‘The government’s real case is that expansionary monetary policy will offset any contractionary influence of the Budget.’
Financial Times, 20101
‘The problem with QE is that it works in practice, but it doesn’t work in theory.’
Ben Bernanke, 20142
‘While monetary policy . . . provided the necessary emergency medicine after the financial crash, we have to acknowledge there have been some bad side-effects. People with assets have got richer. People without them have suffered.’
Theresa May, 20163
‘I find it hard to reach the conclusion that, over a longer time-frame, the outcome of our policies has been – or will be – to redistribute wealth and income in an unfair or unequal way.’
Mario Draghi, 20164
The withdrawal of fiscal stimulus in 2010 left only one expansionary tool – monetary stimulus. Quantitative easing (QE) – buying up government debt in order to put more money in the hands of private business – was the inferior substitute for fiscal expansion, and the offset to fiscal contraction. This is the straightforward economics of the matter. It may be that politically it was the only thing that could have been done. But no one should pretend that it was superior. The chosen vessel for watering parched economies was much more leaky than the rejected alternative.
Throughout the Keynesian ascendancy, the Bank of England had demanded that it be given ‘operational independence’ to prevent democratic governments from inflating the money supply. In 1998 the Bank finally got what it wanted.
The Bank of England Act mandated the Bank of England: ‘(a) to maintain price stability, and (b) subject to that, to support the economic policy of Her Majesty’s Government, including its objectives for growth and employment’.5 The Bank’s Monetary Policy Committee (MPC) was empowered to set the level of the official interest (‘base’ or ‘policy’) rate6 independently of Parliament, a break from post-war practice when the policy rate was determined by the government: Margaret Thatcher, for example, used to veto rises in interest rates on the ground that it would ‘hurt our people’. In the new regime, the Bank would control inflation by varying Bank Rate. Inflationtargeting was from the outset ‘conceived as a means by which central banks could improve the credibility and predictability of monetary policy. The overriding concern was . . . to reduce the degree of uncertainty over the price level in the long run because it is from that unpredictability that the real costs of inflation stem.’7
Having learned from the experience of the failed monetarist experiment of the 1980s, the Bank of England did not directly target money, yet ‘for each path of the official rate given by the decisions of the MPC, there is an implied path for the monetary aggregates’.8 Thus the monetary aggregates remained the most important indicator for monetary policy. The MPC’s preferred measure for this was broad money (M4), which included bank deposits. In addition, the Bank retained its traditional role as lender of last resort, a role denied to the European Central Bank.
Bank Rate, less familiarly the ‘base rate’, is the interest rate or ‘price’ that the central bank charges for lending money to member banks. The theory is that a change in the base rate pushes the yield curve upwards or downwards. It is immediately transmitted to the interbank lending rate. Banks will then adjust their own lending rates, both short-term and long-term. This will affect how much income is saved and invested. In 1930 the Bank of England had denied that it had such power over commercial lending rates, and uncertainty remained about the impact of the short-rate on the long-rate.9
The supposed transmission mechanism from the base rate to the level of spending and prices in the economy can be summarized by Figure 38. The channels work as follows:
• Market rates: changes in the official rate affect the structure of market rates.
• Asset prices: ‘Lower interest rates can boost the prices of assets such as shares and houses. Higher house prices enable existing home-owners to extend their mortgages in order to finance higher consumption. Higher share prices raise households’ wealth and can increase their willingness to spend.’10
• Expectations/confidence: Changes in the policy rate influence expectations about the future course of the economy. Expectational effects are unpredictable. Take, for example, a rise in the policy rate. On the one hand, this might be taken as a sign that the central bank wishes to slow down the growth of the economy to stop it from ‘overheating’, dampening expectations of future growth. But it could also be interpreted as a sign that the economy is growing faster than the central bank had previously predicted, which might increase confidence in the economy.12
Figure 38. The transmission mechanism of monetary policy11
• Exchange rate: An unexpected decrease in the interest rate relative to overseas would give investors a lower return on UK assets relative to their foreign currency, tending to make sterling less attractive. That should lower the value of sterling, increasing the price of imports and lowering the price of exports. At first glance, this would appear to increase UK output, but the effects of exchange rate changes can be unpredictable. For example, if the change in export and import prices have a negligible impact on demand (in technical terms, if UK import demand and demand for exports are ‘price inelastic’), then output will fall.*
The Bank’s approach can be captured by the Taylor Rule (see Appendix 7.3): when inflation is above target, this signals that spending is growing faster than the volume of output being produced, so the Bank of England should increase the base rate to make savings more attractive relatively. Conversely, if inflation was below target, the base rate should rise.
The framework of policy was Wicksellian rather than Friedmanite: bank rate should be set to achieve the target rate of inflation. But ‘flexible inflation targeting’ incorporated the New Keynesian feature of allowing for (small) shocks to Wicksell’s ‘natural’ rate. The policy framework also emphasized the importance of policy rules to anchor expectations. In normal times the Bank would ‘set interest rates such that expected inflation rate in two years’ time is equal to the target’. But in the face of a shock its aim should be to ‘bring inflation back to target over a period of more than two years and explain carefully why the heuristic has changed’.13 In this way the Bank could adapt its policy to changing circumstances and evolving knowledge, ‘so that the policy regime as a whole is robust to changing views about how the economy works’.14 At least, that was the theory. The contradiction between setting a policy rule to anchor expectations, and explaining why it could not be relied on, was never resolved.
The Bank’s preference as between inflation and output can be captured by the following ‘loss function’:15
Losst ≡ (πt − π*)2 + λ(yt)2
Here π represents current inflation, π* the inflation target, and so πt – π* gives the gap between desired and current inflation. Similarly, yt represents the output gap, λ is a term representing the Bank’s concern with output. If λ = 0, the Bank does not care about output and will attempt to curb inflation at all costs. If λ is high, the Bank might tolerate higher inflation if this avoids a fall in output and employment. Finally, the inflation and output gap terms are squared to show that (a), deviations from target inflation and output in either direction are equally undesirable and (b), large deviations are much less desirable than smaller ones.16
A much-praised feature of the British arrangements was the symmetrical nature of the inflation target.17 Policy was set to avoid the evils of both inflation and deflation. An inflation rate expected to run above target would indicate that aggregate demand was running ahead of aggregate supply; an inflation rate below target would indicate a shortage of demand relative to supply. Targeting the inflation rate was thus a way of balancing aggregate demand and supply, with the inflation target replacing the Keynesian full employment target. This reflected Milton Friedman’s view that unemployment would normally be at its ‘natural’ rate if prices were kept constant. Varying bank rate to meet a pre-set inflation target was the monetary version of fiscal fine-tuning.
This pared-down version of macroeconomic policy rested on the view that the expectation of stable inflation (together with ‘prudent’ fiscal policy) would cause the real economy to be stable, barring large shocks. Certainly the Great Moderation years saw a decent correlation between growth and low inflation, in apparent vindication of central bank policy.
