Transcript of Question & Answer Session On Lags

Thank you very much for a very illuminating lecture. Just had a question in relation to your views on the potential lags between possible asset price bubbles and long run price stability and therefore what implications, if any, you believe there may be for monetary policy in looking at those lagged outcomes, and/or macroprudential policy.

Thanks for the question, there's a lot packed into it. I think one of the things I've made clear over the years is that price cycles for assets and for other goods and services can happen for fundamental reasons, particularly because of stock–flow lags, and so it becomes quite difficult to disentangle whether something is a hog cycle, or whether it's genuinely speculative. And usually when there is a speculative element to increases in prices, that's something that started from something real and something fundamental. There are not that many bubbles in asset prices other than Bitcoin that I can think of where there wasn't something fundamental that explained that – and tulips as well. But, you know, two in four centuries.

But it's absolutely right and one of the things we're very aware of is, other than Bitcoin again, these things usually take a few years to build up and it does get very hard to disentangle how much is real and will stay there even after the speculative element goes away and how much of it is then people getting on the bandwagon. And oftentimes that's not speculation in the usual sense, it's panic buying. It's ‘if I don't get in now, I'll never get in’, particularly when we think about property markets, it's people panic buying more so than necessarily people speculating with the intent of making a capital gain later. So, those things can build over a number of years and certainly within my previous role, those longer-term dynamics are something we're very mindful of.

In terms of how macroprudential policies, or indeed any policy in the financial stability policy toolkit, again that's more about my old job than my current job, but those things can play out over long periods of time and we've certainly seen that in other jurisdictions. But my general observation is that many of the quantitative restrictions that have been placed in other countries … oftentimes they're not actually directed at changing price dynamics and certainly in this country it's been quite explicit, APRA has been quite explicit in saying what they've been doing has not been about changing pricing dynamics; it's about getting good lending standards.

But in jurisdictions where they've been thinking more about price dynamics in the setting of policy, what tends to happen is they end up doing several rounds of it. It works for a little while and then it all builds up again. And on some level that's telling you a couple of things. And one is there may well be a fundamental element to what is going on but just the pricing dynamics you get a hog cycle. And the other one is often, if the price of credit is lower than your macroeconomy needs then that's what's going to happen and so, the advantage of being inflation targeting central bank with a floating exchange rate is you do get to set monetary policy according to your domestic circumstances.

Thank you. I really like the concept of lags and trying to understand them better because we do ignore time, comparative static models are very dangerous. I just want to ask you about path dependency. You mentioned it briefly but when I think about long term unemployment you're immediately, or even short-term unemployment, you're thinking about hysteresis, deskilling and hence it being even harder to get those long-term unemployed people. And there's another recent example which is the NDIS where, we have to develop a whole supply side of providers that weren't there and the problem is, the challenge is that the policy could change because it's seen as failure before that supply is developed. So path dependency comes in a lot particularly in social policy and so how much does it come in to your monetary policy thinking? Because I guess it also, will it all be right? It'll all bounce back eventually if we're just patient enough? Wages will really rise? Or do we need to worry about it a bit more and be a bit more proactive?

Well that's basically the fundamental debate about the policy decision and certainly how important hysteresis is in the current situation of the labour market is something we've been debating. I mean, it's important to remember, yes, long-term unemployment has legged up. That's unfortunate, it's still one and a half percent. One of the things we also know is youth unemployment has actually come down quite a bit in the last few months. Certainly, you would at the very least need to make sure that you are pushing on that, and making sure that the flow of people out of unemployment back into paid employment is above average so that you can have that happen.

One of my favourite anecdotes actually comes from the late '90s. And for whatever reason I was reading an IT trade publication and there happened to be a story about various computer firms basically camping outside low-security prisons waiting for the former prisoners to be released who had done computer courses and giving them jobs. Now normally convicted felons are considered not particularly employable, but in a hot labour market, they're absolutely employable if they have done the computer course and people will give people a go if labour is scarce.

So I do wonder how much of the long-term skill atrophy idea is real. We did get there: it took a long time but we got there in the end, down to a half percent, but it did take a long time. It would certainly be less path dependent depending on if there is hysteresis. There's definitely social policy, things that can then happen to widen the spigot coming out. Make it easier to turn that spigot, if you like. It's not something you can fix with monetary policy directly if that's the case.

If I look at monetary policy, the thing that strikes you is that it seems more stable now, that is the Bank seems more reluctant to change interest rates. Even when they sort of say they're expecting one, it doesn't happen. But I think, I'd stick to that point I think that's true, that monetary policy has become, more stable.

