A credit fund holds NCDs in three real estate projects with different loan-to-value, interest cover and unsold inventory. Rates have risen 100 basis points. Rank the three exposures and say which you would cut.

Rank Pune first, Nagpur second and Hyderabad last, and cut Hyderabad. After the shock Hyderabad's cover falls to 0.86x, so sales no longer pay its interest, and at the slower pace its flats take about 35 months to sell against a 24-month maturity. Pune keeps cover near 1.8x. Nagpur holds at 1.25x with its sales finishing right at maturity, so keep it only with an escrow and a cash sweep.

A family with a house worth twice its mortgage can still miss payments if the earner's income drops; the equity in the house only helps once the house is sold. Loan-to-value tells you how much cushion exists if the project is sold; interest cover and sales speed tell you whether the project reaches that point without defaulting. A rate rise barely moves the first and hits the other two directly, through a higher coupon and through buyers who now face dearer home loans.

Work Hyderabad through. Interest at 11% on Rs 110 crore is Rs 12.1 crore, and cover of 1.1x means collections of about Rs 13.31 crore a year. After the shock the coupon is 12%, Rs 13.2 crore, while collections fall 15% to Rs 11.31 crore. Cover is 0.86x and the project is short about Rs 1.9 crore a year, which it can only fund by borrowing more or cutting prices further.

Unsold stock measured in months of sales is a clock. At 15% slower bookings, Pune needs about 16.5 months, Nagpur about 23.5 and Hyderabad about 35.3, against a 24-month maturity. Hyderabad therefore cannot repay from sales; it has to refinance in a market where lenders have just become more cautious. That is refinancing riskThe danger that a borrower who cannot repay from its own cash flow also cannot find a new lender when the old debt falls due., and it is the usual way real estate credit fails.

Now loan-to-value earns its place, as the recovery cushion. Assume a forced sale fetches 70% of the softened value. Pune's land and flats would cover its NCD 1.48 times, Nagpur's 1.21 times, Hyderabad's only 0.98 times. And recovery through enforcement in Indian real estate typically takes years: received three years out and discounted at 12%, Hyderabad's recovery is worth about 70 paise per rupee today.

Cut Hyderabad, even at a price. An illiquid NCD may only sell at a discount; say 6%, about Rs 6.6 crore or 0.27% of the fund's NAV. Holding instead carries an expected loss of roughly 9% of face if you put the default chance at 30% and the recovery at 70 paise, before counting the bigger risk: an open-ended fund holding a paper it cannot sell while investors redeem. If no buyer appears, negotiate a cash sweep of every booking receipt and stop any further lending to the project.

Keep Nagpur with conditions: an escrow over sales receipts, a cover covenant at 1.25x that triggers a sweep, and monthly booking data. Keep Pune. And note the limit of the analysis: the 15% slowdown and 5% price fall are assumptions, so run the ranking again at 25% and 10%. If the order does not change, the decision is robust.