Summary:

Portfolio managers around the world regularly engage in efforts to improve. Improvement is motivated, by among other forces, the manager’s internal desire to excel, the expectations of client/investors and the demands of an ever changing market. Yet, managers report little confidence in their ability to improve. They invest tremendous amounts of time, energy and other resources each year in improving but their payback is modest at best and more often disappointing. This is further borne out by the large fraction of active managers that underperform each year.

This improvement dilemma contrasts sharply with professionals in other high performance industries, such as: pilots, surgeons, orchestra musicians and elite athletes. Top performers in these and other professions are able to improve continuously throughout their careers and do so with high confidence that their efforts will translate into better outcomes.

This discussion will describe why portfolio managers are highly constrained in their capacity to improve currently. The culprit being inadequate feedback provided by traditional portfolio analytics. It will then go on to discuss new analytics that provide the level of feedback necessary for managers to learn appropriately about and improve their skills. A case study will be provided.

Outline 

–         Skill versus outcomes

Why traditional portfolio analytics are inadequate for enabling portfolio managers to improve.

–         Defining skill

Quantifying how good a manager is at buying, selling and sizing positions. Description of analytic methodology used to separate skills from outcomes.

–         Improving with clear, rigorous, granular feedback

Presentation of a case study describing one manager’s success at learning and improving.

Includes novel analytics for quantifying skills (How much alpha comes from buys, sells and sizing decisions uniquely?); describes method for quantifying effectiveness of selling under a variety of conditions (i.e. old and young positions; winners and losers); and presents behavioral finding based on out-of-sample analysis and P-value ranking.