Valuation of Financial Assets and Income Tax Base

The target base of the corporate income tax is the net income earned by the taxpayer during some conventionally defined time period—typically one year. The determination of net income involves matching expenses to the income to which they relate. That is, the taxable base is determined by subtracting from gross income those costs incurred in generating the income. Thus, loan losses must be deducted from the tax base since they are a necessary business cost of engaging in lending activities.

Gross income and its associated expenses, however, may be defined in several ways with potentially important implications in terms of tax revenue and economic incentives.¹¹⁸ Specifically, income and expenses may be recognized upon accrual or upon realization, affecting the timing of tax liabilities, Under a definition based on realization, transactions and other relevant operations are dated and valued according to the timing and amount of actual payments in cash or its equivalents. In contrast, under an accrual definition of income, relevant events have income repercussions whenever a claim or liability arises, or when their value changes.¹¹⁹

The definition of income based on accrual matches more closely the timing and magnitude of the underlying economic event—namely, the change in the taxpayer’s net worth according to a market valuation. Thus, when measuring accrued income, changes in the value of claims and liabilities entail current-year income repercussions even if their realization—that is, their materialization into actual cash flows—may occur beyond the end of the current year. Similarly, changes in the value of claims and liabilities are taken into current income as they accrue although their ultimate realization value may be subject to uncertainty.

The accrued income from an asset (or group of assets) equals, by definition, the net cash flow derived from it plus the change in the asset’s value. In the case of a portfolio of loans, for example, the associated cash flow comprises interest and repayment of principal, less new loan issues; the change in the value of the portfolio encompasses, inter alia, loan losses prompted by deterioration in the collectibility prospects of the loans in the portfolio. The cash flow from a loan portfolio, as well as from most assets, can generally be easily assessed.¹²⁰ In contrast, however, a precise measure of the change in value of some assets, such as loans and other financial claims, may not be readily available. Hence, the determination of the corporate income tax base and, consequently, of tax liability, requires the specification of a methodology for the valuation of these assets.

The accrual method of accounting attempts to approximate a measure of income based on economic, or accrued, income. Standard accrual-based accounting methods, however, incorporate in fact a combination of accrual- and realization-based valuation methods. The deterioration in the value of a portfolio of loans prompted by loan losses is generally recognized upon accrual—that is, when the loan losses are identified; in contrast, similar changes in the value of the portfolio caused by interest rate swings or other capital gains or losses are often recognized only upon realization. Compared with cash-based accounting, the accrual method offers less scope for tax planning and tax avoidance through intertemporal tax arbitrage—for example, by delaying payment of liabilities. In most countries, corporate income taxes are levied on a measure of income that is determined, to the extent possible, on an accrual basis. Pure cash accounting is generally either disallowed or restricted to small taxpayers, owing to the simplicity of this accounting method. In particular, most banks and financial institutions are required to assess their taxable income under the accrual accounting method. Nevertheless, some countries allow cash-based accounting for some financial operations of small banks. For example, tax accounting on a cash basis is allowed for small banks in the United States (banks with average annual gross receipts of $5 million or less over a three-year period), Japan, and Canada. Still, recognition of loan losses when they are identified (rather than when they are realized) is often possible (or mandatory) for banks under cash-based accounting.

Measuring income on an accrual basis poses some problems specific to the financial sector. Accurately assessing the value of financial assets typically depends on the existence of well-developed markets in which those assets are traded.¹²¹ The value of an asset can be established at the time of its first acquisition—for loans, the initial value is given by the amount originally disbursed (usually equal to the face value of the loan)—or on the occasion of subsequent transactions. Moreover, even when an asset is not traded for a long period, but there nonetheless exists an active market in assets of that class, an asset’s value can be presumed to be the market value of similar assets. Thus, for example, frequently traded securities can be valued at market prices. For many assets, however, mark-to-market methods are likely to be impractical since their markets either are limited or do not exist. This is often the case with bank loans, given that in most countries, particularly developing countries, secondary markets for loans are either very thin or nonexistent. Other financial assets such as over-the-counter securities may also lack a market with sufficient depth to allow mark-to-market methods. That said, the lack of a precise asset valuation does not imply that investors cannot assess the value of a bank’s financial assets throughout their life until maturity; circumstances affecting realization values are reflected in the overall value of the bank (e.g., through share prices) and in its creditworthiness, depositor’s confidence, and so forth.

