How should we calculate taxable income?

In the November/December 2003 issue of accounting & business, Christopher Nobes considered the advantages and disadvantages of basing taxable income closely on accounting net profit and concluded that the UK Inland Revenue's well-meaning attempts to achieve convergence were misplaced. In this article, he writes on the need for a conceptual framework for taxable income, rather as accounting standards are now based on a framework

The financial reporting framework has three levels: an overall objective (investor decisions), principles (e.g. prudence) and rules (e.g. the prohibition of LIFO). For taxation, I suggest that the overall objective is the collection of an equitable amount of tax from each business (once the overall tax take has been fixed by parliament), with as little cost as possible.

Income measures

For tax purposes, the most suitable principles for the measurement of income might differ from those increasingly used for financial reporting. For example, the old-fashioned concept of realisation is being gradually abandoned (e.g. IAS 39, IAS 40 and IAS 41 take unsettled gains to the income statement). However, unless the assets have been sold, the gain is an estimate. Also, a tax on unsold assets might impose an unreasonable liquidity burden on the company. This would not be equitable. One problem of ignoring increases in value until assets are sold is that the company can choose to postpone tax by choosing not to sell. Would it be relevant if the assets could easily be sold? Is 'readily realisable' almost as good as 'sold'?

On balance, my conclusion is that concepts other than 'sold' are too vague for tax purposes. Furthermore, taxation could not reasonably be based on the valuations currently used for financial reporting because they are an incoherent mixture of depreciated costs, net realisable values, fair values and discounted cash flows.

If one is trying to measure income, it makes no sense to revalue some assets and liabilities but not to revalue others. However, any universal revaluation basis (including market prices) would be expensive for taxpayers and tax auditors. The conclusion is that the need for objectivity and certainty argues in favour of the tax system's traditional reliance on transactions and realisation.

Minimum total cost

Cost to the taxpayer is reduced by simplicity and objectivity. Judgement is expensive, and this argues in favour of objectivity, assisted by basing taxable income on recorded transactions of purchase and sale rather than on valuations, as explained above.

From the point of view of the tax collectors, too, simplicity and objectivity help to reduce cost. The reduction of tax evasion is also aided by simplicity and objectivity, and by perceptions of equity.

Comparison with a financial reporting framework

Some assumptions, concepts and constraints of the IASB's framework have a different meaning in the tax context. For example, in a tax context, neutrality means lacking in unintended side-effects rather than unbiased. However, the latter concept is also useful for tax. The concept of relevance changes its significance as objectives change: what is relevant for decision-making would not be as relevant for assessing ability to pay tax. Similarly, the benefit/cost comparison needs to be re-written in terms of cost minimisation. Some concepts are less important in the tax context, and some may be harmful.

I will now go through the framework's concepts with tax in mind. The accruals concept may be harmful in the context of taxation because it lacks objectivity. The going concern assumption is less relevant because the tax basis moves away from predictions about future use. Understandability is less relevant when there is only one user of the information and when the objective is not decision-making. Less attention needs to be given to faithful representation when objectivity is already being stressed.

The need for objectivity also militates against materiality, substance over form and prudence. Materiality introduces judgement and inaccuracy. Substance over form implies departure from the readily observable legal position towards some more complex assessments. In principle, the search for substance would be useful in the field of taxation. However, despite some moves in this direction, it would require a major change in approach by the legal system. Prudence introduces the need to compare cost with subjective values, such as net realisable value or value in use (i.e. discounted expected net cash flows).

However, a form of completeness is important: the tax authorities need a full set of relevant information. Generally, this set can be specified more readily in the context of tax than for financial reporting because of the single user and the backward-looking purpose. A version of comparability is also important in that the taxpayers' incomes need to be properly compared in order to ensure equity.

The concept of timeliness is important for financial reporting because delay damages the relevance of information for decision-making. However, although the Government would benefit from speed in tax collection, the passing of some weeks does not make the data any worse for a tax purpose. Lastly, as noted above, the benefit/cost comparison is important for tax once its meaning has been adjusted.

