Modeling a refinance at a stale rate is how trapped capital happens on a BRRRR
Take a single family BRRRR: purchase 158k, rehab comes in at 41k against a 36k budget, closing and holding 11k, all in 210k. Appraised at 268k, a solid appraisal only 7k under estimate. The failure in a case like this is usually upstream of the numbers themselves. If the model was built with a refinance rate quoted many months earlier on a different property and never re-quoted before going firm, the rate by the time of the actual refinance can be meaningfully higher. But the deeper problem often isn't the rate itself, it's that the lender sizes the loan to the property's debt service coverage, not to the loan-to-value ceiling the model assumed. At 75 percent of 268k the loan would be 201k. Say rent is 2,100 and taxes and insurance run 610 a month. At a higher actual rate, the payment on 201k plus taxes and insurance can push coverage below a lender's 1.15 minimum, sizing the loan down to something like 170k instead, that's 31k of capital still sitting in the house that was mentally allocated to the next purchase. The constraint that binds in this scenario isn't the one most operators underwrite for. Worry tends to go toward the appraisal when the appraisal is usually fine. The fix is to model both the value ceiling and the coverage ceiling on every deal, take the lower of the two, and re-quote rates right before going firm rather than at the start of the search, and avoid pre-spending cash-out proceeds before they're actually in hand.
