Buying High P/E Stocks: Was Benjamin Graham Wrong?

I recently made a blog post on the concept of ‘Foresight Valuation’ in order to identify whether or not the P/E Ratio (Or the relates Price Ratios) mean anything and why:

Do share your comments and criticisms. If you regularly use the P/E Ratio to Value stocks, I am most interested in starting a conversation with you. I may change your mind, but perhaps, you may change mine too. In any case, a good discussion is always worth it.

1 Like

I can provide my input on the basis I was someone who swore by PE alone for a decade. But my investment amount was not meaningful in this time to have any impact in my life and also returns were sub par just above the bank interest rates.

I have to say that I used PE as a single important valuation metric as Buffett was my icon and also I wasn’t fully evolved as a sophisticated investor(still not). For some reason, the only book I learnt on investing which was about Buffett’s method was heavy on his early years and had almost nothing on later years where he followed a combination of growth and value investing.
In my humble opinion value investing by itself is all about protecting the downside than about multiplying the returns. So it suits a rich investor than someone just beginning to build.

Thanks to valuepickr my methods have evolved although still untested. While I never went beyond 15 PE (my biggest hits were India bulls housing and Dewan housing bought below 15 PE), the highest PE stock I am holding now is 55 with an average PE of 28 across my portfolio.

So my current opinion about PE is it is only a screen. A part of 20 or more point checklist you may have and nothing more. It does not say anything about cyclicals, turnaround or growth stories which probably makes the quickest wealth.

In my short investing journey a major criterion for stock selection was the earnings multiple. And, it has held me in good stead.
One of my first investments was in the company Gujarat Ambuja Exports Ltd at an earnings multiple of 7.
GAEL, in a relatively short period gave impressive returns.
Nandan Denim was another low earnings multiple stock that performed well.
But, I’ve also had some painful outcomes in low PE multiple cases.
I had a gut wrenching experience in a company called Ahmednagar Forging. Fortunately, I didn’t average it on the downside.
Like all other strategies no approach works always.
What’s of supreme importance is relatively consistent performance.
And, staying invested in low multiple stocks has augured well for me.

Also, I’m a big believer of the role of antecedent as well as consequent luck in our decisions.
We may perform calculations being conservative, careful but there are factors just beyond our control. And, it’s crucial to be aware of it.


I have no qualms with people using the P/E Ratio as a filter. That’s exactly what Ben Graham intended for the P/E when he introduced it – as a thumb rule.

But I don’t understand the idea of Valuation using the P/E Ratio. Phrases like ‘P/E Expansion’ or ‘Costly P/E’ are like Greek to me. But almost every TV show on Stocks or every forum related to Stocks has a lot of P/E conversations floating around.

How it is remotely possible that the P/E, intended to be an upper limit, could be used to Value a stock? How could a Stock’s Value be correct if the Valuation did not consider the PV of Future Cash Flows, since that’s exactly what buying a stock represents?

In fact, if anyone uses the P/E to Value a stock, they’re starting with a flawed assumption that the market is correct (i.e. The numerator is correct) – that the market has accounted for all the possible future cash flows of the company. This is what I have disproved in the blog and actively dispute against.

In my understanding the PE multiple is an indicator of the time duration required to recoup one’s investment. If a company is able to grow faster than its PE multiple, investment will be recovered faster.

A company is available at an earnings multiple of 10 it’ll take 10 years to get your investment back without any growth.

If the company earnings grow at 25 % CAGR
investment will be recouped in less than 6 years.

A PE will be costly if the company grows slowly and one takes longer to recover the investment.

So, let’s say a heavily indebted company growing at 5% CAGR is available at 30 PE that is, in my opinion, expensive if there’s no asset play in consideration.


Agreed. But once again, this assumes that the Price is correct and consequently, the P/E is even roughly correct. How do you verify that the Price is correct unless you consider the PV of future cash flows (A DCF) or at the least, the PV of future dividends (A DDM)?

I show in the blog post that the biggest wealth creators in the last decade have traded at P/Es almost doubly lower than their actually justified P/E.

But, even when we perform DCF we’re making a lot of assumptions. Agreed, we can be conservative in our calculations. But, how do we even know that the company we’re performing calculations for will exist in the future.
Will the economy do well?
Will there be a demand for the products?
We just don’t know and will never know.
So, whenever we’re trying to predict the future there are bound to be errors. Why?
Because we, as human beings, as extraordinary as we are, are good at linear thinking but strictly ordinary at configural thinking.
There are too many variables in a company’s and an economy’s functioning that no one tool is even remotely capable of assessing a fair value- Nothing whatsoever.
It’s incredibly difficult to understand the pulse of economies as diverse as ours.
And, at the end of the day, a fair value is what the buyer is willing to pay.

