Some more detail on the outlook for interest rates and inflation

Andrew Lilico is Policy Exchange's Chief Economist.

My previous blogpost attracted quite a bit of attention (probably more than it deserved).  Curiously, attention focused on those parts of the post in which I was repeating opinions I’ve expressed since early 2009 (indeed, have argued in a more general form since 2002, as for example here), and hence that I didn’t defend in any detail.  As a consequence, much of the subsequent press discussion seems to me to have missed a number of important points, which I propose to explore a little more now.

1)      In my view there are three broadly plausible paths for the UK economy from here over the next few years: deflationary slump (let’s call this the “US 1930s” scenario); robust sustainable recovery (let’s call this the “UK 1990s” scenario); unsustainable nominal boom followed by bust (let’s call this the “UK 1970s” scenario).  The scenario I don’t consider plausible is what we might call the “Japan 1990s” scenario — the scenario widely covered in the press and City circles, in which the economy suffers an extended period of very low nominal growth, with growth rates consistently at about 1% and enduring risk of deflation that never quite materialises.

The reason I don’t consider this plausible is that I don’t believe it could constitute an equilibrium.  Specifically, Japan was able to sustain this scenario in the 1990s because it entered the scenario with very high levels of household savings.  That meant that households were able to endure an extended period of low growth with relatively little proximate impact — the Japanese losses were mainly in output foregone (and of course lost asset valuations and huge rises in government debt).

By contrast, the UK entered the post-2007 crisis with very high levels of household indebtedness.  UK households have gambled on obtaining very high levels of nominal wage growth in order to be able to service their debts.  Without that wage growth, the burden of servicing those debts will start to drive widespread defaulting, leading to a snowball effect driving down prices, triggering nominal wage deflation, more rapid defaulting and further financial sector problems, dragging the economy into the deflationary slump scenario (the US 1930s scenario).  Unless we want deflationary slump, then there has to be nominal wage growth — simple stagnation is not an option; our households are too heavily indebted to tough this out.  (For more detail on UK household indebtedness, see this report, especially Figure 2.)

The near-term risk of falling into deflationary slump will become acute as we enter double dip later this year or early 2011.  But the monetary authorities have the power to prevent us falling into deflationary slump: they can print more money (do more QE).  That will be the correct response to double dip.  And even without a double dip, that will be the correct policy to accompany the fierce fiscal tightening that is planned (especially since the government has chosen to raise taxes (specifically VAT) early in the tightening by more than I had hoped or recommended).

Quantitative easing (QE), though it can certainly prevent deflation, is subject to what is called the “ketchup-in-the-bottle” effect.  In quantitative easing we expand what is called the “monetary base” — the narrowest concept of the money supply.  Now, there is a relationship between the monetary base and what really counts for determining the level of nominal activity — broad money.  Let’s call this relationship a “money multiplier”.

In the medium term, that money multiplier must be determined by real factors, such as the number of cashpoints in the economy, people’s habits for the use of credit cards, and consequences such as the prudent level of liquidity for banks to hold.  In the short term, however, the money multiplier can deviate very considerably from its medium-term level.  For example, if households fear deflation, or if the banking sector is inadequately capitalised, then the money multiplier can fall dramatically, but temporarily.

So when we use QE to try to offset deflationary pressures and financial sector problems, we are raising the monetary base to try to maintain broad money growth in a period in which the money multiplier is temporarily low.  But then if we ultimately do enough to lead people to stop fearing deflation and for the financial sector to recover, we should expect a consequence to be that the money multiplier will snap back to its medium-term equilibrium level — like ketchup surging out of a bottle when we finally hit it hard enough, once enough money is finally printed the result will be a sudden surge of broad money growth.  Now, it could well be that when there is a sufficiently severe financial crisis, even the medium-term money multiplier falls — for example, the banking sector might increase its view as to the prudent level of liquidity to hold.  But even if that is true, it remains likely that the medium-term money multiplier is higher than the short-term multiplier.

