I am one of eight signatories of a letter that has just been sent to the Financial Accounting Standards Board, urging it to repeal APB 23, the rule that allows U.S. companies to designate particular foreign earnings as indefinitely reinvested abroad, thus allowing U.S. deferred tax liabilities to be ignored rather than being deducted from reported earnings at full value.
APB 23 has terrible tax policy effects, as it creates lock-in for foreign earnings insofar as managers who have taken advantage of it don't want to create negative adjustments (or to undermine their ability to make other APB 23 designations in the future). But it also is absurdly discontinuous (causing deferred U.S. taxes to jump from being deducted at full value to being wholly ignored) and excessively discretionary - as I feel I can say, despite not being an accountant, both from having talked to accountants and from the overlap between accounting and legal rule design with respect to income.
I would be (pleasantly) surprised if FASB took notice of this letter in any way. But I see it as a useful contribution to public debate about the issues (including before Congress), and wish to thank those who took the lead in creating this letter.
Anyway, here is the text of the letter (which, it is not hard to tell, reflected the lead input of individuals more knowledgeable about financial accounting than I am):
August 31, 2015
Financial Accounting Standards Board
Dear FASB Members,
We encourage Members of the Financial Accounting Standards Board to repeal Accounting Principles Board Opinion No. 23, the rule that allows a company to make a designation of indefinitely reinvested earnings (IRE) to suppress a U.S. deferred tax liability (DTL) that otherwise would be reported as contingent on the repatriation of deferred foreign earnings.
We are concerned that APB 23 IRE designations undermine accounting credibility and contribute to bad tax policy. The designations are an incentive to reduce domestic economic activity and the U.S. tax base by encouraging investment in low tax jurisdictions. Further, the designations invite real or perceived management conflicts of interest, creating vulnerability for IRE reversals (including reversals that have already occurred) that damage public accounting. 
While pending legislative action could moot company interest in IRE designations, it is important for FASB to recognize that the accumulation of foreign deferral attributable in part to APB 23 has contributed to the advocacy for another repatriation holiday and/or replacing current law with a more territorial system. There may be no better example of the power of accounting than this case in which suppression of a DTL for book purposes (about which some accountants and others have had concerns from the beginning) could end up forcing a corresponding tax law change to exempt foreign earnings from U.S. tax, a result that might not be a possibility if APB 23 had not been adopted or the Opinion had been implemented more rigorously.
In addition to APB 23’s effects on tax policy, we note three non-exclusive accounting concerns, none of which was addressed in detail when APB 23 was approved 43 years ago or since:
1. As the “Quad B” dissenters to APB 23’s adoption warned in 1972, the suppression of DTLs under APB 23 may misinform investors looking at book income by mixing restricted earnings (i.e., income for which IRE designations are made) with unrestricted earnings.
2. APB 23 requires management to assert the unknowable in order to achieve the book income advantage of suppressing a DTL. Companies cannot reliably assert, whether as a probability or a possibility, that certain income will not be repatriated in the next 20 or 30 years (which is how “indefinitely” needs to be defined for accounting consistency). Changes in management, business circumstances, and shareholder needs make such assertions impractical (as demonstrated by recent big and small reversals of IRE designations by companies including Avon, eBay, Pfizer, and General Electric). Prudent accounting requires use of the DTL, which is ideal for accommodating long-term book/tax differences, to remind investors of the inevitable cost of repatriation.
3. The inconsistency of ABP 23 with Financial Accounting Standards No. 52 further muddles book income reporting and creates inequities across companies. FAS 52, which requires currency translation of certain foreign-denominated items for book reporting, does not permit the kind of broad company discretion to suppress a bad book result that is allowed by APB 23 (which can enhance book income by suppressing DTLs) even though there are similarities in company decision-making for repatriation and currency conversion.
Because of the interaction between accounting rules and tax law, we believe both would be served by repealing APB 23, and we would be happy to discuss ideas for transition that would minimize disruption and complexity. At the very least, FASB would well serve the public and itself by addressing tax and accounting controversy surrounding the Opinion.
Revenue Estimator, Federal Government (retired)
J. Clifton Fleming, Jr.
Ernest L. Wilkinson Professor of Law
J. Reuben Clark School of Law
Brigham Young University
Jeffery M. Kadet
CPA (retired) and Adjunct Lecturer
University of Washington School of Law
Edward D. Kleinbard
Johnson Professor of Law and Business
University of Southern California Gould School of Law
David L. Koontz
Robert J. Peroni
Fondren Foundation Centennial Chair for Faculty Excellence and Professor of Law
University of Texas School of Law
Daniel N. Shaviro
Wayne Perry Professor of Taxation
New York University School of Law
Stephen E. Shay
Harvard Law School
 FASB’s decisions on February 11, 2015, to require disclosures of pre-tax earnings sources and further information about tax expense, APB 23 reversals, and IRE designation amounts for certain countries are helpful but in our opinion do not address fundamental issues.
