The quantitative easing (QE) programme launched by European Central Bank (ECB) has been hailed as a long awaited decision, and the latest ECB press releases are pretty happy with the results. However, QE raises nontrivial questions regarding its medium-term effectiveness net of concomitant facors. Firstly, assessment of experiences of QE forerunners (United States, United Kingdom, Japan) is mixed (e.g. Cecioni et al., 2011; Gambacorta et al., 2012). Japan has not yet escaped form its long lasting stagnation, and while US and UK have been doing better than the EZ over the last five years, the specific impact of their large QE programmes is unclear. Secondly, and perhaps more importantly, how QE is expected to work, or what the “transmission mechanism” is, remains unclear. Examples of the issues involved are in the Economonitor posts by Bossone and Valiante.
To begin with, QE ends and means should be defined appropriately. The central banks that have so far engaged in QE have also communicated somewhat different ends: foster the recovery of economic activity, prevent a deflationary spiral, raise inflationary expectations, spur credit supply, … , depending on their mandate, institutional framework and communication style. For the ECB, the QE rationale is to stop a deflationary drift in the Euro Zone (EZ) and realign inflation expectations with the 2% target (e.g. Draghi, 2014). This communication strategy is clearly in tune with the single mandate of the ECB for price stability. As to the means, QE is in fact a catch-all word that covers a number of different interventions of the central bank in the money market (e.g. Bernanke and Reinhart, 2004; Borio and Disyatat, 2010). However, these interventions do have one common feature in that they inject additional base money into the system (hence the qualification “quantitative”) which is reflected into an equal expansion of the central bank’s assets (QE is often presented in terms of a target on the latter). Looking at this common feature, one may say that QE is nothing but a reincarnation of the traditional textbook treatment of monetary policy, the “LM model” for short, whereby the central bank controls “the quantity of money” (or, more precisely, the monetary base of which total money supply is a multiple).
Indeed, traditional textbooks also distinguished between monetary policy by way of control of the quantity of money (which was then considered standard) and by way of setting “the interest rate” against which money is demanded. In the former case, the interest rate is determined endogenously by the money-market equilibrium, given the money supply created by the central bank. In the latter, it is the money supply that is determined endogenously by the money market, given the interest rate set by the central bank. As shown by Poole’s classical paper (1970), for a given policy goal in terms of economic activity, the two techniques are equivalent. If e.g. the policy goal requires a cut of the interest rate by x%, it is equivalent for the central bank to cut the interest rate directly or to create base money to the extent that the interest rate falls by x%. The amount of money creation is also the same. One technique may be preferable to the other only if the “efficiency” of the instrument is introduced in consideration of the type of shock hitting the money market.
However, this is not how QE is today explained and communicated. The case for QE is that the central bank wishes to achieve a policy goal that it can no longer achieve owing to the “zero lower bound” (ZLB) problem. As Bernanke and Reinhart (2004) put it, “even if the price of reserves (the overnight rate) becomes pinned at zero, the central bank can still expand [the monetary base] beyond the level required to hold the overnight rate at zero” (p. 87). In fact, in the EZ case, De Grauwe and Ji (2015) and Orphanides (2014) show that while the ECB policy rate has been dwelling at the ZLB since the end of 2012, clear symptoms of monetary restriction have developed as witnessed by falling creation of base money, broad monetary aggregates and credit, and by higher real interest rates due to deflation. QE is thus seen as a means to correct this undesired restrictive monetary stance by direct creation of base money. But at this point the question is: how can QE achieve what the direct control of the policy rate cannot? If the policy rate is at the ZLB and the final policy goal is not achieved, the transmission mechanism from QE to the policy goal should bypass the policy rate.
The New Keynesian workhorse model for policy making provides and oft-heard narrative. An over-restrictive monetary policy stance manifests itself in the following inequality in the so-called IS curve:
policy rate – expected inflation > equilibrium real rate
which generates a negative output gap. The problem at the ZLB, with policy rate = 0, is that
– expected inflation > equilibrium real rate
which may arise from various combination of factors such as very low or negative equilibrium real rate, too low inflation target of the central bank, expected inflation below target or negative. Then the so-called Phillips curve says that the ensuing sequence of negative output gaps feeds back onto negative inflation gaps, which fulfils low or negative expected inflation in a vicious circle.
Beginning with Krugman (1998) and then Eggertsson and Woodford (2004), the leading idea is therefore that QE mainly operates through raising expected inflation, which is indeed the preferred rationale for QE in the ECB institutional communication. However, there should also be some transmission mechanism between the announced rate of creation of base money and the increase in expected inflation. This is hardly spelt out with clarity by the ECB communication. One possibility is of course the monetarist commandment that inflation is always and everywhere a monetary phenomenon. But if the macro-policy framework is the one outlined above, the transmission can only go through closing the EZ output gap. This is transparent in the institutional communication of the Federal Reserve, where, in force of its dual mandate, inflation and output (or unemployment) targets go hand in hand. Actually, according to Eggertsson and Woodford (2004) proposal known as “forward guidance”, the solution to the problem is not QE per se, but the central bank’s credible promise to overshoot the inflation target. That is to say, the central bank should commit itself to keeping the policy rate unchanged (and provide all the money demanded) as long as necessary to create excess output and inflation. “The expectation that a boom will be created later should stimulate spending now, through permanent-income and accelerator mechanisms” (p. 77).