But whether the anti-inflation commitment was the main cause of low inflation is doubtful. There was a large downward pressure on prices following the entry of hundreds of millions of low-wage workers from China, East Asia and Eastern Europe into the global labour market.19 Mervyn King acknowledges the help from this factor when he talks about a ‘nice’ environment for monetary policy.20
Figure 39. Output growth and inflation in the advanced economies during the Great Moderation19
But, with a rogue elephant in the corner, the whole system is liable to crash down, and this is what happened in 2008–9. The rogue was the financial sector. Deluded by their apparent success in keeping inflation low, policymakers ignored the troubles brewing in the banks. With the unexpected collapse of the financial system in 2008–9, monetary policy faced a challenge not seen since the Great Depression.
The Bank of England was slow to respond to the growing signs of banking crisis. In Howard Davies’s words, ‘[it] lectured on moral hazard, while the banking system imploded round it’. Unlike the US Federal Reserve, the European Central Bank also worried about ‘imaginary inflationary dangers’.21 But following the collapse of Lehman Brothers in September 2008 the policy rates of the main central banks were rapidly slashed towards zero.
Figure 40. Cutting interest rates: central banks’ base rates22
This was the traditional response. With the economy still in free fall, interest rate policy could do no more. An extra tool was needed. Alistair Darling, Britain’s Chancellor of the Exchequer, announced on 18 January 2009 that the Bank of England would set up an asset purchasing facility (APF), which would be ‘useful for meeting the inflation target’. Quantitative easing had arrived.
Two days later, the Governor of the Bank, Mervyn King, explained the thinking behind it:
The disruption to the banking system has impaired the effectiveness of our conventional interest rate instrument. And with Bank Rate already at its lowest level in the Bank’s history, it is sensible for the MPC to prepare for the possibility . . . that it may need to move beyond the conventional instrument of Bank Rate and consider a range of unconventional measures. They would take the form of purchases by the Bank of England of a range of financial assets in order to expand the amount of reserves held by commercial banks and to increase the availability of credit to companies. That should encourage the banking system to expand the supply of broad money by lending to the private sector and also help companies to raise finance from capital market.23
The theoretical case for QE was built on the idea of a liquidity trap.
The situation which produces the ‘trap’ is one in which the expected rate of return on investment (Wicksell’s ‘natural rate of interest’) is lower than the lowest rate of interest banks are willing to charge for loans. The zero bound is the limit of what interest rate policy can achieve to lower commercial banks’ lending rate. At the zero lower bound (ZLB) the demand for money to hold becomes perfectly interest elastic (expands without limit).* This is because the sense of security from holding cash, even at zero interest, trumps the cost of forgone expected financial returns. Once the zero lower bound is attained, central banks must turn to other means to lower loan rates in the market.
QE was called unconventional monetary policy because the conventional pre-crash policy of controlling credit by price was no longer available. As a consequence, central banks had to gamble with the Fisher–Friedman version of monetarism which had broken down in the 1980s. Willy-nilly, central bankers became quantity theorists.
The Fed was quickest off the mark. The need for large-scale QE was the lesson Ben Bernanke drew from the Friedman and Schwartz story of the Great Depression. Shortly before he became Chairman of the Federal Reserve Board in 2006, Bernanke wrote: ‘By allowing persistent declines in the money supply and in the price level, the Federal Reserve of the late 1920s and 1930s greatly destabilized the U.S. economy.’24 Equipped with this historical lesson, Bernanke and most other central bank governors were determined to avoid this mistake when the crisis hit in 2008. The Fed announced its first asset purchase programme in November 2008.25 ‘Extraordinary times call for extraordinary measures,’ declared Bernanke.26
In its initial round of purchases (QE1), between November 2008 and March 2010, the Fed bought $1.25 trillion of mortgage-backed securities (MBS), $200 billion of agency debt (issued by the government-sponsored agencies Fannie Mae and Freddie Mac) and $300 billion of long-term Treasury securities, totalling 12 per cent of the US’s 2009 GDP. Its second round of purchases (QE2) – $600 billion of long-term Treasury securities – ran between November 2010 and June 2011, and its third round (QE3) started in September 2012 with monthly purchases of agency mortgage-backed securities.27 The programmes were wound up in October 2014, by which point the Fed had accumulated an unprecedented $4.5 trillion worth of assets,28 equivalent to just over a quarter of US GDP in 2014. In the composition of its purchases, the Fed, as we shall see, was more adventurous than its British counterpart.
In the UK, QE has come in three bites. The Bank of England injected £200 billion of electronic money into the British economy between March 2009 and January 2010 (QE1), and £175 billion between October 2011 and November 2012 (QE2 and QE3), making £375 billion in all, or 22.5 per cent of 2012 GDP. The majority of its purchases were of highly liquid gilts, though the Bank also bought a small amount of commercial paper and corporate bonds. After the Brexit vote in June 2016, the Bank of England decided to resume QE in August.
For the ECB, ‘repo’ operations, known as LTROs or long-term refinancing operations (designed to refinance banks), remained its main source of balance-sheet expansion until it started its asset purchase programme in 2015.29 That is, it was bank salvage, not monetary policy. In 2012, the ECB President, Mario Draghi, promised to do ‘anything it takes’ to preserve the euro. This pledge, which was opposed by Jens Weidmann, President of the German Bundesbank, saved the European Monetary Union. In March 2015, the ECB started to buy €60 billion of euro-area public sector debt per month. A year later, this monthly amount was increased to €80 billion and high-grade corporate bonds became eligible for purchase. The amount dropped back down to €60 billion in April 2017, and to €30 billion in January 2018. In July 2017, the ECB held assets to the value of 40 per cent of 2016 Eurozone GDP.30 For each of the three central banks, the scale of their balance-sheet expansion was unprecedented.31
Three strong arguments backed the new programmes. The first was that they were simply an extension of the ‘open-market operations’ technique practised by all central banks as part of their normal money-market management. Open-market operations (OMOs) were the means by which the central bank supplied the banks’ marginal liquidity needs on a daily basis, either by buying or selling government securities or by means of ‘repo’ transactions, so as to keep the inter-bank lending rate close to the policy rate. However, QE was ‘unconventional’ in the sense that the technique had never been used outside Japan in a situation in which the total supply of liquidity had dried up. Nevertheless, the fiction persisted that QE did not mark a permanent expansion of the money supply, since the bonds which were bought would be sold again as soon as the economy was back to ‘normal’.
The second argument was pragmatic: fiscal policy had been ‘disabled’ by the huge expansion of public deficits in the first six months of the crisis, and conventional monetary policy by the zero lower bound. QE was the best of a waning number of options.
