Is QE still working?
February 8, 2012 3:24 pm
by Gavyn Davies
The Bank of England meets on Thursday with expectations running high that the MPC will announce a further large dose of quantitative easing. Even if they pass this month, which seems possible, this is likely to be only a temporary postponement. Whenever it comes, the next move will be another bout of “plain vanilla” QE, involving the purchase of £50-75bn of government bonds, and taking the overall Bank of England holdings to over one third of the total stock of gilts in issue.
Meanwhile, the Fed is still debating whether to increase its holdings of long dated securities, and if so whether to focus once again on government debt, or to re-open its purchases of mortgages. Any further QE would be contentious on the FOMC, but there is probably still a majority in favour.
Central bankers, unlike many others, have not lost faith in the efficacy of QE. The vast majority of them not only believe that additional asset purchases can further reduce long term bond yields at a time of zero short term interest rates, but also that this can increase real GDP growth, compared with what otherwise would have occurred. Are they right?
Most of the empirical evidence published since QE started in 2008 is on the side of optimism. Admittedly, a lot of it is published by the central banks themselves, who are scarcely the most independent source on this matter. But the weight of evidence is still impressive, and runs counter to those who believed from the start that QE would be a complete waste of time, if not worse.
First, there is little doubt that QE has significantly reduced the level of bond yields in the US and the UK, which is what it was primarily intended to do. The BIS reckons that the impact on the average bond yield across the curve has been fairly minor, amounting to about 25 basis points for each of the three doses of QE in the US, and to a total of only about 25 basis points for the much larger episode of QE in the UK in 2009/10. But two separate studies by the Bank of England (here and here) conclude that the impact of the UK action was about 100 basis points or more, and several other US studies suggest that the BIS estimates are on the low end of the range. The graph below shows how the yield curve has flattened markedly on both sides of the Atlantic since QE started:
The success of the central banks in reducing bond yields has come as a surprise to some economists, who believed that bonds and cash would be perfect substitutes when short rates hit zero. If that had been the case, then the central banks would not have been able to nudge bond yields downwards, no matter how much cash they had offered in exchange for them.
But in reality it turned out that bonds and cash were not perfect substitutes, so the central banks were able to raise bond prices (and cut yields). They needed to spend a considerable amount of extra cash to do this, but not the infinite amount which would be implied by perfect substitutability.
It is true that there is a lower limit to sustainable bond yields set by the Keynesian liquidity trap (explained in this earlier blog). Japanese experience suggests that this baseline is around 1.3 per cent, and the graph shows that the yields on maturities out to five years are already at or below these levels. So the central banks have now done all that they can do with QE in that part of the curve.
However, that still leaves the rest of the yield curve, especially the part between 10 and 30 years, where there is plenty of scope for a further decline in yields. And, of course, the central banks could choose to buy mortgage debt, corporate debt, or other private securities, where spreads could be substantially reduced by official purchases.
Admittedly, any of these options would imply that the central bank balance sheets are taking on even more risk. But that is the whole point of the strategy. As the private sector attempts to restore its overall risk levels to the levels held before QE, they bid up the prices of other risk assets, like equities. The Bank of England study quoted earlier suggests that the immediate impact on UK equity prices may have been as much as 20 per cent.
That leaves the question of how the strategy affects the rest of the economy. The empirical evidence on this (which is well summarised in this Banca d’Italia overview) is also supportive of the policy, so far. Key research papers suggest that real GDP in the UK may have been boosted by about 1.5 per cent, while that in the US may have risen by 0.6-3.0 per cent, compared to what otherwise would have occurred. Inflation also rose, by more than 1 per cent, but again that was the deliberate intention of the central bank, not the reverse.
While encouraging, this evidence does not prove that future injections of QE will have the same benign effects, either in scale or even in direction. Much of the evidence seems to indicate that the first bout of QE had the most significant impact on bond yields, with subsequent bouts having far less bang for the buck.
There are various reasons, including the increasing importance of the liquidity trap, and the waning impact of signalling effects about future central bank policy, that suggest this drop in efficacy may continue to be the case.
Nor should anyone feel entirely confident that the long term effects of QE on inflation are well understood. The historic correlations between the growth of central bank money and inflation in the long term are a cause for concern, as this earlier blog argued.
However, the growing consensus among central bankers is that their experiment with QE is still working. It was a shot in the dark, and a rather desperate one at that. But up to now it has had the desired effect, which is certainly a far better outcome than the alternative.
The success of the central banks in reducing bond yields has come as a surprise to some economists, who believed that bonds and cash would be perfect substitutes when short rates hit zero. If that had been the case, then the central banks would not have been able to nudge bond yields downwards, no matter how much cash they had offered in exchange for them.
But in reality it turned out that bonds and cash were not perfect substitutes, so the central banks were able to raise bond prices (and cut yields). They needed to spend a considerable amount of extra cash to do this, but not the infinite amount which would be implied by perfect substitutability.
It is true that there is a lower limit to sustainable bond yields set by the Keynesian liquidity trap (explained in this earlier blog). Japanese experience suggests that this baseline is around 1.3 per cent, and the graph shows that the yields on maturities out to five years are already at or below these levels. So the central banks have now done all that they can do with QE in that part of the curve.
However, that still leaves the rest of the yield curve, especially the part between 10 and 30 years, where there is plenty of scope for a further decline in yields. And, of course, the central banks could choose to buy mortgage debt, corporate debt, or other private securities, where spreads could be substantially reduced by official purchases.
Admittedly, any of these options would imply that the central bank balance sheets are taking on even more risk. But that is the whole point of the strategy. As the private sector attempts to restore its overall risk levels to the levels held before QE, they bid up the prices of other risk assets, like equities. The Bank of England study quoted earlier suggests that the immediate impact on UK equity prices may have been as much as 20 per cent.
That leaves the question of how the strategy affects the rest of the economy. The empirical evidence on this (which is well summarised in this Banca d’Italia overview) is also supportive of the policy, so far. Key research papers suggest that real GDP in the UK may have been boosted by about 1.5 per cent, while that in the US may have risen by 0.6-3.0 per cent, compared to what otherwise would have occurred. Inflation also rose, by more than 1 per cent, but again that was the deliberate intention of the central bank, not the reverse.
While encouraging, this evidence does not prove that future injections of QE will have the same benign effects, either in scale or even in direction. Much of the evidence seems to indicate that the first bout of QE had the most significant impact on bond yields, with subsequent bouts having far less bang for the buck.
There are various reasons, including the increasing importance of the liquidity trap, and the waning impact of signalling effects about future central bank policy, that suggest this drop in efficacy may continue to be the case.
Nor should anyone feel entirely confident that the long term effects of QE on inflation are well understood. The historic correlations between the growth of central bank money and inflation in the long term are a cause for concern, as this earlier blog argued.
However, the growing consensus among central bankers is that their experiment with QE is still working. It was a shot in the dark, and a rather desperate one at that. But up to now it has had the desired effect, which is certainly a far better outcome than the alternative.
0 comments:
Publicar un comentario