect book-tax differences in the same manner. Interestingly, Plesko analyzed confidential Treasury Department data and found that the aggregate book-tax difference declined in 2000 and 2001—and even turned negative in 2001. This negative book-tax difference indicates that the amount of income reported to the IRS was greater than the amount of income reported on public financial statements. The negative amount casts doubt on the claim that the book-tax difference reflects increased tax shelter activity because the ideal tax shelter reduces the amount of income reported to the IRS without affecting the amount of income reported under GAAP.
How Accurate Are the CTJ Estimates?
The book-tax difference issue is not fully discussed in the CTJ report. However, its methodology section suggests that financial statement profits and taxes were adjusted to calculate the average effective tax rates for its sample of companies. Oddly, the report fails to acknowledge any limitations to such an analysis, even though there is no way to accurately adjust book values to tax return values without access to confidential tax return data. Therefore, judging the accuracy of the CTJ estimates is impossible. Nonetheless, they are unlikely to represent the true tax bills of these companies.
Indeed, several companies have already taken issue with the CTJ estimates. The report states that SBC Communications paid $1.58 billion in taxes between 2001 and 2003, with a refund from the government of $475 million in 2003. However, the company’s Vice President of Finance, John Stephens, told The Wall Street Journal that SBC paid $15 billion in taxes over this period. Additionally, SBC spokeswoman Anne Vincent said that “It’s just not true that we got a half-billion dollar check back from the federal government.”
A spokeswoman for Pepsi Bottling Group also took issue with the CTJ report, which claims that Pepsi paid $189 million in taxes on $1.55 billion in profit, for an effective tax rate of 12.2 percent. Pepsi’s Kelly McAndrew said, “It is impossible for us to determine how they calculated the tax rate, which is also incorrect.”
These comments are not surprising because the report does not provide a detailed methodology and because the CTJ estimates are derived using incomplete data. In fact, at least one of CTJ’s previous attempts at estimating corporate tax rates using only financial statement data has been proven grossly inaccurate.
In a report dated January 17, 2002, CTJ claimed that Enron received a net refund of $278 million on its federal income taxes in the year 2000. However, the Joint Committee on Taxation reported in 2003 that Enron paid $63.2 million in federal income taxes in 2000, consisting of $21.3 million in regular federal income tax and $41.9 million in corporate alternative minimum tax (AMT). The CTJ estimate was not only off by $341 million, but it suggested a net refund when the company actually paid over $60 million in taxes.
This error can be attributed to CTJ’s methodology, which consisted of adjusting for only book-tax differences that can be observed on financial statements. In the case of Enron, CTJ estimated the net refund by subtracting its publicly reported employee stock option expense of $390 million from its publicly reported current federal income tax expense of $112 million. Making this adjustment is not improper, but making only this adjustment does not yield the company’s actual tax liability.
Regrettably, the new CTJ report appears to have repeated the same mistakes, adjusting for only selected book-tax differences that can be verified using financial statements. For instance, the CTJ methodology note (in its entirety) for Coventry Health Care reads as follows: “Federal and state tax benefits from stock options in 2003, 2002, and 2001 were $19 million, $15 million and $2 million.” This note implies that stock option deductions were subtracted from Coventry’s publicly reported (under GAAP) current federal income tax expense, and that no other adjustments were made. The CTJ estimates of Coventry’s actual tax payments are seriously suspect if only this adjustment was made (as was the case with CTJ’s estimate of Enron’s 2000 taxes).
The truth is that the information needed to calculate these companies’ tax rates is simply not publicly available. Consequently, CTJ’s methodology renders all of their estimated corporate tax rates questionable.
Even if most of the CTJ estimates are reasonable, the reported figures are misleading for other reasons. For instance, the study all but ignores the statistical variation in its own averages and it omits any discussion of marginal tax rates.
Variation in Tax Rates
The CTJ study purports to show the average effective tax rates for a “representative” sample of U.S. companies. In other words, CTJ claims that their 18.4 percent average tax rate represents the tax rate of the “typical” U.S. corporation. The study states:
From the list of just over 300 profitable companies…we gleaned the list down to 275, based mainly on an effort to maintain a representative sample of companies in each industry. The overall trend in pretax U.S. profits reported by our 275 companies is virtually identical to the trend in total corporate pretax profits compiled by the Commerce Department.
Despite matching the trends reported by the Department of Commerce, the CTJ estimate does not appear to be a precise representation of the typical U.S. corporation’s average tax rate. Put differently, if the CTJ sample’s average tax rate is representative of the typical company’s tax rate, then virtually any tax rate can be considered typical. The evidence for this conclusion is in the variation of the average tax rate in CTJ’s sample—a measure that is not reported in their study.