1 Like


DCF and DDM at least try to predict the future (Like you say, even conservatively). The P/E does not. The P/E is just a massive opinion poll (“what the buyer is willing to pay”). As Warren Buffet liked to put it:

“A public opinion poll is no substitute for thought.”

And just because we cannot consider all the moving parts that goes into the Value of a company, is no excuse to throw one’s hands up in the air and give up, succumbing to cheap fixes like the P/E. A DCF or a DDM is as close as a human can possibly get to finding out the true Value of a company. Of course, someone like Warren Buffet might to a better DCF than most of us, but the argument still stands.

The question of a company not existing in the future or having bad financials going forward is a question of research and knowledge about an industry. If an investor doesn’t have these, there’s no point in even beginning to Value a company. So that point is moot. I understand that anything can happen in the future. But a good knowledge about the industry, a broad lattice framework and a minimal understanding of Accounting should increase the probabilities of predicting that future. This is why the concept of ‘Circle of Competence’ becomes so important.

Besides, if an investor truly believes that there is no way to understand the real Value of a company, then he should not be investing at all. If current Prices truly reflects the Value of a company, there’s no profit to the made. This is the line of argument in the Efficient Market Hypothesis, which has been debunked by good investors (Even mediocre investors) time and time again.

1 Like

Before I express my thoughts I must reiterate my stand- The future is simply unpredictable.
Now, we may adopt various tools to assess the value of a company. Most analysts employ strategies that you’ve outlined in your answer to assess the value of a company.
So, let’s find out how accurate analysts have been at predicting future earnings.
A study was performed by a financial magazine in 1995 to examine the accuracy of analysts’ earnings estimates between 1974 and 1996.
Estimates for the quarter were made in the prior three months and analysts had the liberty to revise it upto 2 weeks before the quarter end. Almost 500000 analysts’ estimated were included. And, this was at a time when abundant information was available.
The average error was mind boggling- 44% was the error of analysts.
Let’s give analysts in the 1970s and 1980s the benefit of the doubt that they didn’t have access to enough information.
But, in the 1990s when Internet was easily accessible in the US the error was more often than not above 40%.
I don’t understand DCF as well as you do but I’m sure that if the error in earnings estimates is 40 % the consequences will be disastrous.
Now, it’s possible that high variation in some companies’ results skewed the outcome.
Even after adopting an ultra conservative approach the error was an astounding 23%.
Just 58% of estimates were in the +/- 15% band.
We often feel that more information will improve our accuracy.
But, more information improves the coherence of the story our brain fabricates, not necessarily the accuracy. We often are victims of overconfidence and the illusion of control.
There’s nothing wrong in trying to predict. But, let’s be aware that we’re painfully weak at making predictions.
Our estimates are dependent on inflation, interest rates, unemployment, production, home government policies, foreign government policies. Can anyone predict so much? An unequivocal no. These are dynamic, rapidly changing.
Even economists fail to predict these developments.
Now, it’s possible that some may feel that analysts can predict better in sectors with ‘‘visibility’’. Results, yet again were devastating. Tremendous divergence was reported.
There is seldom any industry where predictions can be made with certainty.
And, these were analysts who were dedicated, hardworking, diligent. They’re not to blame for such an unremarkable performance. It’s just the nature of the world we live in.
We’ve glorified leading investors and they’re worshipped by us.
But, have we ever wondered that the success they achieved could have been an outcome of being incredibly luck?
We like to believe that we’re in control of our actions. But, there’s a tad too much out of our control.
We are remarkable beings but we do have our limitations. Let’s appreciate the randomness of the world we live in.


Do you have the link to the study? I would be interested in reading it.

Besides, a 23% error rate is acceptable, given how unpredictable life is, like you say. This means that if you diversify enough (Say, 20 stocks) and you make a 20% CAGR in general on all your investments, that means you may make 20% on 15-16 stocks and perhaps 10% or below on 4-5 of them.

If that’s not good investment, I don’t know what is. I’m not saying we can perfectly arrive at a Value for any Asset, but it would be folly to not try. And as they say, half-knowledge is dangerous. P/E is just that, half-knowledge.

I came across that study in a book I read.
Let me search on the Internet the study.
If I’m able to locate it I’ll share here. Hope that works.

