There are 35T’s driven passively while 65T’s are driven actively. The former is the Asset owner - the pensions, the endowments, the sovereigns, the families, and the individual investor sitting on the trade screen looking at his 401K account go up and down daily. The latter is the Asset Manager - the rocks, the guards, and their fidelities. The power equation is skewed in favor of the managers. The owners have to queue up for service [1]. One could stand in a queue but the principal-agent problem is bigger. There is conflict in purpose. The manager’s purpose is short-term aka fees, the owner’s purpose is servicing the sponsor, in many cases solving the pension’s crisis [2] - how to operate in a low yield inflationary world while funded status stretch and longevity risks increase.
Every crisis is an opportunity. Machine investing has a 65T bridge-building opportunity. A bridge above the managers, giving DIY asset management tools to owners while the managers suffer fee compression, unrelenting compliance, negative alpha, idiosyncratic world, leaving them with little time to innovate. Managers could shop for ML startups, prepare, be open, but won’t that change how nature works. The second mover has to replace the first mover. ML is the second mover here and it’s building the bridge to decentralize investing, the app for investing shouldn’t be far away.
History suggests that impactful innovations are not necessarily complex. If 'Page Rank' [3] could organize the world’s information, we may not need a master-algo to invert the 99/01 iva statistics which for years have claimed that it is better to be standard and poor than to be exceptional and rich. Despite aggressive academic resistance stock markets remain a scientific challenge [4] (behavior is scientific) which suffers from Campbell Goodhart’s observer effect [5]. An algorithm that can work around this hump can build the bridge.
The machine sits between the investor and domain (financial markets in this case). The investor (Owner) needs customization while the machine has to work on the trade-off between scalability and customization. The investor may assume or avert risk, the machine has to focus on the spectrum of interactions of the investor with the market. The investor is biased but the machine can’t be. The investor is an amplifier, the machine can’t be. The investor is a storyteller, the machine has to be explainable. The investor is buried in content, the machine has to operate content light and context heavy. The investor lives in an irreplicable world, the machine has to replicate. The investor breathes causality, the machine can’t be confounded and has to anticipate.
If an algorithm can do the above, it can build the bridge.
[1] G. L. Clark, A. H. B. Monk, Asset Owners, Investment Management, and Commitment: An Organizational Framework, The Journal of Retirement, Winter 2019.
[2] The Nation’s Fiscal Health, An Annual Report to Congress, April 2019
[3] J. MacCormick, 9 Algorithms that Changed the Future, 2012
[4] M. Pal, How Physics Solved Your Wealth Problem!, SSRN, 2016
[5] P Dirac, The Principles of Quantum Mechanics, 1930