Capitalmind
Capitalmind
Actionable insights on equities, fixed-income, macros and personal finance Start 14-Days Free Trial
Actionable investing insights Get Free Trial
Concepts & Tutorials

Of Underreported Sensex P/E Ratios

I write at Yahoo, Of Underreported Sensex P/E Ratios:

When it comes to stock indexes, the fact that they are based on a collection of stocks makes it difficult for us to understand their characteristics. I had explained how the Sensex and Nifty EPS (Earnings per share) were calculated, and how we weren’t really seeing them grow quite as much.

It turns out there is a problem with the way the index P/E (Price to Earnings Ratio) is calculated. What the index creators do, for both the BSE and the NSE, is to add up earnings of all the companies in the indexes to get the total earnings. They divide the total market capitalization — each company’s total shares multiplied by its share price — by the total earnings to get the P/E.

The figure that comes along tells us how richly or poorly the index is valued. For instance, even after a reasonably good earnings season and a drop in the indexes by about 10% from the January peak, the current P/E of the Sensex is reported to be 19.32. This is higher than the historical average of 17 to 18, and thus might lead us to believe we are overvalued.

But a recent bit of research has led me to the conclusion that all is not right on the mathematical front. The problem is in the kind of earnings that the index owners use to calculate the P/E.

Companies report different kinds of earnings. At one level they report the results of the company itself, called the standalone results. And then, they announce consolidated results that includes earnings of subsidiaries, appropriately adjusted if the company owns less than 100% stake.

So when Tata Motors owns Jaguar and Land Rover, the earnings it reveals at the standalone level is only valid for Tata Motors, the Indian company. The JLR earnings are only revealed in the “consolidated” results. You might think the difference is small, but see this:

Tata Motors Net Profit (Standalone): Rs. 1,812 cr.

Tata Motors Net Profit (Consolidated): Rs. 9,274 cr.

The difference can be staggering. I did a quick calculation on the entire Sensex, and:

Total Net Profit (Standalone): Rs. 142,914 cr.

Total Net Profit (Consolidated): Rs. 168,944 cr.

Note: As of May 30, 2011.  JP Associates and Cipla have not announced consolidated results yet. RCOM has announced consolidated but no standalone results. ONGC has no Q4 or consolidated results available right now.

The consolidated results are 18% higher. And this is a relatively recent phenomenon, where companies are increasingly diversifying through subsidiaries, and the growth sits in their consolidated results.

Now, how does this impact the Nifty or Sensex P/E? I called both the IISL (which maintains the Nifty) and the BSE Index division (which maintains the Sensex). They both confirmed that the index P/E calculations only use standalone earnings.

So what we’re being told as the Nifty or Sensex P/E is off by a considerable margin. Most infrastructure and power players operate through subsidiaries, especially where state or central assistance is used, in a Public-Private-Partnership for example. (Reliance Infrastructure reported 50% higher profits in the consolidated results than standalone) Banks now don’t just do banking; they run insurance, mutual fund and foreign banking subsidiaries (ICICI Bank had a 6,093 cr. profit in consolidated results compared to 5,151 cr. standalone) IT Companies have subsidiaries abroad, Metals and Mining players have been buying companies all over the globe, and telecom companies have a global footprint.

The Sensex P/E revealed is simply the full market capitalization divided by the full earnings. By that yardstick:

Sensex P/E, Standalone: 19.3

Sensex P/E, Consolidated: 16.3

Figures as of May 30, 2011

The 16.3 figure is indeed more acceptable and the Sensex isn’t quite that overvalued.

Note: I would again beg to differ on this figure. When the Sensex is calculated it is actually weighted differently by company — so a 1% move (Weight: 11.6%) in Reliance affects the Sensex more than a 1% move in Hindalco (Weight: 1.77%). To account for that we must apply the weights to the earnings and market cap as well, giving us a consolidated P/E of 17.54.

Why, then do the indexes not use consolidated results? Their answer is that they don’t usually have all the consolidated results. Companies, by SEBI rules, aren’t required to report consolidated results every quarter; only standalone earnings are required. Many companies, like Infosys and Tata Motors, do so anyway because investors need the real picture; but others like ONGC choose not to reveal consolidated results until the end of the financial year. Since we are now getting the results of March 2011, which is the end of the Financial Year 2010-11, we are able to collate consolidated results properly. But that isn’t available each quarter, so the index creators don’t want to use them.

