The Swiss National Bank’s January 2015 decision to abandon the Swiss franc’s peg to the euro led to short-term chaos in exchange markets and had a dampening effect on the Swiss economy. Some economists suggested Switzerland was poised to enter a sustained period of stagnation à la Japan. The decision also reignited policy debate on the benefits and drawbacks to central bank intervention in currency markets. While such intervention can be justified in certain situations, such as if the market is producing the “wrong rate”, it can also impose significant economic costs. The ECB’s recently implemented quantitative easing programme has been regarded by many as a thinly disguised attempt to weaken the euro in order to improve the eurozone’s competitiveness. However, the euro’s recent weakening began well before the ECB announced its programme; moreover, previous rounds of quantitative easing by other central banks have had minimal impact on exchange rates. Show
Intervention in the Foreign Exchange Market: Rationale, Effectiveness, Costs and BenefitsKeith Pilbeam Prior to the move to generalised floating in 1973, the adoption of floating exchange rates had long been advocated by eminent economists such as Milton Friedman and Harry Johnson.1 However, the experience with floating rates over the last four decades has shown that they are not the panacea that many advocates had presupposed. This has led many economists to propose schemes designed to limit exchange rate flexibility, such as John Williamson’s target zone proposal.2 In practice, central banks have frequently intervened in the foreign exchange market in a bid to influence the exchange rate at which their currency is traded, hence the term “managed” floating. In this paper we look at the economic rationale behind central bank intervention in the foreign exchange market. We then proceed to discuss the effectiveness of foreign exchange intervention, making the point that the theoretical and empirical literature overwhelmingly suggests that in order to be effective in the medium term, exchange market intervention needs to be non-sterilised, that is, it must result in a change of the domestic money supply and the short-term interest rate. Sterilised intervention, whereby the impact of the intervention on the money supply is offset by an open market intervention by the central bank, can at best have only a very short-term impact on the exchange rate. Finally, we consider two case studies of foreign exchange intervention and the associated costs and benefits in practice, namely, the cases of the People’s Bank of China and the Swiss National Bank. Managed floatingSince the advent of floating exchange rates in 1973, it has become evident that authorities have not always allowed their currencies to float freely but rather have frequently intervened to influence the exchange rate. A number of rationales have been put forward to justify such intervention. Before examining some of the most frequently used arguments for intervention, it is necessary to assume that the authorities can influence the nominal and/or real exchange rate in their desired direction; without such an assumption, no rationale for intervention can exist. Further, exchange market intervention can only be justified if it can be demonstrated that foreign exchange intervention has a superior benefit-to-cost impact when compared with other policies – or that constraints prevent the use of superior policies. In the following discussion, it should also be remembered throughout that exchange rate management can vary in degree from occasional intervention in order to influence the exchange rate to a permanent pegging. The arguments for some degree of discretionary intervention to some extent overlap, but they fall into three main categories: (i) the authorities can choose an exchange rate more in line with economic fundamentals than the market can; (ii) intervention is required to mitigate the costs of exchange rate “overshooting”; and (iii) intervention is an appropriate instrument for smoothing necessary economic adjustments. Authorities might be able to produce a more appropriate exchange rateFor a variety of reasons, the exchange rate produced by the market may be the “wrong rate” compared to underlying economic fundamentals. The market may use the wrong model, it may have incorrect perceptions about the future and it will have difficulty in interpreting the implications of news relevant to the exchange rate. However, the fact that the market may produce the wrong rate does not justify intervention by the authorities; it is necessary to demonstrate that the authorities can choose a more appropriate rate. There exists a case for intervention if the news or information available to the market is efficiently used, but the news itself is either inadequate – increasing risk – or misleading, and the authorities are in possession of superior relevant information. Intervention in such circumstances can prove both stabilising and profitable. However, it could be argued that a superior policy is for the authorities to abstain from intervening and instead release the relevant information to the market. Nevertheless, there may be circumstances under which such an information-release is not considered desirable, and even if the authorities were to release the relevant information, there is no guarantee that the market would believe them. Connected with the above argument is a far more convincing reason for the authorities to intervene. While it may be the case that the authorities do not know any more than the market regarding the “correct” rate, they should know better and sooner what they themselves are about to do (in most cases!). In other words, the authorities should be more capable than the market in predicting the future course of their policies, and this is of relevance to the correct exchange rate. Given this, intervention in the foreign exchange market may be interpreted by the market as a commitment by the authorities to adopt a given course of action; if this is the case, economic agents may more readily lend their support to the new policy, helping to make it more effective, and more speedily so, than would otherwise