Against this background, if tax neutrality across assets and alternative investments is to be preserved, losses stemming from a deterioration in the value of a bank’s financial assets—and in particular of its loan portfolio—should be recognized in determining taxable income as these losses accrue rather than postpone their tax recognition until realization. In that manner, the tax base will encompass changes in the economic value of assets as they occur. A deficient coverage under the corporate income tax of changes in the value of a portfolio of financial assets would prompt tax-induced gains or losses across assets according to their different levels of risk, cash flow profiles, and so forth, arbitrarily distorting their market values. This distortion, in turn, might have potentially detrimental consequences on bank capitalization and hinder the efficient allocation of financial investments.

Accounting Aspects of Loan-Loss Provisioning

The two main financial statements for banks, as for most businesses, are the balance sheet and the profit-and-loss statement. The balance sheet is intended to reflect the value of the stock of assets, liabilities, and equity at a moment in time. The profit-and-loss statement reflects the flow of income and expenses during a specified accounting period as they arise from financial and real transactions and other relevant events. Thus, the result of the profit-and-loss statement must equal the change in assets less the change in liabilities and, correspondingly, the change in the value of equity before transactions with equity participants, such as equity contributions or profit distributions.

Many financial standards and regulations are based on historic cost accounting principles. With historic cost accounting, assets and liabilities are recorded at the actual value at which the original transactions took place. Strict historic cost accounting would require that the changes in value of balance sheet elements be reflected only upon realization. Nevertheless, in order to preserve the integrity and information content of the financial statements, accounting standards and regulations force or allow the recognition of selected gains and losses as they accrue, resulting in practice in a methodology that is somewhat of a hybrid between the two extremes of historic cost accounting and market value accounting. Thus, for example, most banking financial regulations require that unrealized losses on a bank’s trading portfolio be recognized through mark-to-market methods. In contrast, the recognition of unrealized gains or losses on a bank’s investment portfolio of securities may, under some conditions, be deferred until realization.

Changes in the notional or actual market value of a bank’s loan portfolio are, likewise, only partially reflected in the financial statements. Thus, for instance, value changes arising from movements in the market discount rate are rarely reflected in the balance sheet. In the case of credit (default) risk, regulatory financial standards provide different means of recognizing loan losses. Specifically, there are two main accounting methods of recording these loan losses: loan write-offs and creation of provisions for loan losses (or bad and doubtful debts provisions).

A loan is written off when the claim is redeemed through repayment or when it is deemed irrecoverable and therefore worthless. Some countries (e.g., Australia and the United States) also allow partial write-offs when only part of the debt is considered uncollectible. Under the provisioning method, receivables are recorded at their full face value until they are written off. However, a provision for loan losses is set up as an allowance against the eventuality that some of the receivables may prove to be uncollectible. This provision either is shown separately as a liability in the bank’s balance sheet or is netted off against the impaired asset to which it relates. Generally, allocations to the provision for bad and doubtful debts are recorded for financial purposes as charges against the profit-and-loss account and, consequently, they reduce financial income. When a loan is considered worthless, it is written off against profit and loss or instead, if a provision for loan losses had previously been made, it is charged against the provision without direct income consequences. Conversely, in the event of recovery of a loan already written off or provisioned, a corresponding income item is recognized by the amount previously written off and subsequently recovered or by the amount of the provision that needs to be reversed.

There are two basic types of loan-loss provisions, namely, specific and general provisions.

(1) Specific provisions are linked to particular loans and are assessed individually after distinct circumstances impairing the loan collectibility have been identified. In some countries, however, specific provisions against loan pools are also established; in that case, the loans in the pool are required to show some indication of uncollectibility (e.g., payment arrears). These pools may be defined according to various criteria including country or industry groups, length of delinquency period, and so forth. Specific provisions are often netted off against the face value of the related loan or loan pool. They are not generally counted as regulatory capital for capital adequacy purposes.

(2) Genera] provisions are not linked to any particular loan and constitute an allowance against unidentified loss potential on loans that have not been specifically provisioned. Countries differ in the accounting and regulatory treatment of general provisions. They can be netted off against gross assets or shown as a liability item in the balance sheet. Some countries’ regulations consider general provisions for loan losses as part of the regulatory capital base toward meeting capital adequacy requirements.¹²² General provisions can be classified according to their loan base and according to the method employed in determining their level. Their base can be the whole loan portfolio—although some countries exclude loans to the central government and other “safe” loans—or specific categories of eligible loans. The methods employed to ascertain the adequate level of a general provision range from a fixed percentage of eligible loans, to a moving average of the percentages of bad loans to total loans within the group of eligible loans based on a bank’s past experience (“experience method”), or to more judgmental approaches based on loan analysis. Additionally, some countries avoid the adoption of a fixed method and leave the choice to the discretion of the bank (e.g., Switzerland and Austria).

