Market developments in 2007/08 made it increasingly likely that bank lending to the private sector would get cut back sharply in 2009. But loans create deposits. So, a cut in bank loans could be expected (other things being equal) to cause a commensurate decline in bank deposits, and hence in the liquidity of the non-bank private sector. That, however, could have set off a further adverse, and self-amplifying, round to the deflationary spiral. Quantitative easing was necessary to try to break the connection between cuts in bank lending and resultant falls in liquidity outside the banking system.
How Quantitative Easing Operates
Central bank actions that can be roughly categorised under the banner ‘quantitative easing' fall into several categories:
• So-called credit easing - central banks respond to falling bank lending by lending direct to the private sector. This is largely the form of quantitative easing seen from the Fed.
• Purchases of government securities from the non-bank sector (as in the Bank of England).
• Vastly extended open market operations (OMOs), as seen from the ECB and earlier from the BoJ.
The Effects of Quantitative Easing
The effects on broad money of the various types of quantitative easing have been small - the multipliers have been tiny. There are at least four possible explanations for this: 1) the collapse of wholesale markets left banks with ‘funding gaps' unwilling to lend regardless. 2) The collapse of the interbank market left banks unwilling to lend to each other. 3) Lending to households and corporates on anything but tough terms is perceived as too risky. 4) Banks were even unwilling to purchase debt securities given a fear of future interest rate increases. Lowering remuneration on a portion of reserves might help a bit here.
Fortunately, QE has also helped to lower bond yields, helped risky assets to rally and, in the US and UK, encouraged exchange rates to weaken (although obviously at the ‘expense' of others - exchange rate weakness being a channel which can't work for everyone).
The Effect on Risky Assets Has Been Sizeable
In the first place, asset purchases by the central banks raise the price and reduce the yield of the assets bought, and the secondary effect of the generalised increase in liquidity should raise asset prices more broadly (more in the case of CE and QE than traditional OMO). Again, one of the problems of analysis is that QE (and OMO) is just one factor influencing the path of asset prices. One way of trying to get around this problem is to look at event studies, i.e., the change in asset prices immediately after an unexpected announcement of a change to QE. This is normally a rapid effect, so one can hope to be able to isolate and identify the effect, but that effect will depend on expectations, which in the context of a previously untried unconventional measure may be subject to considerable revision. Also one should recall that the announcement may be made to try to offset some other adverse shock.
Nevertheless, we have looked at the effect on government bond yields (2, 5, 10 years), corporate debt yields (BBB) and foreign exchange rates of a set of QE-related announcements in our four currency areas. This analysis suggests sizeable announcements effects in some instances. Analysis done by the IMF (Panacea, Curse, or Nonevent? Unconventional Monetary Policy in the United Kingdom, Andre Meier, August 2009) suggested that the introduction of QE lowered gilt yields in the order of 40-100bp.
More generally, one cannot fail to be impressed by the speed and strength of the recovery in asset prices and in confidence more widely since the nadir in March 2009. If there is one word commonly used to explain why this has happened, it is ‘liquidity'. The main source of this injection of liquidity has been the QE programmes. One certainly cannot prove that, without the various QE programmes, this recovery would not have happened, but their introduction and operation was surely a major overall benefit, even without a ‘multiplier'.
When to Stop QE
But how do you know when enough is enough? How do you calibrate the programme? There are various ways. In particular, one can use intermediate targets or final targets. Let us start with intermediate targets.
Different Approaches to Choosing Amount of QE
M4 as an intermediate target in the UK: QE in the UK was originally intended to raise M4 by a similar magnitude to the required increase in nominal GDP (where projections suggested a shortfall in nominal GDP of at least 5%). 5% of GDP is around £70 billion. The Bank of England recognised that strains in the financial system meant banks were less likely to increase their lending substantially following an increase in reserves and acknowledged that it was "also possible that the eventual increase in nominal spending might be somewhat smaller than the increase in the stock of broad money resulting from the asset purchases". It agreed therefore that reserves should initially be increased in the range of £50-100 billion and finally decided on £75 billion. Nominal income growth of 5%, comprising 2% inflation and 3% real growth, was the ultimate objective.
Borrower needs and market impacts determined size of Fed programmes: The size of bank reserves was not a focus of the Fed in determining the size of its ‘credit easing' programmes. The Fed's policies focused more on reducing credit spreads and improving the functioning of private credit markets. Bernanke summarised the approach to the programme's scale as follows in a speech at the LSE in early 2009: "When markets are illiquid and private arbitrage is impaired by balance sheet constraints and other factors, as at present, one dollar of longer-term securities purchases is unlikely to have the same impact on financial markets and the economy as a dollar of lending to banks, which has in turn a different effect than a dollar of lending to support the commercial paper market. Because various types of lending have heterogeneous effects, the stance of Fed policy in the current regime - in contrast to a QE regime - is not easily summarized by a single number, such as the quantity of excess reserves or the size of the monetary base. In addition, the usage of Federal Reserve credit is determined in large part by borrower needs and thus will tend to increase when market conditions worsen and decline when market conditions improve."
