The dollar has more than one price, and central banks are pricing it
Three items from the past fortnight look unrelated. Read together, they describe how the price of getting dollars is set - and who is trying to change it.
The market price is not one price. A Bank of England staff working paper compares FX forwards in the same currency pair, half hour and maturity and finds that the part of the price specific to the dealer varies with a standard deviation of about 4 basis points a year. Dealers also charge 2.5 basis points more to clients buying dollars forward than to those selling, widening to 11.8 basis points at tenors under a month. The authors attribute the spread to dealers' clienteles and pricing power more than to their own funding costs.
A central bank lowering the price. Bank Indonesia held its policy rate at 5.75% on 23 September but raised the discounts it offers on hedging foreign inflows with it - to 25% on twelve-month swaps and 30% on twelve-month domestic non-deliverable forwards. In effect, it is subsidising the cost of the hedge that a foreign investor pays to hold rupiah assets.
A backstop agreed in advance. Japan and Malaysia signed a third bilateral swap arrangement of up to $6 billion, effective 18 September, letting both sides exchange local currency for dollars and Malaysia also draw yen. It is a price guarantee of a different kind: access at pre-agreed terms when the market price becomes unreliable.

The common thread
For anyone outside the United States, the relevant cost of the dollar is not the Fed's policy rate but the price of obtaining dollars - through a forward, a swap or a credit line. That price has three layers: a market basis, a dealer-specific margin on top, and in stress, whether any price is available at all. Each item above acts on one layer. The research measures the dealer margin; Bank Indonesia pays down part of the hedging cost for the flows it wants; the swap line insures against the day the market layer disappears.
What to watch
The BoE finding implies that stress indicators built from quoted rates can understate what end users pay, especially at short tenors where the premium for buying dollars widens. For Indonesia, the test is whether hedged inflows respond. And for swap lines, the only real test is the one no one wants: a period in which they are drawn.