The Curvature of Money

A physics idea about exchange rates points to a sharper early-warning system for currency policy

By Ginanjar Utama

Indonesia has spent years considering whether to redenominate the rupiah — cutting three zeros from the currency so that a 20,000-rupiah note becomes a 20-rupiah note. If it ever happens, nothing real will change. Salaries, prices and exchange rates would all shrink by the same factor. The number printed on the bill would be different; purchasing power would not.

That ordinary observation hides a powerful idea. Money has no absolute value. It only has value relative to a unit of account. Change the unit, and the numbers change, but the economic reality does not.

Physicists have a name for systems whose description can be relabelled without changing the underlying reality: a gauge symmetry. It is the mathematical structure behind electromagnetism. Applied carefully to foreign exchange, it offers more than a metaphor. It gives a unified way to read arbitrage, funding stress, carry trades and the cross-currency basis — and it suggests a sharper early-warning system for central banks managing exchange-rate pressure.

Value is relative. Arbitrage is not.

In a web of exchange rates, most numbers depend on the unit chosen. The dollar price of the rupiah, the rupiah price of the yen, and the euro price of the dollar all shift when one currency is redenominated. But one thing survives every relabelling: the result of a closed loop.

Start with dollars. Convert them into euros, then yen, then back into dollars. If you end with more dollars than you started with, the profit is real. No choice of numéraire can make it disappear.

That is the foreign-exchange version of curvature. In physics, curvature measures the failure of a round trip to return an object to its original state. In currency markets, triangular arbitrage measures the failure of a round trip through exchange rates to return money to its original value. When there is no arbitrage, the space is flat. When there is arbitrage, it curves.

High-frequency traders eliminate ordinary triangular arbitrage almost instantly. But the same geometric idea becomes more interesting once time, interest rates and funding constraints are added.

Time adds another direction

Holding a currency over time earns an interest rate. That means time itself acts like a conversion link. Once this is included, the familiar machinery of international finance — covered interest parity, uncovered interest parity, the carry trade and the forward premium puzzle — can all be read as different forms of curvature in a currency-time lattice.

The most important object in that lattice is the cross-currency basis.

In a textbook world, investors should be able to obtain dollars in two equivalent ways: borrow dollars directly, or borrow another currency and swap it into dollars. Covered interest parity says the two routes should cost the same. Since the global financial crisis, they often do not. The gap is the cross-currency basis.

When dollars are scarce, the basis turns more negative. Synthetic dollar funding becomes expensive. For emerging markets, that matters because the basis is not just a technical footnote in swap markets. It is part of the cost faced by foreign investors who buy local-currency bonds and hedge the exchange-rate risk.

In the geometric picture, the basis is temporal curvature: a deviation in the loop that moves across currencies and forward through time.

The same framework also clarifies the carry trade. The return from borrowing in a low-yielding currency and investing in a high-yielding one can be split into two pieces. One is the basis — the riskless covered-interest-parity deviation. The other is the currency risk premium — compensation for bearing exchange-rate risk. The distinction is essential. The first is a funding-market friction that policy may be able to ease. The second is a market price of risk that policy cannot simply arbitrage away.

Why mispricing persists

If the basis is a riskless deviation, why does it not vanish?

Here the physics analogy becomes useful again. In a plasma, mobile charges rearrange themselves to screen an electric field. In a superconductor, they can expel the field almost entirely. Currency markets have their own screening medium: dealer balance sheets, arbitrage capital, funding lines, collateral capacity and risk limits.

When that capacity is abundant, mispricing is screened away quickly. When it is scarce, arbitrage capital cannot reach far enough. A residual deviation remains pinned open.

That residual is the basis.

It persists not because the arbitrage is imaginary, but because the balance-sheet machinery needed to close it is constrained. Economists call this limits to arbitrage. The geometric language makes it more operational: the basis is the part of the curvature the market’s arbitrage “plasma” cannot screen.

