Is QE losing its effectiveness? 1. An overview

Posted by Tim Congdon in News Archive | 0 comments

Has QE lost its bite? Do large-scale asset purchases by the state no longer boost the economy? Contrary to the impression given by today’s media, this is an ancient question in economics. It flared up in the 1930s and now seems to be doing so again. (I refer to the 1930s, because it was the disagreement between Keynes and Ralph Hawtrey, in effect the UK government’s top economic adviser and exponent of the so-called ‘Treasury view’, that provoked Keynes to write The General Theory. Hawtrey claimed that large-scale asset purchases by the state [i.e., QE] could boost an economy with no need to unbalance the budget; Keynes denied this, and urged increased government expenditure and budget deficits.) A good statement of a common argument was  made  by  Jeremy  Warner,  the  influential  commentator  on  The  Daily Telegraph, in a piece on 19tSeptember. Although a supporter of the ‘initial bout’ of QE in early 2009, he suggested that ‘now QE seems to have hit the law of diminishing returns; it appears pretty much ineffective in getting the economy going again’.

QE is to be understood as ‘deliberate action by the state (the government or the  central bank or the  two acting in concert) to increase the  quantity of money’. In other words, the claim that QE is subject to ‘diminishing returns’, and is therefore ‘ineffective’, is tantamount to the claim that increases in the quantity of money have a diminished (or even little or no) relationship with nominal national income. This claim is false. It has indeed been thoroughly refuted by the events of the Great Recession. The Great Recession is readily interpreted within the standard framework of monetary economics, in which the demand to hold money is a stable function of a small number of variables, including income, and changes in the quantity of money alter the equilibrium level of national income. The note below sets out some relevant evidence from UK experience in the last few years.

Claims of QE’s ineffectiveness

Famously, monetary policy is supposed to be ‘boring’. But, unfortunately and perhaps even infamously, commentators on monetary policy have to find something interesting to say. Despite the constraints imposed by the very technical subject matter, they must be as exciting as possible. Since the introduction of ‘quantitative easing’ in early 2009 the commentariat has launched a variety of polemical missiles against it. At one extreme is Liam Halligan, with comparisons between QE in the British context during the Great Recession and the money printing for so long endemic in Latin American  banana  republics.  As  I  have  explained  in  several  places,  Halligan  does  not  have  an organized theory of the relationship between money and national income, while he seems more concerned to maximize the noise made by his weaponry than to specify and hit a target. More recently, a quite different claim has been put forward. Whereas Halligan sees QE as having the power to generate a catastrophic hyperinflation, the latest contention is that it has ceased to be able to affect anything much.  Specifically, it has been said that QE is subject to ‘diminishing returns’ and is now less effective as a means of ‘boosting demand’. (There is a general acknowledgement that QE was needed in early 2009, to avert another Great Depression, and that it did have favourable effects in boosting the UK economy at that stage.) Jeremy Warner set out this point of view on 19th September in his often excellent column in The Daily Telegraph. Perhaps more surprisingly, one of the most influential figures in UK policy-making circles, Professor Charles Goodhart of the London School of Economics, has proposed a similar argument in an academic paper.

This note presents contemporary evidence against the doubters, but a few preparatory words are needed to situate the current debate in the evolution of macroeconomic thought. Until the 1930s the notion of ‘macroeconomic policy’ had not been developed properly. Decisions were taken by central banks (which had emerged as a type of institution separate from commercial banks only in the late 19th century) usually by appeal to a rule of thumb of some sort (‘the Bank rate tradition’ in the UK), but sometimes with a meaningful theoretical rationale. By the 1920s the dominant strand of monetary analysis was ‘the quantity theory of money’, particularly in the version espoused by Irving Fisher in his 1911 The Purchasing Power of Money. Fisher’s ideas were further amplified at Cambridge, England, with the emergence of a school of thought in which national income was seen as being determined by the quantity of money (understood as meaning, in practice, ‘the quantity of bank deposits’).