The dividing line between general and specific provisions for loan losses—and occasionally between provisions and write-offs—although clear from an analytical viewpoint, is sometimes fuzzy at the operational level. When reviewing banking practices, it is easy to find many instances where the distinctions are purely conventional. In particular, specific provisions relating to some loans or loan pools are often indistinguishable from general provisions. For example, most banks determine the level of their specific provisions on a pool of consumer loans with small individual face value on the basis of statistical methods or other predetermined formulas without resorting to an individualized examination of loans. Also, some banks would create specific provisions on most loans to an industry if the industry is undergoing unusually severe economic difficulties and there has been a surge in defaults by borrowers from that industry. Likewise, the deterioration of general economic or political conditions in an emerging economy may prompt international lenders to create specific provisions against all loans to debtors from that country. In turn, partial write-offs may be considered similar to specific provisions in economic implications if not in their accounting.

Existing Methods of Tax Deduction of Loan Losses

Existing methods of tax deduction for loan losses pertain to two broad categories: (1) the charge-off method requires that a debt be uncollectible and written off at the time the tax deduction is claimed; and (2) the provisions (or reserve) method allows the tax deductibility of eligible provisions for bad and doubtful debts without requiring a book write-off of the underlying asset. Paralleling accounting practice, tax-deductible provisions may be general (general provisions method) or specific (specific provisions method). A description of the tax deduction methods used in a sample of countries for which detailed information was available can be found in the appendix to this chapter.

The charge-off method is employed in Australia and the United States.¹²³ Under the charge-off method, an expense is recognized for bad debts only as they actually become either wholly or partly worthless and are written off the books (both Australia and the United States allow partial write-offs). When a receivable is determined to be irrecoverable in whole or in part, its value is reduced by the amount deemed irrecoverable and an expense is recognized for tax purposes in an equal amount. The tax deduction must be claimed in the year in which the debt loses its value and is written off. Thus, evidence must exist that the debt had some value at the beginning of the period and that some event occurred during the year that prompted the loss. In determining worthlessness of a debt, all pertinent evidence, including nonperformance, adequacy of collateral, and debtor’s financial condition, may be considered. The debt, however, must not be merely doubtful or otherwise substandard; rather, it must be considered uncollectible. In the United States, debt worthlessness can be presumed for tax purposes in the case of write-offs pursuant to a regulatory order.

The provisions method may take several forms according to the type of provisions that qualify for the tax deduction and the degree of conformity between allowed tax provisions and regulatory provisions and reserves. Under the specific provisions method, additions to regulatory specific provisions are often tax deductible. Nevertheless, in some countries the tax authority chooses to modify the timing or level of the specific provisions that are allowed for tax purposes. For example, in the United Kingdom, the Inland Revenue matrix employed to compute tax-deductible specific provisions against loans to selected developing countries is different from the corresponding matrix of the Bank of England used for regulatory purposes. Other countries, while taking regulatory specific provisions as a basis for the tax deduction, set quantitative caps (e.g., in Canada, country-risk provisions for loans to developing countries are limited to 45 percent of the principal) or additional qualitative requirements.

Under the general provisions method (also called the “general reserve method”) the tax deduction is computed as the necessary addition to the general provision to bring its opening balance, adjusted for write-offs that occurred during the period, to the allowed closing balance. As mentioned above, the allowed closing balance is usually determined as a percentage of eligible loans. This percentage may be a fixed coefficient across taxpayers or may be computed on the basis of the taxpayers’ past experience. It should be noted that the prescription of a maximum closing balance of the provision does not imply any constraint on the amount of bad debt identified and written off during the period that can be charged against the taxable base. This is because when bad debt is written off, it is charged against the provision, reducing its balance. This, in turn, increases the tax-deductible addition necessary to bring the closing balance of the provision to any given level.

By and large, specific provisions appear to be the most common form of determining tax deductions for loan losses. However, in France, Italy, Japan, Luxembourg, Switzerland, and Spain, various forms of general provisions are allowed for tax purposes. Nonetheless, in most of these countries, the tax authority sets restrictions on the deductibility of general provisions, often establishing a ceiling as a percentage of relevant assets (e.g., Italy and Japan) or explicitly disallowing some general provisions (e.g., France and Spain). Although no single method of determining the tax deduction for loan losses has been uniformly adopted internationally, a tendency can be observed to constrain or disallow for tax purposes precautionary general provisions for future loan losses and to deny tax deductions for hidden reserves¹²⁴ in those countries where these reserves exist.

