You can overweight one of two sectors by 3 points: private banks with 16% loan growth, 16% return on equity and 2.6 times book, or specialty chemicals with 22% earnings growth, 18% return on equity, 38 times earnings and 15% of downgrades over three years. Which, and what would change your mind?

Private banks, because at about 16 times earnings they are priced for their growth, while chemicals at 38 times are priced for more growth than the downgrades suggest. On simple building blocks, banks offer about 14% a year over five years against about 7% for chemicals if their multiple settles at 25 times. I would change my mind if bank credit costs start rising, or if chemical earnings estimates turn upwards while the multiple is still falling.

Put them on one yardstick first; comparing 2.6 times book with 38 times earnings is comparing kilometres with litres. For a bank, earnings are book value times return on equity, so price to earnings is price to book divided by ROE: 2.6 over 0.16 is about 16.2 times. Now the gap is visible. Chemicals grow earnings faster, 22% against about 16%, and earn slightly higher returns, but cost 2.3 times as much per rupee of profit. Adjusting for growth, the banks trade at 1.02 times earnings per point of growth and the chemicals at 1.73.

Think of a friend who says every year that he will run a marathon in three hours and always finishes in three and a half. After a while you adjust his forecasts yourself. Estimates cut by a cumulative 15% over three years, about 5.3% a year, suggest the 22% consensus has been too high, so a fair working number is closer to 15.6%. That matters twice for a stock at 38 times: the growth is lower, and a sector that keeps missing forecasts rarely keeps its premium multiple.

Use building blocks: earnings growth, plus the change in the multiple spread over five years, plus the dividend yield. For banks, assume growth of 16% and the multiple easing from 2.6 to 2.3 times book, costing 2.4% a year, plus a 1.2% yield. For chemicals, use the revision-adjusted 15.6% growth, a multiple falling from 38 to 25 times, which costs 8.0% a year, and a 0.5% yield. That gives about 14.4% a year for banks and 6.8% for chemicals; for chemicals to match banks, their multiple would have to end near 35 times, close to where it starts.

Name one signal on each side. For banks, the risk is credit cost: a 16% return on equity that falls to 13% as bad loans rise would cut the justified price to book, and loan growth of 16% needs deposits growing as fast to fund it. Watch slippages into bad loans and the gap between loan and deposit growth. For chemicals, the signal is the revision trend turning positive while the multiple has already fallen; at 25 times with estimates rising, the case would flip. Say also that both overweights are relative bets: 3 points of the fund is a view on which sector beats the index, not a call that either will rise.