The third argument was ideological. Monetary expansion was preferable to public investment, since it avoided a ‘government role in the allocation of capital’.32
Tim Congdon explains the expected real balance effect by invoking Fisher’s Santa Claus: agents finding themselves with excess money balances at the existing rate of inflation spend the excess by increasing their purchases. The cumulative attempt of recipients to get rid of the extra money raises all prices to a level at which the desired ratio of money-holding to expenditure has been restored. Thus a stable demand for real balances is brought into equilibrium with the increased supply of money through a rise in nominal income. How this rise will be shared between output and prices will depend on the size of the output gap.33
How much extra money will Santa Claus need to spray round the community to achieve a given inflation target? In the Fisher theory the answer was given by the money multiplier: the amount of new bank loans which can be created by an increase in reserves (‘base money’) in a fractional reserve banking system. If the reserve requirement is 10 per cent, an injection of £1,000 will enable additional loans of £900, leading to additional spending and deposit creation, with the total of new money summing to a multiple of the original injection.34 If the money multiplier is known, then so will be the effect of any given amount of QE on nominal income (output plus prices). However, if the money-multiplier mechanism is leaky, the amount of new money needed to raise nominal income to a desired level is unknown. For example, the excess could ‘automatically be extinguished through the repayment of bank loans, or what comes to the same thing, through the purchase of income yielding financial assets from the banks’, leaving the quantity of money (deposits) the same.35
Keynes had pointed out the problem when he warned in 1936 that, ‘if . . . we are tempted to assert that money is the drink which stimulates the system to activity, we must remind ourselves that there may be several slips between the cup and the lip’.36 He identified two such slips or ‘leakages’ from the circular flow. First, creating extra bank reserves would have no influence on spending if ‘the liquidity-preferences of the public are increasing more than the quantity of money’.37 In other words, the effect of money on prices depended on the amount spent, not on the quantity created. In his earlier Treatise on Money he had identified another slip. Even if demand were to be stimulated by cash injections, it might not be demand for currently produced output. Recipients of the new money might use it to buy existing assets, such as stock exchange securities or real estate or Old Masters.38 In this event QE would have to rely on an indirect wealth effect on consumption to achieve its desired impact on nominal income.
It was considerations of this kind that led Keynes to conclude that the only secure way to get new money spent in a slump was for the state to spend it itself.
How did the Bank of England expect QE to work in practice? The answer is, it didn’t quite know. Its chosen route, in the Bank’s own words, was ‘the creation of central bank reserves . . . by buying outright from the private sector assets that have either a longer duration and/or higher credit risk than the corresponding liability’.39 In non-Bank speak, it would create riskless cash reserves for the banks by buying their riskier assets.
What, in the Bank’s view, would this achieve? In its earliest presentations, the Bank of England specified two main transmission channels from these reserves to spending. The first was the ‘portfolio substitution’ channel; the second, the ‘bank funding’ channel. They are illustrated in Figure 42 below.
The bank funding or, more familiarly, lending channel was a straightforward substitution for the inability of the Bank to get its base rate of interest below zero. As a result of QE, commercial banks would hold significantly higher levels of reserves. This would induce them to lower the interest rates they charged on loans. This would increase their loan portfolios. The spending of the loans would expand the economy.
Figure 42. Four key monetary debates40
In practice, the Bank of England didn’t much believe in this channel, and believed in it even less after a short period of experience. Only 30 per cent of government securities were bought from banks; the rest from non-banks. The reason is understandable. Given the impairment of banks’ balance sheets, and the collapse in the confidence of borrowers, there was not much hope for a rapid increase in bank lending. Therefore QE1 was explicitly designed to get round the banking system, not through it.
The Bank of England (like the Fed, but unlike the ECB) put its main hopes in portfolio rebalancing/substitution. This was to be activated by buying government bonds from private investors, like pension funds and insurance companies. As the Bank put it:
Insofar as investors regard other assets – such as corporate bonds and equities – as closer substitutes for government bonds than money, we might expect them to re-balance their portfolio towards these assets if their money holdings are boosted by temporary bond purchases . . . This would tend to put upward pressure on the prices of those assets.41
Keynes had thought that if bond yields fell too low, people would prefer to hold cash than buy bonds. But the Bank reasoned that a policy aimed at reducing the excess demand for bonds would cause investors to switch not to cash but to financial assets like equities, which promised higher, if riskier, returns. The increase in the paper wealth of the new asset holders would encourage them to spend more.* In other words, the Bank, following Friedman’s lead, implicitly jettisoned the speculative demand for money from Keynes’s liquidity preference function. The desire for liquid assets might go up, but there would be no leakage from the circular flow of money.
As time went on, the Bank discovered extra channels. In particular, it started to attach increasing importance to the effect of its announcements in activating the required responses. At first it hoped to take advantage of their ‘surprise’ effect. When it discovered that the surprise soon wore off, it started to emphasize signalling and ‘forward-guidance’. When the Bank acts, its actions give clues to what it will do in the future, and these clues are signals; ‘forward-guidance’ is an explicit commitment to act in a certain way under specified conditions. In its most explicit form, the forward-guidance channel works through policymakers making long-term commitments to keep interest rates exceptionally low. The policy boasts a placebo effect – self-fulfilling prophecies producing a recovery without undertaking the significant risks of expanding the central bank’s balance sheet.
Hence, the commitment to continue the low bank rate and asset purchases for a definite length of time was considered crucial to achieving the hoped-for effect of the policy, i.e. raising the inflation rate. Like similar pronouncements from the Treasury concerning time-limited deficit-reduction targets, signalling and forward-guidance were attempts to boost the credibility of the policy.
In 2013, Mark Carney, the new Governor of the Bank of England, signalled the Bank’s intention to keep bank rate at its then current level of 0.5 per cent until unemployment had fallen to 7 per cent.
The Bank can only directly control the short-term interest rate. But this rate has already been cut to the lowest level that the Bank feels comfortable with . . . another way for the Bank to support the economy has been to offer this indicator, by which companies and mortgage borrowers can estimate for how long such low interest rates may be around for in terms of months or years. Forward guidance is thus a way of converting low short-term interest rates into lower long-term interest rates. The thinking is that if the High Street banks can be convinced that they will be able to borrow overnight from the Bank of England at just 0.5% for many nights – indeed many months or years – to come, then they will hopefully be willing to lend money out to the rest of us for the longer term at a commensurately lower interest rate as well.42
There is a trade-off between credibility and pragmatism. Bank Rate was kept at 0.5 per cent until August 2016, even though British unemployment had been below 7 per cent for the previous two years. However, commitments to keep a policy in place for a period of time cease to be credible if circumstances point to a change of policy. In October 2017 base rates started to come off the floor for the first time since the crisis began. How long it will be before they reach what is regarded as normal depends on the momentum of recovery, about which no one can be certain. However, it could be argued that the emergency short-term rate of close to zero set in the winter of 2008 is now well below the equilibrium rate for a recovered economy – its only effect being to sustain ‘zombie’ companies which should exit economic life.
It should be noted that the explicit purpose of the whole exercise was to raise inflation to its target of 2 per cent. In fact, the expectation of higher inflation was a crucial part of the mechanism for increasing spending: if households and firms expect prices to go up (or, equivalently, the real rate of interest to fall) they will increase their current purchases of goods and machinery to get them at a cheaper price. Who would not buy today, if they expect higher prices tomorrow? However, if higher prices were expected to boost investment, it was soon realized that, if this was achieved, inflation would depress consumption by increasing goods’ prices. As far as increasing output was concerned, raising the rate of inflation was a doubleedged sword.
How does one assess the achievement of QE? As with any assessment of policy, a fundamental problem lies in the difficulty, indeed impossibility, of isolating the impact of the policy from contamination by external factors. It is relatively easy to evaluate the impact of QE on financial variables such as interest rates, bond rates, stock exchange prices, and so on. But what is the effect of such changes on real GDP? There is no particular virtue in achieving financial targets as such. It matters not whether interest rates or asset prices go up or down, except in terms of their effects on output and employment. These financial events were simply transmission mechanisms to the real economy. If they fail to transmit recovery the policy is useless.
