Normally, we think of frontiers as boundaries that mark and often regulate our passage from one realm to another. In finance, though, the one frontier worth mentioning – the Efficient Frontier – isn’t a place through which we pass. It’s a destination, supposedly the only place we should want to be. Whatever traveling we do is back and forth along the frontier with the aim of picking a one spot at which we’ll settle down (and move to a new one later on if or when our preferences change).

 NOTE: This post focuses on Wealthfront because its web site provides for the greatest level of insight as to how it works. Bear in mind, though, that the other large generalist robo advisers operate pretty much the same way, so most observations made here can be assumed applicable to the others as well.

The Great Wall of Finance

Imagine you’re atop a long narrow structure that looks a bit, like the Great Wall of China. You can move a long way in either direction along its length, but you can’t really go much along the width, since the wall is pretty narrow.

You do not want to go down off that wall.

• Along one side is the mysterious and impenetrable Land of the Impossible, which is inhabited by creatures that don’t – can’t – really exist. If we think we see any, forget about it. They’d be mirages. One example: A risk-free six-month U.S. Treasury note yielding 42.4% in a zero inflation environment. Here’s another: a stock that goes up but can’t go down. If we assume the Great Wall runs east to west, then the land of the impossible is to the north. On a map, it would be “above” the frontier.
• We could, if we wish, go down the other side, the south side, and enter the Land of Inefficiency. But we choose not to. It’s inhabited by undesirable “sub-optimal” creatures, such as high-risk startups with expected annual returns of 0.03%. We can grab as many of these creatures as we want. But no rational person would want to do that. All of these “portfolios” feature more expected risk than we need to assume for a given level of expected return, or they offer less expected return than we need to accept for a given level of risk.

The top of the wall, the frontier itself, is the only place worth being. In finance, this wall is known as the “Efficient Frontier.” It’s not built from bricks. It’s built instead from “portfolios.” But these aren’t just any portfolios Each is an “efficient” portfolio that offers the highest level of expected return anyone could rationally hope to achieve based on the spot along the wall/frontier (the level of risk) you have chosen.

Migration Along The Length of the Great Wall

Imagine you’re back atop the wall. If you take a step to the right, you’ll find yourself standing on a different efficient portfolio. It will be more risky than the one on which you previously stood. But as compensation for tolerating the extra uncertainty, you can rationally expect a higher return. The same sort of change occurs with each new portfolio you encounter as you move one step at a time to the right until you eventually reach the end, the highest level of risk to which you are subjected given the assets that are in the portfolio. That’s not for the feint of heart. But the intrepid souls willing to venture there get, as a payoff, the highest possible level of expected return.

Now, get back to the starting point but this time, move to the left. Now, your portfolio is a bit less risky than the one you started with. That feels comforting, but finance abhors a free lunch. You pay for that extra comfort by being willing to accept a lesser level of expected return. You can keep going to the left, and reducing both risk and return until you reach the end, a portfolio offering the lowest possible level of risk (and, of course, the lowest possible level of expected return).

Constructing the Great Wall

This Great Wall has nothing to do with Chinese Emperors or Donald Trump. Harry Markowitz is credited with being its primary architect (although if you dig a bit, you’ll see that there is some disagreement on this) and there’s been some important renovation work done by Fischer Black and Robert Litterman.

Finance students learn about that in classes named Portfolio Theory, Portfolio Management, etc. with such syllabus topics as MPT (Modern Portfolio Theory) or MVO (Mean-Variance Optimization). But the math is complex, the data needs are considerable, and the computing power needed to put it all together is enormous. Or at least that’s the way it used to be. (When I was in grad school, I had to bring a pile of punch to the computer lab, drop them off, and come back in a day or two to get the printout that had my results).

Now, it’s different. Data is accessible. The math is still complicated, but a free Excel or Google Sheets add-on that you may not know you have (it’s called Solver) can bang out the answers in a few seconds. So today, anybody who wants to get atop the wall, to have an efficient portfolio, can have one. That’s how the robo advisers build portfolios for people they never meet face to face and do so without a human investment manager having to think about what makes sense for the particular client.

When the client registers, he/she answers some questions and voila, the computer decides which spot on the wall is right for that person and after a few more mouse clicks to complete the registration and fund the account, the efficient portfolio is off and running as soon as the robo receives the funds. It’s just that simple, which is pretty much why “fintech” bloggers have been gushing with praise.

The Law of the Land

How closely were you reading the description of east-west movement along the wall? Did you notice that I sort-of fudged quickly through the discussion of maximum and minimum return-risk tradeoffs? I didn’t define how we know what the maximum and minimum are.

That means the minimum level of risk is not zero. It’s defined by the level of risk we observe in the least risky asset from among the collection we’re choosing to look at. Ditto the other side. The maximum level of risk is what the portfolio would hold if we allocated 100 percent to the riskiest asset.

Here’s why that’s important. The possibilities offered by any efficient frontier are limited by the collection of assets one chooses to work with. If Adviser A chooses a great collection of assets while Adviser B chooses to work with garbage, it’s possible that even the most conservative portfolio held by one of A’s clients may outperform the most speculative portfolio offered by B. So for as much chest-pounding as mathematicians may sometimes do when pontificating to ordinary people who couldn’t tell a Lagrange Multiplier from an inverted matrix, the fact is that a large measure of the success or failure of your efficient portfolio has nothing at all to do with the fancy math that tells you where on the wall you’re standing, but on the quality of the assets (ETFs in the case of robo advisers) selected for potential inclusion in the portfolios.

