This article is meant to serve as a companion to “What the Evolving Robo Advisory Industry Offers,” a feature article in the October 2016 AAII Journal. Particularly, it compares the asset allocation recommendations and risk-return profiles that were created by several robo-advisers using a hypothetical AAII member.

Additionally, some of the robo-adviser companies disclose a “rough estimate” of the different allocations they use for specific risk-return profiles. Companies that disclosed their set asset weightings include Alpha Architect, Asset Builder, Betterment, Fidelity Go, Hedgeable, Personal Capital, Schwab Intelligent Portfolios, TradeKing Advisors and Wealthfront.

Keep in mind that just because a robo-adviser lists set asset allocations for various types of investors doesn’t mean that you would be placed in the exact allocation you think you fit into. For example, after going through the Betterment questionnaire as an “average AAII member,” Betterment’s algorithm returned a slightly different allocation than the predefined allocation that seemed to fit our example risk-return profile.

Additionally, these allocations may change over time. Each companies’ algorithm and model may change depending on the current market conditions. A model’s output is a function of its inputs.

Determining Allocation

Generally speaking, there are two different types of asset allocation: strategic asset allocation and tactical asset allocation. Strategic asset allocation compares an investor’s return objectives, risk tolerance and investment constraints to long-run capital market expectations. Asset classes are chosen and then permissible weighting ranges are determined. When a portfolio is rebalanced (typically at specific time intervals), the asset classes are reallocated to their target weights. Tactical asset allocation involves making short-term adjustments to asset class weightings based on short-term expected returns and relative performance among asset classes.

Tactical asset allocation is more of a market timing approach and the manager is often trying to beat a benchmark. Asset classes can be intentionally underweighted or overweighted based on short-term market expectations; tactical asset allocation is active management at the asset class level.

A majority of the robo-advisers employ strategic asset allocation.

With most firms, you will see the terms “mean variance optimization,” “efficient frontier,” “Black-Litterman model” and “modern portfolio theory” mentioned. Describing each of these models is beyond the scope of this article, but if you are curious you can find ample information online.

At their core the models focus on the risk-return profile: The goal is to generate the maximum return for a given level of risk. Risk is typically measured by standard deviation, and return is usually a derivation of historical returns combined with an “expected return.” Correlation—the way different asset classes have historically moved in relation to each other—is also considered. The models will add an asset class to your portfolio if its expected return boosts that of your overall portfolio, without adding a significant level of risk. The lower the correlation between asset classes, the better.

These models are extremely sensitive to their input parameters, and are particularly influenced by expected returns.

It was interesting that although many services cited the same models and key phrases, they often recommended totally different asset allocations. The reason for the difference between companies lies within the assumptions used and how often the assumptions are updated. For example, Wealthfront updates its assumptions annually, while Vanguard, which takes a bit more of a “hands on” approach to its robo-investing service, analyzes each investor’s plan quarterly.

The tax efficiency of investments can also play a role in asset allocation. Figure 1 shows Wealthfront’s explanation of the benefits of each asset class: U.S. stocks, foreign developed stocks, emerging market stocks, dividend growth stocks and municipal bonds are considered by Wealthfront to be tax efficient.

Figure 1. Wealthfront Asset Breakdown

 

 

Robo-advisers generally determine tax efficiency by the cash flow distributed by an investment: the higher the cash flow (dividend and interest payments), the less tax-efficient the security. Alpha Architect, AssetBuilder, Betterment, Personal Capital, Schwab, Vanguard, Wealthfront and WiseBanyan all address “tax-efficient investing.”

Robo Questionnaires

Another reason for the large variation between robo-advisers’ asset allocation recommendations is the way they bucket investors into different risk-return profiles. Not all companies use the same five risk-return profiles, and not all companies ask the same questions to determine where you lie on the spectrum.

Alpha Architect, AssetBuilder, Betterment, Schwab Intelligent Portfolios, TradeKing Advisors and Wealthfront allow you to fill out a questionnaire and get a recommended allocation without actually signing up. To try it for yourself, go to these links.