When the downturn that led to the USA’s Great Depression started in 1930, Keynes – who had reviewed The Purchasing Power of Money for The Economic Journal – used the closing pages of his new book The Treatise on Money to advocate ‘monetary policy à outrance’. The meaning of this phrase – translated as ‘monetary policy to the uttermost’ – was that the state should purchase assets on a large scale from the private sector, in order to increase the quantity of money and to lower ‘the rate of interest’ (i.e., bond yields). Ralph Hawtrey, who worked at the Treasury from 1904 to 1945 and was, in effect, the Treasury’s chief economist, took up the theme in a book called Trade Depression and the Way Out. A champion of sound finance and a strong believer in a monetary theory of ‘the trade  cycle’  (i.e.,  of  national  income  determination),  Hawtrey  was  opposed  to  public  works expenditure and budget deficits. (Like most Treasury civil servants of the day, he regarded public works expenditure as liable to be wasteful.) In his 1936 General Theory Keynes moved beyond the argument in his 1930 Treatise. He suggested that in certain circumstances monetary policy might be ineffective. If so, it was the government’s job to spend more on its own account, even if the result were a large budget deficit. (However, The General Theory in fact said next to nothing about how precisely so-called ‘fiscal policy’ should be organized.)

As discussed in a note attached to this e-mail, ‘monetary policy à outrance’ and QE are more or less equivalent. In this sense Keynes was the inventor of QE, while the exchanges between Keynes and Hawtrey anticipated the contemporary debates. The discussion of QE in the Great Recession echoes the discussion of monetary policy à outrance in the Great Depression. (I want to make this point partly to correct the widespread impression that QE is somehow unprecedented, new-fangled and experimental, and that the phrase ‘quantitative easing’ is a cloak for a large-scale conspiracy to undermine modern society, which is what some of its detractors seem to think. In an article in The Daily  Telegraph  of  7th   November  by  James  Delingpole  ‘Ben  Bernanke’s  quantitative  easing’  is grouped with ‘welfare’ and climate-change measures as a symptom of the never-ending growth of ‘Big Government’.  On the contrary, QE is an alternative to so-called ‘expansionary fiscal policy’, which often has the result of increasing the size of the state.   Delingpole describes himself as ‘a writer, journalist and broadcaster who is right about everything’, as well as being ‘the author of numerous fantastically entertaining books’. Quite so.)

QE is therefore not new. It has long been understood that the central bank can undertake expansionary open market operations and that such OMOs are effective because they ‘increase the quantity of money’, ‘reduce “the rate of interest”’ or whatever. Admittedly, the range of expansionary OMOs available to ‘the authorities’ (i.e., the government and the central bank) is wide and diverse, and many textbook treatments of the subject are incomplete and unsatisfactory. I am not here going to explain how QE increases the quantity of money, as I have already written much about it elsewhere. It is sufficient to notice that, when the state makes payments to the private sector’s bank accounts in excess of the payments that the private sector makes to the state, the sum of all the balance in the private sector’s bank deposits (and hence the quantity of money) rises. In other words, I am taking it for  granted  that  worries  about  QE’s  effectiveness  focus  on  the  supposed  unreliability  of  the relationship between the quantity of money and nominal national income. I am not interested in assertions that QE is unable to boost the quantity of money, because – in my view – such assertions are silly.

Money and nominal national income since 1997

There is insufficient room here to elaborate the monetary theory of national income determination. (Subscribers may be referred to part five of my 2011 book Money in a Free Society.) However, the heart of that theory is simple enough. Agents are understood to have a demand to hold money balances, just as they have a demand for the services of other assets (such as those provided by the housing stock or transport equipment). Further, the demand function for money can be represented as having a small number of arguments, including the attractiveness of money relative to non-money assets and the technology of using money, but with income invariably viewed as the dominant argument. If the arguments in the money demand function other than income are assumed to take constant values, a property of the standard money demand function in economic theory is that money and agents’ desired level of income rise or fall equi-proportionally. It follows that – if, for example, the quantity of money is doubled – agents’ ‘equilibrium’ level of income (i.e., national income and expenditure in the aggregate) ought also to double. Equi-proportionality may not fully hold in practice for at least two reasons,

  1. The value of the arguments in the money demand function may change, altering the desired ratio of money to income, and
  2. The economy may not – after a doubling of the quantity of money – see an immediate doubling  of  national  income,  since  a  period  of  ‘disequilibrium’  (in  which  spending patterns and wealth portfolios are being adjusted) may persist for months, quarters or even years.

