Paper LBO on an oilfield services company whose EBITDA swings 30% either way with oil prices. Buy at 6x with 3.5x debt, exit at 6x in year 4. Work the return at trough, mid-cycle and peak, and say how much leverage a cyclical business can bear.

About 1.6x and 13% at mid-cycle, 2.4x and 25% at the peak, and 0.8x, a loss, at the trough. Debt falls to about Rs 398 crore whatever the exit, so the swing lands entirely on equity. The bigger danger is a trough in year 1, when 3.5x becomes 5.0x. Size debt on trough cash flow: about 2.5x mid-cycle EBITDA here.

A household with a salaried job and one with a seasonal business can both earn Rs 12 lakh in an average year, but a bank should not lend them the same EMI. In a cyclical LBO the debt is a fixed claim on a variable cash flow, so every swing in EBITDA lands on the equity, and the lender's question is what the trough year looks like, not the average. Set the deal up the usual way: Rs 1,200 crore of price, Rs 700 crore of debt, Rs 500 crore of equity. Then run the cash, and only then run three exits.

Free cash flow is EBITDA less interest, tax and Rs 40 crore of capex. In year 1 that is 200 less 70 of interest, less 22.5 of tax, less 40 of capex: Rs 67.5 crore, and it grows each year as interest falls. Over four mid-cycle years Rs 302 crore of debt is repaid, taking it to about Rs 398 crore. In a trough or peak exit year the cash differs a little, but debt at exit sits between Rs 353 crore and Rs 443 crore in every case.

Now apply 6x to the year 4 EBITDA. At mid-cycle, Rs 1,200 crore less Rs 398 crore of debt is Rs 802 crore of equity, 1.60x and 12.5%; at the peak, Rs 1,207 crore, 2.41x and 24.6%; at the trough, Rs 397 crore, 0.79x and -5.6% a year. A 30% move in EBITDA moves equity by about 50% either way, because debt does not move with it. Say one thing about the multiple too: buyers rarely pay 6x for peak earnings or so little for trough earnings, so the peak case is flattered and the trough case is harsher on paper than it may be in a sale.

The real danger is not a trough at exit but a trough straight after closing, before any debt is repaid. If EBITDA falls 30% in year 1, Rs 700 crore of debt is 5.0x, and EBITDA less capex covers interest only 1.43x; a 45% fall leaves 6.4x and no free cash at all. Any maintenance covenant set with normal headroom off 3.5x is breached in the first bad quarter. Size the debt on the trough instead: if trough EBITDA less capex, Rs 100 crore, should cover interest twice, interest can be Rs 50 crore and debt Rs 500 crore, 2.5x mid-cycle and 3.6x at the trough.

Then show what the safer structure costs. With 2.5x debt the sponsor puts in Rs 700 crore. The mid-cycle return falls only from 12.5% to 11.2%, the peak from 25% to 20%, and the trough loss shrinks from -5.6% a year to -1.3%. A point and a bit of mid-cycle IRR buys a deal that survives the year the oil price falls. Close the view there: at 6x mid-cycle with 3.5x debt this is a bet on timing; at 2.5x it is a bet on the cycle averaging out, which is the bet a sponsor can defend to its own investors. The limit is the 6x exit assumption, which carries the peak and trough answers more than any line of the cash flow.