Discussion about this post

User's avatar
Larry Pollack's avatar

Really liked this line regarding MVO modeling: "Those inputs—expected returns, volatility, and correlations—are not objective truths. They are assumptions that, in practice, can be influenced by institutional incentives as well as underlying economics." Too many in the pension industry represent their Monte Carlo simulations based on those assumptions as crystal balls. In practice, it's often not hard to know the results in advance (at least at a high level) based on the assumptions.

And the incentives for public pension plans that use these assumptions include minimizing reported *liabilities* based on discounting projected pension benefit cash flows using expected portfolio returns instead of using (lower) discount rates that reflect the timing and credit risk of the liability cash flows as required under basic finance theory. Consultants whose assumptions and models don't justify high actuarial discount rates, as well as any actuaries inclined to apply finance principles in their discounting of pension liability cash flows, will likely find themselves excluded from the ~$6tr public pension industry. Between the consultants and the actuaries, it's not clear to me who the tail is and who the dog is. In any case, their incentives align.

Michael Aronstein's avatar

Exactly correct. The rage to “prove” market and economic hypotheses through quantification of non-robust systems engulfed academia and has infected investing processes at most institutions. The financial history books are packed with discredited and destructive academic exercises that were promoted by consultants, adopted by major institutional investors and only abandoned after catastrophic failure. The academic literature around the “Nifty Fifty” mania is very instructive. Abandoned at the lows in 1974-5 to shift toward “liability matching” which promoting buying long term bonds right before their worst collapse of the century.

8 more comments...

No posts

Ready for more?