At any rate, this gives us a framework for analysing the evidence on the relationship between money on the one hand, and income and expenditure on the other. Sure enough, the ratio between money and expenditure may change, but such changes may be consistent with the underlying theory as long as they can be interpreted as due to

  1. changes in the non-income arguments in the money demand function, and
  2. the working-out of a period of monetary disequilibrium.

Let us now have a look at some numbers. The chart below shows the annual changes in two series, the quantity of money (as measured by the M4x aggregate) and nominal GDP. The series are of quarterly data and go back to the end of 1997.Money and nominal GDP in the UK from 1997

From a monetary perspective, the period (of almost 15 years) falls into three phases.

  1. The first phase ran from 1997 to 2004, when money growth averaged about 6% – 7% a year and was fairly consistently near that figure, despite some jerkiness at the start. Money growth was higher than that of nominal GDP, but the years were of reasonable macroeconomic stability. As we now know, they could be viewed as the second half of the Great Moderation, which is generally regarded as having begun in the UK in late 1992.

From mid-2004 to the autumn of 2007 money growth accelerated and then slowed slightly. At its peak (in the third quarter of 2006) the annual rate of growth of M4x was just under 12%, the highest rate of money growth since the Lawson boom of the late 1980s. The slowdown in money growth was not well-defined until 2008. For most of 2007 the annual rate of money growth was in the 9% – 10% area. This was a strong period for asset prices, with above-trend growth in demand and output.Money in the UK from 1997

  1. The third phase is from late 2007 to today. It needs to be remembered that, because the chart is of annual changes, all readings cover four quarters, each of which may behave quite differently from the others. The chart shows that the annual rate of money growth collapsed from 9.0% in Q4 2007 to 2.3% a year later, and was then only slightly positive in both 2010 and 2011. The turning-point was in fact in the autumn of 2007, coinciding with the Northern Rock crisis and, perhaps more fundamentally, banks’ re-appraisal of their expansion strategies as it became clear that the wholesale inter-bank market was no longer open. The contrast between the double-digit annual growth rates of money until late 2007 and the low-single- digit annual growth rates thereafter is obvious and indeed eloquent. We have here an apparently  compelling  mono-causal  explanation  of  the  Great  Recession.  However,  on  a quarter-by-quarter basis, or even on a year-by-year basis, the relationship between money growth and nominal GDP in the Great Recession itself is unimpressive.

I will leave it to the reader to decide whether money trends have anything useful to say about the fluctuations  in  nominal  GDP  growth  in  this  15-year  period.  The  best  bit  of  evidence  for  the importance of money was in fact in the downturn from 2007 to 2009. Money growth collapsed and that collapse preceded, by a short period, a similar one in the growth of nominal GDP. In early 2009 only a fool could have denied that the very low rate of money growth then being seen was irrelevant to the economy’s plight. That gave plausibility to the case for QE, which was adopted in March 2009.Money and nominal GDP in the UK in the last major downturn

The implementation of QE in spring 2009 was followed by a big rebound in the economy, just as its advocates envisaged. The last two years have been ‘of recession’, but only in the sense that demand and output have struggled to expand. Employment has increased and, strange to say, the level of corporate and personal bankruptcies has been modest compared with the recession of the early 1990s. However, as noted, the relationship between money growth and nominal GDP in the Great Recession has in fact been unimpressive. (See the chart below.) Why has the assumed link between money and nominal GDP appeared not to survive the last three years?