Comparison of Tax Deduction Methods

The guiding principle regarding the tax treatment of loan losses should be that the allowed tax deduction matches as closely as possible the accrued deterioration in the economic value of a bank’s portfolio of loans. In practice, however, the precise timing and extent of the economic loss is difficult to ascertain. An exact measure of the economic value of a loan portfolio would need to be based on the overall stream of future payments (including principal and interest) that can be expected, given the relevant information available at the moment of the valuation—that is, the expected present value of the portfolio conditioned on the existing information. Nevertheless, for practical purposes, only the deterioration in the expected realization value of the principal of the loans in the portfolio may be subject to a reasonable standardized assessment.¹²⁵ The economic impact of different methods of tax deductibility of loan losses hinges on the timing of recognition of this deterioration in the collectibility of the principal.

In this light, the general provisions method typically embodies an anticipated deduction. Under the general provisions method, the allowed balance of the tax-deductible provision is computed as a percentage of outstanding loans at the time of the closing balance. Therefore, the issuance of new loans during the year will automatically generate a tax deduction (equal to the given percentage of the principal) even when no new information has become available since the loan was disbursed that might indicate a deterioration in the value of the loan. Thus, a deduction contemporary to the issuance of the loan is taken in anticipation of losses that are forecast to occur at some future moment during the life of the loan. In fact, the aim of according an anticipated deduction is often explicitly mentioned in the tax law.

The importance of the tax subsidy embodied in the general provisions method depends, inter alia, on the statutory coefficient used to compute the maximum balance of the provision as a proportion of eligible loans. If this coefficient is a low percentage, the bulk of the tax deduction will need to be taken when loans are written off. In fact, the lower the coefficient, the more similar a general provision becomes to the charge-off method. Consequently, overtaxation may also occur under the general provisions method if the tax norms for loan write-offs are excessively restrictive. Under this method, notwithstanding the initial deduction of a fraction of the value of the loan, the write-off and corresponding charge of the full face value against the provision may need to be postponed until well after the loan has lost its economic value. Restrictions on writing off loans also diminish the tax deduction via a decline in the coefficient used to compute the allowed level of the provision because this coefficient is often computed on the basis of past write-offs. Overtaxation will occur whenever the loss associated with delays in obtaining the tax deduction of the nonprovisioned fraction of the principal exceeds the tax subsidy implied by the initial anticipated deduction.

If a general provision is determined on the basis of a fixed coefficient across asset categories with different levels of intrinsic risk, those assets with the lower risk will be favored over assets with higher risk. For example, when the coefficient is the same for all taxpayers independent of their loan-loss history and of the composition of their portfolio, low-risk portfolios will enjoy the same tax deduction as other higher-risk portfolios while experiencing lower loan losses.

Under the charge-off and specific provisions methods, the tax deduction is triggered respectively by loan write-offs and by credits to the provision account. Therefore, whether over- or undertaxation occurs hinges on the conditions under which the taxpayer may take those actions. When the specific provisions method is used, once a loan has been provisioned in full and the corresponding tax deduction has been taken, the timing of write-off is irrelevant from a tax standpoint. On the other hand, if the provisioning rules are restrictive, the taxpayer may be forced in many cases to write off a loan in order to obtain the tax deduction. In fact, as the provisioning rules are more restrictive, the specific provisions method approaches the charge-off method for all practical purposes. Conversely, if the charge-off method allows for partial write-offs, the difference between the charge-off and the specific provisions method may become reduced in some cases to accounting methodology.

Both the charge-off and specific provisions methods can suitably match the tax deduction to the occurrence of the economic loss since both have the potential to link the timing of the tax deduction to identified events impairing the value of specific loans. Specific provisions, however, may offer more flexibility for incorporating pertinent business practices and advances in accounting techniques regarding loan valuation. In contrast, loan write-offs may be subject to constraints beyond the purview of tax and financial policies. A commonly encountered cause of excessive restrictions on write-offs is the requirement that all legal means of collection and, if applicable, all legal actions to execute the collateral be exhausted before a loan is written off. This may be disproportionately onerous if the judicial system is too slow, inefficient, or overburdened. Also, banks may be reluctant to write off loans if by doing so the lender relinquishes all or part of the legal claim against the borrower or if the bank feels that a high level of write-offs could damage its public image or the confidence of depositors. In practice, countries that adopted the charge-off method are generally considered more restrictive in allowing tax deductions for loan losses.

Tax Treatment of Loan Losses and Incentives to Bear Risk

The tax treatment of loan losses determines the manner in which default costs are deducted from the tax base. A nonneutral treatment of loan losses under the corporate income tax relative to other business costs will either enhance or erode the aftertax return of a portfolio with risk of loan losses vis-à-vis alternative investments.