In Figure 43, the dark grey bubbles are what the authorities wanted to achieve through QE, while the effect of the medium grey bubbles is indeterminate. What they didn’t want were the light grey bubbles: for banks to sit on their reserves and not lend; and for investors to buy financial assets and not spend. There was clearly a risk of asset bubbles, but the Bank hoped that an asset boom would produce increased capital investment and consumer spending through a wealth effect. In this 2013 assessment of Britain’s experience of QE there were five light grey bubbles and only three dark grey ones.
This channel was supposed to work, in the first place, by depressing the yield of gilts. This would induce holders of gilts to switch to equities: ‘If QE successfully raised equity and corporate bond prices, we might expect firms to respond by making more use of capital markets to raise funds. In other words, there would be a positive effect of QE on the quantity of debt and equity raised, as well as its price.’43
Joyce et al. estimate that the first (£200 million) wave of the Bank of England’s asset purchases, from March 2009 to January 2010, reduced gilt yields by around 1 per cent, comparable to a 1 per cent reduction in short-term rates.45 Meaning and Warren (2015) estimate that the total £375 billion of QE reduced yields by around 0.25 per cent through the effects of increased supply of bonds alone (i.e. excluding expectational effects).46 This lowered borrowing costs throughout the economy. The fall in the cost of government borrowing, and interest payments on the national debt, improved the fiscal numbers, enabling budgetary policy to be somewhat looser than it would otherwise have been, given the commitment to austerity. And it lowered, at least temporarily, the cost of finance for companies, which had spiked dramatically in 2008–9.47 External MPC member David Miles believes that ‘a significant part of the fall in spreads on sterling corporate bonds is specifically linked to the Bank of England’s purchases of gilts’.48
Figure 43. Good and bad outcomes of QE44
Over the period from 4 March 2009 to 22 January 2010, the FTSE index rose by 50 per cent. But so did the Euro Stoxx 50 and the German Dax without the benefit of QE. Even the Bank of England, hardly a disinterested observer, concedes that it ‘would be heroic to attribute all of these gains to QE’.49 Nevertheless, ‘the evidence is consistent with [a portfolio rebalancing channel] effect’, though it is ‘impossible to know what would have happened in the absence of QE’.50 The equity and housing markets recovered much more quickly than the rest of the economy, but there is no way of showing how much of this was due to QE.
What is clearer is that QE failed to stimulate bank lending. While commercial bank reserves at the Bank of England (‘narrow money’) rose dramatically (from £30 billion in March 2009 to over £300 billion by the end of November 2013),51 the annual growth rate of bank lending fell from 17.6 per cent in February 2009 to negative in September 2010 (Figure 44). Theory tells us why. The private sector was increasing its saving. Banks were less willing to lend, and firms and households to borrow. The increase in central bank cash was not nearly enough to offset the huge rise in liquidity preference. Even Mario Draghi, the President of the ECB, was forced to admit that the monetary expansion would fail to unblock the bank lending channel if ‘banks . . . hold on to precautionary balances’.52
The consensus view is that the modest recovery in UK bank lending in 2012 was mostly due to the government subsidizing programmes like Funding for Lending and Help to Buy, which were fiscal rather than monetary policies. Funding for Lending was introduced in July 2012, and Help to Buy in April 2013. The first was ‘designed to incentivise banks and building societies to boost their lending to UK households and private non-financial corporations (PNFCs) . . . by providing funding to banks and building societies . . . with both the price and quantity of funding provided linked to their performance in lending to the real economy’.54 The second was designed to help people with as little as a 5 per cent deposit to buy a home; the government encouraged banks to approve such mortgage requests by guaranteeing the repayment of a percentage of the loan. But to this day bank lending is well below the historical average.
Figure 44. Growth in UK bank (M4) lending53
The failure of QE to revive bank lending has led to even more unconventional policy. In January 2017, Mario Draghi started taxing ‘excess’ reserves held by commercial banks at the ECB in order to encourage them to lend. There is a limit to this – commercial banks will turn to other methods of storing money if it becomes expensive to store reserves at the central bank. In early 2016, the Bavarian Banking Association recommended that its member banks start stockpiling physical cash.55
The dilemma is straightforward. If negative rates on central bank reserves do not feed into lending rates, they are useless; if they do, they will hit banks’ profitability unless banks start charging depositors interest for holding their money in banks as well.56 If this happens, there will be a flight into strong-boxes.57
The Bank supposed that part of the extra cash it pumped into the economy would be used to buy foreign securities, forcing down the exchange rate and thus enlarging export demand.
Figure 45 shows that the fall in the sterling exchange rate preceded QE; further, it only very temporarily improved the current account balance.58
It is hard to gauge the impact of signalling. A number of analyses have used ‘event study’ methodology, inspired by the efficient market hypothesis. This asserts that market prices adjust to ‘news’ rather than actual events. Using this method, researchers have discovered announcement effects on bond yields, currency and equity prices.59 But those committed to the ‘surprise’ theory of market behaviour are bound to conclude that central bank announcements will be subject to diminishing returns, and this seems to have been the case. Market participants, having accustomed themselves to unconventional monetary policy, became increasingly acute in guessing the size and timing of the next wave. As a consequence, QE2 had much less impact than QE1. However, central banks played the strategic game. By announcing changes in the composition of purchases, like the Fed’s ‘Operation Twist’ and the Bank of England’s decision to ‘increase the amount of shorter dated securities’, they were able to surprise investors and continue, at least in their own view, to make impacts on yield curves.61
Figure 45. UK exchange rate and current account, and QE60
Through the four channels above, the injection of narrow money (M1) was supposed to influence the movement of broad money and, through broad money, growth in nominal GDP.
Broad money is largely synonymous with bank lending. As we have seen, bank reserves went up while bank lending fell. The same story can be told with broad money.
The presumed relationship between narrow money and broad money (the money multiplier) never emerged, because the decrease in velocity of circulation offset the effect of QE. ‘I accept that the growth of money in the QE period has been much lower than I had been hoping,’ wrote Tim Congdon to the author. ‘Nevertheless, it has stopped a much worse recession.’
Figure 46. Growth in UK money supply and money lending post-crash62
Figure 47. UK broad money (M4) growth64
The Bank of England estimated that the level of real GDP was boosted 1.5–2 per cent by QE1.64 There is huge uncertainty about this: we can be reasonably confident about the sign of the effect but not its magnitude. What is clear from the table overleaf is that the monetary injection over the period 2009–12 far from offset the depressing effects of fiscal policy, as the Treasury had expected.
In 2012, the Bank of England stated that: ‘Without the Bank’s asset purchases, most people in the United Kingdom would have been worse off . . . Unemployment would have been higher. Many more companies would have gone out of business.’65 It is impossible to say.
