Most likely to succeed
A quantitative report from Citi has placed Australia at the top of the pile in relation to global markets. Paul Chanin, managing director and head of Asia quantitative research, speaks to Business Spectator's James Frost from Singapore and says:
.
James Frost: Paul, you’ve produced a quantitative study that places the Australian stock market at the top of a list of markets to invest in. Can you tell me a little bit about how you came to this conclusion?
Paul Chanin: Well, to start with this is a quant model. It’s a process which we’ve been running for about five years and use to rank many of the largest markets around the world using very objective criteria. The modelling process itself is rather unique as it tries to account for the differences between markets by building models that try and judge the attractiveness of each market, only using factors that are traditionally associated with predicting outperformance or underperformance of that market.
So, what we are saying now is that based on the factors that are important for Australia, things like dividend yields, long term price momentum and earnings stability, Australia looks relatively good. Better in fact than almost all the other countries in the world. So, therefore we find Australia is looking good.
JF: The report mentions that you have ranked 22 markets for which exchange traded funds were available. Does this mean the report is looking at Australia as a place to invest in terms of an index ETF or is it broader than that?
PC: We created this process to help investors globally manage their international exposure and therefore we wanted large liquid markets. Now, what quite a lot of people do is use this to help create an country allocation overlay using ETFs.
The universe of countries we chose was countries for which there were MSCI ETFs available when we launched – plus add the US into that because there’re are so many ways of playing that through derivative and other fund index level instruments. So, it is an index level approach.
JF: So which index have you used for the US?
PC: We look at the MSCI indices for every country with the exception of the US where effectively we look at a broad index, the S&P 500, and the tracking areas between the MSCI and these broad benchmarks like the S&P in the US is pretty low so we feel comfortable doing that. But for the others everything is based on the MSCI index returns, but generally for these large markets there aren't huge difference between the local benchmarks and the MSCI anyway.
JF: What are the single biggest inputs into your quant screens?
PC: The way we started was by looking and identifying about 25 useful factors … indicators which tend to be helpful when you are making a country selection decision. So when we look across a universe of countries, something like dividend yield helps. Over long periods of time, markets which have a high yield tend to outperform markets that have a lower yield. Other indicators that go in are things like the yield gap measure – often called the Fed model – the difference between the earnings yield and the bond yield. Generally speaking when it’s positive the stock market tends to outperform.
We look at other valuation metrics too, things like return on equity … profitability is a good thing and generally markets with higher profitability outperform ones with lower profitability. And the same with things like momentum, earnings revisions and other analyst indicators and things like changes in real effective exchange rates. Now, what makes this process rather unique is we don’t use a one size fits all model. Although a factor may be generally useful there can be reasons why it won't be applicable in the context of a specific country.
The example I often give is the dividend yield factor. The decision for a company to pay a dividend comes down to a number of things; obviously the ability to pay, the earnings environment and so on, but it’s also affected to some extent by the tax regime. If the tax regime in a country is a barrier which makes paying dividends inefficient, companies in that country will tend to pay a low yield even when that country is doing very well and that market is doing very well and it’s an attractive investment.
JF: And you can account for that in your study?
PC: In essence, yes. What we do is we measure how useful each factor is in the context of each country: we look at the observable skill of each factor. So we aren't simply saying that dividend yields are useful; we look to see if dividend yields are a good predictor in each specific country before we decide to use it in that countries model. So the process creates multi-factor models for each country and those multi factor models take the five most useful factors and weight those factors in the context of that country.
The other thing about this process is it doesn't create static models. Every month as markets evolve, as things change as look at the data, we recalculate all these relationships (what we call the country/factor IC's) and as the relationships change the factors that are most useful may also change. So if a country in 1995 was very heavily dependent on capital intensive industries, then clearly certain sorts of indicators would likely be useful. But as the dynamics of the market evolves, the sorts of metrics that are likely to be useful can also change – and we hopefully pick that up.
JF: So when are you next going to update this report?
PC: The model is run every month and it has been since early 2004. Australia has actually looked attractive for a while. It was a market we liked in 2006 and 2007, were neutral on for much of 2008 (it was sort of middle ranking) and since January this year we started to like it again and it has moved back up into our top quintile of favoured markets.
JF: Is Australia likely to become less attractive as companies reduce their dividends, or is this expectation already factored into the model?