Leaving it up to the banks in the Eurozone: In the Eurozone, in contrast, no attempt appears to have been made by the ECB to assess what volume of liquidity to add to the system. Instead, the ECB has encouraged the banks to bid for whatever volume of reserves that they chose, on the easy terms and increasingly long maturities offered by the ECB.
Trial and error in Japan: During its own earlier experiment with QE, the March 2001 BoJ minutes reveal that the BoJ was uncertain about the ultimate effects of expanding reserves balances and must therefore have been uncertain on the ultimate size required to achieve its objectives. How the BoJ arrived at its initial targets for reserve balances is not clear. According to the minutes, if the BoJ set the outstanding balance of current accounts at the Bank as its operating target, "and increased its amount sufficiently" then a few members expected the overnight call rate to decline to close to 0%. They also thought that the increase in the current account balance itself might ease monetary conditions further "although uncertainty remained as to the extent". A few members thought that the economic effects "were uncertain in some respects", but that the Bank, "in view of the economic situation, should try out the measure as long as no significant harmful effect could be anticipated".
In terms of the BoJ's more recent actions, in his regular press conference on October 14, 2009, our Japanese team note that BoJ Governor Shirakawa suggested the termination of commercial paper/corporate bond buying, and of the special fund supplying operation to facilitate corporate financing. In the case of the former, there have been virtually no bids from financial institutions so that outstanding balances are negligible.
But QE Has Succeeded Without Meeting Monetary Objectives
While the increase in commercial bank reserves at the central bank has in all cases been massive, there has been virtually no follow-through into faster expansion in the broader M4 stocks, nor in bank lending to the private sector. The intermediate monetary objectives for the UK have not been achieved and recently broad monetary aggregates in the US have been weakening. This does not, however, mean that the overall exercise has by the same token been a failure.
Circumstances are somewhat akin to the experiments with monetary targeting in the UK and the US between 1979 and 1982. Then, monetary targets were adopted as a means to lower inflation. The relationships between the targeted monetary aggregate and its supposed determinants then went haywire, and the targets were often missed by a significant amount, but inflation was brought down, just as intended.
Similarly now, liquidity has been injected into our systems, asset markets have recovered (almost miraculously since March), and exchange rates have fallen most in those countries (the UK, the US) pushing QE with greatest vigour. Now this may, of course, be just coincidental, and not causal. But it would also be a mistake to assume that it was not causal.
Asset Markets to Determine the End of QE
So where do we go from here? The problem is that the recovery so far has been centred on asset markets (and in the support from the public sector fiscal deficit), rather than on private sector expenditures. Personal savings ratios have been rising and, with much spare capacity, business capital expenditures low (and the inventory rebound will be temporary). In this context, central banks are, in most cases, rightly indicating that interest rates will probably need to remain at present rock-bottom low levels for an extended period.
At the same time, however, equity markets have rebounded vigorously, risk-aversion has fallen (admittedly from extreme levels) and the decline in exchange rates in the UK and US has gathered momentum. In this context, would it be right for central banks to inject yet more liquidity fuel into the system?
At the FRB Chicago Conference on September 25, FRB Governor Kevin Warsh warned that the Fed could terminate and even reverse its credit easing faster than markets currently appreciated, so long as economic and market conditions remained reasonably favourable.
Similarly, ECB President Jean-Claude Trichet has indicated that the one-year financing exercise of the ECB is likely to be the last one of its kind, so that the ECB is on track for winding down its programme of special expansionary measures.
In some contrast, the MPC of the Bank of England voted at its November 5 meeting to extend the QE programme by a further £25 billion, though this runs at half the pace of its prior operations (and one member voted against any increase). This contrast is fully explicable by the difference in recorded output changes between the UK, where the second estimate for 3Q was for a fall of 0.3%, and the US and the Euro-zone, which recovered by 0.7% (non-annualised) and 0.4% respectively, according to their own initial estimates. We rather doubt whether this recorded under-performance of the UK will persist, for a variety of reasons, including the greater policy stimulus here, the depreciation of the exchange rate, the strength of the labour market, and the apparent stabilisation of the housing market. Thus, we expect the UK to join in the recovery, albeit a fragile recovery, alongside the other developed economies. As and when this happens, we expect the BoE to follow the lead of the other central banks and announce a pause in its QE programme (probably at the February MPC meeting).