In a 2016 assessment, the Bank concluded that it was not asset purchases as such which boosted activity, but their effect on sentiment.66 Keynes, too, had written that ‘a monetary policy . . . may prove easily successful if it appeals to public opinion as being reasonable and practicable and in the public interest, rooted in strong conviction, and promoted by an authority unlikely to be superseded’.68
Figure 48. UK output and unemployment67
QE was meant to have a joint effect on prices and output, but there was considerable confusion about the relationship between the two. Was it the effect on output that was supposed to bring inflation up to target? Or was it the rise in inflation (more accurately, the expected rise in inflation) which was supposed to lift output? Targeting inflation presupposed that inflation governed output: people would spend more because they expected prices to go up. This is how the real balance effect was supposed to work. Keynesians reversed the causality: it was people spending more that caused prices to go up. Therefore the target should have been output, not inflation; and the tool fiscal policy, not monetary. The Bank’s failure to boost inflation (except possibly in the first bout of QE) was due to a deficiency of aggregate spending.
Who was right? Figure 49 shows that the period 2008–16 demonstrated no better correlation between money (narrow or broad) and inflation than did the monetarist experiment of the 1980s. The best correlation during the Great Recession was with oil prices (Figure 50).
Figure 49. UK CPI inflation and QE69
The Keynesian conclusion is clear. The inability of QE to get inflation up to target ‘in the medium term’ was due to the government’s failure to get output up to trend in the short-term. This was true not just of the UK. The Bank of Japan has been using QE for nearly four years without getting inflation anywhere near its 2 per cent target. In the circumstances Governor Haruhiko Kuroda’s pledge to deliberately overshoot the target in order to raise inflation expectations was somewhat lacking in credibility.
The effects of QE were supposed to be distributionally neutral. It wouldn’t be true to say that savers were bound to lose and assetholders bound to gain from QE, as many savers own pension funds. Nevertheless, the balance of gain went to the rich. The median or typical household in UK held only around £1,500 of gross assets, while the top 5 per cent of households held an average of £175,000, or around 40 per cent, of the financial assets of the household sector held outside pension funds.70 By enriching the already wealthy, QE increased the well-documented concentration of private wealth in ever fewer hands. But richer households have a much lower marginal propensity to consume – that is, they spend a lower proportion of new income than poorer people. So enriching the already wealthy had a much smaller impact on overall spending than if the same amount of money had gone to lower-income groups.
Figure 50. Oil prices and UK CPI inflation71
Figure 51. Distribution of UK household financial assets, 201172
This distributional effect is not a generic consequence of QE but of the way it was done. The political neutrality of the Bank was thought to be its great advantage in conducting macroeconomic policy, because it would not be tempted to direct money for political ends, i.e. to secure the re-election of the government. In a speech at the LSE in 2017, Mark Carney repeatedly claimed that the central bank was an agent of ‘the people’.73 But the chain of accountability is not clear. Theoretically, the central bank acts on a mandate from the government, which depends on renewable popular support. This larger accountability is jammed, though, because only a small group of insiders understands the technique of monetary policy. In practice, the bank’s accountability is to the financial system, which means to existing asset owners.
Let’s look again at the diverging recovery rates between the UK, USA and the Eurozone. In the last chapter it was suggested that these can be correlated with the impact of fiscal policy. Can we find a similar relationship with monetary policy? Or, more plausibly, was it the combination of the two which explains the different outcomes?
There is general agreement that QE was more successful in the United States than in the UK, and less successful in the Eurozone than in either. The broad explanation for these discrepancies is that there was more ‘stimulus’ from both fiscal and monetary policy in the USA than in the UK, and more stimulus from monetary policy in the UK than in Europe.
Figure 52. Post-crash outcomes: UK, USA and Eurozone74
Studies of US ‘credit easing’ show that it achieved a bigger ‘bang per buck’ than asset purchases in the UK. Whereas in the first round of QE in both countries (2008/9–10) the Fed injected only half the amount of money relative to GDP as the Bank of England (7 per cent to 14 per cent), it is estimated that the injection had double the effect on GDP (4 per cent as against 1.5–2 per cent).75 If this is so, the probable reason is that the Fed’s QE programme was overwhelmingly targeted at the most distressed parts of the financial system and purchased riskier mortgage-backed securities, whereas the Bank of England bought virtually only Treasury gilts. However, one cannot segregate this supposedly ‘bigger bang per buck’ from the simultaneous $800 billion fiscal stimulus enacted by President Obama in February 2009. What seems clear enough is that the US authorities, both monetary and fiscal, were together willing to take bolder action to get the US economy moving again than those in the UK and the Eurozone.
The euro was afflicted by two original sins – the disconnect between fiscal and monetary policy and its neo-liberal monetary constitution. The European Central Bank was technically debarred from buying government debt. As a result, the monetary response to the crisis can be summarized as ‘too little, too late’. Its first response to the storm signals was actually to raise interest rates in July 2008. It was then slower than the Bank of England and the Fed to cut them as the Great Recession unfurled. Similarly, it only arrived at QE on the UK and US scale in 2015.
The consequences of the ECB’s passivity before then were dire. Whereas in the UK monetary policy was used deliberately to offset the effects of fiscal austerity, in the EU there was no offsetting action from the ECB. By 2011 US real GDP had recovered to its pre-crash levels; the UK followed in 2013, but the Eurozone not until 2015, after suffering a double-dip recession. Only since 2015, with the Juncker investment programme (see above p. 257), have expansionary monetary and fiscal instruments both come into play.
Why was the ECB was so slow to act? The three central banks have somewhat different mandates but this was not decisive.76 A more important institutional constraint was that the ECB’s rules forbade it from holding more than a third of any specific bond issue, or more than a third of any one country’s debt. Without a single eurobond jointly guaranteed by all members, this limitation was inevitable.
An even more important explanation is the ECB’s misreading of the crisis. It saw it as temporary – in February 2008, ECB President Jean-Claude Trichet was warning of the risk of an inflationary spiral.77 This partly reflected the theoretical framework of the day in which inflation was seen as the main obstacle to steady state, marketled economic growth. In addition, until the sovereign debt crisis hit the Eurozone in 2010, the financial impact of the US collapse was limited. But the ECB’s passivity also reflected a particular historical mindset. For the ECB, heir to the Bundesbank, the supreme danger to avoid was a repetition of the hyperinflation of the early 1920s. By contrast, it was the Great Depression of 1929–32, and the need to avoid a repeat of that, which had the biggest historical impact on Ben Bernanke and other US policymakers.
Governments whose policies fail to achieve their promised results always claim that they were pursuing policies that would have succeeded had it not been for unexpected ‘headwinds’. Thus MPC member Spencer Dale, speaking in 2012:
Some commentators have pointed to the weakness of growth over the past couple of years as evidence that the impact [of QE] has been relatively limited. But this seems a silly argument. The scale of the headwinds affecting our economy over this period – in terms of the squeeze in households’ real incomes stemming from the rise in commodity and other import prices, the fiscal consolidation, the tightening in credit conditions, and the fallout from the Eurozone crisis – has been huge. These headwinds have to be taken into account when assessing the effectiveness of the policy actions taken to offset them. There is a legitimate debate as to exactly how effective our policy actions to date have been. But I have little doubt that without them our economy would be in a far worse state today.78
Figure 53 below, taken from a Bank of England paper, claims to show what would have happened to broad money and output growth without QE1.
Just as economic models are provable only ceteris paribus, so all empirical assessments are relative to counterfactuals. But which headwinds to blame and which models to use depend on one’s theory of the economy.