A 23% error in estimating earnings is significant. Eg-The expected earnings were Rs.100 but the results are around Rs. 80. Wouldn’t that change a lot when we’re trying to estimate a company’s value.
Sure, there’s nothing wrong in trying to assess a company’s value.
But, to be certain that the calculations performed are correct and outcome will be in sync with our expectations is delusional.
Confidence, at the end of the day is the story that’s woven by our brain which makes us feel better.
And, no strategy is perfect. It has its weaknesses. PE strategy also has its weaknesses. So does DCF.
No strategy is effective in isolation. When used together they work best.
Both the right and left hemispheres of our brains are impressive. But, if they function without associating with the other part the outcome will be unpleasant.

1 Like

I don’t understand how the right and left hemisphere of the brain comes into question.

The DCF includes both stories and numbers. At the end of the day, that’s all there is to a stock: stories and numbers that keep the stories grounded in reality. This is an insightful comment on the same by Prof. Aswath Damodaran

To answer your question of how to account for randomness in a DCF, please read my earlier blog post (Scroll right to the end if you are in a hurry):

So, really, in a DCF, we do the following:

  1. Tell a story about the company based on our understanding of the industry and the business
  2. Come up with conservative numbers to correlate those stories
  3. Account for randomness in estimates using simulation
  4. Use Margin of Safety to account for extraordinary change in any of the above steps

The P/E or the other Price Ratios offer very little in any of these steps and if they do, it’s to a very minimal effect. The tools available to us are limited: Price Ratios, DCF or DDM. I’m not sure if another study was conducted on Price Ratios, but I am very sure that the error here will be way off.

If you have only three options to logically understand a problem, which one will you choose? The one with the highest hit rate or the one which is the easiest to do?

1 Like

By right and left hemispheres I was referring to the importance of association.
Results are often better when association, collaboration take place instead of working in isolation.
DCF is a great tool. But, it’d be imprudent to pass disparaging remarks on the efficacy of PE multiple investing strategy.
The apparently simple PE investing strategies have performed very well in the past. Often outperforming other approaches.Even the fancy, complicated strategies.
Enough studies have been performed to attest to this effect.
Just because something is easy doesn’t necessarily mean it’s wrong, inaccurate.
PE investing may not be in vogue because of the emergence of other fancier valuation assessment strategies.
But, I’m fairly certain that most fund managers adopt methods other than PE investing. They sure use tougher strategies.
Then, why do they more often than not fail to outperform indices?
Leading investors have rubbished the efficient market hypothesis.
But, how do we know leading investors know the truth?
They could be mistaken.
All I’m saying is I don’t know.
It’s always good to be accepting of new ideas.
It’s possible that our previously held beliefs were totally wrong.

Can you please provide 2-3 such research articles? I really suspect your claims, because even Ben Graham’s Screens have failed to give superior returns:

Because they charge a higher fee. This has been the major critique of PEs and Hedge Funds. Their pre-fee returns actually beat the index by quite some margin, but they charge a very high fee and ultimately under-perform the index. Another major contributing factor in how often an investor in Hedge Fund or PE fund buys/sells his investments.

Consider the case of a simple SIP in Indian Mutual Funds. Do you think an investor who buys/sells Mutual Funds as and when he wishes will perform better than an investor who keeps on investing without incurring the additional frictional expense related to buying/selling frequently?

So if you remove the high fee and frictional expenses, the true performance of Hedge Funds and PEs are actually respectable.

They don’t know the truth. They are the truth. When someone like Warren Buffet with a solid track record says that the EMH is rubbish, you can’t deny it, because he has proof. He is the proof. But of course, you don’t even have to be as good as Warren Buffet to know that.

In India, the long term returns on the index is close to 11% now IIRC. We have no access to how much each individual investor makes on their investments, but making a 11% CAGR is fairly easy, given the investor is astute. Even ignoring that, Mutual Funds in India have been able to return 15-16% in the same period. Indian Mutual Funds are proof that the EMH is rubbish. Again, this links back to your previous question.

In the 1960s some researchers wondered, just like you, if the PE ratio was a measure of subsequent market performance.
In a comprehensive study done in 1968 the relative performance of high versus low PE stocks was analyzed. 189 companies in 18 industries over the 25 years between 1937 and 1962 were assessed.
The stocks were divided into 5 equal groups solely according to their PE rankings. The quintiles were rearranged by their PE ranking for periods of one to seven years. Recasting the quintiles annually on the basis of new PE information resulted in the stocks most out of favour showing a 16% annual rate of appreciation over the total time span. On the other hand, switching in the highest PEs on the same basis resulted in only 3% annual appreciation over the period.
Even though the performance discrepancies were reduced with longer investment periods, even after the original portfolios were held for seven years the lowest 20% did almost twice as well as the highest.
Source- Financial Analysts Journal ( 1968)

Paul Miller Jr. presented similar results.
He divided companies not quintiles according to their PE.
Highest PE stocks generated 7.7% CAGR between 1948 and 1964 while the lowest PE stocks generated 18.4% CAGR.
It was also found that the lowest 20% of stocks, ranking according to PEs did best in 12 of the later 17 years. The highest 20% did best in only 1 subsequent year and worst in 8.
Source- Drexel & Co

Another study divided the 6000 stocks on the Compustat tapes into five equal groups as per market size, for 21 years ending in 1989.
Each group was then divided again into five subgroups according to PE rankings.
The low PE stocks outperformed the high PE groups for all market sizes from nano cap to mega cap.

