Thanks to the Internet era, we columnists receive instant feedback from you. It is both a joy, and a challenge, when you ask me to explain myself. In December, I said that buying and selling decisions were made easier if one had a target price, set by one's understanding of the company's business.

This led to a large number of email queries like "How do you set a target?", and "Is the P/E (price to earnings) ratio the only benchmark for setting a target?"

The answer to this question has filled many books. Yet as a stock-picker in the public domain, I must try and answer it. Please bear with me if you know all that I am saying, and also with the fact that my answer will spread over two issues. Let me start by saying that there is no escaping from P/E ratios.

Let's begin with 'E', or earnings. A share in a company is a share in its earnings as long as the company exists (of course, if you sell the share, this entitlement passes to the new owner).

The dividend, or what the company pays you, is only a fraction of this earning since most companies retain much of their profit to finance further growth; but then that growth leads to greater earnings, in which you still hold a share. 'P', or price, is what the market is asking you to pay for this stream of earnings.

Clearly then, when buying a share, we should calculate whether its price is justified by what the company will earn in the future.

This raises two issues -- one is estimating this future stream. The second question is -- what will you pay today for returns in the future? For example, what is the present value of Re 1 every year from now until the end of the century?

We know that the actual value of this sum is Rs 93, but clearly no one will pay this amount. For two reasons -- firstly, most of us would rather have Re 1 today than Re 1 tomorrow (and Re 1 to be received in 2100 seems practically worthless today); secondly, if the interest rate is 8 per cent per annum, I can 'buy' Re 1 per year by putting Rs 12.50 in the bank.

Now, if interest rates go up overnight to 10 per cent, that same annual income of Re 1 per year will cost only Rs 10. Since fixed income streams suddenly became cheaper, shares should also become cheaper. The inverse relationship between interest rates and share prices generally holds true, but in our markets, stock prices have not reacted to the recent hike in interest rates. This is because growth in earnings has been so strong.

This means that the amount we are willing to pay for a share depends firstly on its current earnings (the P/E ratio); but more importantly, it depends on our expectation that earnings will grow over time. As the Indian economy has gathered steam, companies across sectors have shown earnings growth year-on-year (and quarter-on-quarter). This has increased the willingness of investors to bank on growth, and value shares at higher and higher multiples of current earnings.

Thus, in May 2003, shares in the Nifty were valued at 10.84 times earnings. Today, the same index is valued at almost 23 times earnings. This is despite the fact that the price of a fixed income stream has come down (which is what happens when interest rates go up), as interest rates have climbed by about three percentage points in these three years.

This seems to be illogical, but is not because the investor in shares has seen the earnings in shares go up by over 25 per cent each year during the time period. Now this is only an analysis of the past. What about the future? And, moving from the theoretical to the specific, how do we decide what we are willing to pay for a specific company's shares?

These are questions I will attempt to answer in my next column.

The author is an investment advisor to a select group of clients. He can be reached at msatyanand@yahoo.com