Figure 53. Bank of England estimates of effect of QE on UK growth rates79
Taking his cue from the Friedman and Schwartz explanation of the Great Depression of 1929–32, Tim Congdon believes that the relative failure of QE was due to not printing enough money. ‘We know’, he argues, ‘both that governments can print money and that economic agents have a finite demand for real money balances. We therefore believe that policy-makers can engineer whatever inflation rate they choose. The generation of inflation, and the prevention of inflation, seem extremely easy: just print the right amount of money.’80
In contrast, by 2014 the Bank of England had more or less given up on QE:
the relationship between reserves and loans typically operates in the reverse way to that described in some economics textbooks. Banks first decide how much to lend depending on the profitable lending opportunities available to them – which will, crucially, depend on the interest rate set by the Bank of England. It is these lending decisions that determine how many bank deposits are created by the banking system. The amount of bank deposits in turn influences how much central bank money banks want to hold in reserve (to meet withdrawals by the public, make payments to other banks, or meet regulatory liquidity requirements), which is then, in normal times, supplied on demand by the Bank of England.81
Thus the Bank sought to exculpate itself both for responsibility for the crash of 2008 and for the weakness of the recovery.
QE offers as good an experiment in macroeconomic policy as we are likely to get, which is not that good. Attempting an empirical assessment of its effects is bedevilled by the omnipresence of counterfactuals. We are trying to compare what happened with what might have happened had policy been different – had there been more QE, or had it been done in a different way, or had it not been done at all, or had something else been done, or had fiscal policy not been contractionary.
So the best we can do is to compare what it set out to do with the actual outcome. On this test the conclusion is reasonably clear. It promised to boost output by raising the rate of inflation, while being neutral on distribution. In fact, over five years (2011–16) it failed to get inflation up to target; it had, at best, a weak effect on output; and it was far from being distributionally neutral. After nine years of emergency money, the financial system remains as dangerously stretched as it was before the crisis, and the economy as dangerously dependent on debt.
Economic theory can help explain why.
The first generation of monetary reformers – Fisher, Wicksell, the early Keynes – believed passionately that the way to prevent booms and slumps was to keep the price level stable. The QTM seemed to give the monetary authority a scientific basis for doing this. To guarantee monetary autonomy the reformers were willing to jettison the erratic control of the gold standard. But they were no more willing than the gold enthusiasts to entrust monetary policy to governments. Monetary policy should therefore be independent both of the gold standard and of the state.
The main disputes at this stage concerned the transmission mechanism from money to prices. This harked back to still-earlier disagreements about the nature of money. Was it cash or credit? For Fisher, money was cash: control of the monetary base or ‘narrow’ money was key to control of prices. Since even at that time most transactions were financed by credit, there needed to be a determinate relationship between money and credit, which was found in the monetary multiplier. This depended on the existence of a ‘real balance effect’. Enter Fisher’s Santa Claus, sprinkling the cash equivalent of goodies round the house. Milton Friedman and the American monetarists were Fisher’s heirs.
Wicksell saw money as credit, not cash. The key to control of the money supply was the control of bank credit. This could only be done by regulating the price of credit (or interest rate); the terms on which banks made loans. The early Keynes was a Wicksellian; and central bank policy in the Great Moderation of the early years of this century, with its reliance on Taylor rules, owed more to Wicksell than to Fisher or Friedman.
However, Wicksell raised a troubling problem for those who relied on monetary therapy alone to keep prices steady. As Henry Thornton had already noted, there were two interest rates needing attention, not one. The first was Bank Rate, and the structure of commercial lending rates which supposedly depended on it. The other was the ‘natural’, or ‘equilibrium’ rate, the expected real rate of return on investment. The task of the central bank was to keep the market rates equal to the natural rate.
This was the point of entry for the Keynesian revolution. Keynes came to see that the crucial element of volatility in market economies was not in fluctuations in the price level but in fluctuations in Wicksell’s natural rate. So policy should be directed not to stabilizing prices, but to stabilizing investment. Fiscal policy had to be the main instrument of ‘demand-management’, since it was spending, not money, which needed to be managed.
The economic collapse of 2008–9 showed that monetary policy directed to the single aim of price stability was not enough either to maintain economic stability or to restore it. The economy collapsed, though the price level was stable.
QE was an attempt to apply Friedman’s lesson of the Great Depression, as learned by Bernanke, to a situation where nominal interest rates had reached their zero lower bound. Preventing a collapse in the money supply was to be achieved by what was euphemistically called ‘unconventional’ monetary policy, but was really just a re-run of Fisher’s Santa Claus. Pump enough cash into the economy and the extra spending it produced would soon lift it out of the doldrums. But this supply-side monetary therapy took no account of the collapse of investment demand. The recipients of the central bank’s cash either did not spend it, or did not spend it on currently produced output, so ‘broad money’ – bank deposits – fell, even as narrow money (reserves) exploded. In the language of Keynes’s Treatise on Money, the money got stuck in the ‘financial circulation’. At best it achieved about 20–25 per cent of the expected output gain, but at the cost of pumping up unstable asset prices and producing a finance-led recovery.
The crisis left the relationship between fiscal and monetary policy unresolved. If push came to shove, most policymakers in 2009 would have said that fiscal consolidation would restore sufficient ‘confidence’ to allow monetary policy to raise the rate of inflation. In fact, confidence was not restored. This left monetary policy ‘overburdened’. It was now expected to push up output as well as prices, with no more agreement than before about which pushed up what.
The best that can be said for QE is that it was a default position. Central banks were right to reduce Bank Rate to the zero bound. But the main effect of their reliance on portfolio rebalancing to boost output was to boost the portfolios of the wealthy, with minimal effects on output. One doesn’t need headwinds to explain why.
Professor Congdon occupies an important but lonely position in the history of monetary thought and current debates about monetary policy. He can be called a Keynesian monetarist.
He is a monetarist in that he believes that the level of (nominal) national income is determined by the money supply, i.e. that changes in the money supply are the primary cause of changes in national income. (He also adds ‘and wealth’ from time to time.) Further, he believes that changes in the money supply have an equi-proportional impact on income; if the money supply increases by 20 per cent, then income will increase by 20 per cent too.82
All of which is to say he believes in the Quantity Theory of Money. But he is a broad money monetarist. He believes that broad money (cash and bank deposits, roughly speaking) is the relevant measure of the money supply. As such, he stands in contrast to Fisher and, at some points in his career at least, Friedman, who thought that national income was determined by the amount of ‘base’ or narrow money in the economy (cash and central bank reserves), as these in turn determine the level of bank deposits via the ‘money multiplier’ effect.*
As far as policy is concerned, Congdon believes that (a) the central bank can directly control the level of broad money in the economy, and (b) that as long as money growth is kept stable by the central bank, economic disaster can be avoided. In his account, the 2008 crash was caused by a fall in the quantity of money, and if central banks had simply pumped more money into the economy, then we could have been spared the worst of the recession.
