Transcript of Question & Answer Session Economic Forecasting at the Reserve Bank of Australia

Thank you very much Chris, I’m Graham Wells and I’m the President of the Economic Society here, and it’s my job to lead the questions and comments and I’ll abuse my position. I guess the thing that was emphasised in your slides and which can’t be emphasised too much, is the uncertainty around forecasts and the users of forecasts tend to overlook this, and particularly politicians. In the Treasury forecasts would have had the same range of uncertainty as you’ve indicated in your monetary policy statements but because that uncertainty is minimised or overlooked that causes all sorts of problems in saying that we’re going to achieve this or we’re going to achieve that and so I think you’ve done a great job in trying to push this idea of having forecast uncertainty rather than just point forecasts. It doesn’t play out quite so obviously in the Reserve Bank as it does in the Treasury because budgets obviously rely on them and they’re kind of more point things but good luck with trying to press the forecast uncertainty button.

So we’ve got some roving mikes and I’d like the questions to be related to economic forecasting rather than a whole range of other topics that Chris might be able to talk about, so inclined, but so questions; down the front here.

Chris thanks for that of course. One of the things that you see in the theory of economics is that there is an understanding that the way in which the central bank assesses information transmits important information to the economy itself, so when the central bank says oh look we’ve changed our forecast from 2 per cent to 3 per cent for inflation let’s say, if there’s nothing changed in the underlying public information that the rest of the economy has then that’s either telling us that the central bank has changed something in the way it thinks about the economy, so it’s changing its preferences, or that it’s got some other source of information like the liaison that you were talking about gives us … When you’re considering how you put your forecasts out, does the Bank actually think about that impact on the markets? Because you can sort of see a nasty circle eventuating there if that is the case. Or do you think that the theorists do have it right in that is actually what is going on?

That’s a good question that probably has quite a philosophical element to it that if we think deeply about could become quite painful. But I think it’s a good question, maybe I’ll just refer back in answering that question to sort of this point about uncertainty. I think one of the things we did in trying to sort of generate those fan charts, these ones here is think carefully about how well do we forecast relative to how well others forecast and it turns out that more or less compared to consensus economic forecasts actually we weren’t … We couldn’t beat them in fact often they would beat us with regards to GDP. It’s not clear to me that any individual within the consensus beat consensus all the time and so there’s this tricky challenge where any sort of forecasters you can’t really just … The best is usually the average of a whole lot, probably including us and I don’t think as I said we don’t beat that average, we don’t have a tendency to beat that average. So that says we’ll just use the average, and if everyone used the average well a useful average actually wouldn’t exist. So this is the point in which I think it becomes a bit chicken and egg like.

I think there is the point you make though that there’s possible that when you change our assessment about developments and if that’s driven by special information we have maybe that’s of use, but I think that would probably go over state, even though our liaisons particularly valuable, in certain areas more so than others right, where you can meet or get on the phone and talk to a few number of large players in a particular industry and have a good understanding what’s happening there, but you’re not going to be able to do that for the economy as a whole. And typically our liaison doesn’t throw up something in terms of what’s likely to be coming that’s very different from the business surveys which are doing the same sort of thing. But the value of the liaison is occasionally on industries where there are just a few players it can be extremely valuable and then it’s really just trying to understand the mechanisms and what it is that sort of drive the motivations behind things, what are sort of the various cause or elements. So I don’t know if that’s a very satisfactory answer, I doubt there would be a time where we would be able to sort of say we are convinced that we’re going to beat everybody else, because in fact what we’re convinced about is we don’t know very much about the future, lots of things can happen.

Turns out on inflation I think it’s true, it was true at the time we did this that we do tend to beat consensus which is nice given we’re inflation targeting central bank, but probably not by much. So let me leave it at that, there was a question up the back there. The mike went that way.

Paul.

All right thank you Chris I was just wondering whether the reviewers or indeed staff at the Reserve Bank or yourself thought there was any role for regional modelling with an increased variation there seemed to be a great dispersion of economic activity across the states and across the nation, than there used to be in the past and I was wondering whether that was the reason why some of the relationships between the variables are changing in the models and whether Adrian and David thought that there was some benefit in … Had obviously systems of equations but also modelling them in certain regions to try and pick up differences in mining boom states and the like?

