Cagan’s report does take into account CLTV. On page 4, he explains how he computes the loan amounts, and he makes very conservative assumptions: he adds first and second mortgages, and then adds all the HELOCs on a property (since he doesn’t have access to current balances on those HELOCs, he assumes the full amount has been drawn).
So I believe his database is rather good, and, to my knowledge, this is the only public report that has statistics on CLTV amounts (stuff coming from NAR and CAR only reports LTV, which, as you point out, is quite meaningless).
As for his analysis, there he stops presenting facts and starts making some assumptions and projections. Naturally, projections and forecasts are debatable, and different people making different assumptions would arrive to different conclusions. Assuming 20% or 30% price drops would certainly change the results, and I think the reader is free to look at the data and see what would happen under different scenarios.
In my opinion, the greatest value of the paper is in making public the loan data. If anybody knows of any other public resource that has equal or better data, I would surely like to look at it.