So much for Congdon the monetarist. Congdon is also a peculiar kind of Keynesian in that he takes his Keynes from Keynes’s A Treatise on Money, not from The General Theory. Like Keynes, he believes in the possibility of autonomous collapses in the money supply (e.g. following a shock to investment), leading to falls in nominal income, but believes that these can be successfully offset by the monetary authority pumping money into the economy – if necessary without limit. Congdon’s spiritual home, that is, is with Irving Fisher, Ralph Hawtrey and the monetary reformers of the 1920s who tried to use monetary policy to prevent the oscillations of the business cycle. But he condemns the Keynesian attachment to ‘fiscal policy’ as at best redundant, and at worst (the more general case) pernicious.
Thus Congdon rejects equally the fiscal element of the Keynesian revolution and the money-multiplier mechanism of most monetarists. So he is something of an outlier. I have benefitted enormously from my exchanges with him, as well as from his published writings, but I always end up not quite understanding why he holds the positions he does – and so passionately. So the object of this note is to ask: is his position coherent? Are his prescriptions useful?
The interrogation can be grouped into three parts: his use of evidence; the gaps in his theory; and his rejection of any sort of fiscal policy.
Evidence is of utmost importance to Congdon. In contrast to mainstream work in economics – ‘unscientific and shoddy’83 – he believes that the monetarist approach is on the side of logic and facts, and that the evidence for his position is so ‘overwhelming’ that monetarism can be treated as a ‘true proposition’.84 So we might start by seeing if the evidence he presents can meet this high bar.
Congdon’s central piece of confirmatory evidence is the correlation between the rates of growth in nominal income and broad money over time. In one of our (many) exchanges, Congdon wrote, ‘the evidence is overwhelming – from all countries in all periods of more than a few quarters – that changes in [the money supply] and [nominal income] are related’.85
Could such evidence, by itself, secure the monetarist position? Surely not. Congdon’s claim is that changes in the money supply cause changes in national income. But we know that correlation does not imply causation, and in a fiat money economy there are compelling reasons to believe that the arrow of causation can run in the opposite direction. Nearly all money in the modern economy is created by commercial banks making loans,86 and it is plausible that banks’ lending behaviour is caused by changes in the real economy.
Congdon knows this. In contrast to his statistical over-confidence, he recognizes elsewhere that ‘the citing of numbers does not establish a definite causal link or prove a rigorous theory beyond contradiction’.87 Moreover, the faith he has in his evidence is not especially consistent. Indeed, he can veer from certainty to circumspection in the space of a page. In the Introduction to his Money in the Great Recession (2017), underneath a figure showing the behaviour of broad money in the 2000s, Congdon writes that ‘it is immediately clear that a decline in the rate of change in the quantity of money must have had a role in the Great Recession, just as it did in the Great Depression’.88 Yet later in the very same paragraph he cautions: ‘more research and analysis is needed before strong statements about causality can be ventured’!89
Interpretation aside, what about the evidence itself? In his contribution to Money in the Great Recession, Bank of England economist Ryland Thomas disputes the evidential backing for monetarism. First, he notes that ‘the behaviour of nominal spending in the early years of the [Great Recession of 2008–9] . . . did not conform to a simple monetarist relationship where spending follows broad money growth with a lag’.90
Such a finding is uncomfortable for Congdon. Nevertheless, he tries to circumvent this genre of criticism by conceding that, in the short-term, the causal link between broad money and nominal income/wealth can break down because of Keynesian-type ‘animal spirits’91 – a notion which elsewhere in the book he castigates as ‘woolly’, ‘imprecise’ and ‘journalistic’.92 Similarly, he emphasizes that changes in the money supply determine the ‘equilibrium’ level of nominal income and wealth, but that actual values can fluctuate around this point.93
Keynes’s rejoinder – ‘in the long-run, we are all dead’ – is apposite here. How long or short is the short-run? What happens in the short-run – in a recession, for example – has an enormous impact on people’s lives over a long period. Equilibrium theory is no use for analysing short-run fluctuations, since it excludes these by assumption. Yet Congdon has no qualms using the QTM to support his short-term policy prescriptions,94 even though it is an equilibrium theory.
In fact, Thomas’s statistics pose an even more fundamental problem for Congdon. Using data stretching from 1870 to 2010, Thomas notes that there is no evidence of a stable monetarist relationship ‘where contractions in money lead contractions in nominal GDP . . . in many periods broad money growth appears to move contemporaneously with or even to lag nominal spending’.95
That is to say, changes in nominal spending have often occurred before changes in broad money. In contrast to Congdon’s view, Thomas rightly concludes that ‘the relationship between money and spending within and across business cycles [i.e. in both the short- and long-run] is complex’.96 The evidence, then, does not prove Congdon’s case, as he seems to believe. It does not disprove it either. Highly abstract theorems like the QTM are so enfiladed with ceteris paribus conditions that they are neither provable nor disprovable. Thus it is always possible to say that quantitative easing in the UK in 2009–10 failed to boost broad money growth to the expected extent because of a misguided simultaneous tightening of banking regulations.97 A robust theory should not require too many qualifying conditions.
Congdon relies on theoretical argument – as all economists must – to support his monetarist hypothesis. Specifically, he proposes a transmission mechanism from money to nominal income/wealth based on the ‘real balance effect’.98 Congdon calls this the ‘hot potato argument’;99 it is the necessary assumption on which his theory hangs.
The basic argument is that agents have a desired ratio of money to expenditure. In the event of a monetary shock – if the central bank expands the money supply, for example – then agents end up with ‘too much’ money relative to this ratio.100 As a result, they increase their spending to get rid of the excess. The process continues until the excess is ‘extinguished by a rise in sales [output] or prices’.101 Which it is depends on whether there is any spare capacity in the economy, but either way nominal income increases.
The main criticism of Congdon’s transmission mechanism is that it is leaky. Take the equation of exchange, the identity at the heart of the QTM:*
MV = PT
Congdon argues that purchases of securities from the non-bank private sector directly increases broad money (deposits), which, according to him, will lead to an equi-proportional increase in nominal income. In other words, it has no impact on velocity. But this simply ignores the leaks. I focus on three here.
In order for the real balance effect to work, agents have to respond to an increase in their deposits by actually spending their extra money; if they hoard it, the transmission mechanism breaks down. In terms of the equation of exchange, an increased propensity to hoard is reflected in a fall in the velocity of circulation.
Congdon may dismiss any such increase in liquidity preference as a short-term phenomenon. But quantitative easing has further implications for the behaviour of velocity. When a central bank engages in QE by buying securities and assets from private sector agents, most of it will go to the wealthy minority that owns substantial assets. The wealthy have a much smaller propensity to spend – they save a larger proportion of any increased money they get – than the poor. The consequence of such an exercise will therefore be to slow down the velocity of circulation, as a single given unit of money will change hands fewer times. The decline in velocity will at least partially offset the attempt to increase the quantity of money. The equi-proportionality condition is violated.
Similarly, the wealthy are much more likely to spend new money on buying assets and on financial speculation. Does this matter for Congdon’s transmission mechanism?
In the equation of exchange, T is composed of a mix of transactions that contribute to the real economy, and other transactions, mainly financial. The evidence presented in this chapter gives us reason to believe that a disproportionate amount of QE money will be spent on financial speculation, and not in the real economy, meaning that asset prices will rise. Should we worry?