be the case. Thus, there exists a case for official intervention on the grounds that the authorities have better knowledge of their future policy intentions than private market participants. Official intervention in the foreign exchange market can literally “buy credibility”, convincing economic agents that the authorities intend to fulfil their stated domestic policy targets by committing the assets of the central bank in support of its declared future policy. A key postulate of the rational expectations literature is that the authorities will only be able to achieve their short-run inflation objectives painlessly if economic agents are convinced that the authorities intend to carry out their stated objectives. The opportunity to purchase some credibility by intervening in the foreign exchange market could prove to be a useful policy tool. Intervention to mitigate costs of exchange rate overshootingThe Dornbusch overshooting model shows that a move to monetary restraint can lead to a short-run real exchange rate appreciation, while an expansionary monetary policy can lead to a real depreciation.3 These real exchange rate movements leading to over- and undervaluations in relation to purchasing power parity (PPP) will exert effects on the real economy. In what follows, we shall refer to substantial and prolonged deviations from PPP as exchange rate misalignments. Misaligned exchange rates distort the allocation of resources between tradables and non-tradables as well as consumption patterns between the two. Undervaluation, by raising the domestic price level and placing downward pressure on real wages, may spark inflationary pressures, while overvaluation, by squeezing the tradables sector, may result in increased unemployment. Misalignment complicates and inhibits investment decisions because uncertainty as to the duration of the over- or undervaluation will affect the profitability calculations concerning whether to invest in tradables or non-tradables; such uncertainty is particularly inhibiting to marginal investment decisions. Misalignments almost certainly exert a ratchet effect on protectionism. In periods of undervaluation of the currency, resources that would ordinarily not be viable enter into the tradables sector, but as the rate corrects itself, they come under increasing pressure and may then seek recourse to protection. Alternatively, if the currency is overvalued, this tends to lead to knee-jerk protectionist cries due to the pressure on the tradables sector. It should also be remembered that undervaluation for one currency involves overvaluation for another and vice versa, so that one could expect protectionism to be a global and persistent phenomenon so long as exchange rates are misaligned. Since an over- or undervaluation must necessarily eventually be corrected, this will involve the various adjustment costs arising because of factor immobility, both occupationally and geographically. Retraining of labour involves costs and time, and aggregate demand cannot be painlessly varied at will. Foreign exchange intervention designed to reduce the costs and extent of exchange rate overshooting can be justified. It is worth noting that the case for intervention in this instance is not in any way due to inefficiency in the foreign exchange market. The rate produced by the market is the correct rate, but because of “sticky” goods prices, there are short-run real exchange rate changes. Intervention to smooth the economic adjustment processThere may exist a rationale for intervention in the foreign exchange market to achieve a preferable exchange rate in the short run in order to permit a smoothing of the necessary adjustments that the economy, for various reasons, must undergo. The rationale for smoothing the adjustment process is that it is a painful process for those who have to adjust, and it is more acceptable at a controlled pace than at a market-determined pace, which can be quite abrupt. Suppose that a country has a persistent balance of payments surplus because the traded goods sector is too large relative to the non-traded sector. There will consequently be a tendency for an appreciation of the real exchange rate, which will encourage factors to move from the traded goods sector to the non-traded sector. If the authorities are concerned about the possibility of large transitional unemployment resulting from such an appreciation, they may try to moderate the appreciation to allow time for the traded goods sector to contract and the non-tradables sector to expand, so as to avoid what they consider to be excessive transitional unemployment costs. Corden coined the phrase “exchange rate protection” to describe an exchange rate policy whereby a country protects its tradable goods sector relative to the non-tradables sector, for example by preventing or slowing down an exchange rate appreciation that would otherwise take place.4 Exchange market intervention can compare favourably to other methods of protection for the purpose of slowing down the necessary adjustment, such as tariff protection. This is because exchange rate protection, which involves influencing the real exchange rate and with it the accumulation of reserves, must necessarily be a temporary method of protection, whereas tariffs and subsidies have a habit of becoming permanent features and, because of their explicit protective nature, tend to invite retaliation. It is worth emphasising that the adjustment arguments advanced for exchange rate intervention involve smoothing the adjustment process, not preventing it. Ideally, the exchange rate should be allowed to adjust towards its equilibrium rate at an optimum pace. It is the acceptance of the principle of exchange rate adjustment that ensures that the required changes in the economy do take place. Figure 1 |
Indicator | Rank | Change 2013-2014 |
---|---|---|
Direct investment flows inward ($bn) | 59 | -32 |
Direct investment flows inward (%) | 57 | -22 |
Cost-of-living index | 57 | -1 |
Compensation levels ($) | 57 | 0 |
Remuneration of management ($) | 56 | 0 |
Exchange rates | 50 | 5 |
Portfolio investment liabilities ($bn) | 50 | 4 |
Source: IMD World Competitiveness Yearbook 2013, 2014.