A loan is a contract stipulating a flow of payments to be made by the debtor to the lender. This flow of payments (the service of the debt) includes payment of interest and repayment of principal. The value of a loan for a financial institution at any point in time is generally computed as the expected value of the future cash flows discounted at the prevailing (or expected) interest rate on alternative riskless assets—such as, for example, treasury securities. The expected value of future cash flows depends on the loan interest rate and the probability of borrower default. If the probability of borrower default is perceived by the lender to be high, the loan interest rate will be increased to compensate for the higher risk. The difference between the loan interest rate and the corresponding interest rate on a riskless contract with the same time profile of principal amortization is called the risk premium.

The risk premium compensates the lender for possible defaults—-at least a priori, and possibly also a fortiori, on average across loans and over time. That is, the risk premium is the price paid to lenders in order to induce them to assume risk. Market forces tend to ensure that the overall level of risk taken up by the economy as a whole strikes a balance between the income-risk preferences of lenders and the return-risk profile of investments that are effectively undertaken. A desire by the lenders as a whole for higher levels of risk (at the initially prevailing risk premium) would increase the supply of loanable funds for risky investments, driving down the risk premium charged on loans and encouraging borrowers to undertake riskier investments. Conversely, a higher aversion to risk among lenders would prompt the opposite result: a higher risk premium would prevent the undertaking of riskier investment projects that, at the increased interest rates, would become unprofitable.

The mechanism of tax deduction of loan losses may alter the risk premium via changes in the relative prices of assets with different levels of risk, thereby distorting the incentives to undertake risky investments. A distortion of the relative price of safe assets versus risk-bearing assets can be expected because, on the one hand, the tax treatment of loan losses will not alter the price of safe assets that, by definition, are not susceptible to default. On the other hand, an ill-designed tax treatment of loan losses will distort the price of risky investments.

The relation between the timing of the tax deduction for loan losses and the timing of the associated economic loss is perhaps the most relevant characteristic in assessing possible biases introduced by the tax treatment of bad debts. Indeed, it can be shown that the value of a loan portfolio will not be altered by the income tax (or by its rate) if, and only if, a deduction for loan losses is allowed at the time when the actual impairment in economic value takes place—that is, when the loss accrues.¹²⁶ Otherwise, if the tax deduction is advanced or deferred relative to the accrual of the loss, the value of a loan portfolio will be respectively enhanced or impaired.

In one case, when the tax deduction is given before the occurrence of the associated economic loss, an interest-free loan from the treasury to the bank effectively takes place. Thus, an accelerated tax deduction on future loan losses increases the return on risky investments, creating an incentive to substitute risky loans for safer investments within banks’ portfolios. Correspondingly, the risk premium will be driven down until it reaches a new equilibrium level at which banks become again indifferent between assets with different levels of risk. In the end, more financial resources are likely to be devoted to risky lending than in the absence of the tax incentive. This economic logic runs contrary to some arguments that advocate early tax recognition of loan losses and full tax deductibility of precautionary reserves on the grounds that such practices would lead to a sounder and safer financial system by encouraging reserve formation.

Conversely, in the opposite case, when the tax deduction is unduly deferred relative to the actual economic loss, the size of the eventual loss incurred upon borrower default will be enlarged. Thus, lenders will increase the risk premium in order to cover the additional loss caused by overtaxation. In this case, resources will be driven away from risky lending and toward safer investments—where the tax-induced cost is lower. In either case, a nonneutral tax deduction for loan losses can be expected to introduce a wedge between the social preference for risk and the risk-composition of investments actually undertaken.

Tax Treatment of Unpaid Interest

Unpaid interest on debt constitutes a source of financial losses closely related to loan losses. Uncollectibility of a loan is most often indicated by the debtor’s failure to service the debt. In fact, unpaid capitalized interest may greatly exceed the value of the principal. This is particularly likely during episodes of unusually high nominal interest rates such as those associated with high inflation, adjustment policies, external imbalances, or high country risk.

Most tax systems determine the taxable income of banks and financial institutions mainly on an accrual basis. The accrual method of accounting demands that income and expenditure items be recognized as the claim or liability arises rather than when it is settled. Consequently, interest on loans and other investments becomes taxable, in principle, as it accrues rather than when payment is actually received. Thus, in general, the fact chat a loan has incurred arrears in payment of interest does not necessarily email per se a deduction of the unpaid interest from the taxable base. The rationale for including accrued unpaid interest in taxable income is that the lender acquires upon accrual a claim on the borrower similar in nature and legal status to the claim on the principal. The value of the claim adds to the bank’s assets and hence to its income.

It is apparent, however, that the claim on the unpaid interest can often be presumed to have a value lower than the full interest due or even nil. As in the case of the principal of a nonperforming loan, the valuation of the claim on the unpaid interest hinges on an assessment as to its collectibility. When circumstances indicate that the accrual of unpaid interest no longer reflects a genuine increase in the lender’s assets, the loan is reclassified to a nonaccrual status for tax purposes and the interest ceases to be recognized as taxable income. The reclassification typically occurs after interest has not been paid for 30, 60, 90, or 180 days.

