QuantumRisk Advisories
Vintage/Triangular Matrix Method (2008, BTS)
The Vintage/Triangular Matrix Method used by the residential (RMBS) and commercial (CMBS) mortgage industry research to estimate defaults and losses is a biased tool. The worksheet, 35KB Excel file, allows you to see for yourself how the method alters the input default behavior such that the averaged behavior no longer matches the input characteristics.
The QuantumRisk Advisories are provided free to clients, and the public, on the condition that there are no legal liabilities to QuantumRisk, its consultants, or associated parties with regard to use or application of these Advisories.
Join Our Mailing List
Email:
For Email Marketing you can trust
FICO Scores Ability to Predict Long Term Defaults (2007, BTS)
Our research suggests that FICO Scores are not good long term predictors of defaults. Therefore, we suggest that you double check your residential mortgage bond investment decisions. More info here.
Statistical Biases in the Quant Algorithm Testing Methods (2009, BTS)
Quant algorithms are tested on the basis of how often they turn in a profit versus how often they lose money. This is an insufficient criterion as upside risk (profits) and downside risk (losses) are not symmetrical distributions. More info here.
Regime Change is Model Misspecification (2005, BTS)
Our research suggests that regime change is not a real phenomenon in finance, so take care when building such models. More info here.
Preserve Your Principal (2010, BTS)
In our chase for profits we often forget to seek ways to preserve our principal. The Wall Street Journal article on 1023 Cherry Road is an eye opener on how not to do things. Therefore be open to high LTV capital structures & match your duration correctly. More info here.
Using Portfolios for Loss Containment (2010, BTS)
Portfolios should be used to mitigate losses. First we have to understand how & why loss tails behave the way they do. Second, understand how effective are portfolios in reducing the impact of tail loss. More info here.
The Risk Return Function is Non-Linear (1995, BTS)
Our research demostrated beyond a shadow of doubt that the risk return relationship is non-linear. That if reward seeking becomes excessive (for what ever reason) the returns diminish significantly to negate any reward seeking behavior. More info here.