How to conduct a quantitative easing

By Alistair Milne

Central banks are worried about falling rather than rising prices. By early next year, it is possible that central banks’ target policy interest rates will all be reduced to their minimum possible level of zero. Does this mean that central banks will then have lost control over monetary policy and be unable to prevent a cumulative debt deflation?

Many, including Ben Bernanke, US Federal Reserve chairman, point out that central banks can then use further unorthodox tools to further loosen monetary policy.

Once interest rates are at zero, the central bank is relieved of the responsibility for draining reserves to stop overnight interest rates falling below the policy target rate.

It loses control over interest rates but gains control of the quantity of reserves and can use this to increase its balance sheet to an almost unlimited extent, buying securities, matched by increases in both wholesale deposits with commercial banks and commercial bank reserves at the central bank. By this policy of ‘quantitative easing’ the central bank increases the money supply even when interest rates hit their zero-bound.

Here is an illustration. To conduct a quantitative easing, a trader employed by the central bank buys a government bond for £1000 from an investor such as a pension fund. To settle the trade, the pension fund’s cash account with a commercial bank is increased by £1000 from the central bank, and to settle this payment the commercial bank’s reserve with the central bank is in turn increased by £1000, matching the £1000 increase in central bank assets.

But it is doubtful if this particular transaction does much to increase bank credit. The commercial bank has more short- term deposits, so monetary aggregates have increased, but it is unlikely to lend this money out, when as now banks have too many short-term liabilities and too many illiquid and undervalued long-term assets.

When quantitative easing was attempted in this way in Japan from 2001 until 2005, the main impact was to increase reserve assets rather than bank credit.

The central bank has, though, changed the composition of net public sector debt, broadly defined to include the debt of the central bank. There is less long-term and more short-term debt in the market and long-term interest rates fall somewhat.

The central bank is then likely to lose money, buying bonds at a premium high price and then, when the easing is unwound, selling them at a discounted low price.

This has economic effects because the loss-making trade subsidises long-term borrowing by the private sector. The effect is similar to that achieved when government subsidises long-term borrowing.

Quantitative easing will be much more effective if the central bank uses its balance sheet to buy not government bonds but better quality illiquid and undervalued structured and mortgage-backed securities. This eases bank funding constraints and so directly expands the stock of credit. Moreover, as the economy recovers, credit spreads will fall and so the central bank can make a profit.

Quantitative easing will be more powerful still if the central bank takes pure credit spread exposures, using interest rate swaps to remove its exposure to fluctuations in nominal interest rates.

It can also conduct equivalent synthetic transactions, purchasing government bonds alongside an interest rate swap and the acquisition of negative net worth credit default swaps. Unlike a private sector participant, as the monopoly supplier of outside money it can always meet margin calls and so cannot be squeezed out of credit default swap trades.

Finally, to guide expectations, it should set forward targets for credit spreads.

Perhaps the clearest way to present this point is to put the question in another way: what is the most appropriate alternative instrument of monetary policy, during the period when money market interest rates are reduced to their zero floor?

Aggregate bank reserves or money stock are poor choices, since in present circumstances they can increase by huge amounts without impacting credit or expenditure. A better choice is market credit spreads. The Bank of England’s monetary policy committee can use its regular meetings to announce its preferred levels for average market credit spreads. Bank monetary operations can enforce this decision.

By setting credit spreads at appropriate levels the bank will put a floor under market values, restore credit market liquidity and economic activity and make a handsome profit to boot.

A potential problem is the transition back to positive nominal interest rates, but this can be handled by a more permanent but less generous government-backed scheme for systemic credit insurance, such as been proposed by Laurence Kotlikoff and Perry Mehrling and myself on this forum.

Alistair Milne is reader in banking, Cass Business School, City University, London

Copyright The Financial Times Limited 2017. All rights reserved. You may share using our article tools. Please don't copy articles from FT.com and redistribute by email or post to the web.