Central Banking – the consequences of QT

In our first two articles in this series, we have looked at the policy options facing central banks as they continue to normalise their operations, and in particular the rationale behind slimming swollen central bank balance sheets. In this article we shift focus and start to consider the consequences for economies and markets of the approaching normalisation.

Although Janet Yellen, the Federal Reserve Chair, was studiously silent on the subject at last week’s Jackson Hole[1] symposium, the Fed is likely to announce in the near future, and possibly as soon as next month, when it will begin the process of quantitative tightening (QT), the disposal of at least some of the assets purchased as part of quantitative easing. The ECB is equally likely to announce at some stage during the autumn how it intends to proceed down a similar path; and other central banks with swollen balance sheets (Bank of England, Riksbank, Bank of Japan, Swiss National Bank) will eventually follow suit.

It therefore becomes of paramount importance to have some idea of how this process will affect economic activity and asset prices. But, as we have noted in our previous articles, this is very much unchartered territory, at least in terms of affecting economic activity, and there are few if any guides for us to use.

In such circumstances it is tempting, for policy-makers and commentators alike, to start from the basis that since QT reverses QE, the effects of QT should in some senses be the reverse of the effects of QE.  Intuitively, if QE led to lower long-term interest rates, an injection of liquidity and a rebalancing of portfolios away from fixed income to equities, then its reversal should have the opposite effect.

But this is by no means certain. For one thing, QE involved buying bonds, from both the banking system and the non-bank private sector. By contrast, QT – at least as envisaged by the Fed – does not (yet) involve selling bonds; for the moment, all that is planned is that a share of maturing bonds will not be rolled over but will be presented to the issuer for redemption. The eventual impact on activity therefore depends in the first instance on whether the original issuers of the bonds the Fed bought need to borrow the same amount again to fund the redemptions; and, if so, from whom the borrowing is done.

This is at heart a monetary matter, affecting the flows of money between the central bank, the official sector (if the original borrower was government), the banking sector and the non-bank private sector.  Now monetarist analysis is not (outside the Bundesbank) held in such high esteem these days as it once was, but it is still the most valid and formal framework for analysing such flows and so – with apologies to those who prefer a more discursive approach – we will therefore base our assessment of the issues in the rest of this article on monetarist principles.

The flows arising from QT either will or will not affect the stock of broad money in the economy. (‘Broad money’ here refers to the total liquid assets – cash but above all bank deposits – of the non-bank private sector, ie, households and non-financial corporations). This, in turn, is connected with the pace of economic activity.

To explain:

Assume that the Fed presents the US Treasury with a maturing $100 bond. The Treasury pays the Fed $100 from its account with the Fed. But the Treasury has at some stage to raise that $100 from the private sector in taxes. The net effect is that the private sector’s money holdings (broad money) fall by $100.

If this is all that happens, QT will eventually – since it is counted in billions per month – have a contractionary effect on the economy[2]. As the private sector’s money holdings shrink, there is less money available for consumption.

However, there may be a further step. The original issuer (the Treasury or, say, Freddie Mac or Fannie Mae), may need to borrow the same amount again. They can then either borrow from the banking system or from the non-bank private sector.

If they borrow from the non-bank private sector, the net effect on broad money is still a contraction. All that’s happened is that – in effect – the issuer has taken $100 from households and companies and given it to the Fed.

However, if the borrowing is from the banking system, the original impact on broad money is neutralised. This is because banks create money out of nothing. So here, what happens is that the issuer borrows $100 from a bank, thus increasing money supply by $100; and then drains money supply by $100 through its repayment to the Fed.

In either case, however, the Fed’s balance sheet has shrunk, as have the excess reserves of the banking system held with the Fed.  In other words, QT will in principle have a contractionary effect on the economy by way of shrinking the stock of money. However, this can be neutralised if the original borrower needs to borrow the same sum again and does so from the banking system.

But if this is – at least on the surface – fairly straightforward, the impact of QT on asset prices is less clear.

One object of QE was to bring down long-term interest rates. Broadly speaking, this has worked, although whether as a result of central bank purchases or central bank holdings is not entirely clear. (This relates to another issue, namely whether what matters for asset prices is the flow of central bank purchases or the stock of central bank holdings. Economic theory is very clear that it is the stock that matters. But there are strong arguments that this theory, when confronted with practice, is wrong and that the flow matters more[3]).