But then, why not use the consolidated results of at least those companies that regularly release them? Again, the index creators argue they need to be consistent — they cannot take standalone results of one company and consolidated results of another. I’m not sure I buy that theory; the only correct result is the consolidated picture, and if you are using just standalone data you are horribly wrong. If there are consolidated figures available and you use them, you become less horribly wrong. If one has to be wrong, less wrong is preferable.

The real solution, according to me, is to force companies to reveal consolidated results each quarter. Companies may not like it — no one likes more work — but there’s no point in inaccurate or incomplete result declarations. Additionally it’s easy to hide losses or costs in subsidiaries if you have to only reveal the true picture once a year. Currently companies get a month from the quarter end to reveal results — that time frame could be extended to two months for the consolidation. Eventually, companies must be required to reveal full balance sheets and cash flow statements as well, every quarter. (You can hide a lot of losses in balance sheets nowadays — an article for another day)

If our Sensex P/E is indeed in the 16s and 17s and not the 19s as we have been told, how does it change the view? At this point, with rising interest rates and super-hot inflation, I believe even our 16s and 17s are too high for a slowing economy. Inflation pushes up costs as it spreads into wages and secondary products, and rising interest rates are a measure of controlling inflation through slower growth.

Yet, it makes you think — if the data at the index P/E level requires rework, then our assumptions of the Sensex Earnings Per Share (which is the Sensex value divided by the P/E) can go completely wrong. While I bemoan the inability of the Sensex EPS to reach the 1,000 mark — that people predicted would happen three years back — at the consolidated level, the EPS has just crossed 1,000 even according to my revised calculations.

At another, and perhaps more worrisome level, if such data, the data of the WPI (Wholesale Price Index)  has glaring errors, is it right to use the data to make business or policy decisions? Shouldn’t we, as a nation, focus on more comprehensive, transparent and trustable data first? Like in the fight against corruption, we are more interested in the rhetoric than in finer details; but it’s obvious where the devil resides.

More Yahoo Columns:

  • Ganesh says:

    Interesting post. Had never occurred to me that the Index P/E reported on the website could be fraught with so many errors!
    The index P/E is an important piece which determines my asset allocation – However I use a percentile of the P/E within a 5-year population as against judging the market by an absolute value (e.g. overvalued >16). So, I might be okay as long the index P/E is consistent, even if is erroneous.

  • Bhupesh says:

    But then how Index EPS are calculated in other countries/market?

  • Amit says:

    >> the current P/E of the Sensex is reported to be 19.32.
    >> This is higher than the historical average of 17 to 18,
    >> and thus might lead us to believe we are overvalued.
    But historical p/e was also on standalone basis.
    If standalone earnings as a percentage of consolidated has not changed over past decade – we need to compare apples to apples
    So today’s 19 standalone p/e vs historical 17 standalone p/e – and we get an overvalued market.
    Amit

    • True. But I had done some similar calcs in approximately 2007 or so, and had found that the results (cons vs. standalone) didn’t vary that much. But you’re right, we need to get the consolidated P/E data and then work it out.

  • Nash K says:

    I believe even our 16s and 17s are too high for a slowing economy. Inflation pushes up costs as it spreads into wages and secondary products, and rising interest rates are a measure of controlling inflation through slower growth.
    Wouldn’t it be more appropriate to look at PEG (P/E to Growth, with the growth being that in underlying earnings) instead of PE to take a call on whether the SENSEX is over/under-valued? Of course, all the limitations you mentioned, would also apply to the calculation of growth-rates…

  • kora says:

    it is advisable to judge valuation on standalone results that are without any support of promoters, parent or sister concerns … a standard practise across the research analyst’s community and the quantitative investing community as well …
    .
    also the consolidate results can be rigged with the creative accounting practices …

  • Gyan Roy says:

    Deepak,
    Thanks for this post. Lately I had been noticing a lot of discrepency in EPS numbers quoted by analysts and the numbers on the BSE website. This was not the case till 2008 at least. I was wondering why there is so much discripency and went on to do the excercise myself. I found that current pe(consolidated) of sensex is around 14.8 and the EPS comes to around 1093. Am I right in my calculation.