What moves currencies — course checkpoint

You began with a plausible but shallow answer — "interest rates" — and you finish able to say which rate, over what horizon, against what expectation, and where the relationship stops working. Let us assemble it.

The four units, in one breath

Rates and the carry trade. An exchange rate is the relative price of two monies, and money pays interest, so the differential is the reward for holding one rather than the other. The level of that differential is already priced; what moves spot is a change in the expected path, which is why two-year yields track currencies better than today's policy rate. Borrowing the low-yielder to hold the high-yielder is the carry trade: at AUD/JPY 100.00 with a four-point differential, it earns 4.0% a year and breaks even if the pair falls just 3.83%, to 96.17. Carry returns are strongly negatively skewed — small gains for years, then a mechanical, positioning-driven unwind that can erase 7.5 years of carry in weeks. A large nominal differential is usually compensation for inflation or devaluation risk, not a gift.

Parity conditions and long-run anchors. Covered interest parity is arithmetic: F = S × (1 + i_terms × t) / (1 + i_base × t), enforced by riskless arbitrage. The forward rate is a spot rate adjusted so nobody is paid twice — not a forecast — and 96.17 was the one-year forward all along, which is why a hedged carry trade earns exactly zero. Uncovered interest parity claims the high-yielder depreciates by the differential; Fama's regressions returned coefficients near zero or negative instead of one — the forward premium puzzle, and the carry trade's entire historical return. Purchasing power parity anchors the very long run: deviations decay with a half-life of three to five years, and no model reliably beat a random walk under twelve months. Trade flows are three days of FX turnover per year of world trade; the stock they accumulate over decades is what matters.

Central banks and pegs. FX trades the gap between two reaction functions, and it trades surprises against what was priced — which is why a hike can weaken a currency. Intervention is sterilised or not, and that decides whether anything real changed; a $20 billion operation is about 1.2% of a day's yen turnover — overwhelming for twenty minutes, small over a year. The impossible trinity forces a choice among a fixed rate, free capital and monetary independence. A central bank selling its own currency has unlimited capacity; one buying it back has a countable stock — which is why 15 January 2015 happened as it did: the SNB never ran out of francs, it stopped being willing to spend them.

Risk appetite and the dollar. In stress, currencies sort into risk-on and risk-off blocks through at least four distinct mechanisms — funding unwind, repatriation, safe-asset depth, and self-fulfilling convention. Commodity currencies move with their terms of trade because national income does, but it loses to the rate channel regularly. And the dollar sits on one side of roughly 89% of all trades, prices about 40% of world trade, and is owed by borrowers everywhere — so global fear produces global dollar demand even when the fear is American in origin.

The three habits worth keeping

1. Ask "compared with what was priced?" The level of a differential, a delivered hike, an existing surplus — all already known. Only the revision moves the price.
2. Ask which horizon you are on. Flows dominate a day, rate paths dominate a year, prices and stocks dominate a decade. Most bad currency commentary is a correct statement applied to the wrong horizon.
3. Ask what the premium is paying for. Every persistent excess return here — carry, the forward premium puzzle, the dollar's convenience yield — compensates for something specific. Name it before admiring the return.

And what this course does not give you

It does not tell you where any currency is going. It has not presented the carry trade, or policy divergence, or any other driver, as a strategy to use — the carry trade in particular is documented here precisely because its return profile is asymmetric and its unwinds are violent, and stop-losses protected nobody on 15 January 2015. Every episode here is a factual account of a documented past, not a template for a predicted future. This has been education about how a market works, never advice about what to do in it, and nothing in it is a recommendation to buy, sell, or position for anything.

Before you sit it

Each of these is a minute at your desk. Any one that is not names the lesson to reopen first.

• Say what a carry trade earns while it works, and what it risks when it stops — Why do investors borrow in yen to buy Australian dollars?
• Say why a forward rate is arithmetic rather than a forecast — Why is the forward rate arithmetic, not a forecast?
• Name the three things a country cannot have at the same time — What does it actually cost to hold a currency peg?
• Name two currencies that tend to rise while everything else is falling — Why do some currencies rise when everything else is falling?

Try it now

1. Build a one-page dashboard for the latest date the page holds. Every input is below: a year of EUR/USD, USD/JPY and AUD/JPY; the policy rates of the Fed, the ECB and the Bank of England; and annual inflation and then net trade for two of those economies, the United States and the euro area. The macro rows are annual, so note how old the newest one is beside the date of the prices.

2. Write four sentences about that date — one per unit theme — using only what is on the page.
3. Then write a fifth sentence naming something important these numbers cannot tell you. Keep the dashboard: the next course in this domain moves from what drives an exchange rate to how that market is structured and traded.

Checkpoint quiz next. Nothing in this course was a recommendation, a forecast, or a view on any currency — you have learned to read the drivers, which is a skill, not a signal.