Depreciation Is Industrial Policy

Industrial policy is argued about in the language of grants and tariffs, which are visible, contested and comparatively small. The largest instrument is a schedule in the tax code that almost nobody discusses.

The cost recovery rules determine how quickly a business may deduct the cost of an asset. A machine written off over three years and the same machine written off over twenty are different investments, and the difference is created entirely by the schedule.

The effect is on the after-tax price of capital. Accelerating a deduction is worth real money because a deduction taken now is worth more than the same deduction taken later, which is the whole mechanism.

The history explains the incoherence. Recovery periods were set in different decades, under different theories of what tax policy should encourage, and each revision addressed the concern of its moment rather than the structure as a whole.

A government that shortens a recovery period has cut the price of that asset without appropriating a dollar.

The result is a table where the assumed life of an asset frequently bears no relationship to how long the asset actually lasts. That is not a drafting failure. It is the accumulated residue of many separate decisions to encourage or discourage something.

That is why the instrument is attractive and why it escapes scrutiny. It appears in the accounts as revenue not collected rather than as money spent, which is the treatment least likely to attract a hearing.

Consider what this does to a firm choosing between two ways of producing the same output. If one method uses equipment in a favourable class and the other does not, the schedule has decided the question before any engineer looks at it.

The published fiscal statements carry the aggregate effect of these provisions, and the aggregate is large enough that any comparison with a grant programme is unflattering to the grant.

That is precisely what industrial policy is meant to do. The difference is that a grant programme announces its preference and defends it, while the schedule expresses the same preference silently and permanently.

The asset classes are the policy. Which category a thing falls into determines its recovery period, and the categories were drawn at different times for different reasons and have been amended piecemeal ever since.

The permanence is the underrated part. A grant programme expires or gets reauthorised, which forces a periodic argument about whether it still makes sense. A recovery period sits in the code until someone removes it, and nobody is assigned to notice.

So the schedule encodes an implicit judgement about which industries are worth encouraging, assembled over decades by people addressing unrelated problems, and it has never been reviewed as a whole.

There is a distributional consequence that follows from the mechanics. The value of an accelerated deduction depends on having taxable income to deduct against, so the benefit concentrates among firms already profitable.

The comparison with visible support is instructive. Award data records grants and contracts in detail, by recipient, searchable by anyone.

A new entrant with losses gets nothing from the same provision, which means an instrument described as encouraging investment functions in practice as an advantage for incumbents over challengers.

Nothing comparable exists for the schedule. There is no register of who benefited from a recovery period, because the benefit is the absence of a payment rather than the presence of one.

The rules that allow losses to be carried forward soften this without removing it, because a deduction taken years later is worth less, and the whole value of acceleration is in the timing.

The statutes that set these periods are public, and reading them tells you the rule without telling you its distribution. Those are different questions and only one is answerable from the text.

None of this argues for lengthening every period. Recovery rules that badly understate the real economic life of an asset genuinely do discourage investment, and correcting that is legitimate.

The reform that would change the argument is not a change to any period. It is a requirement that the revenue effect of each class be estimated and published alongside the appropriations that fund the visible programmes.

It argues that the correction should be made deliberately, as policy, in public, rather than accumulated by amendment until nobody can say what the schedule as a whole is trying to achieve.

That would put the two instruments on the same page for the first time, which is the precondition for anyone deciding whether the balance between them is the one they wanted.

At present a country can hold a full public argument about a subsidy programme while a larger transfer runs underneath it, permanently, decided by a table that has never been debated as policy because it does not look like policy.