Old Switzerland vs. New Switzerland
Our discussion above has described the prospects of the Swiss economy in the context of the extraordinary circumstances of the world economy. In this section we show that, except for the fact that interest rates have never been negative in the history of the Swiss National Bank, the current economic circumstances are not exceptional.
Figure 6 depicts the performance of exchange rates, interest rates and inflation in Switzerland in the period 1979-2014. Between September 1979 and January 1985, the Swiss franc appreciated from USD 1.55 to USD 2.67. This followed a massive quantitative easing programme by the SNB coupled with lower interest rates, ultimately triggering a severe property bubble, which in turn forced the SNB to raise interest rates to ten per cent by 1990. The property bubble was of course accompanied by massive inflation, which peaked at more than seven per cent in 1989, as shown on the right side of Figure 6.
In other words, the Japanisation of the Swiss economy already occurred, but in 1985, not in 2015. More importantly, instead of following the path that Japan followed in the 1990s, the Swiss economy was resilient enough to restructure itself. By 1997 Switzerland was already ranked 12th in the IMD World Competitiveness Rankings, up from the 20th position in 1990.
Figure 6 can also be read in the context of the macro events of the last five years. As the SNB has stated several times in the last few months, when the decision to intervene and maintain an exchange rate floor of CHF 1.2/EUR 1 was announced in September 2011, the international context was much worse than it is now. Ireland had been bailed out in December 2010, and the US Federal Reserve had begun its second quantitative easing programme. Therefore, the pressure on the Swiss franc at that time was coming from both sides of the Atlantic (and not just from eurozone). Nonetheless, the question of whether the current conditions would be better had the SNB not intervened remains open.
Figure 6
Exchange rates, interest rates and inflation in Switzerland, 1979-2014
Source: Datastream.
Conclusion
This paper argues that Switzerland is suffering the effects of a currency war in which Switzerland is not one of the combatants. Currently, in an environment of a strong dollar and a weak euro, the US is partly satisfied because this allows oil prices – which are priced in dollars – to remain low (which gives the US a geopolitical advantage vis-à-vis Russia and Saudi Arabia) and makes imports of raw materials to the US inexpensive. For the eurozone, a weak euro is great for exporting countries, such as Germany, and it will allow Southern European economies to improve their domestic demand. Moreover, a strong euro is not sustainable in the long run if it requires higher interest rates and leads to a loss of competitiveness. Switzerland is caught in the middle because it prefers a strong euro (which allows it to be more competitive with respect to its major trading partner) and a weak dollar (Switzerland is an oil importer).
Any monetary intervention by the Swiss National Bank in the near future is currently unlikely, since the SNB has clearly lost its credibility by abandoning the Swiss franc’s peg to the euro. Given the size of the Swiss economy, Switzerland is not in a position to start a currency war with the US. The Swiss franc is not yet a “big currency” in world markets, especially compared to the dollar, euro, yen and renminbi. Therefore, the chances of the Swiss franc destabilising currency markets are slim.
Consequently, the near-term prospects for the Swiss economy are deflation and a slowdown, at least in 2015. For the most part, companies will resort to cost-cutting and trying to become more efficient rather than shedding staff. Switzerland has proven to be an innovative country and should weather this storm in the long run. Unlike Japan in 1985-1995, the prospects of a prolonged stagnation of the Swiss economy are slim.
- 1 JPMorgan Chase, Striking Facts 2015, January 2015.
- 2 Introductory remarks by Thomas Jordan, Chairman of the Governing Board of the Swiss National Bank, Media News Conference of the Swiss National Bank, Zurich, 19 March 2015.
- 3 Observatory of Economic Complexity.
- 4 Although the IMD World Competitiveness Rankings have been produced since 1990, data collected with the current methodology is only available since 1997.