PC: The answer to that is that it all depends on what happens: both to the other markets and to the other factors within Australia. So there’s actually two things to consider. Firstly, the model is designed to rank countries in a relative manner, and we do this by looking at how well each country ranks on those factors that are important in that market relative to how other markets look on that factor. So if Australian companies reduce their dividends more than companies in other markets then we would say that will likely be a negative for Australia's prospects. But if Australian companies reduce their dividends: but they reduce them less than companies in other markets do, that would be a relative positive.
Secondly, it depends on how other things that are important to Australia change – as the dividend factor receives less than 20 per cent weight in the overall Australia model. So it is quite possible that Australian dividends are cut more than elsewhere (and hence the relative yield deteriorates) but improvements elsewhere more than offset that – or vice versa. In terms of is it already factored in: explicitly no. However, one of the other key factors we find is important is the Earnings Revision Ratio which explicitly looks at the number of analysts that are up-grading their earnings numbers vs the number downgrading earnings. And hence the deteriorating broad earnings environment is already being factored in – and this currently is a negative for Australia.
JF: Commodity prices play a big role in the performance of particular companies in Australia and a large segment of the economy. To what extent do commodity price forecasts play a role in your research and have you taken into account some of the recent reported reductions in thermal coal prices?
PC: Again, we don’t take these into account explicitly, but the inputs to this model include things like forward earnings and changing analyst forecasts – hopefully the broader environment is considered by the analyst community! Also, we look at things like real effective exchange rates (which picks up terms of trade) and momentum; and again, price action should capture some of this.
JF: Just on momentum, I was under the impression that quants can be parochial and belong to particular camps like the PEG (price-to-earnings-growth) ratio camp or the momentum camp. Do you identify yourself as being part of any particular quant subgroup?
PC: No, I don’t and I would probably dispute that strongly. Quants and quant processes should be objective. A good quant process can draw on a huge amount of data and can consistently examine the facts and trends in that data. So while from a personal investment standpoint I probably have more of a value bias (partly because every time I look at the data I find value makes sense) in terms of our core processes we always incorporate both value and momentum metrics. I generally think anybody who wishes to come down firmly in one camp or another, and either ignores certain types of factor – or starts to rely too much on one group – is being a bit short-sighted. I try and ensure we are able to exploit information from all the major style indicators: value; growth; momentum; quality etc.
JF: Apart from the one I’ve just raised, what’s the biggest single misconception people in the finance industry have about quant models and quant analysis?
PC: Well, I think one of the things is that people sometimes think that we do stuff which is radically different from fundamental analysts – and I think there’s a little bit of a misconception because we don’t – we just do things in a rather different way. When we talk about a factor, and if you remember back in August 2007 there was talk of the perfect storm for quant investors, is work out the behaviour of the markets by looking at the observable attributes associated with stocks that outperform or underperformance – the same things really fundamental analysts follow too.
So if you look through the factors we incorporate into this country selection model, they are all the things that traditional market strategists would look at. They would generally look at interest rates and interest rate trends – and so do we. They look at charts to gauge momentum, while we look at price momentum more objectively, using more mathematical ways to look and compare momentum. So we just do things more objectively. So instead of looking at the earnings outlook in general terms we objectively look at forward PE ratios, and analyst earnings revision ratios etc.
Strategists look at the outlook one way: look at things like inflation rates and inflation rate differentials and so on while we look at the same thing through following changes in a real effective exchange rate measure, same but different!
JF: There’s a line of criticism floating around that quants are largely responsible for the mess the world finds ourselves in now …
PC: Well, that’s the August 2007 fallacy, I think.
JF: Given that feeling, do you think there will be a smaller or larger role for quants in the future?
PC: The pendulum swings both ways. Back in August 2007, a lot of quant funds had done phenomenally well and maybe people were getting a little overconfident and were relying too much on quant processes: not necessarily just processes to identify outperformance, but processes to control for risk: and there were definitely some people who then had used that to lever up too aggressively. There was maybe a perception that there was no risk in the markets and with volatility low, you could control for risk very well. And a lot of people use quant techniques to do that. So while I wouldn't say quant were to blame, they weren't blameless either. They were and are a part of the broader investment community at large and I think that basically at that point in time investors generally had started to become fearless. A lack of hubris was maybe to blame – and we are all paying the price!
Related Industry Sectors
Related People