Portfolio Risk
A book of loans is not the sum of its loans. Its risk lives in the correlation between them — in how many of them fail together — and that is a completely different question from how any single one performs. This is the last thing to learn before method, because it is where individual reading becomes portfolio judgment.
Everything so far has trained you to read one loan well. But nobody holds one loan. They hold a book, and a book has a property no single loan has: its loans can fail together. Ten loans that each look independently fine can be, collectively, a single bet wearing ten disguises — same city, same sector, same vintage, same sponsor, same interest-rate assumption. When the thing they share turns, they don’t fail one at a time; they fail at once. Portfolio risk is not “how risky are these loans on average.” It is “how correlated are their fates,” and averages hide correlation completely.
The tool for seeing this is concentration. You measure how much of a book’s exposure piles into the same buckets — the same geographies, the same property types, the same sponsors. A book spread across twenty uncorrelated markets is a genuinely different animal from a book with the same number of loans all in one metro, even if every individual loan scores identically. The most useful way to express this is effective-N: not the raw count of loans, but the number of truly independent bets the concentration implies. A hundred loans that are really one big geographic wager might have an effective-N of three. That gap — a hundred positions, three real bets — is the whole point, and it is invisible if you only read loans one at a time.
This connects straight back to a perception failure you already met — the flattened dimension. A portfolio can look beautifully diversified on the axes you happened to plot and be lethally concentrated on one you didn’t. Diversified by property type but all one vintage. Spread across sectors but all financed at the same rate that’s about to reprice. Concentration analysis is only as good as your willingness to check the axis you’d rather not look at, because the concentration that kills you is usually the one your dashboard wasn’t drawn to show. The discipline is to keep asking: what do these loans secretly share?
From concentration flows monitoring and watchlists — the operational half. Once you know where a book is concentrated, you know where to watch hardest: the metro that carries a quarter of the exposure, the sponsor who shows up on eight loans, the vintage facing the maturity wall. A watchlist is concentration turned into attention: a standing list of the positions whose trouble would matter most, checked against the early-warning signals from the spine of the course. You are not watching every loan equally — you are watching the ones that move the book.
The platform’s concentration surfaces compute exactly this — HHI, effective-N, top-N exposures by geography and sector — for a portfolio or a securitized trust, and feed the watchlists that tell you where to point the rest of the toolkit. Read a book this way and the individual-loan skills you built earlier snap into their proper place: they are how you read the positions the concentration told you to watch.