You want and need alpha whether you know it or not and every investment managers job is to get you net alpha. What is net alpha? It is simply getting more bang for your investment dollar. Think of higher returns with lower risk where the net of fees (that you pay) return exceeds that of a representative benchmark’s.
This might seem like a lot to unpack but we figured out how to deliver on this basic concept after much trial and error. It is the reason we exist to help you and if we cannot deliver on this basic yardstick, then you (or your adviser) need to search for it.
Let’s keep in mind that the US stock market is approximately $35T, that’s T as in trillion! So, US markets being as smart and developed as they are, we had to evolve our thinking to attempt to get aligned with it unlike the majority of active managers that feel they are smarter than all those other investment managers behind that $35T.
Against that $35T task master, the Achilles heel of most any model is backtested results since any computer with sufficient data can find a profitable answer regardless of whether it is the right answer. Unfortunately, such backtested models almost never survive in real money trading in the unknown.
Nothing new here to most investors but the ‘aha’ moment for us was that human-based, non-systematic approaches suffer from the same dilemma if not more so as systematic, computerized approaches. That is, standard approaches to making investment decisions, whether fundamental or technical analysis in nature, will all suffer from the same issue of back testing, curve fitting or biasing. So, what worked in the past will not necessarily work in the future.
Moral of this story:
1) Everyone has access to the same data and can data mine it to no end,
2) be careful with respects to the type and quantity of data going into your investment decisioning construct,
3) more data is not necessarily better, and
4) using market prices is the most effective means by which to arrive at an investment decision if you do the right thing with it.
Against this backdrop, we started here.
Next question to raise– what’s a bias, how to minimize it? Biases exist in all forms of investing; it boils down to the degree of bias.
For example, standard market capitalization (MC) investing is size-biased (calculated as current market price of stock multiplied by the company’s number of common shares outstanding), meaning if you construct a portfolio of two stocks where one stock has an MC of $9 and the other has an MC of $1, then essentially you invest 90% of your money on one and 10% on the other. You then hope that the stock that has the higher MC goes up because you are now biased to it.
Go back to the early 2000s, when size biasing resulted in underperformance (to an equal weight portfolio) as Cisco (remember that tech name?) was the big kahuna inside of the SP500 index and, given its underperformance through the tech wreck, made a large outsized drag on the SP500 index performance. That is just one form of a bias and there are many.
When a "technician" looks at a chart of moving averages (MA) and concludes that buying a stock every time its 9-day MA moves above its 50-day MA results in great gains: that’s a bias. Because now the technician’s decision-making is predicated on a specific rule. While anyone can find as many of these biases/rules as necessary to get a profitable decision model (based on backwards looking data), there is no statistical basis that such a model will ever result in sustained alpha. The technician simply believes they have an answer when in fact all they created was a set of biases.
And, fundamental analysis is rife with such biases too as any amount of information gleaned from financial statements and disclosed economic data of any form can provide significant fodder for biases that still result in momentary but unsustainable delusions of grandeur in the seek of pure alpha.
So, without getting into citing all the academic research and actual data that supports such contentions, we surmised that the highest probability of generating pure net alpha had to exist on the less-biased end of the investing spectrum. If an equal weight portfolio construct results in the least amount of bias, then that ended up being our starting point for development of a true next generation form of investing.
Equal weighting a portfolio is nothing new and is widely used in many funds. Whilst not as prevalent as MC, both share a common key element in that the risk weighting schemes start with the market price of a stock. That’s right, the adage “the market is always right” is alive and well in this context. In this way, we conclude that for our approach to be similarly efficient and low in bias (like EW and MC), it too can only resort to using market price as its sole data input.
Moral of this component: eliminate biases in your investment process or suffer the fate of most traditional active strategies of no sustained alpha generation.
Now armed with only equal weighting and market prices, we had to figure out how to construct a more learned approach to calculate an appropriate
risk weight.
This required a complete rethink of traditional mathematical and statistical mathematical analysis. Then, figure out how to make the mathematical construct understandable to a computer such that it can learn how to draw a unique, derived return stream for each stock.
No, it is not HAL9000 in capabilities (for you 2001: A Space Odyssey types) but a relatively narrowly focused mathematical construct using just market prices for a computer to determine the risk exposure itself.
Moving on, now that we have a workable investment construct, we put it into live. Interestingly, against the above backdrop, much early live development revolved around a portfolio of ETFs.
Back in 2010, the then marketing speak of the big issuers of ETFs (read: Blackrock, State Street Global Advisors, Vanguard, etc.) was that ETFs somehow were the final bastion for easily attaining a low-cost, diversified portfolio construct. This drew us in as our original all ETF portfolio was balanced across sectors (using sector spiders), commodities, and major market indexes. That portfolio product helped to hone several key learnings that were central to making a diversified model translate into live and getting closer to sourcing that much-ballyhooed idea of sourcing pure net alpha for our clients.
The key learning from live development using ETFs: the ETF structure dampens out most of the individual stock volatility. So, while our investment process can certainly show great results on a basket of ETFs, we found significantly more alpha potential by simply investing in the underlying stocks that make up the same ETFs. There are two significant takeaways:
1) ETFs, even low-cost ones, are inherently very expensive to the end investor because they basically reduce the potential for pure alpha
2) We can design a portfolio that closely meets the needs of the client rather than meets the needs of the investment manager
Summing up, early foray with our core investment process laid the foundation for creating unbiased models that really work with live money and increasing our pure alpha seeking abilities by eliminating ETFs.
Armed as such, we simply use the same process to give the investor what they want knowing beforehand that they will benefit from the core process no matter whether they want ESG, Next Generation, Fixed Income, 60/40% Balanced, NASDAQ100, ADRs, Microcaps - you name it. (Click this link if you want to learn more)
Just like the old Burger King slogan, you can "Have it Your Way!" and seek that alpha by going the Passive 2.0 route.