Look I don’t think they did not to my knowledge. It’s a good question though. I think the trouble when it comes to modelling is a model’s going to fit reality only so well and if it gets too complicated then it doesn’t become very useful, one because it probably might fit a particular episode but then it’s going to suffer from the problem of overfitting and down the track it will give you some sort of strange result that you hadn’t sort of contemplated, so it’s really important that models that do well just as forecasting instruments tend to be fairly simple, in fact it’s quite hard to beat single equation models of variables that are amazingly simple, but they don’t have any economics in underlying them and they don’t provide particularly deep insights that are useful for policy makers as to the nature of what is the best way to respond to shocks in order to try and help stabilise the economy one way or another.

Now when you’re talking about splitting the economy up into different sectors or different regions or states, in a way you’re adding a lot of complexity and sometimes that could be useful I think but I think it’s going to be hard to use that to sort of forecast the future. I think it could provide useful insights and try and understand what’s happening but I think that’s what we have been doing quite a lot of, of late, and that’s what a liaison program does all the time by industry and region look at things, and there is really is quite a divergence in conditions both during the upswing to the mining boom and then the down swing, between the mining states of Queensland and WA and then much of the rest of the economy; so I think it’s useful to be aware of those things but whether it’s useful to sort of model them and whether you get any really deep insights from doing that I’d have my doubts because you’d probably introduce as much noise as you do sort of generates useful signal.

One over there.

Okay and that gentleman too.

Chris how you going? This is a bit tangential to your modelling but it sort of relates to having unemployment forecast in your quarterly statement. The labour market at the moment we had a lot of strength late last year and it seems to have been coming off in terms of what the ABS has come out with in terms of job ads and that sort of thing, do you have a strong read on that? I mean is there a sense that it is decelerating or it’s just not clear at the moment, thanks.

Well for what they’re worth our forecasts I think in the February statement would have had I think conveyed the sense at least in the text that what we recorded up to the later part of 2015 was some very, very strong labour market outcomes in terms of employment growth in the months of October and November for example and we didn’t forecast that being sustained at those sort of growth rates. So we always forecast employment growth was likely to come off in the near term, just to sort of come back to more reasonable sort of trends. We had forecast a pretty stable unemployment rate for a time being because we thought growth would be still I’m talking over 2016 and a bit beyond, a bit below trend. I think it’s probably too soon to say and I wouldn’t just focus on something like job ads the ABS job vacancies which is a valuable measure of a forward looking measure for employment, that was a pretty healthy number the latest one that came out from that, so I think it’s sort of too early to say and the media release that came out yesterday sort of hinted at one of the sorts of new pieces of information that the board would be considering was whether the improvement in labour market conditions that was evident last year whether that’s continuing and I think it’s too early to say.

But it did seem to be that I guess a little bit more doubt perhaps had been a month or two earlier when the …

Oh I think it was just a reflection of the fact that often what we’re commenting on there is the facts, the employment, there hasn’t been much employment growth for a few months now. A weaker employment growth was to be expected after those very strong numbers but like I said sure a few measures of job ads have been sort of flat after having had a big trend rise, but then job vacancies were stronger so I think it’s a wait and see is the key there.

Chris you should have pointed to the dark blue patch there and said well we’re still in that range.

Ah well we are, in fact I think, but I’ve just tried to dissuade you from focussing on the central forecast but we’re actually probably pretty close to those right now but we haven’t moved very far along the line yet so time will tell.

Chris, I was just wondering whether these confidence projected forward contain the policy statements so reaction to that so, you find yourself on the extremity the Bank is going to do such and such and that’s going to carry forward with your projection is that already built in or are these just sort of a blind statistical projections?

It’s more the latter I don’t know I’d emphasise the blind part but I’d emphasise they’re really a statistical artefact so what we’re really doing is not generating a whole set of alternative internally consistent paths, there aren’t lots of little blue lines that are quite specific and central ones that happen under different scenarios and with interest rates responding and inflation doing something and unemployment doing something else, and then we do another one and we do some untold number of those and put them together and generate these forecasts, what we do here is we generate sort of our best guess the central forecast if you like and then just take a record of forecasting and then add some width to those to be attuned to that.

