Wednesday, August 26, 2009

Ronald McKinnon: a concerted modest increase in interest rates?

Ronald McKinnon (Stanford), by way of comment to my post on exit strategies [Exit Wounds, 3 August] sent me an article of his, "Liquidity Traps and the Credit Crunch", that appeared in the Financial Times Forum of 14 August and "emphasizes the importance of getting out of the liquidity trap in order to restore the normal flow of credit even before full 'exit'".

In brief, "...starting from a position where interest rates are already low, say 2 per cent, reducing them to zero has only a second-order effect on expanding aggregate demand. But going from 2 per cent to zero leads to a tightening of the credit constraint on the supply side. Although there may be a “dead cat bounce” to the economy on the demand side in 2009, leaving the Fed funds rate at zero makes it impossible for the resumption of “normal” bank credit to support growth in future. A condition for restoring normal borrowing and lending in the interbank market is to have positive rates of interest at all terms to maturity. Only then will banks that are liquid lend to those that are illiquid. But if the risk free (i.e. federal funds) rate is close to zero, banks with excess reserves will not bother parting with them for a derisory yield." This - he argues - might lead to a trade contraction by causing supply constraints in the export sector, starved of credit.

I readily accept Ronald's general concern for the implications of low interest rates; I never thought - unlike Willem Buiter - that negative interest rates were the right policy to get out of recession, not least because they could not survive in globalised financial markets. But there are two points in RMcK's argument that I found difficult to accept. (1) While undoubtedly he is right on the importance of foreign trade credit, I could not bring myself to consider the cut in trade due to stricter credit constraints as a supply constraint (which term I would reserve for capacity constraints); (2) I could not see why low policy interest rates should deter banks from raising their margins between lending rates and policy rates to the point that it pays them to lend.

Faced with these objections, Ronald sent me an excellent reply, and developed the second point into a full-fledged argument that I reproduce below with his permission, for which I am most grateful. The bottom line of Ronald's argument is that "The Exit Problem: When to Reduce Monetary and Fiscal Stimuli?", has a "Partial Solution: concerted modest increase in interest rates?" [from his paper on "The Global Credit Crunch: China and the United States", Beijing 29 August 2009, which he also kindly sent me but is not reproduced here other than for figures 9 and 10 below].

"On your two points: (1) it is a matter of semantics. You may or may not like my referring to a credit constraint as a "supply" constraint. But credit is an input into working capital. If working capital is unavailable or restricted, then dumping more liquidity into the system will not increase output as if there was a constraint on physical capacity.

(2) Your second point: in the liquidity trap, why don't banks just raise their interest rates to final (retail) borrowers to maintain their profit margins and willingness to lend? This is an important and very subtle point which I did not explain very well in my FT post.

The willingness of banks to make forward commitments to lend at "retail ", to nonbank firms and households, depends very much on the wholesale interbank market. If the wholesale interbank market works smoothly without counter party risk, then even currently illiquid banks can make forward loan commitments--say 3 months in advance--to their retail customers. These banks know that if they are still illiquid when three months are up, then they can always bid for funds in the wholesale market at close to the "risk-free" interest rate to cover their retail commitment.

Now suppose some upsetting event, such as a crash in home prices that makes all mortgage related assets on bank balance sheets suspect. Then counter party risk becomes acute, and banks become less willing to lend to each other unsecured. The LIBOR market is unsecured. So illiquid banks become less willing to lend forward at retail because they don't know what their wholesale cost of capital will be 3 months hence. Indeed if any one bank tries to bid more than the going interbank rate of interest for funds, it is immediately suspected of being a more risky counter party!

Now enter the central bank concerned with the drying up of retail bank credit. It reacts in a text book fashion of flooding the interbank market with liquidity so that the interest rate on federal funds is driven to zero. Bank trading in federal funds is collateralized--usually through short-term re-po agreements, where the collateral is usually U.S. Treasury bonds.