Additionally, when a loan is placed on a nonaccrual status, the value of previously accrued unpaid interest may be reassessed and, if warranted, a tax deduction taken commensurate to the estimated loss. For this purpose, the lender’s claim related to the unpaid interest is often considered of a lower quality than the claim on the principal—particularly in the case of collateralized loans if the value of the pledged collateral does not cover capitalized interest. Thus, the tax deduction of previously taxed unpaid interest and the reclassification of the loan into a nonaccrual status are often accorded before the tax deduction of the principal (e.g., before the full provision of the principal is accepted for tax purposes).

Loan Rescheduling

The rescheduling of loans is not necessarily associated with an impairment in the value of the debt. In countries characterized by volatile financial conditions, banks may choose to extend most loans as short-term debt as a means to maintain the liquidity of their portfolio.¹²⁷ Regardless of their short maturity, many of these loans may be routinely used by borrowers to finance what is essentially medium-term investment or working capital. Thus, in regular times, refinancing of loans may constitute a common business practice and even a sign of good loan performance. During times of financial distress, however, troubled institutions may resort to loan refinancing and rescheduling as a means to avoid the recognition of a sudden deterioration in the value of their portfolio.

In general, for tax purposes, the rollover of an existing loan is equivalent to the recovery of the original loan—and associated capitalized interest, if any—and subsequent issuance of a new loan. Nevertheless, for financial purposes, the regulatory authority may preclude such treatment in order to avoid overstatement of income by troubled institutions via an overstatement of the quality of their portfolio. In that case, so as not to tax fictitious income, the tax system should not request the reversal of previously made provisions inasmuch as these provisions are also kept for financial purposes.

Mergers, Acquisitions, and Loan Sales

Mergers, acquisitions, and sales of impaired loans are often an essential part of the restructuring and consolidation of a troubled financial sector. On these occasions, the tax value of the transferred assets tends to be significantly higher than the value actually realized in the transaction. Specifically, loans may have been insufficiently provisioned prior to the transaction in an attempt to postpone the recognition of insolvency or as a result of over-restrictive tax norms.

Generally, any differences arising upon sale between the book and realization value of assets are incorporated in the tax base at the time of the transaction. Specifically, the sale of a loan portfolio below its book value implies the realization of a loss by the seller. On the other side of the transaction, the buyer restructures the composition of its assets without tax consequences—for example, by swapping cash for loans with a net value (after provisions, if applicable) equal to their purchasing price.

This type of operation may result in revenue losses to the treasury because the seller is allowed to recognize the loss, upon realization, and is entitled to the associated tax deduction. This loss and the associated tax deduction, however, should be considered legitimate. They are the result of reconciling an overstated—and thus, overtaxed—book value with the actual economic value of the assets as priced by the market. This market-driven treatment of loan portfolio transactions becomes even more crucial during episodes of financial distress, when reluctance to recognize the actual value of the assets for tax purposes may hamper a necessary consolidation of the banking sector—which, in turn, may prompt potentially larger budgetary consequences in the future.

A similar situation may arise on the occasion of a bank merger if the new company is not allowed to recognize for tax purposes the true value of the loan portfolio. The valuation of assets, however, presents more difficulties in the case of a merger since no assets are sold as such and, consequently, a directly observable market price for the loan portfolio is not available.

In either case, the tax treatment of these operations should not stand in the way of banking sector consolidation by, for example, according a more detrimental treatment to banks than is generally available to other taxpayers. Tax credits accumulated by the failing bank (e.g., losses carried forward) can be allowed to be transferred to the acquiring or newly formed bank to the extent that this treatment does not excessively complicate the administration of the tax or introduce significant inefficiencies. For example, inefficiencies would arise if the acquiring bank values the tax credits substantially below their nominal value owing to incomplete financial markets or uncertainties regarding their future applicability. In that case, the actual incentive provided by the tax treatment would be substantially below the cost of the incentive to the treasury.

International Practice

Banks and other financial institutions are generally subject to tighter financial regulations and supervision than other enterprises. Indeed, bank regulations are often more precise than general tax regulations, and the banking supervisory body can usually mobilize substantially greater administrative and staff resources than tax administrations can typically devote to the banking sector. Partly as a result, the regulatory treatment of banks has influenced—and often determined—their tax treatment.

Most countries take regulatory accounting as a reference point to determine taxable income and introduce adjustments as needed. The departure from regulatory income is most prominent in countries that follow the charge-off method (e.g., Australia and the United States) but substantial differences also exist in other countries. A tendency can be observed to disallow for tax purposes some general reserves for unspecific banking risks (often including general provisions for loan losses) and hidden reserves.