Recent uneven relationship between money and GDP

Since early 2009 the relationship between money and nominal GDP has been uneven, to say the least. Despite the money-creating effect of the QE operations, M4x growth slowed in 2009, and the quantity of M4x actually fell in late 2010 and early 2011. By contrast, nominal GDP increased in late 2009, and its rate of rise (4% – 5% a year) in 2010 and 2011 was not much less than that seen for most of the Great Moderation. On the face of it, the money/nominal GDP relationship has ‘broken down’ yet again. However, this conclusion would be too hasty. Supporters of the monetary theory of national income determination rightly emphasize that the relationship is medium-term and rather imprecise in nature. In the short term (i.e., the periods of one to two years for which most macro forecasts are prepared) numerous other influences buffet demand and output.Money and nominal GDP in the UK during the Great Recession itself

The stagnation of M4x in the two years to mid-2011 may seem peculiar, since this came after the big supposed money-creation effects of the QE operations. However, the explanation is simple enough, that banks were obliged to shrink their risk assets because of the wave of new regulations to which they were subject. As they shed loans and securities under regulatory pressure, their balance sheets – and hence their deposit liabilities (i.e., the quantity of money) – failed to grow. Nominal GDP nevertheless moved ahead in 2010 and 2011, partly because of the positive effects of non-monetary influences (i.e., the strong global recovery of 2010 and the dynamics of the inventory cycle), and partly because the dramatic fall in interest rates altered the relative attractiveness of money and non- money assets. Because money was less attractive to hold (compared, say, with corporate bond funds), nominal GDP could and did growth faster than the quantity of money.

For all the problems it is worth emphasizing that – over the period of the downturn and the Great Recession combined, i.e., from mid-2007 to today – the quantity of money and nominal GDP have in fact risen at not dissimilar rates. Let us take it that the last quarter of the Great Moderation was Q2 2007, and so the period from Q2 2007 to Q3 2012 constitutes the downturn and the Great Recession. In the 21 quarters from Q2 2007 to today, the M4x quantity of money rose by 13.5% (i.e., at a compound annual rate of 2.4%), while nominal GDP rose by 11.8% (i.e., at a compound annual rate of 2.1%). The medium-term similarity of the growth rates of money and nominal GDP has held in the recent past, despite the wobbles in shorter run periods of years, hinting once again at the validity of the monetary theory of national income determination.

Conclusion: QE remains effective, because the monetary theory of national income determination is valid

The last paragraph is a challenge to sceptics of the monetary theory of national income determination. If an analyst had known in mid-2007 that, over the next five years, the rate of money growth would be a mere 2½% a year, what would he or she then have said about the likely associated rate of increase in nominal GDP? Economists who believe that ‘money matters’ (i.e., who believe in the monetary theory of national income determination) would have made a confident forecast that this period would also see the lowest rates of increase in nominal GDP since the 1930s, probably in association with a recession. They would have been right. Economists who pooh-pooh money and banking as influences on macroeconomic outcomes would have seen no significance in the imminent collapse in money growth to very low levels. They would have been focussed instead on fiscal policy, the exchange rate, real wage movements, etc., Uncle Tom Cobley and all, just as – for example – the National Institute of Economic and Social Research did in 2007 and still does today. Admittedly, in the decade or so to 2007 the quantity of money grew faster than nominal GDP. But this was in line with a trend which dated back to the early 1980s and was to be explained by the higher returns on bank deposits. These higher returns, which made money more attractive to hold, had been made possible by the competitive environment and technological improvements (the spread of credit cards, computerisation, etc.) of the 35-year period from the Competition and Credit Control reforms of 1971.

The facts of the relationship between money and nominal GDP since 1997 are consistent with the standard monetary theory of national income determination. The theory implies that a medium-term relationship holds between the rates of change of money and nominal GDP, and that the equilibrium level of nominal national income is a function of the quantity of money. A necessary consequence is that – if the state engineers deliberately large changes in the rate of money growth (as it undoubtedly can do) – it is taking powerful actions to affect macroeconomic outcomes. Claims that ‘QE has become ineffective’ are contrary to evidence and at variance with a large body of well-established theory.

Next week I will look at some of the particular ‘channels’ whereby changes in the quantity of money have affected the behaviour of companies and financial institutions in the UK during the Great Recession. That will be intended as a further rebuttal of those who deny the ability of monetary policy operations, such as QE, to help in macroeconomic management.

Tags: , , , , , , , , , ,

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

Time limit is exhausted. Please reload CAPTCHA.

News Archive

Article Archive