Friday, 16 October 2015

Currency and Index Correlation

Another piece of coursework. Quite long this time, 1700 words.
This a great subject, the correlation between stock markets and currencies, and indeed there are many correlations to be found. Obviously the useful correlations are only found when the currency lags the index or vice versa. A real-time correlation is of no particular value. There are other indicators which can be found by plotting the charts of world indices in USD, where pivots appear, but of course being single plots, do not offer lagging and leading signs.
The correlations work in a strange way. In a 'risk-off' environment (or bearish, or low volume, often seen in Summer), we often see a strong correlation between currency and index, which means that if the index were expressed in USD, price/action would be flatter. Regard the German DAX overlaid with the Euro. The pair reversed, so USD comes first, meaning that DAX rises are merely Euro falls.

You will see that during summer, leading up the August 24th crash, there was an extremely strong correlation, in other words the DAX if denominated in USD hardly moved at all.
After the crash, the correlation continued, although not as strongly until the morning of September 18th, at which point the correlation went precisely into reverse. What happened? On 16/17 September, the US FOMC[1] announced for the first time some forward direction on interest rates, namely that rates were likely to rise before the end of 2015". This is real news, and 'trumps' any correlation.
A rate rise is risk-on, viz. good for the dollar (obviously) but not good for the stock markets. The German DAX is well known as being the most closely linked to the US dollar, probably because it's component companies are all very big multinationals[2], (Daimler AG [Mercedes-Benz], Adidas, Bayer Pharmaceuticals..) who have global supply chains and customer channels, and although they may publish accounts in EUR, to all intents and purposes huge proportions of their supply costs and sales receipts are either in USD, or hedged to equivalent.
I should say at this point that large corporations who source and sell globally invariably hedge their currency risks by running forex dealing departments the size of a small bank. Consider sales in emerging markets, where currencies fluctuate wildly against USD. To meet periodic budget targets and avoid profit warnings, this activity is essential.
The point is that the tenuous link between currencies and indices is easily broken by firm news. The FOMC dollar 'euphoria' kept the indices in reverse correlation for the best part of a month. Of course the depressed indices then suggest to the market that there will be no rate rise, and we have seen in the last few days the correlation return.
Incidentally, the CME Group have a useful web page[3] which shows the market's sentiment on the likelihood of a rate rise. The current view is almost no chance in October, and only circa 30% chance of a rate rise in December or January.
To complete the picture on the Eurozone, let's look at the other major European indices, the French CAC40, the Spanish IBEX35 and the Italian MIB40

The pattern on these three is similar to the German DAX but not so pronounced. However, because the link is not so strong, there is sometimes a possibility of an index (or the currency) lagging, which presents a low-risk trading opportunity. Consider for example the July price/action on the CAC40, where the index clearly moved ahead of the EUR, whereas the DAX was more or less in lockstep.
This time lag is important. If there is great correlation but it is simultaneous, there is no particular advantage, other than the relatively difficult exercise in spotting pivot points (support/resistance) in the currency where none exists in the index, or vice versa.
The Swiss market is interesting. The index is a 'minnow', a small non-volatile index which is hardly traded at all outside Switzerland (i.e. by non-CHF investors). However, the currency is a major safe haven, as seen by the August 24 spike. The chart for this shows poor positive correlation, but does reflect the FOMC dollar 'euphoria' since the September announcement. My view is the same as most global traders, the index is irrelevant, and so therefore is correlation.

I will pass on the Scandinavians for now, mainly because it's a similar situation to Switzerland, the indices are only lightly traded and not very volatile, and unlike the Swissie, the Scandie currencies are not that volatile either. There is also a very strong link between NOK and the price of oil, which is outside the scope of this article. This leaves the final European denomination, GBP and the FTSE 100. Here's the chart.