In statistical terms, the study presents the average tax rate for its sample, but it disregards the standard deviation of this tax rate. This omission is critical because the standard deviation provides a range within which most companies’ average tax rates would be expected to fit. Because CTJ claims that its sample is representative of most U.S. companies, the average has to be evaluated within this range. Table 1 reproduces the CTJ sample’s average effective 3-year corporate tax rate (18.4 percent) and provides the standard deviation of this average tax rate (11.71 percent).
If the CTJ sample is representative of the typical U.S. company, this standard deviation of 11.71 percent can be used to estimate a range within which most companies’ average tax rates fall. The low end of this range is calculated by subtracting 11.71 percent from the average, and the high end is determined by adding 11.71 percent to the average. Consequently, the standard deviation of 11.71 percent implies that any company with an average tax rate between 6.72 percent (18.4 – 11.71) and 30.14 percent (18.4 + 11.71) can be considered “typical.”
In other words, the typical U.S. firm’s average effective tax rate, as estimated in the report, can vary substantially from 18.4 percent. This much variation in CTJ’s average tax rates—combined with CTJ’s imprecise methodology for computing the tax rates—sharply limits any usefulness of the 18.4 percent average. Furthermore, CTJ does not provide any discussion of marginal tax rates, a key rate for decision-making purposes.
Total Income Versus Additional Income
Average tax rates are calculated by dividing taxes paid by total income. Marginal tax rates are those that are applied to additional increments of income. Similar to personal marginal tax rates, corporate marginal tax rates are applied to various income brackets, with different tax rates applied to different amounts of taxable income. The CTJ study estimates the average effective tax rates for its sample of companies, but it does not discuss the companies’ marginal tax rates.
This omission is critical because marginal tax rates—not average tax rates—are the relevant measure for decision-making purposes. For example, marginal tax rates are used to evaluate investment opportunities because marginal rates determine the net amount earned on the next dollar invested. The higher the marginal tax rate faced by the firm, the less income a project will be expected to earn and the less likely it is that the firm will undertake the project. This concept is completely ignored in the CTJ study, which misrepresents an average tax rate of 18.4 percent as proof that companies do not pay the statutory tax rate of 35 percent.
However, just as with personal tax rates, many companies face a marginal tax rate that is higher than their average tax rate. For instance, after all credits and deductions, an individual with taxable income of $150,000 may face a marginal tax rate of 33 percent, but have an average tax rate of only 18 percent. This individual would face a lower average tax rate because some income is exposed to lower marginal tax rates and because deductions and credits reduce the amount of taxes due. Corporate taxes work the same way, only with a different rate schedule.
The corporate federal income tax rate schedule, presented in Table 2, shows that any company earning at least $75,000 in taxable income faces a marginal tax of at least 34 percent. According to the CTJ estimates, all 275 companies in their sample have a taxable income of at least $75,000, which means that they all face a marginal tax rate of at least 34 percent. As seen in Table 3, even after eliminating all of the companies that CTJ claims received a tax refund (in either 2001, 2002, or 2003), almost 90 percent of the companies in their sample face a marginal tax rate of at least 34 percent—nearly double the average effective tax rate estimated by CTJ.
What New Tax Breaks?
Another key omission in the CTJ study is its failure to name any tax law changes, which it blames for lowering companies’ tax rates. The report states that “President Bush signed new business tax breaks totaling $175 billion over the 2002–04 period.” Presumably, the CTJ report refers to the Job Creation and Worker Assistance Act of 2002 (JCWAA), which was signed on March 7, 2002. The report does not name any legislation, but the JCWAA addressed two of the tax provisions discussed in the CTJ paper.
The JCWAA altered the tax treatment of accelerated depreciation and of net operating losses, two provisions criticized in the paper as new tax breaks. However, accelerated depreciation and NOL tax deductions have existed for many years. The JCWAA affected the timing—not the magnitude—of these tax deductions. The CTJ report also lists employee stock options, tax credits, offshore sheltering, and the failure of the corporate alternative minimum tax as reasons that corporations are paying lower tax bills.
The U.S. tax code has been filled with special tax credits and “corporate welfare” provisions for many years, and these should all be abolished. Still, implying that these problems with the tax code are the fault of any particular President or Congress requires, at minimum, a list of laws that were passed during their tenure. The paper falls short of this standard because it does not list one section of any law passed between 2001 and 2003.