While our view is that asset market, and foreign exchange, developments will thus lead to a cessation of active QE programmes soon, this still leaves the commercial banking sector with a massive expansion in its reserve base, at the same time as it still needs to restore capital ratios and wishes to delever. Under these circumstances, pushing up banks' reserve base still further is largely useless, just going further into a liquidity trap. But at some stage banks might wish to utilise such reserves to buy assets and/or make loans. How central banks might manage this aspect of the exit strategy will be the subject of a future note.
Conclusion: Declare Victory and Prepare to Withdraw
To conclude, much of the intended monetary expansion from QE was short-circuited by running into a banking liquidity trap.
Nevertheless, the first-round effects of pumping much more liquidity into the non-bank private sector, as a result of the more ambitious US and UK schemes, has probably played a significant part in the asset market recovery that, in turn, was essential in stabilising the economy and preventing a second Great Depression. Perhaps on this front it is time to declare Victory and prepare for withdrawal.
Summary
Statistics South Africa published its revised and expanded GDP data set on November 24. The new series, which has been rebased to 2005 prices, shows that South Africa's 3Q GDP rose by 0.9%Q (seasonally adjusted and annualized) to deliver a -2.5%Y print. This was higher than consensus estimates of 0.3%Q (-2.6%Y), but lower than our admittedly bullish 1.8%Q forecast (-2.2%Y). An undershoot in manufacturing value add and the finance and real estate sector explained most of our forecast error. Retail and general government services surprised on the upside, however. Although the 3Q09 reading undershot our forecast, we maintain our full-year GDP growth forecast of -1.7%Y, thanks in part to historical revisions.
Minor Data Revisions This Time Around
In accordance with international best practice, an update of the base year and accompanying revisions to the GDP estimates have taken place every five years in South Africa. The 3Q09 data follow on from that standard, resulting in an upward revision of some 0.4%Y since 2005 (although from 2002 the improvement is just a modest 0.1%Y). In earlier exercises (1995 and 2000), GDP estimates were revised up by some 0.5% and 0.3%, respectively.
At the sectoral level, the 2002-08 revisions appear rather sizeable in some instances, with the largest revisions occurring, on average, within the construction (-2.0%Q) agriculture (-1.4%Q), manufacturing (+0.8%Q), finance (+1.0%Q) and electricity (+1.7%Q) sectors.
Welcome Recovery in Manufacturing - Although Below Expectations
Despite a historically high level of consistency between monthly industrial production data and the quarterly manufacturing GDP data, the 3Q09 GDP manufacturing reading of 7.6%Q was significantly lower than the 10.5%Q suggested by monthly IP data. As the manufacturing sector constitutes as much as 15% of GDP, this deviation was enough to explain roughly half (0.4pp) of our forecast error. According to Statistics South Africa, the 7.6%Q growth in this sector was driven primarily by strong performances in petroleum, metals and steel, as well as food production. It is encouraging to note, however, that the growth rate of manufacturing production has been revised upwards - in line with higher levels of aggregate economic activity.
Looking forward, we expect additional support from sectors such as vehicles, paper products and furniture as overall economic recovery gains traction, and maintain our full-year overall GDP forecast of 1.7%Y.
Finance and Real Estate Disappoint
A 1.5%Q contraction in finance and real estate was rather disappointing too, considering the sharp bounce in the domestic equity market, stabilization and early signs of recovery in the local property market, and a lower CPI deflator in 3Q09. Exceptionally weak growth in domestic credit uptake appears to have more than offset such improvements.
Agricultural Production Still Negative, 4Q Snapback Expected
Elsewhere, agricultural production remained in deep negative territory, in direct contrast to data published by the Department of Agriculture showing improvements in selected food crops (mainly cereals). We had expected a technical bounce in agricultural production after the rather dismal performance posted in 2Q09. That the sector contracted by a further 9.8%Q is concerning, as this may have implications for food prices in 2010. We look for a sharp snap-back in 4Q09 as activity normalizes.
Encouraging Signs Emerging
On the positive side, wholesale and retail trade printed ahead of our expectations, as seasonal adjustments provided a boost to a tepid set of monthly sales statistics. Transport and communication, general government and personal services maintained their positive growth momentum, with construction activity providing its usual support base too.
Conclusion
In conclusion, the expanded and rebased GDP series shows that South Africa's growth performance is a touch better than previously expected. Going forward, we look for a technical bounce in agricultural activity, as well as mining production, as strike activity unwinds. We also expect manufacturing activity to remain in positive territory, joined by finance and real estate. We believe that the improving fortunes of these sectors point to a broad-based recovery that should give the SARB enough comfort that its aggressive monetary policy easing since December 2008 is already yielding a growth dividend, thereby obviating the need for further policy stimulus.
Read Full Article »