Suppose the ECB knows this, though it cannot say this. Then it also knows that the EZ output and inflation gaps are nothing but the result of the gaps in each country. The cause of negative gaps at the country level should lie in its own IS inequality, i.e. the country real interest rate exceeds the equilibrium real rate. Unless financial markets are perfectly integrated and arbitraged, and all countries are equal, each country real interest rate is typically different from any other and from the policy rate, as a result of each country’s spread over the policy rate charged by lenders, and expected inflation. Therefore, the ZLB problem in the EZ is that, for a given distribution of country spreads and inflation rates, the ECB is unable to lower the common floor of the country nominal interest rates any more.
Available data provide a clear picture of this problem. Since it is widely agreed that bank credit is the primary source of private expenditure in the EZ, let us look at the average interest rate of bank loans to non-financial corporations in each EZ country (excluded the latest members, Latvia and Lithuania) provided by the ECB. To begin with, figure 1 shows the monthly observations of the mean and standard deviation of these interest rates across the EZ countries form 2003:1 to 2015:3. The mean shows a well-known time profile: low and declining in the early years of euro-optimisms, skyrocketing as the financial global crisis was mounting, plummeting after the first “conventional” ECB interventions, then rising again with the sovereign debt crisis, and eventually brought again under control by the second wave of ECB measures. What is relevant here is the concomitant evolution of the standard deviation. It remained low and stable around 50 b.p. until the outbreak of crisis. In spite of the success of the ECB in curbing the bank rates on average, the crisis created a stepwise surge in the standard deviation which grew by almost 100 b.p. over the subsequent years and appears remarkably resistant to the easing of the central monetary conditions. Computing the standard deviation of real bank rates (on the basis of the average y-y inflation rate) does not change the picture substantially (the step-up occurs one year earlier). This evidence is in fact a symptom of the notorious phenomenon of the dis-integration of the EZ financial markets. Hence, country interest rates can be, and remain, quite different from the average.
Figure 1. Mean and standard deviation of interest rates of bank loans to non-financial firms in the EZ countries (monthly observations, 2013:1-2015:3)
Source: ECB Statistical Warehouse, Interest Rate Statistics
Figure 2 focuses in greater detail on the relationship among the ECB policy rate (the Main refinancing rate), one of the key money market rates for banks (the 3-months Euribor) and the average spread between each country’s bank rate and the Euribor. The data tell two quite different stories. On the one hand, apart from the freezing of the interbank market in September-October 2008, the ECB has steered the Euribor quite effectively. On the other hand, the crisis ushered in a clear break of the tendency of the average spread to fall, it jumped from 100 to 250 b.p. at the end 2009 and then it started to rise steadily for four years up to 337. Hence, country interest rates can be, and remain, much higher than the policy rate of the ECB.
Figure 2. ECB Main Refinancing Rate, 3-m. Euribor and average spread of EZ country bank interest rates with the Euribor
Source: elaborations on ECB Statistical Warehouse, Interest Rate Statistics
Figures 3 and 4 provide a first insight into the differences among country bank rates by comparing the average bank rate and spread in three EZ areas: the seven more financially stable countries (EZ7: Austria, Belgium, Finland, France, Luxembourg, Germany, Netherlands), the five more financially distressed countries (EZ5: Greece, Ireland, Italy, Portugal, Spain) and late accession countries (LEZ: Cyprus, Estonia, Malta, Slovenia, Slovakia). Figure 3 shows that the average bank rates of the three areas tend to move together; however the crisis has generated a yawning gap between the EZ5 and LEZ areas and the EZ7 area. This pattern is reflected by the average spread of bank rates above the Euribor in the three areas in figure 4, which in the EZ5 and LEZ areas peaked at abut 450 b.p. in October 2013. A reversal seems to be in place since then.