- 5 Employment in the World Competitiveness Yearbook refers to the percentage of the population working relative to the total population, so it should not be interpreted as an inverse measure of unemployment.
The Fall in Long-term Interest Rates: Quantitative Easing Effect or Trend?
Cinzia Alcidi, Mikkel Barslund, Willem Pieter De Groen and Daniel Gros
When it was finally announced in January, there was little suspense left: the ECB will pursue quantitative easing (QE) starting March 2015 to the tune of €60 billion per month for at least 19 months. The announcement had been expected for some time, even if the size of the intervention was somewhat larger than expected. It follows through on President Mario Draghi’s intentions, laid out in September 2014, of adding €1 trillion to the ECB’s balance sheet. The ECB now joins the central banks of other major developed currencies in what can no longer be called “unconventional measures”. The Bank of Japan (BoJ) has conducted monetary policy this way in several rounds since 2001. The US Federal Reserve embarked on the first of its three QE rounds in 2008 with its programme of large-scale asset purchases aimed at driving down the yields at longer maturities. In the spring of 2009, the Bank of England began transactions within its asset purchase facility programme.
The motivation for the recent ECB actions have been largely debated and should be found in the eurozone’s prolonged weak recovery and very low inflation, with headline numbers entering the territory of outright deflation. With short-term policy rates already at the zero lower bound (or lower) and a disappointing take-up of the cheap bank funding under the longer-term refinancing operations programme, the outright purchase of government and private securities, that is QE, was a logical – if not uncontroversial – next step.
The question commanding the most attention now is: will it have an effect on the real economy?
The initial impact (i.e. news effect and/or discounting expectations) on financial markets appears substantive for both long-term interest rates and currency exchange rates. A lower value of the euro relative to other large currencies will benefit European companies in international competition. However, the euro began its slide against the dollar in the fall of 2014, making it unclear how much of the recent drop is the doing of QE. Lessons from previous episodes of QE are diverse. The dollar nominal effective exchange rate appreciated around the announcement of the initial QE in the US, while the moves later on were ambiguous and significantly smaller: depreciations around the announcements of QE2 and QE3, and appreciation around the announcement of the “operation twist”. The most recent Japanese announcement of quantitative and qualitative monetary easing caused the yen to depreciate, but past episodes of QE had no unambiguous effect on the currency. The impact of the Bank of England’s asset purchases was also ambiguous. The first QE was associated with a small depreciation of the pound sterling, while a small appreciation materialised at the time of the second QE.
In general, in a context of high policy “activism”, it is quite natural to attribute asset price movements to changes in the monetary stance. However, in reality they may also merely constitute a swing around fundamental trend values. This piece focuses on this point. We start with the observation that when looking at global trends for long-term interest and exchange rates over a long period, it emerges that large asset purchase programmes led by the US Federal Reserve and the Bank of England have not resulted in fundamental shifts in financial prices. In particular, a downward trend in long-term interest rates has been present since the early 1990s. This is to some extent at odds with economic literature focusing on episodes of QE which suggests that QE policies in Japan, the US and the UK had an impact on long-term interest rates and hence on key macroeconomic variables – although the findings indicate that the magnitude of the effect is very small in some specific cases and in general uncertain.
The transmission channels of QE
Broadly speaking, QE can be said to work directly by flattening the yield curve. In particular, the purchase of longer-dated government or private securities will compress long-term interest rates. Lower interest rates should induce private spending, thereby affecting income and inflation expectations. The effect of QE programmes on interest rates can be channelled by different transmission mechanisms, which in some cases can have a direct effect on the real economy, e.g. equity prices. Figure 7 illustrates these channels, which include expectations (confidence), learning about future policies (policy signalling), reallocation of portfolios towards alternative assets (portfolio rebalancing), direct injection of liquidity, increase in money through the credit multiplier and exchange rates.1
Figure 7
QE transmission channels
Source: Adopted from M. Joyce, A. Lasaosa, I. Stevens, M. Tong: The Financial Market Impact of Quantitative Easing in the United Kingdom, in: International Journal of Central Banking, Vol. 7, No. 3, 2011, pp. 113-161, here p. 201.
