Early payoff risk is the structural feature of performing notes that the yield math rarely foregrounds. You modeled a 13-year duration and got nine months. The annualized return held up, but the assumption doing the most work in your original model was duration, not rate, and the borrower just removed it.
What you ran into is called prepayment risk, and for performing note buyers it cuts in a direction most people do not account for: a payoff feels like a win, and technically it is, but you were priced for a long, compounding income stream and instead received a lump sum you now have to redeploy. The discount-to-face model amplifies this. A big chunk of your effective yield comes from that 16-cent discount being amortized across the life of the loan. Compress the life, and the annualized number still looks fine, but you lose all the forward cash flow you were counting on.
On the reinvestment side, if the tapes you are seeing are priced tighter than your risk profile allows, that is a real constraint. One thing worth examining is whether your note sourcing has been reactive to what comes to market, or whether you have been working relationships with originators and smaller portfolio lenders who might sell closer to par but with more predictable borrower profiles and lower prepayment likelihood. The borrowers who pay off fast in month nine are usually refinancing or selling, and some of that is visible in the underlying property and borrower profile before you buy.
Also worth confirming with your servicer or a note-side attorney: whether the original note had any prepayment penalty language, and whether that flowed through to you on assignment. Some do not. Most residential notes originated in the last decade are penalty-free. Just confirm what you actually hold before the next one closes.
What does your current sourcing funnel look like, and are you buying off tapes or through direct relationships?