Alpha Architect: Click here and select “invest now” at the top of the page.
AssetBuilder: Click here and select “find your portfolio” toward the top of the page.
Betterment: Click here and select “get started” toward the top of the page.
Schwab Intelligent Portfolios: Click here and select “open an account” at the top of the page. (You don’t need to open an account to get the recommendation.)
Wealthfront: Click here and select “invest now” at the top of the page.
TradeKing Advisors: Click here and select “get started” in the center of the page.

Generally speaking, the robo-advisers factor in the following metrics:

• Age
• Income
• Liquid assets
• Investable assets
• Desired investing term

Each question helps the robos determine your ability to take on risk and your willingness to take on risk. Although they sound similar, these two metrics are very different. Ability to take on risk has to do with your net worth, liquid investments, financial needs and age, while your willingness to take on risk has more to do with your tolerance for and aversion to risk. Ability usually trumps risk. For example, even though a 25-year-old may be comfortable losing $50,000 in a given year, that doesn’t mean he or she actually has the $50,000 to lose.

The process attempts to mimic the typical method used by a traditional adviser, where the adviser creates an investment policy statement (IPS) for each client. An IPS looks at the investors’ risk and return objectives as well as their constraints, which typically include liquidity requirements, time horizon, tax considerations, legal and regulatory considerations and unique circumstances.

Throughout the robo questionnaire process, however, you are not provided with a full IPS. The result is usually just a basic profile of risk and return that allows the robo-adviser to determine which asset classes and respective weightings to employ. An exception to this is Vanguard Personal Advisor Services, which mentions that an IPS will be created for each investor (but not for free!).

Some of the robo-advisers employ goals-based investing, which alters asset allocation based on the goals you have selected. Instead of solely asking questions about your net worth and age, goals-based investing allows you to choose from different goals such as saving for a car, saving for a retirement account, building wealth and saving for an emergency fund. Based on our analysis, some of the robo-advisers that offer goals-based investing include Betterment, Schwab Intelligent Portfolios and Vanguard Personal Advisor.

As an aside, not one of the robo questionnaires examined for this article asked about the investor’s liabilities or fixed expenses.

Asset Allocations

In the spreadsheet attached here, you can see the recommended asset allocations for different levels of risk.

Keep in mind that a majority of the robo-advisers do not invest directly in each asset class; instead, they invest in asset classes through exchange-traded funds (ETFs). So while a particular company might have 20% allocated to U.S. stocks, it typically means U.S. stock ETFs. To see which companies include securities other than ETFs, see our Overview Robo Table. 

As a reminder, these asset allocations may change over time; just because they are mentioned on the robo-adviser’s website doesn’t mean you will definitely be bucketed into one of the respective weightings.

Also, certain asset classes can be changed over time as the robo-adviser reanalyzes its long-term outlook, expected returns and correlations.

One example is WiseBanyan’s statement regarding emerging market bonds, “Currently, we do not include emerging market bonds in our investor portfolios. This is because our model and investment process has determined that money is better allocated to emerging market equities and U.S. bonds. Were we to include this in our portfolio, we would use iShares USD Emerging Markets Bond Fund ETF (EMB).” WiseBanyan also does not include municipal bond or natural resource ETFs in its investors’ portfolios at this time.

It may also be difficult to compare “apples to apples” because several robo-advisers separate their asset classes differently. Some are more specific in their allocation breakdown than others. Schwab Intelligent Portfolios appears to be the most specific in their asset class breakdown, listing as many as 21 different asset categories. The company also goes even further into detail by explaining which securities they use as “primary” choices and which they use as “secondary” choices within each asset class grouping. Schwab’s software uses primary and secondary groupings in order to employ tax-loss harvesting; the algorithm switches between market-cap-weighted ETFs and fundamentally weighted ETFs to avoid wash sale rules. Schwab also explains why each asset class was selected, what role it plays in the portfolio, when it performs well and poorly, and why a specific security was chosen above others.

Not all of the robo-advisers disclosed the specific assets or ETFs that are used in each asset allocation. AssetBuilder, Betterment, Fidelity Go, Hedgeable, Schwab Intelligent Portfolios, TradeKing Advisors, Wealthfront and WiseBanyan are those that did disclose this information. Click here to download a Word document with a list of securities employed by each of the robos mentioned above.

