I was asked by an Advisor friend, what to do with the portfolios. My first response was not to book losses. Knowing that portfolios were mauled, there was limited salvage value. The conversation continued and I told him there are always opportunities to invest, be it Agricultural Commodity ETFs, frontier markets, and/or some emerging markets. The conversation ended there but steeped somewhere in my subconscious and kept bugging me for a systematic Machine Investing solution.

I realized that markets had reached a Nowhere To Hide (NTH) status because equity V-shaped recoveries looked less likely. Though anything could happen to a market headed for capitulation, anticipating the quality of the bounce was like guessing the number of pennies in the glass jar at the local mom-and-pop bookstore in the neighborhood. Guessing is not investing. And with Oil an inch away from $100, WEAT more than 100% from 2020 lows, Cryptos still looking for a bottom, Fixed Income in a “Touch Me Not” state, Real Estate in a house of interest rate cards phase with employees refusing to come to the office, the world of investing looked so different and calling a bottom could be like catching a falling sword. There are exciting and profitable opportunities all around, but without a long-term holding period, one could be in for increasing risk before getting into any profitable reversion.

Teaching DIY investing to the man (woman) on the street, is a difficult job, but not thankless. I had done it before, starting the year 2000, calling different multi-year peaks and troughs, but never had I taught Machine Investing. The subject is new and trusting machines is a journey, dramatically different from the human approach centered around NEWS and information. Unlike humans, machines don’t get stressed, if there is NTH as their job is to find and hunt for investment opportunities, build portfolios, reduce risk, optimize holding period, reduce noise, and beat the respective benchmark.

There are many differences between Machine and human Investing.

1 – Geographical Bias

2 – Asset Bias

3 – Cash

4 – Value and/or Growth

5 – Chaotic Statistics

6 – Portfolio of Portfolios

7 – Smart Beta

8 – Currency Bias

9 – Variance and Mean

Geographical Bias – Unlike Humans, the Machine has no Geographic Bias. Give it G30 scope and it will hunt for opportunities in G30 economies. Give it G10 and it might tell you the limitations in the mandate but still work to conjure up a solution, illustrating the risks. Both Institutional and Individual investors suffer Geographical Biases because familiarity seems knowledge, which in most cases creates risk. Familiarity also creates confirmation bias. Knowing does not translate well into Alpha (beating the respective benchmark).[1]

Asset Biases – Machines don’t need to be told twice that Real Estate is not a hard asset, it’s a paper asset and with more than 50% of Global Wealth in Real Estate, Machines understand what damage a reversion can do. Machines don’t have the challenge of assimilating history, as history is their unique advantage. They can learn from history and create a new alternate history, an informational structure, a language, only they can understand.

Even if Equity was not one of the oldest assets and instruments, it is the most popular and just because something is popular does not make it the most return generating, especially, when the popularity has reached the proverbial shoeshine boy. When the ‘Rich Gets Richer’ [2] seems infallible, that is when it starts it descent. Equity is popular culture now, the very reason, the element of surprise. All people can not become rich together at all times. 

“When the ‘Rich Gets Richer’ seems infallible, that is when it starts its descent.”

Cash is an Asset – Machine understands that Cash is not trash [3] and is a variable that should be carefully used [4] to avoid the herd and reallocate. 

Value and Growth – Machines understand that Value and Growth, Momentum and Reversion [5], Divergence and Convergence are all different names of the same natural phenomenon, which explains how trends persist and reverse. The Machine is focused on understanding the transformation of one force into the other and how behavior can be mapped for better tactical and strategical allocations for Alpha generation.

Chaotic Statistics – Machines understand the Campbell [6] and Goodhart’s law [7], which suggests that “When a measure becomes a target, it ceases to be a good measure”. “The more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor”. In simple words, an indicator that seems to work, is only preparing to fail. The Machines are designed to assume that indicators, variables, and statistics are dynamic, designed to be chaotic, destined to be noisy and hence relying on them is a fool’s game. Machines’ job is to look at statistics in their chaotic context, not as a dependable measure.

“An indicator that seems to work, is only preparing to fail”

Portfolio of Portfolios – Machines understand that the only way to reduce risk is by building a portfolio. And the future is not about buying stocks, but buying portfolios. Because reduced risk is intrinsic to the stability of returns. If the risk of a stock can be muted by adding it to a portfolio, the risk of an asset can be muted by building a portfolio of portfolios of assets. Hence Investing is about a portfolio of portfolios, which does not naturally give up on returns because there is super diversification. Machines understand that systematic asset allocation is just another name for portfolio of portfolios.

“Reduced risk is intrinsic to the stability of returns.”

Smart Beta – Machines understand that Intelligence in essence is about understanding an incumbent basket method [benchmark] and building a new improved basket method [A new Index] that is similar or lower in risk and better in returns. This new method that is open, transparent, and continues to beat the old method across baskets is the definition of intelligence as it extends the idea from stock market data to any other data and hence builds a layer of intelligence above all information. A scalable method to build a better basket for anything Data. In Investing terms this means that a machine does a better job building an Alpha generating portfolio than a human portfolio of selections because informational content is a popular and poor indicator that works and fails. A Machine’s sole preoccupation is to find a scientific method that can find a golden ball from the urn of balls, without looking inside the urn.

Currency Bias – Currency is yet another portfolio. “To be in Dollars or not to be?” is not the question for the Machine. The question for the machine is how to build the basket of the currencies that can answer, “To overweight the dollar or not to do?” [among the other currencies]. Such a currency model can look at a composite and manage currency risk while assuming global macro investment exposures.

Variance and Mean – New age Machines don’t think about how to maximize returns for a certain given risk, but how to reduce the risk for a given return. Such inverted thinking is how Machines become intelligent and to conceive such inverted problems. The Machine has to assume that risk and return are not connected in a linear way but in a non-linear and chaotic way, where they transform into each other, relentlessly.

These are some of the things Machines can do so that you don’t have to worry about Where to Hide. Because your job is about testing, understanding, and experimenting with such machines so that they can free you from your fears and anxieties as it educates you how to invest for the next 12, 24, 36 months and maybe longer into a portfolio of portfolios that may have an Agro 10, a Crypto 10, an India 50, a Currency 10, a global 100 and maybe a few more portfolios.

Welcome to the world of Machine Investing.

AlphaBlock Team

Bibliography

[1] Obstfeld. M, Rogoff. K, “The Six Major Puzzles in International Macroeconomics: Is There a Common Cause?”, NBER, July 2000

[2] Rieman. J, “The Rich Get Rich and The Poor Get Poorer.” Wiley. 1979.

[3] Dalio. R, “Cash is Trash”, Bloomberg, September 2021

[4] Pal. M, “How to Cash?”, AlphaBlock Newsletter, Edition 10, May 2022

[5] Pal. M, “Reversion and Diversion”, SSRN, November 2015

[6] D. Campbell, “Assessing the impact of planned social change”. Evaluation and Program Planning, 1979

[7] C. Goodhart, “Problems of Monetary Management: The U.K. Experience”. Papers in Monetary Economics. Vol. 1. Sydney: Reserve Bank of Australia, 1975