Rules

We can now work out some rules from the above concepts. Comprehensiveness requires the inclusion, at least eventually, of income in the form of interest, dividends and capital gains, as well as from trading. However, objectivity/certainty suggests that tax-relevant revenues should not rest upon judgements. This means that any revaluations of assets should be ignored. The same applies to 'revaluations' caused by currency movements affecting unsettled monetary balances. Similarly, profit on contracts should be taxable on completion.

The implications for depreciation and impairment of tangible and intangible assets need to be considered. I suggest that none of these re-measurements should be relevant for tax. If taxpayers are allowed to wait until there is a transaction before paying tax on a gain, then they can wait before getting a deduction for a loss.

However, for assets likely to lose value (e.g. machinery), the tax authorities could allow a fixed percentage of cost per year. This would be adjusted to actual loss at the time of sale. In UK terms, this means capital allowances with balancing charges. This approach would avoid the large judgements involved in measuring value in use, which is the normal impaired value for an asset (e.g. under FRS 11 or IAS 36). In general, it would also avoid the temptation for companies to exaggerate capital losses for tax purposes.

The same approach could be extended to inventories. That is, for tax purposes, they should be valued at cost until sale. This means not using the 'lower of cost and market' rule, because market value is subjective. Interestingly, this is the approach in some countries (e.g. the Czech Republic). For the measurement of the cost of using up inventory, I would require FIFO for tax purposes, on the grounds that it is likely to be close to the actual cost of inventory used.

Interest and R&D expenses (except those for tangible fixed assets) should be tax deductible, irrespective of any capitalisation for financial reporting. The accruals basis could be allowed for interest and for other straightforward expenses such as wages. This would add simplicity without serious risk of manipulation.

For allowances against bad debts, I would take the same view as the tax authorities in the UK, the US and France. That is, specific allowances that can be checked against particular customers should be allowed, but general allowances should not be.

All leases should be treated as operating leases for tax purposes. This is more objective than following the financial reporting rules, especially as they presently involve capitalisation of only some leases even though all non-cancellable leases entail assets and liabilities. Even if all leases were capitalised for financial reporting purposes, as proposed by some standard setters, it would be better to capitalise none of the leases for tax purposes because the proposals involve substantial judgement in terms of calculating the proportion of the leased asset that should be capitalised.

The current treatment of asset-related government grants under UK rules (SSAP 4) or IASB rules (IAS 20) is to take them to income over the life of the asset. However, it is not clear why the grants are not income immediately assuming that the conditions are met. This is the conclusion in IAS 41 (paras. 34 and 35) for grants related to biological assets. Consequently, to use the financial reporting basis for tax purposes seems unsatisfactory and anyway rests on estimates of lives. To base income on compliance with conditions also introduces judgement. Presumably, the simplest solution is for the granting departments to liaise with the tax authorities and to fix the size of grants on the basis that they are not taxable.

Turning to liabilities, I suggest that the creation of provisions should also be ignored for tax purposes. The arrival of FRS 12 (or IAS 37) has anyway narrowed down the scope for provisions. Many of the remaining provisions (e.g. de-commissioning or pensions) involve great judgement, including discounting.

Income from investments should be treated on a dividend received (or receivable) basis. This means that the equity method (e.g. as used in investor statements in Denmark and the Netherlands) would not be allowed.

The correction of errors (whether 'fundamental' or not) and the results of changes in accounting policy should be passed through taxable income rather than being treated as prior-year adjustments. Otherwise, they would never be taken into account for tax.

Most systems of taxation of corporate income allow losses to be carried backwards or forwards against profits of other periods. The system proposed here would need to do that, especially as it disallows various forms of matching or smoothing, e.g. provisions.

Conclusion

This tax system, which differs from that in the UK but is very close to the existing US system, would be a stable one, designed to be suitable for tax. It would also mean that tax systems would be immune from changes made by the standard setters that took no account of tax issues.

ACCA has published a report entitled A Conceptual Framework for the Taxable Income of Businesses, and How to Apply It under IFRS.

Christopher Nobes is PwC professor of accounting at the University of Reading, UK.