Multi-Jurisdiction Capital in a Fractured Rate Cycle
As central banks diverge rather than move in lockstep, the old assumption of a single base rate against which a private portfolio is priced no longer holds — and covered interest rate parity breaks down faster than most hedging programmes are built to track.
For most of the period in which current private-banking practice was formed, policy rates across major economies moved broadly together. Divergence existed but was modest, and the working assumption of a single reference cost of money was a reasonable simplification.
That assumption does more work than it appears to. It underpins how a portfolio is benchmarked, how a lending facility is priced, how a hedging programme is sized and how a family thinks about where to hold liquidity. When policy paths separate materially, each of those decisions has a different answer depending on currency — and the simplification quietly becomes an error.
Covered interest parity as a maintained assumption
Covered interest rate parity states that the forward premium between two currencies should equal their interest rate differential, because otherwise a riskless arbitrage exists. It is the assumption underneath most hedging arithmetic.
It is also an assumption that has failed repeatedly since 2008, persistently rather than momentarily. The residual is the cross-currency basis, and it is not a rounding error: it widens when balance-sheet capacity is constrained, when there is a scramble for a particular funding currency, and predictably around reporting dates when dealers withdraw capacity.
The basis is not noise around a correct model. It is the price of the balance sheet required to hold the arbitrage open — and that price rises exactly when hedging matters most.
The practical consequence: the cost of a hedge is not the interest rate differential. It is the differential plus the basis, and the basis is widest precisely in the conditions that prompt someone to hedge.
Real rather than nominal differentials
A second error compounds the first. Comparing nominal policy rates across jurisdictions says little about the cost of money in each, because inflation differs. Two economies with identical nominal rates and materially different inflation are not offering the same real return, and a family holding assets in both is not facing the same constraint in each.
For decisions with a long horizon — where to hold reserves, which currency to borrow in, how to match a liability denominated in one currency against income in another — the real differential is the relevant input. Nominal comparison systematically favours the high-inflation jurisdiction, which is precisely backwards.
Sequencing rather than blanket hedging
The response is not to hedge everything. A fully hedged multi-currency balance sheet pays the basis repeatedly and often hedges exposures that offset each other.
The sequence we would apply:
- Map liabilities by currency and date. School fees, tax, debt service, committed capital calls. This is the only hedging requirement that is genuinely non-discretionary.
- Net naturally before hedging. Income in a currency offsets liabilities in that currency. Hedging a gross exposure that is already partly matched is paying the basis for nothing.
- Hedge the residual, matched to tenor. Rolling short-dated hedges against long-dated liabilities is a roll-risk position, not a hedge — the basis is re-priced at every roll.
- Price the basis explicitly. Treat it as a line item with its own expected cost and its own stress case, not as an execution detail.
- Borrow where the real cost is lowest, adjusted for basis. The cheapest nominal rate is frequently not the cheapest funding once hedging cost and real differential are included.
What this means for a private balance sheet
Mainly that "the cost of money" is no longer a single figure. It varies by mandate, by currency and by time horizon, and a structure that assumes otherwise will make consistent errors in one direction — over-hedging the visible exposures, under-pricing the roll, and holding reserves in whichever currency happens to look attractive nominally.
The remedy is unexciting: a currency-and-date liability map, hedges matched to tenor, and the basis priced as a cost rather than absorbed as a surprise.