That makes the basis an early-warning indicator. It reveals stress in the plumbing before the pressure fully appears in the spot exchange rate.

Why central banks should care

For a central bank in an open emerging economy, this reframing has three practical implications.

First, the basis is a tax on stabilising capital. A foreign investor who buys Indonesian government bonds and hedges the rupiah exposure pays the basis as part of the hedge. If the IDR basis widens toward 90 basis points on an annualised basis, that is a material cost on hedged participation. The exact cash cost depends on tenor, rollover and market convention, but the direction is clear: when stress rises, hedged inflows become more expensive precisely when the currency most needs them.

The basis and the capital flow are two readings of the same pressure. A widening basis says that the stabilising money is being priced out before the spot rate fully reflects it.

Second, the effectiveness of intervention depends on screening capacity. When dealer balance sheets are healthy, a central-bank nudge can be transmitted through the market and arbitraged into place. When balance sheets are stretched, the same intervention may drain reserves without restoring confidence. The market keeps testing the line because the arbitrage capacity that would normally close deviations is impaired.

That means the policy question is not only, “How much should the central bank spend defending the currency?” It is also, “Is the market capable of absorbing and transmitting the defence?”

In some episodes, restoring arbitrage capacity — through FX swap facilities, dollar liquidity lines, hedging incentives, repo backstops or better market-making conditions — can matter more than spot intervention alone. It treats the screening failure rather than only fighting its visible symptom.

Third, the framework unifies indicators that are usually scattered across separate dashboards. Triangular consistency, forward points, the cross-currency basis, carry premia, hedge costs, dealer balance-sheet conditions, capital flows and the shape of the basis curve across tenors are often monitored separately. In this framework, they are different curvatures and screening conditions of one market object.

That opens the door to a more coherent funding-stress dashboard: not just the level of the basis, but its term structure; not just spot pressure, but whether hedged investors are being priced out; not just reserves, but whether arbitrage capacity is strong enough for intervention to stick.

The twist in the curve

There is a further, more speculative extension. A basis is not only a single number. It is a curve across tenors — one month, three months, six months, one year and beyond. Different currencies can have basis curves with different shapes: one may be steep, another flat, another stressed in the short end.

If each currency carries a “shape” of its basis curve, moving through currencies and tenors may create a path-dependent twist. You can return to the same currency value but not to the same funding-curve position. In mathematical language, this is a non-abelian holonomy. In practical language, it is a hidden form of basis-curve stress that ordinary scalar indicators may miss.

This should not yet be called a tradable arbitrage. A tradable arbitrage requires a self-financing strategy that survives bid-ask spreads, funding costs, margin, liquidity and execution risk. That has not been demonstrated. For now, the more defensible interpretation is as a diagnostic: a measure of cross-tenor funding stress that may warn of debt-flow fragility before it becomes obvious in spot FX.

For a central bank, that is already useful. Not every warning signal needs to be a trading strategy. Some signals matter because they show where the market’s plumbing is bending.

A lens, not a crystal ball

Two caveats are essential.

The first is data. An operational version requires synchronized, high-quality market data: spot, forward points, cross-currency basis quotes, bid-ask spreads and tenor-by-tenor liquidity. Representative stress levels can illustrate the geometry, but they cannot replace a live feed.

The second is interpretation. Geometry can organize the market; it does not eliminate uncertainty. The framework does not say where the exchange rate must go tomorrow. It says where stress is accumulating, whether arbitrage capacity is impaired, and whether policy is fighting the source of pressure or merely its surface expression.

That is still valuable. Currency crises rarely begin as a single dramatic break. They begin as small bends in funding markets, hedging costs, dealer capacity and investor behaviour. By the time the spot rate becomes the headline, the deeper curvature may already have been building for weeks.

The curvature-of-money lens says: watch the loops. Watch the basis. Watch the term structure. Watch whether the market can still screen deviations. If those gauges start to bend together, the warning is stronger than any one indicator alone.

Money, it turns out, has a shape. The job of policy is to notice when it starts to curve.