F Financial_Lion_7458 · 1d ago

Kelly Criterion & Non-Positive Definite Matrices in Asset Allocation

I am currently dedicating my weekends to preparing for the CISI PWMSP, specifically focusing on the advanced asset allocation theories. I find myself stuck on a theoretical calculation regarding the optimal portfolio boundary when serial correlation exists in the returns data. Does the Kelly criterion alter its leverage strategy during periods of negative skewness, or does it still maximize the geometric growth rate based solely on the mean and variance? I am trying to understand if the forecast error margin remains statistically significant when the correlation matrix loses its positive definiteness.

Verification code

Sent to

Clever_Student_5025 13h ago

Kelly assumes zero serial correlation so your covariance matrix is likely invalid if you ignore it, leading to massive estimation error and regulatory risk if the auditor sees a non-positive definite matrix. The IISI syllabus explicitly says you must correct for autocorrelation or you fail the asset allocation section so be careful check this: /certifications/cisi-icwim/