I don‘t have a problem with that then. You pointed out you take this data as accurate rather than what you’re saying before is that APRA seeing that already so that would carry through to make your uncertainty in your forecasting a great issue.

Well I think the people who’ve put these sort of fan charts together when they came up with the widths were taking account of that naturally but it’s a good point you raise and I think some others combined forward looking fan charts with backward looking fan charts to account for that and it all becomes quite complicated. What we do is we sort of … These are sort of consistent but then we also emphasise the fact, and in fact we did that I think quite recently saying well maybe one possibility is things will be revised. We surmised that it might be possible that the labour market which had been strong will be revised to be looking a bit weaker or the national accounts which hadn’t picked up that same kind of strength they were lagging might be revised one way or another could be revised stronger, they revised a little bit up and they’re a bit more consistent now. But I think we did that in the text just to sort of cut through that complexity.

So do we have two more perhaps.

Hello Chris, international students what interest is the foreign exchange rate because it matters a lot to my tuition and I notice that Australian dollars appreciates recently and I find as you say the foreign exchange rate have a close relationship with the commodity price, can you explain how the commodity price influence the foreign exchange? And I notice that the iron ore keep going down over the past two years and your dollars depreciate keep depreciating and I’m from China then something happen in China it’s like our economy still down and the realistic market is collapsed and we build too many houses but nobody want to buy it, so the governments want to cut the production of steel because steel … If they cut production on the steel then the demand for the iron ore will decrease a lot … A lot and things that a large percentage maybe 70 per cent of iron ore, I can remember export to China if the situation getting bad then what’s the impact that influence the exchange rate of the Australian dollar? And also I notice that yesterday the central bank of Australia release that interest rate remain 2 per cent and I want to know that …

Excuse me could you just cut it a little bit short and get to the point.

… Do you … the governments have space to carry a little bit to stimulate the economy further? That’s all my questions thank you.

Okay all right well I think I got most of that, I’ll try to take a stab at much of what you said. I think the main thing and we’ve published quite a bit about this, it’s very hard … Certainly very hard to forecast in fact I would recommend against forecasting exchange rates, but trying to model their behaviour in the past the best that I think you can do, the most reliable thing for Australia at least is there’s a strong relationship between the real exchange rate and the terms of trade, the ratio of export prices to import prices. So when the terms of trade were booming, when commodity prices were rising the exchange rate appreciated and that was a very helpful device to try to mitigate the inflationary pressures to try and provide a signal for resources to move around to enable the mining investment boom to occur without causing great and long standing difficulties in the economy. It was a helpful equilibrating mechanism and when the terms of trade turned around and started to fall significantly, eventually we’ve got an exchange rate depreciation which again is helping the economy to adjust to those lower commodity prices, and that’s working quite well.

In terms of the iron ore price sort of developments in China I think it is true that there’s been some influence no doubt from the weaker demand conditions in China, the excess capacity in the steel market, steel production has declined there, but at the same time when you look over the history of much of the decline in things like the iron ore price that a lot of that occurred probably before that weakness in the steel market in China became evident and was just the response the expected response in many respects, perhaps overdone a little bit but too lots of extra supply of iron ore coming into the market including from Australia which is a lot cost producer and what’s happened is some of that low cost supply has pushed out the higher cost supply including in China, so supply in China’s adjusting. So it’s quite a bit more complex than just demand is weak and the price has fallen. We’re taking advantage of a low cost iron ore increasing supply a lot.

And finally I mean the board has said for a time, they sort of talk about here that in the statement yesterday, the Governor’s statement that continued lower inflation would certainly sort of provide the scope for easier policy but that’s very much conditional on whether that’s determined by the board to be appropriate to lend further support to demand and that’s what they’re still to determine. As I just suggested perhaps in my answer to the other question some of the things you’re looking at are the outlook for inflation, whether the improvement in the labour market conditions are evident last year whether that’s continuing, so there’s not much more I can add to that in terms of the outlook for interest rates, I might leave that part there.

Yes one over here.

Thank you Chris. How have your uncertainty fans been received and who else is doing them and is anyone sort of following suit? I’m not aware of any Treasuries for example in Australia at least that are doing this as yet, and I don’t think the Commonwealth Treasury is either, and not even just around forecast even current data I know the ABS for example has been trying to get commentators, governments and others to forecast … To focus on trend data as opposed to seasoned adjusted data because of the inherent uncertainty in their data releases, so really just a question of how it’s been received and I suppose it’s looking forward what other policy making bodies might do as well with regard to similar forecasts.