It’s important to find out from any prospective robo-adviser what level of flexibility you have in changing the asset classes or assets used once a recommended allocation is presented to you. For example, Fidelity Go, Wealthfront and WiseBanyan don’t allow investors to alter the asset classes or investments used, but Schwab Intelligent Portfolios allows investors to exclude up to three ETFs from the recommended portfolio.

Another important question is whether the robo-adviser analyzes your portfolio as a whole—that is, does it link to your outside brokerage accounts? Or is its analysis limited to only the investments you hold in your robo account? If the service is only analyzing the investments held with it, can it provide a clear picture of your investment objectives and constraints? For example, Wealthfront and WiseBanyan don’t take into account an investor’s entire portfolio, but Betterment and Vanguard do. Not every service spells out what it does on its website, so be sure to ask.

For comparison, AAII’s asset allocation models for an aggressive, moderate, and conservative portfolio are shown in Figure 2.

Figure 2. AAII Asset Allocation Models

 

 

Hypothetical AAII Member Results

To compare the robo-advisor recommendations, we filled out questionnaires at the sites using an “average” AAII member’s risk-return profile. Tables 1 through 8 display the asset allocations that were recommended for this hypothetical investor. The profile of an average AAII member is: age 65, preparing for retirement, slightly risk-adverse but willing to take on some volatility, and investing for the medium to long term. An initial investment of $50,000 was used, which represents a fraction of the average AAII member’s net worth.

Note that the recommended asset allocations can significantly change if one question is answered differently.

Performance

Although asset allocation concerns many investors, I frequently get questions about how the robos perform. Unfortunately, not every robo-adviser displays their performance.

Even though some services display the performance of their strategies, comparing performance between robo firms is difficult. First of all, each investor has different inception dates, which makes an apples-to-apples comparison rather difficult. Secondly, allocations change over time and only the company and its clients know when that happens. For some robo-advisers, rebalancing is done at specific time intervals. But a vast majority state that a portfolio is rebalanced when the investor deposits, withdraws, or receives dividends. Different rebalancing thresholds also make comparison more difficult.

Additionally, many companies post backtested or model-based performance. This is likely because each client’s portfolio is “customized,” and also because not many of the robo-advisers have been around long enough to post significant longer-term performance figures.

Of the companies I analyzed in my original robo article in the AAII Journal, only a handful post performance figures on their websites: AssetBuilder, Betterment, Hedgeable, Personal Capital and TradeKing Advisors. Of these, TradeKing Advisers was the only one that stated its performance was not backtested or model-based. Alpha Architect doesn’t post its performance figures, but offers to send performance data to inquiring investors.

Robo-advising firm Covestor posts performance figures on its website; however, there are well over 50 strategies and many different portfolios could be included in a particular risk-reward ranking. To see Covestor’s performance, click here.

Hedgeable and TradeKing Advisors are the only companies that display a risk figure.

Tables 9 through 13 show performance figures that were provided by the individual companies on their websites.

In terms of comparing performance, investors can look at data from mutual funds, ETFs and popular indexes. Below are tables sourced from AAII’s Second Quarter 2016 Quarterly Mutual Fund Update and 2016 ETF Guide. Table 14 shows index performance and Tables 15 and 16 shows category average returns for funds and ETFs. Additionally, target date funds can be applicable because of their portfolio compositions and because they adjust allocation based on age. Generally speaking, the further away the target date of retirement, the more aggressive the fund’s asset allocation will be. Wealthfront states, “We believe the next best option to having your portfolio managed by Wealthfront is investing in Vanguard’s target date funds.”

Comparing robo returns to mutual fund returns can also show whether an active manager is able to outperform the robo advisers.

Conclusion

Overall, robo-advisers are a great tool for investors who do not want to deal with managing money themselves.

Most robo-advisers aren’t designed to “beat the market,” so if that’s your goal you will likely be disappointed.

The benefits of robo-advisers include automated investment, automated deposits, automated rebalancing, tax-efficient investing, tax-loss harvesting (in some cases), diversification and simplicity.

Robo-advisers’ fees tend to be much lower than a traditional financial adviser; however, you don’t get the “full” experience that you do with traditional financial advisers. It is difficult to customize the particular asset allocation that a robo-adviser recommends for you, and your brokerage accounts held outside of the robo-adviser may not be taken into account.