To what extent has QE had an effect? A survey of the literature
In practice, it is very difficult to disentangle the effects of QE on prices, i.e. interest and exchange rates, and even more difficult to assess its wider macroeconomic effects. Unscrambling causal effects is complicated by two factors which have to do with timing: first, there may be substantial lags in the transmission from financial market variables (say, long-term interest rates) to increased spending and general effects on the real economy. Second, expectations play a significant role in determining prices on financial markets. Some effects of QE on prices may show up well before the policies are instated if financial markets anticipated a QE programme before it was formally announced. This explains why the literature on the topic has flourished in recent times but has remained focused on specific episodes, despite the fact that the debate on unconventional monetary policy operations has become global.
As illustrated by the overview of the literature presented below, the size of the impact of QE on the real economy varies significantly across countries or regions, depending on the time of its implementation, and is characterised by high uncertainty. There are two main sources of uncertainty when trying to estimate the effects of QE on the economy. First and foremost, a host of factors will have been affecting the economy during the crisis period (when most QE programmes are launched). It is extremely difficult to disentangle them and isolate the effects of unconventional monetary policy. Second, most estimation methods require “heavy” assumptions (e.g. about transmission mechanisms at work) that can dramatically affect the results.
Before looking into specific findings, it is worth noting that a lot of emphasis in the QE literature is on long-term rates. Besides the fact that reducing long-term interest rates is an explicit objective of certain central banks’ QE policies (e.g. those of the BoJ and the Fed), such rates are key transmission channels between QE policies and the real economy. In relation to this point, Rudebusch et al. show that although there is no structural relationship between the term premium and GDP; a reduced-form empirical analysis supports the existence of an inverse relationship between the term premium and real economic activity.2
The main findings of the literature are summarised below by looking at the experience of individual countries.
The impact of QE in Japan
Japan was the first to introduce a QE programme in 2001, and it has had a long experience with QE. During the period 1999-2001, before the introduction of QE, the BoJ had followed a zero interest rate policy. However, when the dotcom bubble burst, the Japanese economy was hit by another negative shock and the risk of deflation emerged again. At that moment, it seemed necessary to respond to it with some policy innovation. The BoJ embarked on the purchase of Japanese government bonds as the main instrument for increasing reserves of financial institutions. The BoJ exited QE in March 2006 amid signs that deflation risks were fading. It is interesting to note that during that period, the central bank directly purchased only a limited amount of government securities. The largest amount of these securities was bought by the postal offices, which at that time were owned by the government. As a response to the global financial crisis, in 2008 the BoJ re-launched its government bond purchases and adopted a number of unconventional measures to promote financial stability. In October 2010, the BoJ introduced its comprehensive monetary easing policy, which differs from the typical QE programmes of other central banks by including purchases of risky assets in an effort to reduce term and risk premia and to respond to the re-emergence of deflation and a slowing recovery. The most recent and largest asset purchase programme was started in 2013 as the second arrow of “Abenomics”. Despite such a long reliance on this policy tool, the jury is still out on QE. Research on the effectiveness of Japan’s quantitative easing experiences has yielded mixed results, and most of them point to very limited effects on economic activity.
Ugai offers an interesting survey of the empirical studies looking into Japan’s QE experiences.3 His overview suggests that there is a certain consensus that QE helped reduce yields, but its effect on economic activity and inflation was found to be small in all studies. The survey also indicates that the BoJ’s commitment to maintain the QE policy shaped expectations towards a lasting zero interest rate, thereby lowering the yield curve. By contrast, the results are mixed as to whether an expansion of the monetary base and changes in the composition of the BOJ’s balance sheet led to portfolio rebalancing. Moreover, those studies focusing on QE’s transmission channels find that QE created an accommodative environment for corporate financing, in particular by containing financial institutions’ funding costs and staving off funding uncertainties. Nonetheless, the effects on demand and inflation are found to be limited, most likely due to a dysfunctional banking sector, which impaired the functioning of the credit channel, and to the banks’ deleveraging.
Berkmen revisits the question of whether QE and other unconventional monetary easing measures in Japan have been effective, extending the period of analysis to 2010.4 His empirical analysis shows some evidence that between 2008 and 2010 monetary easing supported economic activity only feebly and that it supported inflation to an even lesser extent. Similarly to Lam,5 Berkmen finds that QE in 2008-2010 had no statistically significant impact on inflation expectations. While the impact on demand was weak, it appears that this episode was more effective than the previous experience, possibly due to improvements in the banking and corporate sectors. This finding is consistent with the results from Baumeister and Benati, who suggest that the monetary policy transmission mechanism may have been stronger relative to the early 2000s, but Japan’s stable inflation expectations and relatively flat Phillips curve inhibited the effect of monetary measures on the real economy.6 Berkmen also points out that no evidence emerges that the 2008-10 QE had an effect on the exchange rate, leading to the conclusion that other channels transmitted the effect to the real economy.7
The impact of QE in the US
Between December 2008 and March 2010, the US Fed purchased more than $1.7 trillion in assets as part of the so-called QE1 programme. This represented 22 per cent of the $7.7 trillion stock of longer-term agency debt, fixed-rate agency mortgage-backed securities (MBS), and Treasury securities outstanding at the beginning of the programme. In November 2010, the Fed announced a programme to purchase $600 billion of long-term Treasury securities. The programme’s goal was to boost economic growth and bring inflation to levels more consistent with the Fed’s maximum employment and price stability mandate.