The second thing I would say is that it's also clear in the last decade that the forecasting ability of the Bank gradually seemed to deteriorate and the forecast is weighted to the inflation, is a clear example of that. So are these at all connected? Could we say that because things are changing so quickly and lags are moving around so quickly, there's more confusion around? Confusion being inability to forecast, and therefore the Bank wants to sit on its hands much more than it used to, or could it be that the view of the impact of monetary policy has changed so that people think well, in general you want monetary policy to be more stable full stop and don't react until the last moment. Do you have any sort of comments on something like that?

Sure, sorry Bob, I'm not going to paraphrase. No, I'm also just, you didn't have the microphone so I hope that came over on the audio for the recording. But, the question is have we become reluctant to move? I don't think that's the case, I think it's just this is a setting that makes sense and with inflation …

I don't mean today, I just mean if you look at the long term …

Well I can't really speak for some of that period but where we've been sitting, certainly the last 18 months, two years, the narrative really hasn't changed that much. And, this time around I presented possibly the unchangiest set of forecasts to the Board that any Assistant Governor has ever done. Other than the fact that we extended six months you couldn't actually see the difference on our unemployment forecast between the previous one and the current one. We have the two lines – this is the previous one, this is the current forecast – and they were on top of each other, pretty much. You could barely see the difference between the lines.

So, certainly in the last couple of years, you're correct that it was very hard to see or to predict the slowdown in wages growth that's happened, and everybody mis-forecast that. The wages growth that we've seen does not fit with what a conventional econometric model would tell you. So, maybe there is a structural change there and that's one we've kind of said well, you know, and this morning in Parliament, the Governor was talking about it. There may be, some of it may be lags but some it may also be, there's issues about competition and so maybe the lag's lengthened or maybe, the NAIRU has come down. It's kind of hard to tell the difference between the lags have lengthened and that because unemployment is always falling. If you could magically hit the point where unemployment levelled out exactly at whatever the NAIRU was and stayed there you'd get exactly the same result with wages growth as if it kept going. So we just don't know whether it's the lengthening of the lag, or you have to get lower. You can't distinguish those two things econometrically.

To be fair the last couple of, the last two years, our inflation forecasts have been pretty spot on. We have had to, sort of make a revision to next quarter because of some government policy changes but I think we should be allowed. We shouldn't be blamed for that one you know. Childcare, we can't forecast government changes to childcare subsidy and systems, especially when we can't quite understand them. So there are always things that will come along like that that are not predictable. And I think one of things I think has been really difficult to get at, we had an enormous change in this economy. And it's interesting for me because I spent 10 years away from monetary policy world, in financial stability world, and then came back and told myself, told my staff think of me as Rip van Winkle. I was sort of paying attention but not to level of detail you were. I was over here thinking about macro prudential policy.

And, if there was one thing, and in fact in some ways this is a good graph to show, if there was one thing that you'd want to know about the Australian economy, if you'd gone to sleep for ten years and woken up and said tell me, what's the one big thing that you'd want to know about the Australian economy over the last ten years is the terms of trade and the consequent mining investment boom. I think one of the things we've learned over that ten years is, the effects of that, both in the upswing and in the downswing has been far more consequential than just conventionally looking at the direct effects would've told you and I think that's really what's been weighing on the labour market. So we've had resource booms before and big cycles in our commodity prices before, but not at this scale and not for this long and the effects have just been far more consequential than I think people realised. There's a question at the back so I think maybe we should take that.

Hi Luci, just curious, following on a little bit from what Bob was saying. Are there, what are there or are there consequences from an extended period of stable rates, of low stable rates? I mean I've heard people talk of sort of boiling frog syndrome where, people start to sort of lose their competitive juices and that sort of thing because money becomes so easy or so cheap. I mean is there anything to that or is there anything that you worry about, about being on hold for such a long period of time and at such a low level?

It's a good question, thank you. I don't know of any economic model that says that moving interest rates around more than you need to keeps people more awake, except possibly in the financial markets. And I think, the business community in Australia tell us that stability is something that they welcome, because it does make their planning easier. So I'm not sure that it would have the effect that you're describing and make them kind of complacent. I think it makes them more confident to make decisions and invest when there's less uncertainty about at that at least that thing.

