Monday, November 2, 2009


History / Backtesting of Mean Reversion of NIFTY NSE INDIA to the 200 DMA after hitting peak spread of ~ +40% above 200 DMA.
This post is related to my original post ( 3rd Oct 2009) - KILL BULL Vol 1 NIFTY to correct 15% and my follow up post - More evidence for 15% fall posted on 25th Oct 2009.
Backtesting Period for study - from 1993 to till date.   This period is ~ 18 trading years  or ~ 4600 trading days. NIFTY was launched in 1989. The post 1991 - 1994 period can be taken as an `outlier period` due to 2 reasons

1.       1992 - Big bull Harshad Mehta scam – period where NIFTY rose to +100% above  its 200 DMA
2.       Post 1991 the forex crisis in India and the resulting IMF bailout that made India allow Foreign capital in many sectors. This lead to huge inflow of foreign capital resulting in NIFTY reaching upto 60% above its 200 DMA.

But for the study I have included a period in 1993 so  as to capture at least one outlier (due to sudden rush of FII)

The Mean NIFTY 200 DMA spread during bull markets is 13%. (The 13% mean NIFTY 200 DMA spread of taken from 2003 to 2007 when Sensex went up from 3000 to 21000). So when the  NIFTY 200 DMA spread touches peak of 40% , it means that the spread is 3x times the mean!! Now you just have to track the 200 DMA Nifty spread to try to estimate when the mean reversion will take place. It is explained as below

Some remarkable points to note from the results ( Nifty vs. 200 DMA peak spread and mean reversion)

1.      What is remarkable in the table is that in the past 18 years , there are just 12 such instances NIFTY where NIFTY reached the highest spread of around +40% above its DMA. This means that this pattern on an average has emerged just once in of 1.5 years!!

2.      From what I have observed ( but not mentioned in the table) is that the NIFTY can be said to be in normal territory when in spread of +20% above 200 DMA or - 20% below DMA ( normal bull and normal bear market respectively). Can also be a loose proxy for LEVELS of Greed and fear. In this range difficult to predict a sudden trend reversal.

3.      Extremes of spreads of +40% above 200 DMA or -40% below DMA are great trend reversal signals. But for exact timing further filters need to be applied. I have used just a simple indicator that is the measure of consecutive trading days when the NIFTY 200 DMA spread is above 30%. This adds another dimension of INTENSITY of Greed and Fear.

4.      From the table below, we can see that the longer the period when NIFTY is trading for consecutive trading days +30% above its DMA, the following 1 , 2 and 3 months we see the sharper falls.

5.      The current period where NIFTY was trading at +30% 200 DMA spread consecutively ended on 21-Oct-2009 and was the 3rd highest ever ( the other two being in 1993 & 1994). Since 2004 it has been the longest ever. This period saw NIFTY made a high of 5192 with peak spread of +37%.  This date (post 21-Oct-2009) was a great signal for shorting the market as historically it is seen that after such a sustained period of bullishness , the 200 DMA NIFTY Spread shows great mean reversion tendency over the next 1 or 2 months. In fact after 3 months , it can be seen that in many cases the 200 DMA Nifty spread  even goes negative . ( that means NIFTY closing below 200 DMA). Below is the NIFTY 200 DMA spread back testing table added with details of in how many such cases the spread turned negative.

6.      Now lets try to describe what happens when say NIFTY trades , say for a consecutive of 40 trading days +30% above its 200 DMA ?
a.      Complacency sets in as evident by VIX trending lower.
b.     Lots of bullish action can be seen across sectors , large or mid or micro caps.
c.      Games like Bulls Eye on CNBC TV 18  – where traders give picks for day trading can be seen to have active followers. Stocks go up 20-40% in few days for no fundamental reason. Rampant `tips`  for the next multibagger, `assured returns` spam your mobile.
d.     Even companies participate to prop up their lofty share prices by declaring bonus  ( Reliance , J P Associates , Adani , last year biggest IPO Reliance Power). At the same time these firms are taking advantage of the prices by selling treasury stock.
e.     Markets reward a stock that announces equity dilution by bidding prices higher.
f.       All in all it seems very easy making money being a bull.  Like a moth to a flame
g.      This is the exact environment that indicates a trend reversal is very near.

7.      Based on the past results we can see ( in table below) that after 1 month of the end of the period where the NIFTY consecutively closed +30% above its 200 DMA ,the spread (on average) reverts from peak of ~ +40% to +13%. , 6% in the 2nd month and further to -0% in the 3rd month.


Deepak Kumar Prasad said...

You said 14500 on sensex .. but it looks that will never be instead it likely for 18000

The Fixer said...

Hi Deepak,
I still hold target of 14,500 based on the past corrections ( as given in the table). sell signal was on 21-Oct-09 and 90 trading days after will be ~ 28-Feb-10, and in the past, after this period we have seen median fall of 24%.

but yes you are right current trend is up. Markets can touch 18k but it would not be sustainable.Also would like to add that the basis of 14500 S target was a mean reversion strategy that basically assumes calling the top and bottom ( which is not perfectly possible). This strategy is used to identify tops and attractive entry points mainly for the VALUE Investor.
But yes if one follows momentum investing, the trend is still up and can go higher.
You can see lots of evidence over valuation in large caps and lots irrationality. But as they say "markets can remain irrational for longer than one can remain solvent".