Conformity between tax and book accounting appears higher in Denmark, France, Germany, Luxembourg, and Switzerland where virtually all provisions (general and specific) are tax deductible. Nonetheless, this conformity is typically less than complete: in France, Luxembourg, and Germany, hidden reserves and the general provision for unallocated banking risks are not allowed for tax purposes. It is normal practice in Germany to release the excess portion of provisions not recognized for tax purposes.

In other countries, such as Italy, Japan, and Spain, the tax authority disallows selected provisions or sets caps on their tax deductibility. Thus, in Italy, a bank may deduct the increase in provisions up to an annual maximum of Vi of 1 percent of total receivables or until the balance of provisions reaches 5 percent of receivables; in Japan, the provision for sovereign debt is limited to 1 percent of loans; and in Spain, general provisions for mortgages and some specific provisions are not tax deductible.

The departure of tax norms from regulatory accounting seems more pronounced in Canada and the United Kingdom, where general provisions are disallowed and specific provisions and write-offs are determined with substantial independence from regulatory accounting. Finally, some countries choose to define book and tax accounting independently of one another. This is typically the case of countries that follow the charge-off method such as Australia and the United States, where financial provisions are disallowed for tax purposes.

Analysis of the Issues Involved

Since financial and tax accounting share at least one of their primary objectives, namely, the determination of income, it appears logical that, to the extent possible, regulatory financial accounting serve as a basis for the assessment of taxable income. First, this approach presents the advantage of minimizing administration and compliance costs for both the taxpayer and the tax administration. Second, it can be argued that the incentive to underreport taxable income provides a counterbalance to the incentives to overreport financial income (or underreport financial losses) to the banking supervisory authority; thereby, providing a self-enforcing mechanism to both the regulatory and tax authorities.¹²⁸

The existence of well-designed banking regulations complemented by reliable supervision and monitoring constitutes an important instrument in implementing an efficient corporate income taxation of the banking sector. The tax authority can maximize this advantage by designing the tax legislation in a manner that allows the use of accounting categories and information available from the banking supervisory authority. To minimize administrative costs, departures of the tax accounting requirements from regulatory accounting need to be easy to comply with and monitor, and ideally they should be confined to, at most, a few areas without requiring a second set of accounts for tax purposes. Also, tax audits can regularly rely on information, data, and financial assessments provided by the supervisory authority.

Nevertheless, conformity of tax and regulatory provisions cannot be an overriding objective and should be subject to the consideration of specific circumstances. Notwithstanding their common objective of measuring income, the characteristics of regulatory and tax accounting and auditing are not necessarily identical. Differences in the regulatory and tax treatment of banks’ income stem from, at least, two factors: (1) the regulatory measure of income may be biased toward underestimating economic income as a means to buttress the solvency of the banking system; and (2) in the presence of measurement errors and incentives for noncompliance, the regulatory and tax authorities may design differently their corresponding norms attempting to minimize opportunities for over- and understating income, respectively.

(1) Although the supervisory authority is concerned with the correct measurement of economic income, prudential objectives tend to introduce a conservative bent. Owing to a variety of reasons, prudential objectives may prevail in practice over an accurate measurement of economic income when designing some features of the financial regulatory framework. As a result, some aspects of the regulatory measurement of financial income may be geared toward understating income.

Prudential regulations aim, inter alia, to forestall systemic financial crises and to limit the risk borne by depositors and other lenders, minimizing the possibility of bank insolvencies. The primary means to address prudential objectives is through capital adequacy requirements. This is usually implemented by setting minimum capital requirements as a proportion of risk-weighted assets and exposures and by mandatory reserves.¹²⁹

Existing standards on capital requirements may nevertheless be deemed insufficient since they focus mainly on assurances against intrinsic credit risk—the risk that a loan may become uncollectible as a result of circumstances specific to that particular loan. The experience of recent banking failures and banking crises in a number of countries tends to indicate that intrinsic credit risk is but one of the possible causes of banking sector distress. Other factors, such as off-balance sheet risks; volatility of capital flows, exchange and interest rates; macroeconomic and external imbalances; deficient market risk management, and so forth, may have played, at least, an equally relevant role in the development of bank insolvencies.¹³⁰ Proposals to extend capital requirements to encompass some forms of market risk have only recently been advanced—for example, the 1995 amendment to the Basle Accord (Basle Committee on Banking Supervision, 1995) or the EU Capital Adequacy Directive (Council of the European Communities, 1993)—and their implementation is still limited. The definition of strategies to deal internationally with other types of risk, such as settlement risk, are at present in an incipient stage.

