A systematic fund has a momentum loading of 0.4, and the momentum factor loses 25% in a month. What does the fund lose from that exposure, and what would hedging it have cost?

The momentum loading costs about Rs 100 crore, 10% of the fund, and turns an 8% market rally into a -1.5% loss. A 0.4 loading on a factor that falls 25% is a 10% hit. A full hedge, shorting 0.4 units of the momentum factor, would have given up about Rs 30 crore a year in forgone premium and costs. If a crash like this comes once a decade, that is more than its expected Rs 10 crore a year of damage, so trimming the loading beats hedging it fully.

A loading is a sensitivity, like the share of a cricket team's score that comes from its openers. If the openers contribute 40% of the runs and they collapse, the total collapses with them. A momentum loading of 0.4 means the fund moves 0.4% for every 1% the momentum factor moves, so a 25% factor loss becomes a 10% fund loss, Rs 100 crore. That is on top of whatever the market does, because the factor is measured as a long-short spread with the market taken out.

Attribute the return factor by factor. Market beta of 1.0 on an 8% rise adds Rs 80 crore. Momentum takes Rs 100 crore, more than the market gave. Value adds Rs 12 crore, quality costs Rs 2 crore and stock-specific effects Rs 5 crore. The fund ends down Rs 15 crore, -1.5%, in a month the market rose 8%: a gap of 9.5 points that clients will notice before anything else.

After a long fall, last year's winners are the defensive stocks and last year's losers are the beaten-down cyclicals. When the market turns sharply up, the losers rally hardest, and a strategy long winners and short losers is on the wrong side of the move. That is why the damage clusters in sharp recoveries after bear markets, and why the market rising 8% is part of the cause, not a comfort.

Price it for a year, not a month. To remove the loading, the fund would short 0.4 units of the long-short momentum factor, about Rs 400 crore on each side. That gives up the momentum premium the fund holds the factor for, about Rs 24 crore a year, plus about Rs 6 crore of costs: roughly Rs 30 crore a year. If a crash like this comes once in 10 years, its expected cost is about Rs 10 crore a year. A permanent hedge pays three times the expected damage to avoid it.

The better tool is to size the loading by risk rather than hedge it. Research on momentum crashes finds they follow periods when momentum's own volatility is high, usually after a market fall, so a fund that scales its momentum exposure down when that volatility rises keeps most of the premium and sheds much of the crash. State the limit: the once-a-decade frequency is an assumption, and a fund whose clients cannot live with a 10-point gap to the market should hold a smaller loading from the start.