As explained earlier, the long-term interest rate is the key variable which enables the linking of the QE policy to the real economy. Using this as a starting point, Gagnon et al. measure the amount of duration the Fed removed from the market by rescaling the Fed purchases using the concept of “ten-year equivalents”, or the amount of ten-year par Treasury securities that would have the same duration as the portfolio of assets purchased.8 Between December 2008 and March 2010, the Federal Reserve purchased more than 20 per cent of the total stock of ten-year equivalents across the three asset classes mentioned above (longer-term agency debt, fixed-rate agency MBS and Treasury securities) outstanding at the beginning of the programmes. This reduced the supply to the private sector of assets with long duration and increased the supply of assets (bank reserves) with zero duration. This affected the risk premium on the assets being purchased and triggered portfolio rebalancing effects.
Gagnon et al. stress that the purchases of MBS posed the greatest operational challenge to the Fed, owing to their more complex nature and their heterogeneity compared to Treasuries, but most likely also produced the most important results.9 As the purchases of MBS and agency debt began at a time when liquidity in these markets was poor and the spreads of their yields to Treasury yields were unusually wide, the Fed’s purchases helped to improve market liquidity by providing a large buyer for these securities on a consistent basis. As a result, yields narrowed relative to Treasury yields. The authors conclude that the overall size of the reduction in the ten-year term premium appears to be somewhere between 30 and 100 basis points. While the effects appear to have spread from Treasury securities to corporate bonds and interest-rate swaps, the most noticeable impact was in the mortgage market, and the effect was even more powerful on longer-term interest rates on agency debt and agency MBS.
Baumeister and Benati assess the effect of a compression in the long-term yield spread, on both output growth and inflation, induced by central banks’ asset purchases within an environment in which the policy rate is constrained at the zero lower bound.10 In the case of the Fed’s initial QE programme, the model simulations, based on a counterfactual analysis, suggest that in the absence of policy interventions, the US economy would have been in deflation until the third quarter of 2009, with annualised inflation rates as low as -1 per cent. Real GDP would have been 0.9 percentage points lower, and unemployment would have been 0.75 percentage points higher, reaching a level of about 10.6 per cent in the fourth quarter of 2009.
Chung et al. find effects which are also not negligible.11 Based on counterfactual model simulations, they find that the past and (at that time) projected expansion of the Federal Reserve’s securities holdings were roughly equivalent to a 300 basis point reduction in policy interest rates (from 2009 through 2012). Model simulations suggest that such stimulus kept the unemployment rate 1.5 percentage points lower by 2012 than what it would have been absent the purchases. The authors also argue that the asset purchases probably prevented the US economy from falling into deflation.
Liu and Mumtaz, by using a change-point VAR model, estimate that the Fed’s first asset purchase programme reduced ten-year spreads by an average of 90 basis points over the crisis period.12 Without the programme, they estimate that the unemployment rate would have been 0.7 percentage points higher and inflation, on average, one percentage point lower in 2010.
Chen et al. focused only on the second large asset purchase programme (QE2) and, based on simulations run in an estimated medium-scale DSGE model, conclude that the effects of the policy on GDP growth and inflation were moderate but had a lasting impact on GDP.13 They argue that the reason asset purchase programmes are in principle effective at stimulating the economy is the existence of limits to arbitrage and market segmentation between short-term and long-term government bonds. Indeed, unlike QE1, the QE2 period did not experience the high financial turbulence that could encourage stronger financial segmentation. Accordingly, their data provide little support that such frictions are pervasive. For this reason, the overall effects on GDP growth are estimated not to exceed half a percentage point, and the inflationary consequences of asset purchase programmes are even smaller. In the exercise, the authors also assess the effect of higher financial fragmentation. Their results indicate that this would have a larger impact on real GDP, but it is much smaller than the results found in the studies by Baumeister and Benati and Chung et al., which use different methods and assumptions regarding the risk premium.14 In particular, Cúrdia and Ferrero find that the effect of QE1 on GDP growth and inflation increases significantly when combined with a commitment to keep interest rates low for some period of time.15 This would suggest that the magnitude of programme effects depends greatly on expectations for interest rate policy. They also add that these effects are weaker and more uncertain than the effects of conventional interest rate policy. This would imply that communication about future rates could have stronger effects than guidance about the exit from QE.