Straight away you should notice some big differences to the EUR comparisons.
1. I have plotted the chart with USD as the quote currency, not the base. Correlation does exist this way, which is the reverse situation of the Euro. In other words the FTSE quoted in dollars would be even more volatile.
2. Much more importantly, there appears to be a lag between the price/action of the FTSE 100 index and cable. You see this very clearly in the weeks either side of the Aug 24 crash, and again since the beginning of October. This lag of approximately a week is of course a great trading indicator.
3. The correlation appears to have 'caught up' as I write on Oct 16, but if not, there is a little further upside to GBP. However, it is definitely worth watching FTSE price action for further clues.
Obviously GBP and EUR are very closely linked fundamentally, due to Britain being a member of the EU, and the huge volume of trade that results from that. I was going to discuss the relationship between the FTSE 100 and EURUSD, or perhaps the cross pair EURGBP. However, to reduce the UK vs Euro point to it's simplest, why not just correlate the FTSE and the DAX which I have done here.
You can see immediately that the correlation is much stronger than any of the currency correlations. So there is really no need to explore further.
Once we move outside Europe the only markets of significance are Japan and Australia. This is largely because only these countries have heavily traded and therefore volatile stock markets, and highly traded currencies. India has the former but not the latter, New Zealand has the latter but not the former.
Let's look at Japan first.

The correlation with USDJPY is very strong. It has been ever since the 2008 crash, when it was not correlated before, as shown on the 20 year chart below.

Japan of course is well known for it's closed economy, and, say some economists, the stagnation that results from that. You could also add that the country 'missed the boat' on the internet revolution. What this means is that Japanese stocks have hardly risen in price in USD terms since the crash of 2008. Contrast this with the 150% rise in the S&P 500 since that date.
Australia is not an important stock market, and is dominated by miners, who have suffered from the problems with the collapse in metals prices, led by gold in 2012/13. Nevertheless the correlation with AUD exists and like the FTSE has some lagging points, although not as clear, and in some cases the currency leads. The Aug 24 ASX200 crash took over a week to be replicated in the AUD currency. The chart shows a small opportunity to short the ASX over the next few days.

Finally we turn to the US itself. Of course we can compare the S&P 500 (SPX) with any pair. Some strange correlations occur. For example look at the SPX compared to CADUSD, again the USD as quote currency. The correlation is strong. In September the pair led the index (giving trading opportunities on the latter). In October, the index lead for a couple of weeks, but the last couple of days seems to have switched back. This makes spotting any opportunities very difficult. Notice by the way how nothing much happened to CAD on August 24.

Comparing JPY is even stranger.  This time we take the conventional position of USD as base, and JPY as quote. The correlation until the first week of October was very good, reflecting the classic 'flight to JPY' because dollar bearish sentiment ceteris paribus is matched with equity sell-offs.

However, since the first week of October, things have gone the other way, almost certainly following the very poor non-farm payroll figures for October. The market has decided that the NFP figure will postpone a rate rise[3 ibid] and consequently the market has risen, whilst the dollar falls.
This constant switch from 'good news=good news' to 'bad news=good news' has been a feature of the equity markets since the tapering of QE. Each bad economic data item has caused the market to believe that normal priced money is further and further away. Concomitant to this issue is decay, as the bad economic event recedes, the currencies and markets fall back into lockstep. There seems to be little effect of 'good news=bad news'. In other words, whereas Fed pronouncements and other economic news affect USD in exactly the way you would expect, it seems that almost all economic announcements push the equity market up.
This is my view is the reason for the sudden and severe crashes in October 14 and August 15. Only once a rate rise track fully normalises the economy will we see some normality in currency and equity correlation in the markets. My view on the commodity collapse is that prices will recover much more slowly, maybe over 3-4 years, and thus should not affect either the CAD or AUD significantly, nor the relevant indices.
My overall conclusion is that currency/index correlations do exist but there are no constant rules, and the correlations only exist in the absence of real 'risk-on/off' significant economic events.

David Atherton
16 October 2014


No comments:

Post a Comment

Please leave a comment