Decoupling: DOW and PSE Index

One of the terms we used to hear very often in the months before the subprime meltdown was decoupling. It’s a fancy economese word to denote the separation or independence of markets from each other. Although these arguments have been around for some time, the idea became quite popular when the US economy began showing signs of slowing and recession–and the buzz word decoupling became a popular defense for other economies and stock markets, primarily the Asian markets (which includes the Philippine economy, and by association the Philippine stock market).

Conrad de Aenille, in his article Decoupling: Theory vs. Reality writes:

Decoupling holds that European and Asian economies, especially emerging ones, have broadened and deepened to the point that they no longer depend on the United States for growth, leaving them insulated from a severe slowdown there, even a fully fledged recession.

In Decoupling: The Biggest Market Myth Of All, Merryn Webb describes how growth experienced in areas outside of the US became evidence to support the theory:

There were hordes of statistics around to back this up. The IMF raised its global growth forecasts in the middle of last year on the back of strong growth in emerging markets. Meanwhile, double-digit retail sales growth in China, India and Russia added grist to the mill of the idea that domestically driven growth was a reality.

There was much talk from analysts at investment bank Goldman Sachs about how the Bric countries (Brazil, Russia, India, China) added more to global growth in the first half of last year than America. Global investors reacted to this flood of happy news by pouring money into the region, buying the good and the bad indiscriminately and pushing up stock prices across the board. The MSCI Asia Pacific ex-Japan index ended the year up over 35%.

But while the whole decoupling theory sounded plausible and provided the bulls with a reason to do what they wanted to do – keep money in the markets – I’m not sure it ever really added up. The Asian economies have clearly moved mountains since 1998 and it is absolutely true that domestically fuelled growth is coming through.

But that doesn’t mean that Asia is immune to US recession. It isn’t. It might look as though other Asian countries have started to export to China in big volumes, but their exports aren’t necessarily destined for the homes of the Chinese. Instead, they are finished in China and exported on to the West, and are therefore highly dependent on the American economy. And as Stephen Roach of investment bank Morgan Stanley points out, the leading Asian economies, while much bigger and more influential than a decade ago, are still pretty tiny in the great scheme of things.

Of course, economic arguments are one thing, but empirical evidence is another. We can test the decoupling argument by looking at two simple pieces of data: the Dow Jones Industrial Average and our own PSE Index, and see how movement in one index affects the other.

If you have been trading for some time you will have encountered the prevailing notion that the performance of the US markets, which occurs during evening Asian hours, will have an effect on the morning’s trade in Asia and the Philippines. Just how much this effect is can be easily quantified.

Looking as far back as 1997, covering the bull and bear markets since the Asian Crisis, I grouped the PSE trading days depending on the DOW’s performance the prior night. Of course not all days will align perfectly due to holidays between the US and Philippines, but most days will. The result is a table that describes the resulting movement of the PSE Index the next morning after a DOW close:

In the study, 605 days were not aligned. But for the trading days following a down close in the DOW the previous night, the cumulative return of the PSE Index was a loss of 7,958 points with a dismal 39% winning percentage! Contrast this with a cumulative return of 7,885 points and a 60% winning percentage for days following an up close in the DOW.

Clearly on magnitude and frequency, the DOW’s performance has a strong influence on the PSE trade the next day.

We can further examine this relationship through a chart plotting the percentage change in the DOW against the percentage change in the PSE the next day:

The upward sloping trendline is the best fit of all points, and the fact that it is sloping upwards shows the tendency of the PSE index to mimic the return of the DOW the previous night. The resulting correlation of the PSE and DOW indices is 0.309 positive.

Now for the decoupling argument. The data above is all trading days from 1997 to present. If the decoupling theory is correct, by running the same study again but this time only for more recent trading days, the relationships should deviate from the results above. To do this, we simply select trading days from 2007 to the present, which covers the critical events and crash of the subprime meltdown:

Our sample size has shrunk, but amazingly the relationships are still the same. On days following a down DOW, the PSE Index lost 3,035 points cumulatively with a 31% winning percentage. Likewise on days following an up DOW the PSE Index gained 2,955 points cumulatively with a 66% winning percentage.

Comparing these numbers with the original chart, it seems that the relationship of the DOW to the PSE Index has even strengthened. The expectancy figures are even more extreme showing a biased tendency for the PSE to mimic the DOW.

Running the correlation chart on the shortened data:

The trendline is still upwards sloping, and this time the points are even more aligned with the general trend, showing a remarkable consistency between the PSE result and the DOW. The correlation computed on this chart is 0.563 positive, even higher than the total sample.

Clearly the relationship between the DOW and PSE persists, and is in fact stronger in recent months. This evidence flies directly in the face of the decoupling argument.

Insight And Application

There are at least two very important insights for a trader coming from the data presented here:

Firstly, regardless of any compelling economic argument for decoupling, if the data does not support it or even shows an early indication of showing it, then there is really no reason to trade against the relationship. Even outside of the data, it makes more logical sense that with the interconnectedness of global economies through trade relationships, and with the ease of information brought about by the internet, all markets are even more connected than ever, and economic movements in one side of the planet will definitely affect all other areas. This trend can only get stronger.

Secondly, knowing the strength of the relationship of the DOW and PSE, a good part in studying and predicting the moves of the local market has to do with understanding the moves in the DOW. This is inevitable. Unfortunately, much economic and broker research still persists in treating the Philippine economy as a distinct entity. While unique factors do exist for the Philippines (and any other country for that matter)–this does not nullify the correlation of global markets to each other. So any analysis of relative value that is specific to one area, such as the Philippines, should not ignore these relationships, and should even discount it. The alternative is to fall into the trap which gave birth to compelling arguments such as decoupling that sound good, but have very little empirical basis.

In the context of the prevailing downtrends in equity markets, although arguments may inevitably arise of how the PSE will bottom out due to uniquely local factors–this will really be unlikely until the DOW bottoms out as well.