It's time to go bottom-fishing

Sometimes apparently idle arguments can stimulate interesting lines of thought. The mood at the moment is obviously gloomy even for diehard contrarians. Several of us were debating the levels to which the market could drop. Amidst much learned talk of possible retracement levels and rising interest rates, a couple of really frightening arguments popped up.

One of these was based on price-earnings ratios. In November 1998, when the last bear market hit what turned out to be bottom at the Sensex levels of 2742, the average Sensex P/E was a measly 9.8. Right now, it's close to 20. Does this mean that the market could halve from these levels and head for 2000 again?

This attempt to judge a technical retracement level by means of the P/E ratio may seem odd but it can't be dismissed out of hand. There is nothing illogical about it. If you can use the P/E to judge if a market is cheap or expensive, then you can use it as a trading tool as well. It may not be efficient but that's a different matter!

Let us assume you make an immediate and reasonably optimistic adjustment for a rise in Sensex EPS of around 20 per cent for fiscal 2000-01. That still implies the market is trading at about current P/E of 16. If you buy the bottoming at similar P/E discount levels argument, you must assume roughly a downside to 2250! Which is where the horror story starts.

The other major arguments presented may as well be described.

Retraction: At the April 1993 bottom of 1980 points, the market saw a 71 per cent retraction of the previous bull-run from 947 points (January 1991) to 4547 (April 1992). The November-December double bottom of 2742, was also around 71 per cent retraction of the previous bull-run from 1980 points (April 1993) to 4643 points (September 1994).

If that relationship holds, this bottom should come at about 71 per cent retraction of the rise from 2742 points in November 1998 to the top at 6150 in February 2000, and that would mean a bottom about the 3730 level. We have hit that zone almost exactly so if you trust this bit of number-crunching, it's time to go bottom-fishing.

Rolling settlement: Another argument was based on previous market behaviour when rolling settlement was introduced. Volumes in rolling stocks shrank drastically and prices dropped 50-65 per cent. So going by this argument, whenever rolling comes in, the market could have a 50 per cent downside. I'm not even going to try and calculate this level! Let's just say that Dalal Street and Nariman Point would be littered with dead bulls if this behavioural argument proves correct. Prices would be back to April 1993 levels or even lower.

The price-book value or q argument: We could use Tobin's ratio, which was recently employed as a possible top and bottom-fishing tool by two English researchers Smithers and Wright in "Valuing Wall Street". Tobin's ratio is equivalent to the Price-Book value per share, only it's taken across the entire market or as broad a segment as practically possible.

It doesn't tell us much in this instance. At previous market bottoms, q (as this ratio is known) has varied between levels as diverse as 2.4 (January 1991), 4.5 (April 1993) and 1.7 in 1998. At previous tops, it's varied as much as 10.25 (1992), 6.8 (1994), 4.0 (2000). Going by historical shifts in q, we aren't particularly over or under-valued at the moment with a P/BV of around 3.

Broadly, the technicians seem to be optimistic compared to either the P/E or the rolling-behaviour school. The technicians are more concerned about the possible duration of the bear market rather than the further downside. Previous bear markets have lasted a minimum of one year and often a lot longer.

Even the pessimistic technicians are talking about a bottom around 3350. This is only about 12 per cent from here. That calculation incidentally is based on the recent chart formations of the Sensex. The market has just broken down out of a triangle and the downside target could be around 3350-3380 depending on exactly where the triangle baseline is drawn by different analysts.

The rolling behavioural-school is undoubtedly right in that volumes will shrink. But if rolling comes with badla thrown in, a removal of circuit filters, and a possible expansion of the derivatives market, stocks will not get quite so badly hit. Remember the banning of badla in 1993-94? It didn't stop a bull market though it slowed it down. The effect of rolling was undoubtedly worse on smaller caps where liquidity has always been a problem. This time it would affect A-group stocks where big-ticket trades would still be possible.

Operators are getting plenty of time to reconsider their strategies under the new conditions. The impact of rolling settlement will probably be discounted by adverse price moves before it is actually brought in. At least some of the current hammering is dictated by fears of rolling. At any rate, I don't think we have enough history to make a serious statistical call on this.

The P/E argument and P/BV arguments certainly have a lot of history backing them. Prior to 1991, according to Dr L C Gupta's studies, the market had rarely traded above double-digit P/Es. Between 1991-2000, the market has traded below double-digits only during 1998, at the end of a four-year bear market.

Could we be moving back to a pre-1991 scenario of minimal discount for stocks? Doubtful - corporate growth has improved after 1991 and there seems to be simply too much money chasing stocks to revert to a pre-1991 situation. With the primary market depressed, the property market depressed, the debt market illiquid and GDP projections scaled back, will that money disappear back into fixed deposits again?

Which comes back to the argument about using the P/E ratio as a tool for judging tops and bottoms? Looking at the Sensex P/E since January 1991, that seems just as tough as using P/BV. The market bottom in January 1991 came at P/E of 17, while April 1993 came at 27 and 1998 at 9.8. Tops were at P/E of 57 in April 1992, 45 in September 1994 and 31 in 2000. This is according to average daily values released by the BSE. These P/Es are updated on a half-yearly basis and annualised.

The mathematical problem with both P/E and P/BV is that the denominators of these ratios are quite volatile. Book value is less volatile but it does change significantly while earnings per share EPS shifts by huge amounts. In a bull market, a bet on big-ticket EPS growth is normal. Further complications arise from absurd valuations during the Scam when prices surged while EPS stagnated during 1992-93. The aftershock of those statistical aberrations is still being felt.

Nobody has a clue quite where exactly bottom-fishing should start. As a contrarian, I find it encouraging that apparently sensible investors can talk in terms of a possible 50 per cent downside from here. It makes me suspect that the bottom isn't too far off. A little more panic selling should remove all supply.

If we assume that the downside could be a 3350 bottom and the upside could terminate at around 6000 in a two-year timeframe, the risk-reward equation is very much in favour of going long. Even if one assumes a 2250 bottom as per the P/E argument, versus a 6000 upside, the risk-reward equation is still in favour of going long.

Devangshu Datta

Money

Tell us what you think of this column