The rest of this section provides brief explanations of key tax deductions criticized in the report and, where possible, how any legislation passed in the last few years affected these deductions.
Depreciation and Net Operating Losses.Corporations have been allowed to depreciate assets at an accelerated pace for tax purposes since at least 1954. The current form of accelerated depreciation, known as the Modified Accelerated Cost Recovery System (MACRS), was part of the Tax Reform Act of 1986. MACRS, like earlier forms of accelerated depreciation, allows companies to deduct more of their assets’ costs in earlier years than in later years, but does not change the total deductible depreciation for any asset. In other words, accelerated depreciation provides a trade-off, a larger tax benefit up front for a smaller tax benefit in subsequent years. (See Table 4.)
The JCWAA provided companies with the option of accelerating their depreciation even faster than under MACRS. This “bonus” depreciation allowed companies to expense an additional 30 percent of an asset’s cost in its first year of service. However, just as with regular MACRS, bonus deprecation does not change the total amount that an asset can be depreciated. Assets cannot be depreciated below a value of zero.
Table 5 demonstrates that the bonus provision simply provides another trade-off opportunity—a larger tax benefit now for a smaller tax benefit later. All of the accelerated depreciation methods affect the timing—not the magnitude—of tax deductible depreciation.
Another JCWAA tax provision that mostly affects the timing of an existing tax benefit is its treatment of net operating losses. Corporate taxpayers incur a net operating loss when certain deductions exceed their gross income in a given tax year. Because the tax code prohibits companies from reducing taxable income below zero, it allows a “carry back/carry forward” treatment. Prior to the JCWAA, companies were allowed to carry back a NOL to the two (and in some cases three) years immediately preceding the NOL and to carry forward the remaining NOL for up to 20 years. If the carry-back method is chosen, the NOL must be carried back to the earliest year first, with the remainder applied to taxable income in successive years.
Similar rules have been in place since at least 1954, and the carry-back and the carry-forward methods can only be used to offset taxable income. The NOL provision in the JCWAA expanded the two (and three) year carry back to five years, but only for NOLs arising in tax years 2001 and 2002. It did not change the carry-forward rule. This expanded carry back was intended to provide businesses with immediate tax relief following the 2001 recession and the September 11 terrorist attacks.
Regardless of the motivation, the JCWAA provision did not expand the total deductibility of any given NOL; it merely changed the timing of the tax deduction. For example, regardless of whether a $500,000 NOL is carried back two, three, or five years or carried forward for any number of years, the total deductible amount of the NOL remains $500,000.
Furthermore, the decision to carry back the NOL is not straightforward. Offsetting taxable income from earlier years with a NOL can affect other tax deductions and credits as well as AMT liability, making the carry-back decision rather complicated. Regardless, both the NOL and bonus depreciation provisions in JCWAA affect the timing—not the magnitude—of existing corporate tax deductions.
The Stock Option Expense Myth.CTJ also mischaracterized the tax treatment of stock option expenses as a special tax break. For the past 30 years, corporate managers have increasingly turned to stock options as a method for compensating employees. These options allow employees to buy stock at below-market rates and sell their shares for a “profit.” Just as with wages, this stock option profit is deducted from corporate taxable income and taxed as ordinary income on employees’ individual tax returns. In other words, for corporate tax purposes, this compensation is treated like any other compensation expense.
The recent controversy surrounding employee stock options arose because corporations do not have to report the value of this compensation as an expense on their financial statements (according to GAAP). Because employee stock options are employee compensation, some experts have argued that they should be accounted for as an expense on the income statement. Regardless of the best GAAP treatment, this disagreement is about whether the tax benefit will be reported the same way on financial statements as on tax returns, not over whether the tax benefit itself will change. To characterize the corporate tax benefit from employee stock options as a tax loophole equates the deductibility of cash salaries and wages with a special tax break.
Corporations are not taxed on wages they pay and stock options they grant because the recipients are taxed at the individual level. This tax treatment serves as an important reminder that corporations are merely legal entities, and people—not corporations—pay taxes. Ultimately, all corporate taxes are taken out of the pockets of people, either through lower compensation to workers, higher prices paid by consumers, or lower returns to investors.
Corporations Do Not Pay Taxes. Evidence from the past 20 years suggests that many people understand that they bear the burden of “corporate” taxes. An increasing number of business owners have chosen to organize as pass-through entities, such as S corporations and limited liability companies (LLCs). These non-corporate entities afford legal protection similar to that of traditional corporations, but they allow business income to “pass through” to the owners’ personal tax returns. Consequently, owners’ business income from a pass-through entity is taxed at only the individual level, whereas owners’ income from a traditional C corporation is taxed at both the corporate and personal levels.