The impact of QE in the UK
When looking at that experience of the UK, most of the existing studies have tried to estimate the impact of QE following model-based approaches and have focused on the first round of purchases, up to 2011. They quantify the impact on GDP and inflation, simulating the effect of a fall in longer-term government bond yields or of an increase in the money supply. The Bank of England provides an overview of the studies on the subject matter and adds some new results.16
The common starting point is that the peak impact of QE on ten-year gilt yields is estimated at about 100 basis points. In econometric models like structural VAR and times series approaches, this translates into an increase in the GDP of about 1.5 percentage points and around a one percentage point increase in inflation. In the context of a monetary approach, the Bank of England asset purchase programme also leads to a one percentage point increase in inflation but to larger estimated increases in GDP of about two percentage points. Lastly, when trying to account for the portfolio rebalancing effect induced by monetary easing, the effect on asset prices and thus demand translates into a peak impact on GDP of 1.5-2.5 percentage points. As clearly acknowledged by Joyce et al., while all estimates are significant, they are also highly uncertain, as none of the methods are able to capture in a proper fashion the transmission channels at work.17
Baumeister and Benati proposed counterfactual estimates of QE for the year 2009, for which they assume that the spread was 50 basis points higher than it has historically been. They find that without QE, the UK would have fallen into deflation (-4 per cent) and in recession with a trough of -12 per cent at an annual rate in the first quarter of 2009.18
It is interesting to note that the literature above focuses only on the first period of QE. The UK did not have clearly distinguished rounds of QE like the US, but in October 2011 the programme was expanded after about a year of maintenance. If the findings of Chen et al. about different rounds of QE in the US were to apply also to the UK, one could argue that the significant impact on the real economy is not independent from the conditions in financial markets in 2009.19
To summarise, in the case of Japan it emerges that the first QE was ineffective vis-á-vis the real economy, and the second QE had just a very small effect on demand and none on inflation. In the case of the US, evidence suggests that QE1 was the most effective in terms of unemployment and inflation, while QE2 was far less effective. In the case of the UK, most studies seem to suggest that the effects were economically significant both on GDP and inflation in the first phase of the programme (until 2011), but the uncertainty around the magnitude of the impact is considerably high.
Figure 8
Ten-year government bond yields, 1990-2015
Source: OECD.
Development of long-term interest rates
As noted above, inferences about the effects of QE are highly uncertain. However, based on the literature review, an effect on the long-term interest rate of no more than 100 basis points seems to be a central figure which shows up in the assessments of QE in both the UK and the US. In order to have a longer-term perspective of the behaviour of interest rates and highlight the shock induced by QE policies, Figure 8 plots the yields of ten-year government bonds for the UK, US and Germany beginning in 1990 and covering the “Great Moderation” period and the financial crisis until the present.
Three findings stand out. First, there is a clear secular downward trend, which has reduced the yield on German bunds from around nine per cent in 1990 to close to zero today. On average, long-term interest rates have fallen by approximately 40 basis points per year during the period shown. This implies that the impact of QE should appear as a reduction larger than the average trend rate of reduction. Secondly, note that Germany was already close to the zero lower bound for longer-dated government bonds before the start of the ECB purchase programme. Thirdly and most importantly, until rather recently long-term interest rates in the three regions moved in lock-step with one another. This is particularly the case for the periods in which the US Fed and the Bank of England have intervened with asset purchases. As shown in Figure 9, the correlation between the monthly changes in the yields of German bunds and US treasuries increased from 2009 through early 2014.
Figure 9
Correlation of monthly yield changes between US treasuries and German
bunds
Note: Correlations are calculated on a three-year rolling basis ending at the date shown.
Source: Own calculations based on OECD.
