Dubai Stocks Make New Bear Market Low

Bad day for the Dubai market yesterday. The DFM General Index was down 3.83% for the day, finished down 8% for the month and made a new barket market low.

The Index closed at 1410.70.  That's the lowest level since the previous bear market low was made in Februrary 2009.  The first time the Index was trading at the current level was.........wait for it.........back in June 2004! 

The good news is that the DFM General Index can only go down another 1,410 points before it hits the zero level.  I'm fairly confident that it can't go any lower than that.

Enjoy.

DFM General Index: 9 Higher Closes in the Last 10 Trading Sessions

Monday's DFM General Index close was the ninth higher close out of the last ten trading sessions. The chart below shows the previous 138 occurrences of nine out of ten higher closes.

As you can see, most of the previous occurrences were in 2004 and 2005.  Recently, however, there have been fewer and fewer instances.

Historically, the DFM General Index has closed higher five trading sessions later 64% of the time following days when nine of the last ten trading sessions yielded nine higher closes.  The average gain was 0.70%.

When the market did closer higher five days later it do so by an average of 1.76%.  When is closed lower it lost an average of -1.22%.

In conclusion, when the DFM General Index has closed higher for nine of the last ten trading sessions this has been broadly positive for the market over the next five trading sessions.  It certainly hasn't foreshadowed any significant market sell-off in the week following such occurrences. 

Enjoy

Price Projections Using Pattern Matching

When analysing markets and trying to predict future price action there are certain techniques and studies that I pay a lot of attention to (here and here for example).  There are other studies, however, that I don't assign any great weight to but still look at regularly because they're, you know, fun and interesting.  The pattern matching and price projection study I'll describe in this post falls into this category.

Here's how it works.  Let's take the cumulative percentage return of the DFM General Index over the previous 25 trading days:

This 25 day price action can compared with all other historical 25 day periods for the DFM General Index.  In doing so we can identify which past 25 day price periods most closely match the current 25 day price action (note: the matching algorithm I used has been adapted from the gummy-stuff.org website.  It uses a Pearson correlation to determine the level of similarity between the current 25 day cumulative returns and the historical 25 day cumulative returns).

As an example, the chart below shows the current 25 day price return (dark blue line) versus the best historical 25 day match (light blue line):

As you can see, the cumulative 25 day return of the best historical match closely tracks the the current 25 day price returns.  The smaller chart shows at what point the best historical match occurred on the DFM General Index.  In this example the the 25 day period that most closely matched the current 25 day price movement was in September 2007.

Next, we can use the price action that followed the best historical match to make a price projection for the current market.  The chart below shows three such price projections:

 

Rather than rely on the price projection of the single best historical match I've also included the second and third best matches.  Each projection extends for 20 trading days into the future and is determined by the price action that followed the corresponding best match period.

For example, the dotted red line shows the 20 day projected price action, from the current market level, based on the price action that followed the best historical match back in October 2007.  The dotted green line is another 20 day projection but based on the second best historical match (Jan 2005) and the orange line is the projection based on the third best historical match (Aug 2004).

The red, green and orange markers on the smaller chart highlight the three best historical match periods on the DFM General Index.

In the example above we can see that the price action that followed both the best historical match (dotted red line) and second best historical match (dotted green line) was bullish, extremely so for the best match.  However, the price action that followed the third best match (dotted orange line) was bearish.

The rationale underlying this analysis is that periods of similar price action will also share similar future price action.  Of course, the future is unlikely to unfold exactly as it did in the past.  Nonetheless, it's informative to know how similar price movements to the one forming currently played out in the past.

In the next post I'll provide the price projections for other GCC indices based on this pattern matching analysis.

Let me know what you think.

Enjoy.

Dubai. The Week Ahead.

So, we have our post hoc explanation for last week's market rise in Dubai: Dubai World Debt Plan Wins Support. 

Is this good or bad news for the market?  I don't know and, to be honest, I'm not a big fan of trying to interpret how large groups of people will react to news stories such as these.

