Two scorekeepers, two questions
In the last guide you met the idea that a business pays [[income-tax|income tax]] on its profit, and that the government has its own reasons — raising revenue, and steering behavior — for how it defines that profit. Hold onto that, because here is the twist that surprises every beginner: the profit figure at the bottom of the income statement you so carefully built earlier in this ladder is *not* the figure the company multiplies by the tax rate. There are two profits, computed two ways, for two different readers.
The first profit is book income — net income computed under [[gaap|GAAP]] (or, outside the United States, IFRS) and reported to investors, lenders, and the public on the financial statements. Its goal, as the standards-setters frame it, is to *faithfully portray economic performance* for someone deciding whether to invest or lend. The second profit is taxable income — the figure computed under the tax code and written on the income tax return filed with the tax authority. Its goal is entirely different: to measure, by the legislature's rules, exactly how much the government has decided to tax this year.
Why the rulebooks disagree
The disagreement is not random; it grows out of what each rulebook is *for*. GAAP is built to portray performance fairly, so it leans on judgment, estimates, and the [[matching-principle|matching principle]] — recording an expense in the same period as the revenue it helped earn, even when no cash has moved. The tax code, by contrast, is written by legislators who want the rules to be hard to game, easy to verify, and useful as policy levers. Those two missions point in different directions often enough that the two profits routinely diverge.
Think of it as one underlying truth viewed through two lenses. An accountant might judge that a customer probably will not pay, and under GAAP record a bad-debt expense *now* on the strength of that estimate — because faithful portrayal demands recognizing the likely loss. A tax authority, wary of companies inventing convenient losses to shrink the bill, may refuse the deduction until the debt is *actually* written off. Same business, same customer, same reality — but the GAAP lens recognizes the cost earlier than the tax lens allows. Multiply that pattern across estimates, timing, and policy choices, and you get the gap.
Permanent differences: gaps that never close
Now to the heart of the matter. Every difference between book and taxable income falls into one of two families, and learning to sort them is the whole skill of this guide. The first family is the permanent difference: an item that counts for one rulebook but *never* counts for the other, in this year or any future year. The gap it opens is real and it stays open — it never reverses.
The classic example is the disallowance of certain expenses. Suppose a company spends 10,000 entertaining clients at sporting events. Under GAAP that is an ordinary business expense — it reduces book income by the full 10,000. But the tax code, as a matter of policy, may declare entertainment costs *non-deductible*. So on the tax return that 10,000 simply does not exist as a deduction. Book income absorbs the hit; taxable income does not. And here is the defining feature: there is no future year in which the company ever gets to deduct that 10,000. The deduction is gone for good. The difference is permanent.
Permanent differences run both ways. Some income that GAAP counts is tax-exempt — interest on certain government bonds, for instance — so it lifts book income but never enters taxable income. Some expenses GAAP records, like a penalty for breaking a regulation, are deliberately non-deductible so the law does not soften the sting. The unifying idea is simple: a permanent difference changes one profit but never the other, in any year, so it shifts the company's overall [[marginal-vs-effective-tax-rate|effective tax rate]] — the actual tax paid as a fraction of book income — away from the headline statutory rate, and keeps it there.
Temporary differences: gaps that close over time
The second family is the temporary difference, and it is the more interesting one. Here the item *does* count for both rulebooks — eventually, over the whole life of the asset or obligation, both arrive at exactly the same total. They simply disagree about *timing*: one recognizes more this year, the other more later. The gap opens, then in a future year it closes again. That reversal is the whole signature of a temporary difference.
The textbook case is depreciation. Recall from earlier rungs that a machine's cost is spread across its useful life — and that you can spread it evenly (straight-line) or front-load it (an accelerated method). For its financial statements a company often uses gentle straight-line depreciation, which keeps book income smooth. But the tax code frequently lets — or pushes — the company to use accelerated depreciation, deducting far more in the early years. This deliberate split is [[tax-vs-book-depreciation|tax versus book depreciation]], and it is the single most common temporary difference in the world.
Machine cost 9,000, 3-year life, no salvage Year BOOK (straight-line) TAX (accelerated) Difference ---- -------------------- ----------------- ---------- 1 3,000 6,000 -3,000 (tax higher) 2 3,000 2,000 +1,000 3 3,000 1,000 +2,000 ---- -------------------- ----------------- ---------- Total 9,000 9,000 0 <- reverses to zero
Read that little schedule slowly, because it carries the whole idea. In year one the company tells the tax authority a *smaller* profit (it deducted 6,000, not 3,000), so it pays *less* tax that year. It has not escaped the tax — it has merely *postponed* it, because in years two and three the smaller tax deduction makes taxable income *higher* than book income, and the deferred tax comes due. A temporary difference is, at bottom, a question of *when*, not *whether*. The total tax over the asset's life is the same; only its schedule has shifted.
Sorting any difference, and what comes next
With both families named, you can classify almost any book-tax gap with one question: *will this item ever appear on the other rulebook's ledger?* If never — a non-deductible fine, tax-exempt interest — it is permanent, and it bends the effective tax rate for good. If yes but in a different year — depreciation timing, an expense the tax code recognizes only when cash is paid — it is temporary, and it will reverse. That single fork is the [[permanent-vs-temporary-differences|permanent versus temporary difference]] distinction, and it is the spine of business income taxation.
There is one honest loose end to flag before you climb on. A temporary difference does more than shift a tax payment in time — it leaves a footprint on the balance sheet. When the company defers tax by depreciating faster for the tax authority, it has, in effect, an obligation to pay that tax in future years; GAAP insists on recording it now as a [[deferred-tax-assets-and-liabilities|deferred tax liability]]. The mirror case, where a company pays tax early and will save later, creates a deferred tax asset. Permanent differences leave no such footprint — having no future to reverse into, they create nothing on the balance sheet at all. That balance-sheet consequence is exactly where the next guide picks up.