 With over $2 trillion of IRE designations, roughly $500 billion of DTLs have been suppressed just in the last decade under APB 23 since the 2004 repatriation holiday. These numbers are so large relative to other financial statement entries that it would not take much in the way of reversals to cause noticeable effects. The DTL suppressions under APB 23 by many U.S. multinationals exceed their existing DTLs, deferred tax assets, and approach the magnitudes of inventories and accounts receivable entries. For example, in 2014 Apple’s own estimate of $23.3 billion of suppressed DTLs (associated with IREs of $69.7 billion) exceeds its $6.5 billion of DTAs, $20.6 billion of net property, plant, and equipment, $17.5 billion of accounts receivable, and approaches its $29.0 billion of long-term debt. While Apple’s ratio of APB-23-suppressed DTLs to total assets may be relatively large at 10 percent ($23.3/$231.8) compared to other companies, a perusal of companies (e.g., General Electric) suggests that ratios of about 5 percent are routine.
 While 14 Members of the APB viewed the Opinion as an improvement in accounting accuracy, the “Quad B” dissenters to APB 23, Messrs. Bevis, Bows, Broeker, and Burger, cited noncomparability in noting that “(APB 23) validates a practice … completely contrary to the underlying concepts of deferred tax accounting … by sponsoring the idea that certain earnings may be accounted for on an accrual basis while the related income taxes are accounted for on the cash basis” (APB 23: Accounting for Income Taxes – Special Areas, April 1972, section 33, p. 6). Also, the cash treatment of taxes under APB 23 is optional, so a company has total accounting control (i.e., the choice between cash and accrual) of future U.S. residual taxes.
 APB 23 requires “… evidence of specific plans for reinvestment … which demonstrate that remittance of the earnings will be postponed indefinitely” (ibid., section 12, p. 4). It might be reasonable for management in its guidance to say that company value will be enhanced if certain earnings remain unavailable to shareholders for a few years. However, the higher standard that should prevail for suppressing a DTL under APB 23 should be consistent with the maximum time arc of other DTLs such as those arising from depreciation, because for investors looking at above-the-tax-footnote financial statements DTLs are effectively homogeneous. If a company believes it might repatriate in year 19 but under its interpretation of indefinitely for APB 23 only looked 5 or 10 years out and therefore suppressed the potential DTL associated with an IRE, and yet the same company or a competitor is carrying DTLs for depreciation (or, say, pensions) that will not expire for 20 years, that is inconsistent and confusing to investors trying to gauge earnings quality. Also, many companies have added to the distortion by asserting that certain earnings are “permanently” reinvested overseas – this term is not found in APB 23, and its use raises even a more fundamental question of how current company management could ever assert such permanence.
 Once it is recognized that indefinitely needs to cover at least 20 years, it would be difficult for any company to make an IRE designation because current management cannot control circumstances or future management’s actions. Another concern is that the current flexibility that company managements have under APB 23 creates a conflict of interest. This is because the prevalence of equity-based compensation encourages a company’s management to lower tax expense so as to increase after-tax earnings and share price. Also, were this attestation made transparent, it could be Pyrrhic for whomever makes it because if U.S. management in 2015 openly stated that over $2 trillion of foreign earnings would be unavailable indefinitely (which should be defined as a minimum of 20 years, as DTLs used for depreciation can last 20 years or more, ditto for DTAs) to shareholders there likely would be a revolt that would install new management. As another example of a test that is not applied under APB 23, the company should be foresighted about interest rates and how they affect the hurdle rate with respect to repatriation, because the correlation between the secular decline in interest rates and the recent huge IRE buildup is not coincidental. Are low interest rates to be expected for the next 20 years, and if not, how would this affect the company IRE decision?
If certain earnings are indefinitely unavailable to shareholders because of an IRE designation, it might be asked whether the designated earnings should be recorded as unrestricted book income when earned overseas in the first place because of the company’s self-imposed mobility restriction. From an investor’s perspective, APB 23 would be more internally consistent and prudent if instead of suppressing the DTL and thereby mixing inferior restricted earnings with other types of earnings, the ruling required a special designation of overseas earnings not intended for repatriation with the main financial statements excluding (or footnoting) such earnings.
 FAS 52 permits some flexibility in presentation of adverse results, but it does not allow a company to ignore currency translation just by promising that it would not convert currency under unfavorable circumstances (i.e., what APB 23 allows). This inconsistency can lead to the odd result that some companies are badly hurt by hypothetical currency conversion while other companies are helped by APB 23 designation and hypothetical nonpayment of U.S. residual tax. The decision about when to convert foreign-denominated earnings into U.S. dollars seems just as discretionary for companies as the timing of repatriation; earnings currency conversion is also a step in repatriation.