Systematic investment plans (SIPs) concentrate on cost. Every month the investor pumps a designated sum into a financial instrument. This is usually a mutual fund but it could be a stock or commodity.

The SIP offers several advantages over lump sum investments. Apart from enabling the investor to park small sums, it also helps to lower the average cost of acquisition. Every time prices fall, the SIP buys more units. In fact, SIPs are the most powerful argument in favour of open-ended funds versus closed-end funds.

An open-ender can offer SIPs; closed-ended cannot. The supposed edge of a closed-end fund manager who can invest without fear of redemption pressure is negligible compared to the mechanical advantage of a system that automatically averages down.

What about the sophisticated version of the SIP, which concentrates on value rather than cost? Value-cost averaging (VCA) plans aren't offered in automated fashion by mutual funds. Apparently the reconciliation is too much trouble. But it isn't difficult to set up a personal VCA.

All you do is set a monthly portfolio value target. Then you buy or sell the requisite amount of units to meet that target. For example, say you set a portfolio value target of Rs 1,000 growth per month. In month one, your chosen instrument is trading at Rs 13. So, you buy 77 units.

Next month, the price is down to Rs 12. Your portfolio value is Rs 924. You buy Rs 1,076 worth of units -- that's 90 units. You now own 167 units. Month three, the price jumps to Rs 20. Your portfolio is now worth Rs 3,340. Since this exceeds your target, you sell Rs 340 worth or 17 units at Rs 20. And so on.

One difference in the mechanics between an SIP and a VCA is that a VCA incorporates an automatic method of booking profits at high prices as well as a system for buying more units at low prices. The profit-booking can boost returns by a massive degree.

In a strong bull run, a VCA often produces mathematically infinite returns. As prices rise and portfolio targets are successively exceeded, the system books continuous profits which can wipe out all accumulated costs.

For example, look at the attached chart of the Nifty Benchmark Exchange Traded Fund - an instrument which tracks the Nifty and offers units held at a tenth of the Nifty's prevailing level. ETFs can be bought off the exchange just like stocks. This reduces calculation slippage - conventional NAV calculations are always on the last session's closing prices.

Anyhow the Nifty jumped from 1371 in January 2001 to 4015 by December 2006 -- a rise of 92 per cent. Let's say, you implement a VCA of Rs 1000 per month in January 2001. By the end of the 72-month period, you have a portfolio worth about Rs 72,000 (after rounding and slippage) and the net cost is Rs 1086. That's a return of 6,500 per cent.

The costs would turn negative and the return would become infinite if the index continues to rise at this rate for a couple more months.

What's the downside?

The big issue is that a similar 90 per cent drop would lead to an equally sharp rise in monthly costs in order to maintain targets. If you implemented a VCA in a long-term bear market, you would get crucified. 

The second problem is that, since costs are never known for certain, you have a problem deciding on the exact installment target. If you set the target too low, the corpus may not be meaningful.  If you set the target high and the market crashes, you will be stretched.

One way around this: Set a fixed period VCA. Then calculate the final target value. Say for Rs 1,000 over 36 months, the target value is Rs 36,000. Assume a 10 per cent adverse move in month 35. This will mean a total commitment of  Rs 4,500 in month 36. Can you meet it?

So VCA can never be a complete investment method in practice. But if you can set a meaningful target and not stretch the time period too much, it could lead.