Foreign currency, part 3: no one-size-fits-all

Every regime locks in the same realised CHF 200; they differ only in how accurate the books are mid-year and how much work that costs. So which fits your business? This final part lays the choices side by side.

The trade-offs, side by side

Each regime is a point on the same accuracy-versus-effort line; none is universally right. One distinction cuts across all of them: a realised result is locked in when cash moves and is always booked, while an unrealised one lives only on paper. Swiss tax follows the statutory accounts (the Massgeblichkeitsprinzip), so a booked FX result is in principle taxed as booked — but the Federal Supreme Court drew a sharp line in BGE 136 II 88: pure translation differences, from restating foreign-currency books into francs, are "the consequence of a fictitious event" and stay out of taxable profit, while only real transaction differences are tax-effective. The timing you pick still carries a real cash-tax edge.

Who needs which

The right regime follows from how much foreign-currency risk drives the business. For a firm whose margins are made and lost on exchange moves, mid-year accuracy is survival; for one that just invoices abroad now and then, it is overhead. A separate choice surfaces once you hold a running foreign-currency cash balance and spend part of it: which purchase rate is the cost basis — FIFO, LIFO or average. That is a jurisdictional question, not a timing one. And VAT runs on its own exchange-rate regime altogether: turnover is converted at the ESTV monthly-average rate or the daily selling rate, your choice — but kept for a full tax period and applied alike to output, acquisition and input tax (art. 45 para. 3 and 5 of the VAT Ordinance). Assuming the rate in your books governs the VAT return is the classic error.

The same end, a different journey

The result that ties the series together is what makes the choice safe: every regime reaches the identical realised lifetime loss; only the timing and the booking count differ. The interim split is subtler. Under the imparity principle an unrealised loss is always booked, but an unrealised gain only on short-term operating items — so our +300 holds, while a long-term FX position would defer it (where IFRS, IAS 21.28, recognises both directions symmetrically). For direct tax, balance-sheet positions are revalued at the ESTV year-end rate — the closing price of the last December trading day, which counts as the tax value at 31 December (StHG art. 14, 17) — and the income statement at the annual-average Jahresmittelkurs.

An adaptive middle ground

If there is a lesson, it is that the textbook regimes are points, not a spectrum, and most businesses live between them. The interesting design question is not "which regime" but "how to get the accuracy of daily where it matters and the calm of year-end everywhere else." The prior art already gestures at it: a monthly cadence, or a multi-currency ledger that holds each currency in its own account and stays accurate at any instant without an entry on every rate move.

An adaptive approach would, in principle, keep the books quiet for the long tail of small, slow-moving positions and tighten up — revalue more often, realise sooner — only where exposure is large or fast. For a Swiss SME shipping the occasional euro invoice yet wanting a year-end it can trust, a mix of both worlds is usually the honest answer. There is no single right method; only the one that fits.