Chart 1 shows that in 1975, S corporations accounted for only 17.22 percent of all “corporate” tax returns while C corporations accounted for more than 80 percent. By 1996, however, S corporation returns accounted for the majority of the total. The IRS projects that S corporation returns will account for nearly 60 percent of all corporate tax returns in 2004. Table 6 shows that filings for other types of pass-through entities have also increased substantially.
Both sole-proprietor and partnership (including LLC) filings have more than doubled from 1975 to 2004. During this same period, traditional C corporation filings increased by only 23 percent. Table 6 also provides a more recent comparison that shows that the number of C corporation filings actually declined from 1990 to 2004. In contrast, the growth in all three categories of pass-through entity filings increased substantially from 1990 to 2004. Given the enormous relative growth in the pass-through entities subsequent to 1990, the distinction between corporate and individual taxes is much less meaningful than it was only 25 years ago.
An ever-increasing number of individuals are running their own businesses as non-corporate entities. They continue to pay taxes on their business income, just not through the corporate tax system. This trend away from the traditional C corporation has contributed to the decline in the relative share of corporate taxes collected by the Treasury, but the CTJ report does not even discuss this issue.
Corporate Citizens and Offshore Tax Sheltering. The CTJ study makes yet another key omission when it faults offshore tax sheltering for the decline in corporate taxes. The study essentially faults corporations for international tax planning: “Some companies have gone so far as to renounce their U.S. ‘citizenship’ and reincorporate in Bermuda or other tax-haven countries to facilitate sheltering activity.” One problem with this criticism is that corporate directors have a fiduciary duty to maximize shareholders’ wealth, even if it means paying less in taxes to the U.S. government.
Just as important, the CTJ criticism ignores that the U.S. is one of the few developed nations that uses a worldwide tax system rather than a territorial tax system. Most nations tax their companies using a territorial system, in which companies pay income taxes only to the tax authorities of the country in which income is earned. The U.S. tax code, however, employs a worldwide tax system, requiring U.S. companies to pay income taxes on “foreign-source” income to both the U.S. Treasury and to the foreign tax authority where the income is earned. For example, U.S. companies that earn income in Ireland are subject to Ireland’s 12.5 percent tax rate and the U.S. corporate tax rate of 35 percent. However, most foreign companies operating in Ireland are subject only to Ireland’s 12.5 percent tax rate.
The U.S. tax code ostensibly allows U.S. companies to claim a credit for foreign income taxes paid, but even in an ideal situation the credit acts as a ceiling, merely preventing these companies’ marginal tax rates from exceeding 35 percent. These higher U.S. tax rates on foreign-source income place U.S. companies at a competitive disadvantage and have contributed to the recent increase in corporate “inversions,” the process by which companies reincorporate outside of the United States. These tax rules have also contributed to foreign takeovers of U.S. companies, in which the parent company remains headquartered outside of the United States.
Because the U.S. tax code unduly burdens corporate directors, owners, and workers by taxing income regardless of where it is earned, it is not surprising that they have tried to minimize the economic harm caused by these high corporate taxes. The CTJ study does not address the reasons that corporate directors would decide to use corporate inversions. It merely disparages the practice as evidence of poor corporate citizenship. This criticism is completely misplaced because failing to use all legal methods of tax planning—inversions included—would make corporate managers derelict in their responsibilities to shareholders.
The recently released Citizens for Tax Justice study of corporate taxation is filled with errors and omissions. The most serious of these problems is that CTJ misrepresents its own estimates as corporations’ actual tax payments. CTJ develops its figures using corporate financial statements, yet corporations’ actual taxable income and tax payments cannot be determined without access to their tax returns, which are not publicly disclosed. CTJ appears to have made the same mistake as in previous reports, in which it adjusted financial statement figures for only selected book-tax differences.
The differences between tax accounting rules and financial accounting rules cannot be reconciled without access to private tax returns, and these book-tax differences do not arise solely because of “tax-avoidance schemes.” CTJ’s failure to acknowledge these legitimate differences is compounded by the lack of understanding it displays toward corporate taxation in the U.S. Thus, the CTJ paper “Corporate Income Taxes in the Bush Years” presents an incredibly misleading picture of corporate taxation in America.
Norbert J. Michel, Ph.D., is a Policy Analyst in the Center for Data Analysis at The Heritage Foundation.
The nonprofit group Citizens for Tax Justice (CTJ) recently published a report titled “Corporate Income Taxes in the Bush Years.”