Countries with diverse institutional, legal, and business traditions, and at different levels of economic and financial development adopt regulatory and supervisory frameworks that reflect their different circumstances. Some countries may not have fully adopted international standards of capital adequacy or may deem them insufficient. Thus, they may consider it necessary to complement their capital requirements with additional mandatory reserves or strengthened provisioning requirements in order to buttress the soundness of their banking sectors, even at the expense of biasing financial accounting practices toward an understatement of income. Furthermore, it is often politically more palatable to require higher provisions against particular balance sheet items or general reserves than to regulate an equivalent increase in the capital adequacy coefficient. This is particularly so in developing countries where requiring a capital adequacy ratio above Basle standards may be politically problematic while, in fact, systemic risks or the probability of correlated debtor defaults may very well be higher than in the industrial countries for which the Basle standards were designed.

(2) The measurement of income for either financial or tax purposes is subject to potentially large errors. The sources of error are both the intrinsic difficulty in valuing risky assets with uncertain realization value and the existence of incentives and opportunities for misrepresentation of income. Tax and regulatory authorities need to rely on a regular basis on information provided by banks, which may be presumed to have incentives to conceal the true measure of their income. Thus, regulatory and tax norms need to be designed to minimize the opportunities for over- or understating income, as the case may be.

It is often difficult to discern when a provision that reduces regulatory income is made in recognition of an accrued loss or as a form of precautionary reserve-building in anticipation of eventual future losses.¹³¹ For example, this distinction may be buried in the classification of a loan or in the valuation of its collateral. The supervisory authority will tend to design financial regulations with a view to limit more forcefully the probability of overstating income. Thus, it will usually allow to some extent voluntary overprovisioning and will set provisioning requirements mainly in the form of floors. On the other hand, the tax authority will attempt to limit overprovisioning and will tend to set ceilings for provisioning levels. This situation is similar to that of two statisticians trying to design tests for the same hypothesis but aiming to minimize respectively type I and type II errors: it may not be possible to reduce simultaneously both under-and overestimation of income without devoting substantial additional resources—which could be neither feasible nor desirable. Instead, the design of the regulations and the enforcement and supervisory efforts may need to target the main concerns of the tax and financial regulatory authorities, respectively. This is often the origin of curtailment or outright disallowance for tax purposes of some of the most open-ended financial deductions from income.

As argued above, it is often not possible, without an in-depth inspection, to separate what constitute provisions made according to minimum regulatory standards and what constitute essentially voluntary provisions. As a consequence, it is common that the tax administration be allowed to conduct its own audits and to monitor banks’ compliance independently of the supervisory authority. These tax audits may result in adjustments to a bank’s taxable income that may or may not prompt parallel adjustments in regulatory income, adding another potential source of discrepancy between regulatory and taxable income.

As a consequence of the prudential bias of financial regulations, financial income may be reduced by the formation of mandatory or voluntary provisions aiming to bolster own resources in the bank’s balance sheet. To the extent that these provisions do not reflect an impairment of the bank’s assets value, they cause reported financial income to underestimate economic income. Provisions of this nature are, even when mandatory, a form of reinvestment of profits and increase the bank’s net worth—therefore they should be considered taxable income. Examples of these types of provisions are provisions or reserves for unspecific general risks; provisions or reserves for eventual future losses on foreign exchange holdings or on the proprietary trading portfolio of securities; general provisions for loan losses that can be constituted without any specific indication of loan impairment; and specific provisions against impaired loans above and beyond a reasonable estimate of the occurred deterioration in the value of the loan.

In contrast, those provisions aiming to compensate for losses already occurred—although possibly not yet realized—represent a genuine economic loss and should be deductible for tax purposes. Since many assets and liabilities are recorded at historic value until realization or their value is priced to market only occasionally, the accuracy of the balance sheet requires the creation of these provisions to capture variations in the value of balance sheet items before their final realization. Examples of these provisions are depreciation provisions for fixed assets and provisions for unrealized (although already accrued) losses on foreign exchange holdings or on the portfolio of marketable instruments. Depending on prevailing regulations, the regulatory provision for bad debts may share characteristics of both types of provisions: those that increase a bank’s net worth and those that reflect an actual economic loss.

To sum up, a provision should not be tax deductible inasmuch as the provision does not represent an accrued loss but a prudential measure and the provisioning is made before the economic loss in the value of the claim occurs. General provisions for bad loans, made without any specific indication of uncollectibility of the claim often fall under this category. In contrast, if the provision represents the recognition of an accrued loss—however unrealized on a cash basis—a corresponding deduction should be allowed for tax purposes.