If the market zooms up next week we'll be told that the debt plan agreement has restored confidence in Dubai.  Should the market do very little next week this will be because the news of the debt plan agreement had already been "priced in." And if there is any sizable fall in the market that will be due to "profit taking."

By the end of next week the appropriate explanation will be obvious. Retrodicting is easy.  Prediction is a little harder.

So, whilst acknowledging the potential importance of the Dubai debt plan agreement I'm going to revert to my regular analysis approach:  determine the current state of the market and analyse what has happened in the past under similar conditions.

With this in mind, below are a couple of interesting things I noticed about the Dubai market last week and what they might mean for the upcoming trading week.

1.  Volume

On an absolute basis the volume increase that accompanied last week's market rise was nothing spectacular.  Below is a chart of the volume for the DFM General Index.  As you can see, the recent volume has so far not even matched the volume levels seen previously this year in June/July.

However, looking at absolute volume levels doesn't take into consideration the overall volume environment of the market. A daily volume increase of one million shares can be significant in a low volume environment but but less so in a high volume environment.  Normalising the volume data makes it easier to compare relative volume changes over time. Below is a chart showing the normalsed volume for the DFM General Index going back to the beginning of 2006.

There are many ways to normalize volume data but I've chosen a simple method that calculates how far current volume levels are from their 20-day average.  For example, on Monday last week the volume in Dubai was over two hundred million shares which was about 270% greater than the average over the previous 20 days.

As you can see, the level of normalised volume seen last week (highlighted on the chart above) is fairly rare and has only been exceeded twice since the beginning of 2006.

So, in absolute terms the current volume levels aren't very significant. However, when we take into account the current low volume environment of the Dubai market the actual volume increases seen last week were very unusual. The question is does this tell us anything useful regarding future price action.

Below is the performance of the DFM General Index during and after high normalised volume periods.  The red line is the performance of the DFM General Index for the five days following a daily volume level which is 150% greater than the average volume of the last 20-days.  The green line is the same but for volume levels that are 200% greater than the average volume of the last 20-days.

From the chart above, it looks like the five days following high normalised volume days have tended to be bullish for the Dubai market. When volume has been greater than 150% of the 20-day average the market has been higher five days later 75% of the time (21 out of 28 instances) for an average gain of 6.55%.  When volume has been greater than 200% of the 20-day average the market has been higher five days later 83% of the time (15 out of 18 instances) for an average gain of 5.83%.

Whilst the sample size isn't massive there is a definite tendency for a sizable market increase in the week following high normalised volume periods. 

2.  Price Activity

I noticed on Monday that the Dubai market had closed on it's high for the second day running. Exactly on its high, that is.  I looked to see how often this had happened in the past.  The answer?  Never.

Then on Tuesday the same thing. Another close on the high.  Three days is a row.

What to make of this?  An oddity to be ignored?  A manifestation of the strength of the market and an indication of further strength to come?  Or a sign of "irrational exuberance" and that the market has gone too far too quickly?

Because there are no prior instances of the market closing on it's high three days in a row (or even two days in a row) I decided to test what had happened when the market had closed very near its high.  The chart below shows the five day performance of the DFM General Index when it closed in the top 90% of it's daily trading range for two days in a row (blue line) and three days in a row (red line).

When the Index closed in the top 90% of it's daily trading range for two days in a row the following five day performance was positive 78% of the time (25 out of 32 instances) with an average gain of 4.90%.  In the 22% of instances where the Index didn't close higher five trading days later the average loss was -2.97%.

For three days in a row the following five day performance was positive 98% of the time (11 out of 12 instances) with an average gain of 3.34%.  The only instance where the Index didn't close higher five trading days later resulted in a loss of -0.66%.

So, the fact that the market has closed at its highs three days in a row has historically been a sign of strength and that higher levels are to come, for the next five trading days at least.

Conclusion

So, two studies pointing to higher levels for the Dubai market this week. Did I miss all the studies that point to a market crash this week? Probably.

Enjoy.