I think it’s been reasonably well received amongst the general community, there’s an understanding of the fact that the future is uncertain and we have to generate policy decisions today taking that into account and then we have to review those decisions as the economy evolves and we can’t just imagine that we’re just going to set policy on a certain course and everything will work out for the best. So I think these help in communicating that notion. There are certainly many other central banks that do it in various different ways, the Bank of England probably most famous with their Rivers of Blood chart, just because it looked like a river of blood because it was red, I don’t know that anyone’s come up with a term for us here. So we’re certainly not the first to sort of pursue this concept and emphasise it, I think it has been something that maybe has been emphasised more by central banks. I think, I don’t want to say too much I think there’s been some attention by Treasury in thinking about the outlook for the budget, the Federal Treasury to be thinking about look the future isn’t certain, but I won’t say much more than that. But I think the place has been very well received as when I go around and we give presentations particularly to businesses and community leaders and so on that are part of our liaison program, once a quarter we go back out to the various parts of the country where we’ve had all those contacts and we discuss our views of the economy and in putting these sorts of things up I think it’s very helpful and they’re very understanding because they’re in a world where the future for their own businesses is extremely uncertain and they’re dealing with this all the time so it makes a lot of sense to them.

One last.

First and foremost thank you for an excellent briefing, may I focus on those lovely charts, lovely fans and say congratulations but also raise a couple of doubts and I wonder if you could address them. The first point is that whether or not they are well received the question is how they will transform both the debate and the decision making and if one looks at … We’re trying to get around the problem of put into position yep and the use of range rather than points is not new, it’s been done again and again and usually the pattern just from the fields I’ve been involved in seems to be that there’s initial level of interest but I mean ultimately people didn’t go back to the midpoint and it returns to square one, how do you sustain a continued focus on the range rather than the midpoint?

And the second doubt is that it does not of course answer the question of risk which you touched on briefly, this is still a range around that midpoint what of not exactly black swan events, but risks of being outside the range, say for example being a minus number below 90 per cent, below 70 per cent but a non trivial risk of a minus number there, again how does one input a non-trivial risk into media commentary into public discussion, into policy making and decision making, do you have any thoughts on these?

And it’s an excellent question. I think the first thing is if they are very low probability events but could have a significant effect on the economy the first part of the approach I think as a policy maker is to say well I certainly want to be awake to any signs that something like might be in the offering and so that’s where you want to be looking at as much data as you can, you want to not just be waiting for economic data but reading the news, that’s where things like the liaison program is helpful because you’re on the ground talking to businesses, but then sometimes events just happen and then the point is once those are starting to … If it’s the case that you’re getting something that’s really at a low probability but quite an important and very … Potentially adverse but not always adverse effect, you sort of respond at the time right. And I think this is where forecasts can be somewhat limited and probably the Lehmann’s collapse was a good example. Wake up one morning Lehmann’s has collapsed, the general environment is one of concern and financial fragility building and Lehmann’s collapse sort of leads to violent movements in financial markets and prices and you say this isn’t going to be good and in very quick time, before forecasters sat down and said oh I’ve got a new model that will perfectly deal with this thing called the GFC which no one I think really has developed even yet many years after, you just say what’s the starting point? The starting point is one of considerable fragility. What’s the direction? The direction’s down, the economy is not going to do well and policy makers respond quickly and then try and sort of go from there and that’s exactly what happened in our case, so that’s where I think that sort of in those sort of events you want to be attuned to them, you want to be thinking about them, looking for the sorts of triggers or the things might say something particularly good or particularly bad maybe about to happen, and then when it does you respond very quickly and usually models and fan charts and forecasts are probably the last of your concerns and you just react quickly and that’s happened I think during the Lehmann’s and sort of GFC crisis, let me leave it at that I think.

Okay well thank you very much Chris I don’t think I’ve ever been to a Reserve Bank presentation that wasn’t interesting and well prepared and you’re on the top side of that distribution, so thank you very much.

Thanks Graham.

And I’d like you to join in thanking him.