Saturday, July 29, 2006

PEG Stands for Purely Esoteric Garbage

PEG is completely worthless piece of information in analysing a stock. Its The relation to price and earning is non linear. First look at the DCF equations would tell you that. You would have used DCF models. Just plot a graph on the DCF valuation and different values of expected growth. Notice the curvature of the graph. You will see that any approximation of this curve using a straight line plot will lead you to huge errors.

PEG comes handy when you have to convince yourself to buy a stock which has cought your fancy. The decision has been made but the rational mind requires justification. In comes PEG. At stock growing at 20% available at PEG of only 1. Utter bullshit!

PEG would give you wrong result even if your growth estimation was correct. If you are ready to tolerate bit of maths then here is why I'm saying this.

Suppose PEG is a constant.

P = P/E * E
P/E = PEG * growth
P = PEG * growth * E

Suppose you have 3 companies with EPS, of Rs. 1 per year.

  1. Available at price 100, expected to grow at 3%
  2. Available at price 300, expected to grow at 9%
  3. Available at price 900, expected to grow at 27%

PEG of all these choices are same. Suppose you invest Rs. 9000 in each of these companies. Lets also assume that your prediction about growth was correct. Now see what happens in the 20th year from now.

  1. Earning 157.8, accumulated profit 2418.3
  2. Earning 154, accumulated profit 1534.8
  3. Earning 938, accumulated profit 4375

Earnings = EPS* no. of shares

In hindsight company C was obviously better choice than A but A was significantly better than B. Why did we get this bizarre result? This is because PEG is a variable. Its different for each combination of price and growth.

Mathematically PEG is a slope of the curve plotted between growth and P/E applicable for that growth. When you compare PEG of two different companies then you are assuming that linear relationship between growth and P/E and you would get wrong results here even when you correctly predict the growth.

Sourced from:
http://in.groups.yahoo.com/group/lawarrenbuffet/message/2169

Mad rush for high growth

  1. If the investing public in general, anticipates growth then the prices move up. Many times this rise in prices negates any advantage that the investor would get if the growth did materialize
  2. If the prices are high you are essentially paying for a future which has not yet materialized. [compare this with paying for the assets which exist as of now]
  3. As the business conditions are subject to change there is a risk involved in paying for growth. Higher the expected growth, higher the risk.
    1. If you are expecting 27% per annum growth for 20 years and the company grows at 24.3% for 20 years, you have been amazingly accurate in your prediction. Your forecast is just 2.7%(0.1G) off the mark. But the accumulated profits in these 20 years would be 30% less than what you forecasted and the profit in the terminal year would be 35% less than your forecast. If you paid for this high growth you may end up loosing even after this stellar clairvoyance.
    2. For a 10% growth prediction the same 0.1G, i.e. 1% forecasting error would result in only 12% and 17% hit to accumulated profits and terminal year profits.
  4. Its easy to overrate growth expectations. If a company has grown at a rate of 50% p.a. in last 5 years and future looks bright, can I expect the CAGR of 27% p.a. for next 20 years. Its very difficult. If the company grows at CAGR of 27% for 20 years its profits would be 119 times current profits. Given that the company had grown 50% p.a. in last 5 years, the profits 20 years later would be staggering 904 times the profits it had 5 years ago. The question arises
    1. Does the company has management capacity and vision to achieve this? Scaling up 904 times in 25 years is almost impossible(except for startups which have low base).
    2. Even with 10% p.a. productivity enhancements, this would require 17.7 times resources. It can be capital or human resource. Can the company raise so many resources?
    3. If the future is indeed that bright many more competitors would join the fray. The margins would go down due to law of diminishing returns. Can you expect the revenues to grow more than 119 times?
    4. Suppose the company has 10% market share of the industry. If the industry grows at 15% in the same period the market share of this company would rise to 72.7%, almost a monopolistic level. What is the strength that makes you assume that this company would become a monopoly in 20 years.
  5. The more you look at these arguments the more you would realize the fallacy of the growth assumptions. When you do this you become averse to paying high for high growth. I would love to buy a growing company if I don’t have to pay an extra dime for that growth. The profits of M&M, Sterlite, Jindal have grown as fast as market favorites like Infosys, Wipro. But I paid P/E of 2-5 compared to p/E of 25-40 commanded by Infosys, Wipro. When the growth materialized I gained 20 to 30 times appreciation. If the companies were to fail and go bankrupt I would have got out without significant loss because the price paid was 50% less than the market value of the assets. Such valuations look highly improbable now but the markets have history of manic depressive syndromes. So its better to wait patiently when markets are in manic phases.
  6. Finally what makes you think that the underdogs wont turn the tables. In 2002 SAIL was deep in red. It reported loss of 1707 crs. You may have thought SAIL would go bankrupt, but it bounced back to profitability. In FY05 the profits were 6817 crs., in FY06 4013 crs. Needless to say stock price jumped through the roof.
  7. Growth is good. We all want our companies to grow. But we want them to grow more than the growth that’s already built into prices. Absolute growth number is meaningless and its pursuit has been one of main blunders, the money managers have made throughout the financial history[I’ll explain this some other day].

Saturday, July 15, 2006

On Discounted Cash Flow Valuation

If we would have been able to reduce the stock valuation to a set of mathematical equations then equity investing wouldn’t have been fun. While I understand the utility of following a systematic approach to valuation, I give very little weightage(approx. 2.5%) to the results of my DCF models in my decision making process. The reason for that is simple. The process required so many assumptions that the probability of being right is very small. It is, however, important to do this exercise because it makes you aware of the risks you are taking and it can give you some clues about the upside.


I generally prefer doing sensitivity analysis on the results of these models. For example if valuation resulting from a set of assumptions is X then I would like to see how this changes if the variables change. For example what if the growth tapers off? What if the interest rates rise? This gives me some idea of the margin of safety I have, while making the investment decision.

Each long term investor must understand that at the bottom of it, he is a businessman. You may be owning smalls parts of many businesses but that doesn’t change the essential nature of the argument. When a businessman takes a decision, say to expand capacity, he would do rough calculation on the expected returns. He would analyze the output price levels which would be needed to make the breakeven. He would check the investments required, fixed costs, running costs to make an educated guess about profitability of the project. He would do an analysis of competitive advantages and disadvantages. He would analyze the market to see if there is room for expanded capacity. If the plan looks fine on paper, he would check if he has resources, expertise and capital to execute the plan. Finally he would think about fall back plans, exit options in case the things don’t go as planned.

To me it would look pretty odd to see a businessman basing his investment decision solely on the output of discounted cash flow valuation models. If this argument is correct then the utility of DCF model to for long term investors should not be overrated.

There is a difference in analyzing the expected returns from a mature business and analyzing the expected returns from a new project. The mature businesses tend to be more predictable but not predictable enough to e reduced to a set of mathematical equations.

Regards
Kamlesh
La Warren Buffett