We don’t usually move from one spot to another on the frontier by changing the underlying ETFs in the portfolio. We’re stuck with those. We implement our return-risk preferences by changing the weights, the percentage allocations each ETF receives (and by the way, even that isn’t all math all the time: Wealthfront humans set maximum and minimum possible levels, constraints, for each asset in the basket – for the record, you too can do that with Solver).

Looking At Some Wealthfront Efficient Portfolios

After you’ve complete the registration questionnaire, Wealthfront uses your answers to compute for you a Risk Tolerance score ranging from 0.5 to 10.0 with a bunch in between measured in increments of 0.5.

Tables 1 and 2 show portfolio allocations for Risk Tolerance Scores of 0.5 (Cowardly Lion); 5.0 (Average Person), and 10.0 (Fearless Gunslinger). I should mention, for those who haven’t guessed, that the verbal labels are mine.

Table 1 – Standard (Taxable) Portfolios

Foreign Stocks (VEA)

Emerg. Mkt. Sk (VWO) US Div. Sk. (VIG) Nat. Resources (XLE) TIPS (SCHP) Muni Bonds (MUB)

Table 2 – IRA Portfolios

Emerg. Mkt. Sk (VWO)

US Div. Sk. (VIG) Real Estate (VNQ) TIPS (SCHP) Corp Bonds (LQD) Emerg. Mkt. Bnd (EMB)

Moving Along The Length of the Wall

In recent posts, I discussed the materials (ETFs) used by Wealthfront to build the portfolios and what I thought of each component.

Today, let’s consider the changes we encounter as we move along the wall from Cowardly Lion to Average Person to Fearless Gunslinger, and see how the portfolio changes.

Not surprisingly, I suppose, Wealthfront really loads up on fixed income for Cowardly Lion clients. This, obviously, reflects traditional ideas about fixed income offering much lower risk (volatility or standard deviation of returns) than equities.

Much of the rest of what we see is also consistent the Investment Advisor 101 playbook: Increase equity allocations for clients able to tolerate more risk, Beef up exposure to income-producing securities in IRAs, since you can get a bit less volatility without suffering from the absence of capital-gain tax breaks.

The main head-scratcher here is the 14% stake Cowardly Lion gets in Emerging Market fixed income. From a common-sense perspective, I personally find that hard to swallow. Undoubtedly, Wealthfront’s quant-leaning Investment Team has reams of expected return, standard deviation, and correlation data to support what they’re doing and strong confidence in the reasonableness of whatever assumptions they made. In other words, I assume the model is correctly determining that the expected returns are high enough in light of standard deviations and correlations to justify using this asset class to account for expected returns so badly needed by Cowardly Lion (badly needed because of the low expectations from other forms of fixed income) and feel no need to consider non-quantifiable aspects of emerging-market risk.

I have the same worries about the pace at which foreign-market exposure and currency risk (these ETFs are not hedged against adverse moves in other currencies relative to the Dollar) ramps up. The Average Person gets a 27% allocation in the Standard portfolio or a 31% stake in the I.R.A.

The Fearless Gunslinger portfolio likewise depends on the expected returns of the foreign ETFs to combine with the standard deviation and correlation assumption in such a way as to justify it’s being the primary engine for driving the higher expected returns Wealthfront need to generate for these clients seek (50% of the Standard portfolio and 54% of the IRA).

Since Wealthfront treats the U.S. stock market as being almost a monolith, a one-trick pony, it leaves itself with no opportunity to boost expected return other than by globe hopping. (Unlike Betterment, for example, Wealthfront does not give itself an opportunity to boost stateside risk-return with, say, small-caps, and none of the robos seem interested in working with other U.S. subsets that could boost exposure to higher reward-risk factors.)

Summing Up

When you sign on to a robo, you are not just swinging a meat-axe at a 1% fee that you might otherwise have paid to a human who would put you in stocks, bonds or ETFs.

You’re not even being passive, which you could just as easily do for yourself with a plain-vanilla stereotypical portfolio consisting of 60% S&P 500 SPDR (SPY), whose net-of-expenses returns correlate very highly with those of $VTI and say 40% in LQD (and you can probably Google ideas that will help you decrease the $SPY percent as you get older, and I already gave you a fixed-income suggestion that problems caused by the fact that LQD never matures).

You are ditching much of what you likely know or thought you knew about investing. Forget Graham & Dodd. Forget Warren Buffet. Forget Peter Lynch. Even forget Jim Cramer. Forget any and all financial planning books or seminars you may have encountered. Forget anything you may have learned at an AAII or Better Investing Chapter meeting. Forget everything you have seen on CNBC or Fox Business. Forget everything you may have learned (hopefully a lot) on Forbes.com. You can even forget the Fama French variety of quantitative analysis.

When you go large generalist robo, you enter a completely different world. (If you want to study up, use Wealthfront’s White Paper as a starting point.) You go all quant all the way. You no longer think in terms of investable assets, or at least not the way you’re used to thinking about them. (For example, you’re no longer allowed to look at China, the Middle East, etc. and wonder about how risky this makes Emerging Market ETFs.) You now have to think in terms of datasets that pretty much resemble many other kinds that may be subjected to the mathematical optimization process, such as traffic, fluid flows, factory processes, etc.

Is this a better world? We’ll see. That’s why I’m tracking model Wealthfront and Betterment portfolios. And that’s why I focus as I do on Smart Alpha, another form of quant that is more aligned to bedrock financial theory and the particular qualities of the equity asset class, in the same family as but not identical to Fama French. (Actually, I think quat can be a wonderful thing, given the way it allows one to banish emotion, bias, hype, etc. from the investment process. Father Time will be the ultimate judge of what kind of quant is better.)