Another study was performed in which a sample was prepared from Compustat 1800 which featured the largest public American companies. Yet again, it was divided in quintiles. Portfolio was recast as per new PE information. The period under consideration was August 1968- August 1977.
Results: Lowest PE stocks outperformed the index by 50%.
This was a tumultuous period for stocks.
It was the post was period of 1973-1974.

In 1994, in an article published in the Journal of Finance, Lakonishok, Vishny and Shleifer presented the outperformance of low PE stocks.
They used the thousands of companies on Compustat tapes from 1968 to 1990 selecting stocks listed on NYSE and AMEX to eliminate stock selection bias.
An annual recast based on PEs was performed.
Low PE stocks generated the highest return of 17.8% beating middle and high PE stocks as well as the index which generated 15.1%
Also, low Price to Book, low Price to Dividend consistently outperformed the market.

In a study considering the period between 1970 and 1996 stocks on the Compustat 1500 were divided into quintiles based on PE.
The lowest PE stocks outperformed the market by 144%.

What happens in a bear market?
Between 1970 and 1996 markets fell by 7.5%
Low PE stocks fell by 5.7%.
And, low PD stocks fell by 3.8%.

And, let me make it amply clear that I’ve nowhere in my post claimed that the efficient market hypothesis is valid.
But, quoting leading investors and influencing our decisions based on their quotes is an ill considered move.
Hardly anyone knows where the market is headed.
It’s seldom predictable and the tidy success stories of rags to riches in markets are made palatable to the masses by ignoring in totality the role of luck.

The problem with doing a broad based research is that it offers no application for the real world. People don’t find out 100 stocks with a low P/E and invest in all of them equally (That is what all the studies you quoted induce). We need to understand how an individual stock’s P/E is indicative of its Value.

We already know that low P/E stocks may be undervalued. That’s why the P/E makes a good Screening criteria. My question, once again, it not whether the P/E makes a good Screening criteria. It does.

My question is what’s the sense in using the P/E to Value a company at any point in the investment life cycle?

Let’s forget all the confusions. My question going forward is simple. Consider this scenario:

  1. You buy a stock at 14 P/E
  2. 2 Years later, the stock is at 25 P/E
  3. At this juncture, how do you decide if the stock should be sold or whether you should invest more?

Maybe this succinct study by Motilal Oswal might offer a better explanation:

You didn’t. But you did ask how leading investors know the EMH is rubbish. I showed you how:

  1. They have disproved the EMH themselves by making superior returns continously
  2. Hedge Funds / PEs in the US also disprove the EMH by beating the market pre-fee
  3. Mutual Fund in India do that as well, comfortably

You sure seem to detest the efficient market hypothesis.
I’m no proponent of the EMH either.
But, let’s be fair. Out of the hundreds of thousands of investment professionals around the globe how many are able to outperform the indices?
Selecting some investment stars who’ve outperformed the indices is not enough evidence to rubbish the theory.
Isn’t it possible that those who were successful in beating the indices were very lucky.
Let’s look at the percentage of outperformers not absolute numbers.

Why are we deviating from the topic again? I have clearly stated that let’s not discuss tangential things. If that was not clear, I’m sorry about that. We can discuss the EMH in a separate thread if required.

I will repeat my question. Consider this scenario:

  1. You buy a stock at 14 P/E
  2. 2 Years later, the stock is at 25 P/E
  3. At this juncture, how do you decide if the stock should be sold or whether you should invest more?

My apologies if I digressed.
I’ll outline the strategies I’ve decided to follow.
I select stocks from various industries.
Divide stocks in various groups based on PE.
I then invest in stocks in the bottom 20% of the PE ranking in every industry.
Once stocks reach the market PE multiple I book my profits and repeat the process.
By doing so I’m able to get exposure to various industries and prevent erosion of profits.
It’s entirely possible that the stock I sell will go on to become a mega bagger. But, it’s just the approach I follow.
A person doesn’t necessarily have to sell when the PE has increased significantly.
Using Price to book, Price to Dividends and price to cash flow in unison may work well.
Different investors adopt different means.

1 Like