The consequence is that once enough QE is done, we should expect broad money growth to begin to rise very rapidly — that will be a mark of success.  But the monetary authorities cannot know what the new medium-term money multiplier is going to be, so it will be simply impossible to calibrate withdrawal from QE precisely so as to deliver the desired broad money growth on exit from crisis.  Now of course the authorities will form estimates of the new equilibrium level of the money multiplier.  And they could use those estimates to try to withdraw QE so as to avoid overshoot broad money growth on exit.  But is that a strategy that would make any sense?  I think clearly not, because if they were to get their estimate wrong by over-estimating the new money multiplier, then they would withdraw too much QE too early, and the consequence would be strong deflationary pressure.  If, say, a couple of quarters into recovery in 2011 after a double dip we were to over-tighten and create deflationary pressures, that could be disastrous.

The correct policy is to err on the side of too much broad money growth, which we can mop up afterwards, tolerating a temporary surge in inflation.  Not because we want inflation, but because we cannot accept deflation.

However, contrary to certain overblown press reports, I’m not predicting anything that could remotely be described as “hyper-inflation”.  A CPI figure of over 6% is only marginally above the 5.2% figure reached in September 2008.  And a rise to 8% in interest rates to deal with that (implying RPI exceeding 10% because of the rise in mortgage interest payments) should not really be so unthinkable — only two years ago, interest rates were 5.75% and heading up (indeed, a number of commentators were predicting 7% interest rates even at that stage).  In that sense, one way to think about my scenario is as a return to the status quo ante the crisis.

Indeed, if I am wrong and there is no double dip, then the next most probable scenario is probably robust recovery, a la the UK in the 1990s.  If that is what transpires, then by 2012 we will be in the third year of robust recovery-driven growth, and interest rates similar to those of 1996/7/8 (peaking in 1998 at 7.5%) should be very much what one expects.  The “neutral” interest rate for the UK is of the order of 5.5%.  If we are in the third year of boom by 2012, households should certainly be able to tolerate interest rates higher than those of 2008 (after all, not many have taken out new mortgages since 2008, and not many were bankrupted by the 5.75% interest rates seen then).

(P.S. For what it’s worth, the reports of 14% mortgage rates in some papers were nothing to do with me, and I do not agree with that number — 8% interest rates are more likely to be associated with 10% mortgage rates in my scenarios.)

Since I’m not proposing that households will be expanding their consumption rapidly, even in response to QE, where is the growth coming from?  The answer is: investment.  There are four key reasons why.

-          The first is extraordinarily low interest rates.  I believe that, if there is a double dip, interest rates will stay very low through 2011 even though growth will be robust for most of the year.  Interest rates will be so low that investors will be able to afford to have many projects go bad and yet still make money.  Very loose monetary policy can stimulate investment very strongly, and investment flows can be very volatile — capital formation grew 25% in the Heath-Barber boom from 1972Q1 to 1973Q1. 

-          The second reason is that the credit crisis will come to an end.  Europe Economics, as part of its analysis for Ofwat of the cost of capital of the water industry (see p83ff), analyzed the progress of past periods of financial crises (as it happens, I led the Europe Economics team).  The analysis there suggested that periods of elevation in corporate bond spreads in severe financial crises last around 4 years.  So, given that the current crisis began in 2007, we should expect corporate bond markets to decisively normalise (as opposed to there being periods in which spreads fall but a continuous risk of their spiking up again any time) in 2011.  That means that the corporate sector should find it materially more attractive to invest from mid-2011.

-          Thirdly, the corporate sector has already markedly reduced its leverage (with non-financial corporation loans down from their 2008Q4 peak of nearly 25% of GDP to 21% by 2010Q1), and by mid-2011 should be well-positioned to take on more debt if opportunities arise, though it may prefer to issue new equity.

-          Fourthly, if I am correct to anticipate money-growth-driven inflation on exit from the crisis, then investors will want to hold real assets.

As I emphasized previously, there are huge uncertainties here.  I do not pretend that my scenario is certain to materialize — there could be deflationary slump; growth could be steadier and more robust than I envisage, allowing earlier interest rate rises; the inflationary spike could be much greater than I expect.  Furthermore, it is perfectly possible that the scenario I paint will materialise, but will do so only a year or more later than I suggest — forecasting both the scale of economic events and their timing is notoriously difficult.  But CPI and interest rates in two years’ time only a percent or two higher than they were two years ago should surely not be regarded as so unlikely as to be out-of-the-question and the fact that a prediction that they could occur counts as front page news must surely be an indication that we had previously lost our sense of perspective.