QE in Australia: What would it look like?
Speaking Notes for ABE Lunchtime Briefing: QE in Australia: What would it look like?
Sydney, 21 August 2019
It has become very fashionable to dismiss the effectiveness of monetary policy, so I want to spend a bit of time defending the effectiveness of monetary policy in general before talking about the potential effectiveness of QE in particular.
There is actually an overwhelming academic consensus that QE works, but for whatever reason, that academic consensus has not penetrated into the broader public debate.
Some people have lost faith in the effectiveness of monetary policy because of a mistaken belief that monetary policy in the US and Australia has been easy because interest rates have been low.
The fact that central banks around the world have been systematically undershooting their inflation targets has also contributed to this perception that monetary policy isn’t working.
A much better indicator of the effective stance of monetary policy is not the level of interest rates or the quantity of asset purchases, but inflation expectations and inflation outcomes.
On that measure, monetary policy in both the US and Australia has systematically erred on the side of being too tight in recent years.
Inflation expectations implied by Australia’s bond market have collapsed even more so than the US.
The other reason for thinking monetary policy has been too tight despite the level of interest rates is that markets have often been pricing in future tightening, even if that tightening that never actually eventuates.
To give a very specific example from the US, in August 2009, at the end of QE1, Fed fund futures were pricing in 300 bp of tightening over the next two years.
Now, we know that’s not how things turned out, but credit markets had that expectation priced into them at the time. So policy has often been a lot tighter than policy instruments would suggest, usually on the back of guidance from the central bank itself.
Phil Lowe spent much of the last two years telling us the next move in interest rates would be higher.
The undershooting of inflation targets has been enormously damaging to the credibility of monetary policy, because it has promoted a belief that central banks can’t hit their targets, when in fact this failure has been either a deliberate policy choice or a policy error, or a combination of the two.
Both those things are perfectly fixable.
In Australia, the choice component reflects an explicit trade-off the RBA has made against financial stability concerns.
I won’t go into the merits of that trade-off now, but it’s a large part of the reason why we are sitting here talking QE.
The policy error has of course been in relation to the NAIRU. The Fed has made essentially the same error, but it was made worse in Australia by the RBA conditioning the inflation outlook almost entirely on the labour market, to the extent that you could be easily forgiven the RBA is effectively targeting the unemployment rate rather than the inflation rate.
I recall with fond nostalgia Ian Macfarlane being very dismissive of the NAIRU and saying it wasn’t a factor in for Australian monetary policy, which might be one reason why he had no trouble hitting the inflation target.
So we shouldn’t confuse bad policy choices or policy errors for ineffective monetary policy, but that’s what some commentators are doing in Australia at the moment.
It is sometimes claimed that monetary policy transmission has become less effective.
Chris Kent did a very good job of debunking that idea in a speech last week, but even if you think that’s true, the logical implication is that we should be using monetary policy even more aggressively, not less, in order to get the desired effect.
The beauty of QE is that there is no in-principle limit to how much of it you can do. Even if you think QE has only a small effect, you can always calibrate the size of QE to get the desired macroeconomic outcomes.
Guy Debelle claimed in a speech last year that he thought QE was subject to diminishing returns, but he also acknowledged that the literature was very much against him on that question.
So what effect did QE have in the US? We are fortunate that QE in the US was conducted in a number of distinct episodes, each of those episodes had different characteristics and was framed somewhat differently, so we have some exploitable variation we can use to tease out its effect.
QE is an instrument and not a policy, so its effect very much depends on how it is conditioned by the central bank.
The academic literature on QE in the US typically focuses on the portfolio balance effect arising from the Fed’s purchases of US Treasuries, which is how the Fed itself saw QE working.
That literature finds that QE lowered long-term interest rates by around 100 bps and was equivalent to around a 250 basis point reduction in the Fed funds rate, with corresponding effects on macroeconomic variables, based on asset purchases equal to around 25-29% of GDP, depending on what asset purchases you choose to count.
As a rule of thumb, the literature on the US experience suggests that asset purchases of around 1.5% of GDP buys the equivalent of a 25 bps rate cut.
However, these estimates should be viewed very much as a lower bound for QE’s effectiveness, because the Fed went out of its way to short-circuit the effect QE would have, concerned it would lead to excessive inflation, as a result of which the Fed ended up badly undershooting on inflation.
The Fed was fearful that, if anything, QE would be too potent.
The Fed’s decision to pay interest of reserves to put a floor under the Fed funds rate meant that, for the most part, QE just accommodated the increased demand for reserves on the part of financial institutions and didn’t make it out into the broader economy via growth in money and credit aggregates.
That decision didn’t take away from the portfolio balance effect and its transmission to the rest of the economy, but QE could have been much more effective if the Fed had not paid interest on reserves and had not constantly talked about exiting from QE, leading markets to price in future tightening.
If the RBA were to avoid those two mistakes, then there is every reason to think QE could be both more potent and smaller in size. Paradoxically, the more aggressively you use monetary policy, the less you have to do.
While QE is generally conceived to work through lowering long-term interest rates, that is only the static or liquidity effect. The dynamic effect is to raise expectations for economic growth and inflation and this is also reflected in rising longer-term interest rates.
If you look at the US experience, long-term interest rates consistently rose during QE episodes and fell during the intervening periods.
That in itself tells you that QE was working.
This is also why predictions that the end of QE would see rising interest rates have not been borne out.
So how should the RBA go about implementing QE, if needed? Like the US, I think the RBA should target the general structure of interest rates and, through that rate structure, the exchange rate.
The stock of government debt securities in Australia may be an impediment to QE in so far as it would put the RBA in competition with ADI’s in meeting their regulatory requirements for HQLA, although the tightness in the supply of HQLA also works in favour of QE having a more pronounced effect.
But there is no shortage of non-government and non-debt securities the RBA could potentially work with.
Some have suggested a more targeted approach directed at specific points of friction in the cash market or in credit markets. To the extent that those frictions exist, the RBA should work on alleviating them anyway to improve the transmission of policy, quite apart from anything in might do with QE.
Another suggested approach is so-called people’s QE, where the central bank finances direct cash transfers to the public. I have no doubt this would be effective, but it would weaken, if not destroy, the institutional separation between monetary and fiscal policy that has been carefully built up in Australia since 1982, when the RBA stopped directly financing budget deficits.
We shouldn’t need to fundamentally change our institutions to get monetary policy to work, but we do need central bankers who understand that they have a singular responsibility for managing aggregate demand.
Phil Lowe has been calling on other arms of policy to help out monetary policy, but the whole point of having a floating exchange rate and an inflation targeting central bank is that we don’t have to rely on politicians for demand management.