Next Boston QWAFAFEW Meeting:
Tuesday, 18 Jul 2017
Dieciocho de Julio Noche
Time: 6:15 PM sharpe
3rd Floor of the Tennis & Racquet Club, 939 Boylston Street
RSVP to hugh@QWAFAFEW.org
Uncertainty-Penalized Portfolio Optimization
QWAFAFEW discussions led by:
Anish Shah, CFA
It is long known that inaccuracy in forecasts sabotages portfolio construction, and approaches have arisen to address the problem. For example, ‘resampling’ (Michaud 1998) perturbs point-estimated inputs and averages the output. Beyond point estimates, ‘robust optimization’ (Goldfarb & Iyengar 2003) places a range around each individual forecast and finds the solution with the best worst case utility.
This presentation’s approach regards inputs as – rather than a range – random variables with estimates of mean and covariance. These mathematically propagate to the mean and variance of each term in the utility function.
(As an example: Covariance of alpha would capture covariance of bets that make up the alpha forecasts. Risk could be forecasted via an uncertain covariance model, as in https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2616109. Uncertainty in transaction costs could be handled by a transaction cost model that includes
The individual terms’ mean and variance propagate to mean and variance in portfolio utility. And maximizing E – c x StdDev articulates inaccuracy where it matters, at the utility level.
The presentation explores uncertainty-penalized portfolio optimization to improve investment performance.
A question for discussion: Clearly, unless uncertainty creates sudden catastrophic loss, the more frequently one plays (rebalances), the less the concern about uncertainty. What is the right way to incorporate rebalance frequency (or turnover) into utility preference?
DOWNLOAD: IGM Uncertainty Penalized Optimization
Anish Shah is the founder of Investment Grade Modeling LLC based in Boston, MA. Earlier stops in his career include risk modeling and optimization at Northfield and ITG. He is a CFA charterholder and earned an MS in Operations Research from UC Berkeley and an MS in Applied Math from Brown (where he left ABD in the PhD program, concentrating in pattern theory).
Anish’s avocation is flamenco guitar (and if the guitar loved him as much as he loves the guitar, he would be one smoking player). Per QWAFAFEW’s request, as he did two years ago, he will open his talk with a flamenco guitar piece.
Extra Extra Bonus
A temporary menu upgrade, Burritos from Anna’s Taqueria.
Please RSVP hugh@QWAFAFEW.org
All annual dues ($150) should be paid to Boston QWAFAFEW by credit card, Bitcoin or check.
By credit card (PayPal)
Send us Bitcoin, our address is:
Please send a separate email identifying your anonymous payment.
Executive Management Associates
12 Academy Avenue
Atkinson, NH 03811
Please RSVP for all meetings you plan to attend: hugh@QWAFAFEW.org
Guests who do not RSVP may have to wait to enter the meeting, due to space constraints (capacity=100, BFD).
Guest fees ($30) can be paid through EMA (check or PayPal, bring receipt) or bitcoin (at address above). We still allow walk-ins (space permitting). The $30 guest fee can be paid by bitcoin, cash or check at the door, no credit cards at the meeting. Please bring exact change (eg: two twenties = $40 Guest Fee).
As always, if you have names or discussion topics, please forward them to any member of the Steerage Committee
Donna Cool Murphy (Dinners)
The members of the Steerage Committee are responsible for coordinating the program content.
Program suggestions from members are always welcome.
RSVP to: hugh@QWAFAFEW.org