Figures 3 and 4. Average bank interest rate and spread with the 3-m. Euribor in three EZ areas
Source: elaborations on ECB, Statistical Warehouse, Interest Rate Statistics
From this point of view, we can see that the transmission mechanism from QE to expected inflation via re-equilibration of the country IS curves can only go through the reduction and realignment of country spreads. This is the intermediate policy goal that the ECB cannot achieve otherwise. On passing, it is curious to note that while the final goal is in line with the ECB mandate, the intermediate one is not (at least in the “hawkish” view), and this may explain why it cannot be mentioned officially. In this perspective, the QE mechanism also takes on a Tobinian flavour, and it comes to overlap with other “non conventional measures” that not only involve the dimension but also the composition of asset portfolios of both the central bank and its counterparties (Bernanke and Reinhart, 2004; Borio and Dysiatat, 2010). The well-known Tobinian mechanism is that, by exchanging some classes of assets with money in the counterparties’ portfolios, the spread on such assets falls so that they become cheaper vehicles for financing expenditure. If this is the mechanism, it should strongly be stressed that the economy should not be in the so-called “liquidity trap”. The frequent identification of the ZLB with the liquidity trap – even by top scholars like Krugman (1998) or Eggertsson and Woodford (2004) – is highly misleading. In the Tobinian framework the liquidity trap is a situation in which the central bank is unable to manipulate the relative prices of assets or their spreads because the counterparties are ready to exchange whatever amount of assets for money. A variant that focuses on banks is that the latter have an “infinite” demand for reserves, so that whatever amount of base money created by the central bank is hoarded instead of being channelled into loans. These phenomena may well arise even when the policy rate is still positive (as it was in the EZ in 2012-13), and if they do, QE is not the solution no matter whether the policy rate is positive or zero. In fact, in Keynes’s original view, the solution to the liquidity trap is not monetary but fiscal policy. This point has recently been restated in different theoretical frameworks (e.g. Bossone, 2014).
Having clarified that, 1) the problem with the EZ is that of an economy with many different interest rates, 2) the ECB policy rate is at the ZLB but the country interest rates are not, and 3) the economies should not be in the liquidity trap, the final step leads us where the true problem lies. In order to re-equilibrate the EZ aggregate output and inflation gaps, QE should be engineered so as to minimize the country spreads. This means that each country’s banking system should receive the right quantity of base money commensurate with the determinants of its own spread. In other words, critical to the success of the QE programme is the country distribution of the base money creation. As can be seen in figures 3 and 4, the bulk should go where spreads are higher, that is the EZ5 and LEZ countries. But we already know that this will not be the case. Owing to the fact that the main vehicle of base money creation is purchases of government bonds, the distribution mechanism will be in proportion to each country’s share in the ECB capital. The EZ7 share is 42.1%, EZ5 amounts to 26% and LEZ to 1.9%. Likely, the banking systems of relatively small countries in the EZ5 and LEZ groups will receive too little QE. It may be considered that in a monetary union the total stock of money is redistributed across countries through their imbalances of payments. Thus EZ5 and LEZ countries may hope to increase their local money supply by means of large payment surpluses, which is however unlikely because these countries also tend to have payment deficits.
The conclusion is that, as is the case with any monetary policy instrument in a monetary union, QE in the EZ is fraught with the heterogeneity problem of the recipients of monetary policy. The institutional strictures and pressures under which the ECB operates will make the task harder. Of course, easier monetary conditions will help the EZ economies. Signs of reversal in differences among country bank interest rates have already been visible throughout 2014, a welcome side effect of QE - the euro depreciation – is already under way, and sooner or later reflation will take hold on average. However the ECB seems to have little control over the deeper transmission mechanisms of QE across the EZ, and the result may eventually be dim. Anyway, as the basic economic principles of monetary unions teach, when the one size of the single monetary policy does not fit all, supplementary, possibly coordinated, national fiscal policies should also be activated (e.g. Draghi, 2014). For well-known reasons, this is a major drag on QE success in the EZ, where national fiscal policies are tightly constrained and loosely coordinated, especially so in the countries where the QE impact will be smaller.
Bernanke B. S., Reinhart V. R. (2004), “Conducting Monetary Policy at Very Low Short-Term Interest Rates”, Papers and Proceedings of the American Economic Association, American Economic Review, vol. 94, pp. 85-90.
Borio C., Disyatat P. (2010), “Unconventional Monetary Policies: An Appraisal”, The Manchester School, vol. 78, pp. 53-89.
Bossone B. (2014), “Secular Stagnation”, Economics, Discussion Paper, n. 47.
Cecioni M., Ferrero G., Secchi A. (2011), “Unconventional Monetary Policy in Theory and Practice”, Banca d’Italia, Occasional Papers, n. 102.
De Grauwe P., Ji Y. (2015), “Quantitative Easing in the Eurozone: It’s Possible Without Fiscal Transfers”, Vox-EU, January 15.
Draghi M. (2014), “Unemployment in the Euro Area”, Speech at the Annual Central Bank Symposium in Jackson Hole, August 22, www.ecb.org.
Eggertsson G. B., Woodford M. (2004), “Policy Options in a Liquidity Trap”, Papers and Proceedings of the American Economic Association, American Economic Review, vol. 94, pp. 76-79.
Gambacorta L., Hofmann B., Peersman G. (2012), “The Effectiveness of Unconventional Monetary Policy at the Zero Lower Bound: A Cross-Country Analysis”, BIS Working Paper, n. 384.
Krugman P. (1998), “It’s Baaack: Japan’s Slump and the Return of the Liquidity Trap”, Brookings Papers on Economic Activity, vo. 29, pp. 137-216.
Orphanides A. (2014), “ECB Policy and Fed Normalization”, CEPR Policy Insight, November.
Poole W. (1970), “Optimal Choice of Monetary Policy Instruments in Simple Stochastic Macro Model”, Quarterly Journal of Economics, 84, pp. 197-216.