Not according to Congdon. His argument is as follows: ‘a capitalist economy has a range of mechanisms by which arbitrage between different asset markets prevents prices and yields in one class moving out of line with prices and yields in another’. Further, ‘over time . . . the hot potato of excess money circulates from one asset market to another and from asset markets to markets in goods and services’.102
This assumes a perfect fluidity in money flows between the different factors of production. There is no allowance for stickiness. Again, Keynes’s reminder that ‘in the long-run, we are all dead’ is the right response to this line of argument.
Recent experience does not suggest that asset bubbles simply ‘sort themselves out’. Undirected expansion of the money supply, even if its intention is to boost nominal output, risks fuelling the next wave of speculation (cf. the dotcom bubble). Ironically, Congdon’s QE, far from restoring equilibrium nominal income, would be a source of further monetary instability.
People can get rid of their excess money by spending it on imports and the like, so that the money leaves the economy. This, though, does not obstruct the equilibrating mechanism in Congdon’s eyes. When the money leaks abroad, the exchange rate goes down, which leads to currency purchases which offset the previous leak, in a replay of Hume’s price–specie–flow mechanism. Ultimately, this tactic will ‘work’, in that the money will eventually work itself into the real economy. As Hume said, one cannot get water to flow uphill.
Flooding the economy with money hardly amounts to a scientific monetary policy. The truth is that monetarists have no idea how much money they will need to pump into an economy to lift it out of recession. There is no reason to believe that the private sector’s desired holding of cash balances is independent of the business cycle. In short, there is no predictable real balance effect. And one consequence of ‘feeding the hoarder’ is that when the hoarder starts to spend again and velocity approaches its ‘normal’ level, a lot of excess money is sloshing around the economy, setting the stage for a runaway inflation.
‘Forget about fiscal policy. It doesn’t do any good to short-run economic activity . . . and may do a lot of long-run harm.’103
Congdon’s objection to any form of fiscal policy is the hardest part of his position to understand. It is not that he objects to increased spending in a slump. Indeed, he believes that it is indispensable. Nor does he mind much whether it is the government or the central bank which ‘prints’ the extra money: he often uses the two terms interchangeably. It is to the government spending the extra money that he objects. His view is quite different from those of people such as Adair Turner, who have advocated ‘monetary financing of the deficit’. Congdon’s essential point is that the state should have no influence on the way the extra money is spent. Why is this?
Once again, he believes evidence is on his side. In a ‘statistical appendix’ to his 2011 book Money in a Free Society,104 Congdon presents data from a number of countries between 1981 and 2008 which show there is no relationship between changes in governments’ discretionary spending – the spending which results from cuts in taxes or deliberate boosts to spending – and changes in output gaps. Keynesian theory would suggest that an increase in fiscal deficits would cause a shrinkage in the output gap. But there is no evidence of such an effect. Therefore the Keynesian case for fiscal policy falls to the ground.
But the logic is faulty. The fact that changes in discretionary spending and output gaps are not correlated can be seen as evidence of the effectiveness of fiscal policy. Governments tend to respond to negative output gaps by increasing their discretionary spending – all other things being equal, then, one might expect a negative correlation between discretionary spending and budget deficits. But other things aren’t equal; there is no overall correlation, and so the negative correlation must be being offset by a different effect. The missing link lies in the positive effect of government spending on the output gap, i.e. in the effectiveness of fiscal policy!
Empirical support in favour of fiscal policy is at least as strong as the evidence Congdon marshals against it. Countries that responded to the Great Recession with more extensive fiscal programmes performed, on the whole, better than those which didn’t.
If the evidence is inconclusive, we have to turn to theory. And indeed, Congdon appears to reject fiscal policy a priori. He writes: ‘an increase in the public debt, due to the incurrence of a public deficit, is not an increase in the nation’s wealth’.105 This is rhetoric, not science. What if the money is spent on the creation of real assets, such as railways or houses? Following Ricardo, Congdon rejects the possibility of productive state spending.
Indeed, one of the main advantages of fiscal policy is that a government can direct the flow of the new spending in the economy. When a recession hits, private investment spending falls far more than consumption spending, and this cannot be wholly explained as a rational response to a fall in the long-run risk-return profile of investment – ‘animal spirits’ must be at play. Keynes recognized this psychological aspect to investment spending. In this event, the government can use fiscal policy to maintain a ‘normal’ level of investment, in order to avoid the erosion of the economy’s productive capacity.
Even if the government runs a deficit in order to finance its current spending, it can contribute to the wealth of the economy. This can be explained by reference to the equation of exchange. If the government borrows money from the bond markets that otherwise wouldn’t have been spent, and then spends this money, the overall velocity of money increases. Nominal income increases as a result, without any prior expansion in the money supply.
Of course, if the Quantity Theory of Money were the correct theory of macro-policy, there would be no need for discretionary fiscal policy: all the stabilization needed could be done by monetary policy. But the QTM begs so many questions, and attempts to apply it encounter so many ‘leaks’, that dogmatic rejection of fiscal policy seems indefensible to me on scientific grounds.
At one point in Money in the Great Recession, Congdon writes mockingly that ‘at the start of the third millennium economists sometimes pretend to be practising a “science” or at least an intellectual discipline with scientific pretensions’.106 Mainstream economics for him hasn’t been ‘scientific’ enough. When it comes to explaining the Great Recession, for example, the ‘mainstream view . . . is untestable, and deserves to be condemned as unscientific and shoddy’.107
My difficulty with Tim Congdon is that he is constantly invoking scientific ‘proofs’ in a field that defies scientific testing. His scientific efforts arise from a doomed attempt to ‘prove’ passionately held value judgements. He is a monetary reformer because he has an intense dislike of state intervention. As a result he dismisses any evidence that monetary policy may be ineffective and fiscal policy may be effective. Like the monetary reformers of a century ago he turns to money to ameliorate the human lot because he cannot bear to turn to the state.
* This is known as the Marshall–Lerner condition: if the sum of export and import demand elasticities is greater than 1, then a fall in the exchange rate will have a positive impact on the trade balance and increase output. Otherwise, the trade deficit will widen and output will fall.
* See Krugman (1998). While both Krugman and Keynes pointed to the existence of a liquidity trap, their ideas are subtly different. According to Paul Krugman a liquidity trap – and consequently the need for QE – occurs when ‘a zero short-term interest rate isn’t low enough to produce full employment’ (Krugman (2014)). In Figure 41, the expected profit rate has fallen so much that only a sizeable negative nominal interest rate could restore full employment. Keynes’s trap, on the other hand, arises when reductions in the bank rate cannot bring down the long-term rate of interest, because investors who expected long-term rates to rise (and therefore to make a capital loss on bonds) will sell their bonds for cash, forcing up the long-term rate. The zero bound is the limiting case, but ineffectiveness of orthodox monetary policy might occur before that limit is reached, because of uncertainty attaching to future bond yields. Either trap might justify the launch of unconventional monetary policy, but Krugman’s trap became the rationale for QE, because it avoided the problem of having to model uncertainty.
* The Bank also bought a modest amount of commercial paper and corporate bonds. The reason for these interventions was to increase liquidity in the moneymarket. The following year, as the Bank became satisfied with the level of liquidity achieved, these assets were resold.
* See p. 35 for more on the distinction between broad and narrow money, and pp. 258–9 for discussion of the money multiplier.
* See Ch. 3 for more details and explanation.