Normally, without significant counter party risk in banks, there isn't much of a difference between the interest rate on federal funds and LIBOR. (See Figure 9 for January 2006 to July 2007 [reproduced below with RMcK's permission, from his Beijing paper quoted above, 29 August 2009]). Then when the crisis struck beginning in July 2007, LIBOR increased substantially above the Fed funds rate indicating the reluctance of banks to lend unsecured. This inhibited banks from making forward retail commitments .

(Incidentally, we could do a similar parallel analysis for foreign exchange transacting. Forward foreign exchange contracting in "wholesale" inter bank markets becoming difficult when bank counter party risk rises. The constriction in the forward market increases currency risk for both exporters and importers and inhibits the granting of trade credit--including letters of credit. Whence the surprising crash in foreign trade.)

However, by August 2009, we are in a purer form of the liquidity trap where both interest rates have been driven toward zero (Figure 9). Now we are in the situation referred to in my FT paper that liquid banks, those with excess reserves, won't bother parting with them for a derisory return. So I would guess (without knowing) that interbank trading is now at a low level .

The implication for all of this is that retail lending in the U.S. has stagnated (or fallen slightly) since August 2008 even when base money in the U.S. has skyrocketed--as shown in Figure 10 [also reproduced below, from the Beijing paper of 29 August 2009].

Figure 9. U.S. Short-term Interest Rates in the Liquidity Trap


Source: FRED and globalfinancialdata.com. From: Ronald McKinnon, "The Global Credit Crunch: China and the United States", Summer Palace Dialogue, Aug 29, 2009, Beijing

Fig. 10. Loans of U.S. commercial banks, Base Money and M2 (2006 Jan = 100)
Source: FRED. From: Ronald McKinnon, 2009, cited.

3 comments:

D. Mario Nuti said...

James K. Galbraith (University of Texas at Austin), sent me the following comment by e-mail:

"Your exchange with McKinnon is very interesting. Raising short-term interest rates is indeed what pushed the banks out of interest-rate arbitrage and back into Consumption & Investment lending after February 4, 1994 -- setting off the long credit boom. I write about this at length in my 1998 book [Created Unequal: The Crisis in American Pay]. Very glad to see an intelligent discussion of the problem now; it's not obvious that the Fed gets it."

Unknown said...

Dear Mario
your discussion with McKinnon is very intriguing. But, in the McKinnon reply to the second point of your objection, two points still remain unclear. While the argument of a ridiculous yield can account for the liquid banks's refraining from lending in the interbank market, I cannot understand how their refraining behaviour can impair the functioning of the interbank market after it has been intentionally flooded with liquidity by the Fed. Second, as in the FT editorial, I cannot see the reason why a near-zero interbank interest rate can refrain the liquid banks to lend 'at retail' at an interest rate only just lower than that which is expected to prevail in the near future (say, at an almost normalised interbank market conditions).

Maybe, I have negleted something in the McKinnon's line of reasoning... Anyway, brilliant debate! Economics, and economic policy, should build up on debates of this kind.
Best Regards,
Beatrice Gorga

D. Mario Nuti said...

Ronald McKinnon answers:

Mario, Beatrice Gorga is close to your position in asking why banks don't just lend directly at retail at high interest rates.

Now I think the best way to look at it is that retail lending is not just spot lending but also consists of making forward lending commitments--say 3 to 6 months hence--to a bank's retail clients. Indeed, a whole credit line is often made available to the bank's best client with no narrowly specified time for drawing on it. But a bank would only feel comfortable giving such a (indefinite) forward lending commitment if it knew that it could easily cover itself in the wholesale interbank market if the retail credit line was drawn down faster than anticipated.

I now think that the analogy with the forward foreign exchange market is closer than I had previously thought. Banks also make forward foreign exchange commitments at retail to exporters or importers--sometimes fairly indefinite. But they would only feel comfortable in doing so if they could easily cover any unexpected shortage of foreign exchange in the interbank markets between currencies--which is also gummed up by counterpart risk and near zero interest rates.