When China let its interbank market seize up: the 2013 cash crunch

An interbank market is the plumbing of a banking system. Banks end each day with mismatched positions, some holding more reserves than they need and some holding fewer, and they settle the difference by lending to each other overnight. The rate they charge is the price of that spare cash, and in normal conditions it barely moves.

In June 2013 China’s moved a very long way.

The numbers

The Shanghai Interbank Offered Rate, SHIBOR, is the benchmark for this borrowing. Shortly before the squeeze, the overnight rate sat at 2.95 per cent. On 20 June 2013 it reached 13.44 per cent.

The seven-day repurchase rate — the other standard measure of short-term funding cost — hit roughly 12 per cent on the same day, the highest on record at that point.

For scale, a bank paying 13 per cent to borrow overnight is paying an annualised rate on money it needs for hours. Nobody does that voluntarily. Rates at that level mean institutions are bidding for reserves they must have and cannot obtain, which is the definition of a liquidity crisis rather than a tight market.

What the central bank did, and did not do

The detail that makes this episode instructive is what the People’s Bank of China did at the start, which was nothing.

A central bank can end a squeeze of this kind more or less at will, by supplying reserves through open market operations. The PBoC declined to do so at the point the market expected it, and allowed the crunch to worsen. Emergency injections followed, and borrowing costs remained elevated even after them.

The reading, then and since, is that this was deliberate discipline. Chinese banks had come to rely on cheap and continuously available short-term funding to support lending growth, a good deal of it routed through off-balance-sheet vehicles where it was harder to see and harder to constrain with conventional tools. Allowing the price of that funding to spike is a way of making the risk of relying on it unmistakable, and of doing so faster than any regulation could.

Why the tactic is hard to repeat

The problem with using a liquidity squeeze as a warning is that liquidity squeezes do not stay proportionate.

Interbank markets fail in a particular way. A bank that cannot be sure of obtaining reserves tomorrow hoards them today, which reduces the supply available to everyone else, which makes the original fear correct. The mechanism is self-reinforcing, and it does not distinguish between the institutions the authorities intended to discipline and the ones they did not.

That gives a central bank running this play a narrow window. Move too early and the message is not received; move too late and the demonstration becomes the thing being demonstrated against. In 2013 the PBoC ended up injecting funds anyway, which is the outcome the entire exercise was designed to avoid signalling it would provide.

What to take from it

Three things, none of which are specific to China.

A quiet interbank rate is a policy achievement, not a natural state. It stays flat because somebody is keeping it flat.

Short-term funding dependence is invisible until it is priced. The 2013 spike revealed the extent of the reliance more clearly than any disclosure regime had.

Moral hazard is genuinely hard. A central bank that always supplies liquidity encourages the behaviour that makes liquidity necessary. One that occasionally declines discovers that the market cannot tell the difference between a lesson and a failure, and neither, in the moment, can the central bank.

How a central bank actually turns the tap

The mechanism is worth stating, because “injecting liquidity” sounds like a metaphor and is not.

A central bank supplies reserves by buying assets from banks, usually under a repurchase agreement: the bank sells securities now and agrees to buy them back later at a slightly higher price. Reserves appear in the seller’s account immediately, and the difference between the two prices is the interest rate on the operation. Reversing it — a reverse repo — drains reserves back out.

Because these operations are conducted daily and at scale, a central bank can set the price of overnight money to within a few basis points of wherever it wants it. That is the significance of what happened in June 2013: a rate of 13.44 per cent was not a market outcome the authorities were unable to prevent. It was a rate they declined, for a period, to prevent.

The corollary is uncomfortable for anyone reading a rate spike as a market signal. In a system with an active central bank, short-term rates are a policy variable most of the time. When they move sharply, the first question is not what the market believes but what the central bank has decided to allow.