However, in either case, if the central bank diminishes the stock of bonds that it holds, there is both a stock and a flow effect. Hence, the net result should be downward pressure on the price of bonds, meaning upward pressure on yields. However, even that is not entirely clear. Certainly, if central banks sell bonds, the greater supply to markets should push down prices. But, so far at least, the Fed does not intend to actually sell bonds, but rather present them for redemption. So, unless the original issuer needs to borrow the same amount to fund this redemption, the net result is actually fewer outstanding bonds, which could act to push prices up and yields down.

However, given trends in fiscal policy in most developed markets, we assume that the base case is still that bonds presented for redemption will be funded by new issuance. The net impact should therefore be upward pressure on bond yields and downward pressure on prices.

Yet lower bond prices need not necessarily lead to a permanent flight from fixed income. It is true that QE led to a reallocation of assets away from fixed income and towards equities. But that was also because central banks bought up much of the stock of bonds, leaving investors with cash needing to be placed somewhere and thus going into equities.

By contrast, lower bond prices, meaning higher yields, are likely to see a reversal of those flows, perhaps not initially, but as bond yields back up towards levels previously considered ‘normal’ – say 4% for the 10-year US Treasury or 5% for the 30-year – insurance companies and pension funds, investors who need to match income and expenditure streams, and who have been hammered by ultra-low bond yields, are likely to move back to at least lock into these yields. That process will be helped by the fact that equities are currently perceived as overvalued, although the degree of overvaluation varies from the USA (very overvalued) to Europe (less far from fair value).

This eventual shift back to fixed income will put downward pressure on share prices. However, once share prices have fallen to levels where they are no longer overvalued, equities are likely to take over from fixed income as the investment asset of choice again, not least because the downward pressure on bond prices from quantitative tightening is likely to last for many years.

This leaves us with the likely asset price effect from QT evolving in three stages. In stage 1, bond prices will fall, but yields will not be attractive enough to see much of an inflow to fixed income. In stage 2, higher interest rates will make fixed income more attractive, while the outlook for weaker economic activity (partly brought on by those higher interest rates) will depress equity prices. In stage 3, equities will recover, as quantitative tightening continues to put downward pressure on bond prices.

How far can bond yields rise? That is of course impossible to answer. But, bearing in mind that the Fed’s inflation target is likely to remain 2%, that the real yield on the 10-year US Treasury tends to revert to 2% (albeit with long deviations from this mean), and that in getting to the new equilibrium level it might overshoot on the way, it is likely that US 10-year bond yields will eventually peak somewhere in the 4-6% range. However, getting there could take a long time.

As we have commented before, the normalisation of central bank monetary policy is likely to be a long process.  And as we have attempted to show in this article, neither the end destination, nor the consequences of getting there, are as yet entirely clear.


[1]              Jackson Hole is a holiday resort in Wyoming with first rate trout fishing.  It was first used for a central bank symposium in 1982, when Paul Volcker was Fed chairman;  he was notoriously unenthusiastic about such gatherings, but he was also a keen angler and the staff thought they would attract him to attend because of the fishing.  In this they were successful, but if Volcker thought he was in for a small seminar and a decent amount of fishing he was disappointed:  the first symposium proved to be a very serious study weekend and the annual Jackson Hole Symposium is now the premier central bank gathering of its type, attended by all leading central bankers and often the forum for governors to make major policy statements.

[2]              It is important to keep this effect in perspective:  with US broad money around $13-16 trillion depending on definitions, even the more aggressive levels of QT proposed by the Fed will shrink broad money by around 1% per annum only.  And this needs to be set against the natural growth in broad money in a healthy economy of around 5-7% p.a.

[3]              Note that this is not just an abstruse academic point, as it crucially determines what signifies the end of QE.  If the way QE works is through the ongoing flow of purchases, QE ends when the purchases end, which for the Federal Reserve was some time ago, and the US economy is already in a post-QE phase.  But if the way QE works is through the central bank’s stock of purchases already made, then QE technically does not end until the purchases have all been unwound and the assets sold back – which has certainly not happened yet in any country and may in fact never happen.