Hence a natural question arises: if QE was successful in compressing long-term interest rates, should not one have observed a divergence between US and German rates? In order to address this question, a systematic and thorough econometric analysis is needed, yet such a fact could provide an explanation for the behaviour in exchange rates. The limited impact of QE on US yields relative to German yields, given the existing long-term trend, may help to explain why QE had no uniform impact on exchange rates. A widening of the differences in yields would be expected to affect exchange rates. In reality, the dollar effective exchange rate moved little, the yen depreciated and the pound sterling appreciated. The view that QE had little impact on yields in the US, the UK and Japan (relative to Germany) is also compatible with the observation that inflation rates have not increased in a sustained manner after QE.20
Conclusion
It is too early to make an assessment of the effect of the ECB’s decision to undertake a QE programme. The initial impact has apparently been successful in achieving the goals of reducing long-term interest rates and of an exchange rate depreciation. Nonetheless, it is not clear how much of it is a direct effect of QE. The euro’s path of depreciation precedes the ECB’s QE announcement. Lessons from individual QE episodes in Japan, the US and the UK do suggest that the long-term effects of such policies are positive. However, a comparative look at the behaviour of long-term interest rates in the US and Germany indicates that the Fed’s QE had little effect on US long-term yields relative to German yields. A similar observation holds for exchange rates. The case of the recent ECB asset purchase programme, which is being associated with a depreciation of the euro and falling long-term rates relative to those of the US, seems to suggest a different story. If this is the first time QE succeeds in shifting expectations permanently, it would be quite welcome.
- 1 M. Joyce, A. Lasaosa, I. Stevens, M. Tong: The Financial Market Impact of Quantitative Easing in the United Kingdom, in: International Journal of Central Banking, Vol. 7, No. 3, 2011, pp. 113-161.
- 2 G.D. Rudebusch, B.P. Sack, E.T. Swanson: Macroeconomic Implications of Changes in the Term Premium, in: Federal Reserve Bank of St. Louis Review, Vol. 89, No. 4, July/August 2007, pp. 241-69.
- 3 H. Ugai: Effects of the Quantitative Easing Policy: A Survey of Empirical Analyses, Monetary and Economic Studies, March 2007.
- 4 S.P. Berkmen: Bank of Japan’s Quantitative and Credit Easing: Are They Now More Effective?, IMF Working Paper No. 12/2, International Monetary Fund, Washington DC 2012.
- 5 R.W. Lam: Bank of Japan’s Monetary Easing Measures: Are They Powerful and Comprehensive?, IMF Working Paper No. 11/264, International Monetary Fund, Washington DC 2011.
- 6 C. Baumeister, L. Benati: Unconventional Monetary Policy and the Great Recession: Estimating the Macroeconomic Effects of a Spread Compression at the Zero Lower Bound, in: International Journal of Central Banking, Vol. 9, No. 2, 2013, pp. 165-212.
- 7 S.P. Berkmen, op. cit.
- 8 J. Gagnon, M. Ruskin, J. Remache, B. Sack: Large-Scale Asset Purchases by the Federal Reserve: Did They Work?, Federal Reserve Bank of New York, Staff Report No. 441, March 2010.
- 9 Ibid.
- 10 C. Baumeister, L. Benati, op. cit.
- 11 H. Chung, J-P. Laforte, D. Reifschneider, J.C. Williams: Have We Underestimated the Likelihood and Severity of Zero Lower Bound Events?, unpublished paper, Federal Reserve Board and Federal Reserve Bank of San Francisco, 2011.
- 12 P. Liu, H. Mumtaz, K. Theodoridis, F. Zanetti: Changing Macroeconomic Dynamics at the Zero Lower Bound, manuscript, 2013.
- 13 H. Chen, V. Cúrdia, A. Ferrero: The Macroeconomic Effects of Large-Scale Asset Purchase Programs, Federal Reserve Bank of San Francisco Working Paper 2012-22, October 2012.
- 14 C. Baumeister, L. Benati, op. cit.; H. Chung et al., op. cit.
- 15 V. Cúrdia, A. Ferrero: How Stimulatory are Large-Scale Asset Purchases?, Federal Reserve of San Francisco Economic Letters, August 2013.
- 16 M. Joyce et al., op. cit.
- 17 M. Joyce et al., op. cit.
- 18 C. Baumeister, L. Benati, op. cit.
- 19 H. Chen et al., op. cit.
- 20 C. Alcidi, W.P. De Groen, D. Gros: Lessons from Quantitative Easing: Much ado about so little?, Paper prepared for the European Parliament, 2015.