There are a lot of things in the world that are not forecastable, the exchange rate being a big one. We know one of the things we do know to be true from economics is that the exchange rate is best modelled as a random walk and if that were not true I would not be here I would be on my private island. Because if you could make money out of that, that's what you'd do. That's not what we hear from the business community. What we hear from the business community in our liaison program and in all our discussions is that, a period of stability is confidence building for them and it makes it easier for them to plan. So, I'm not aware of any economic model that says giving people little surprises keeps them on their toes, keeps them innovating. I don't think that's the case at all. It's a bit like the whole issue about whether recessions are creative destruction. I don't think recessions are creative destruction; I think they're destructive destruction. I think creative destruction happens in expansions when things are tight and you have to work out ways to be more productive because you can't find workers. That's actually really what we are seeing in Japan at the moment; the labour market's so hot they're finding new ways to innovate and I think that's more likely to be creative destruction than a recession is.

I have a question more specific to, you were talking about structural breaks and how it's, a lot of people just say that there's a structural break at this point. Is a structural break then a hindsight kind of thing or is it something that we can model because from what I understand in your NAIRU predictions, you use 1977 as a structural break but after that you have assumed that the economy should work the same. I think so, if I'm not mistaken.

My recollection is that the labour market data monthly comes in, in 1978 I think it is, so that's probably what is going on there. One thing we do know is that allowing for structural breaks in variance do matter for estimates of the NAIRU and make them a bit more accurate more recently. And again, that's actually a sense in which the economy has become more stable. The variance appears to have come down. I just think as a general research agenda both for policy makers and for other commentators and academics that are interested in economic affairs thinking carefully about whether it's a structural break or whether it's something else, and thinking about how you might tell, is just something we need to take a bit more seriously and think a bit more systematically about.

When I talk about the people who say ‘this time it's different’, it's not usually people who have an econometric model and are doing a break test that are making those pronouncements. It's sort of the people who engage in commentary who will see unemployment is high, oh it must be because of skills mismatch. It's like, well, really? Did we all suddenly forget what we were able to do two years ago? No, we just had a really big recession and it takes a long time to work that down especially when in as was the case in some countries not all macroeconomic policies were working in the same direction. We were lucky that that wasn't the case here but there were countries where things were pulling in opposite directions. Thank you.

Maybe I will ask one more question.

Sure. Come to the microphones just so …

I mean it would be nice, this is a terrible question. It would be nice to give a lecture on lags and talk about how important they are, how they may have varied and end up with a statement which says something like, ‘and therefore what we need to do is’, right and then you say something about collecting more information or slowing up lags or speeding up? Because these questions used to be answered a long time ago, the unemployment data used to come out quarterly and people said ‘oh we need it monthly’. We asked people what they spend not what they planned to spend therefore we knew that sort of data. Is there anything apart from trying to slow down social media, is there anything that we should be thinking about doing?

No, that is a good question and I didn't set out a particular agenda here but I think what I really want to do is raise the question so people can think about what the right answers are. One thing that has come up in our public communication and in recent debates is we're one of the few countries that has a quarterly CPI. So, we have a quarterly CPI but the Board meets eleven times a year so, most meetings I don't actually have new inflation data to tell them about. Would it be better if we had a monthly one? Well yes, but there's a trade-off about cost, the Bureau of Statistics has other things it needs to do, its resources are very limited. And so, we need to think about whether the priority on that relative to something else.

Certainly, it would really help if you were getting things monthly, in theory you should be able to see if it's a change in trend sooner than if you're waiting every quarter. But of course, we also know from the labour force there's a bit of noise that you then have to trade off. So we now have daily house price data. Do we react to every day's … No, no we don't: we generally look at the monthly data. But it's kind of a good to get a, you get a sort of a ‘now cast’ of the current quarter or the current month. That's very useful. So yes, look, more data is always better, computer storage is getting bigger, tools are getting better.

So the capacity to use those big datasets and derive insights from them is directly relevant and I think in some sense what I would advocate, is thinking systematically about particularly where there are process lags, thinking carefully about, well, are the steps along the way that you can learn something about what's going on? So, say for example one of the questions we face, that relates to process lags is relative to the past, the construction of housing that's happening now involves a lot more apartments than it used to. And one thing we know is that apartment blocks take longer to build than detached … you can start building the houses and sell them and people can move in, instead of having to build the whole block of apartments before people can move in. So the lags between the building approval and the completions are changing. But we don't know enough about that.

So they're the sort of things where actually knowing more about the process lags would be helpful, that is literally one of things we grapple with in our forecasting. That is a structural change actually that is genuine, has genuinely happened and it has changed the lags. I probably should have mentioned that in the speech, but to be honest I already had a few examples.

Okay. So, let's all join me together and thank Luci.