Tag Archives: cullen-roche

The Appropriate Portfolio Vs. The Optimal Portfolio

Perfect is the enemy of the good” – Adapted Italian Proverb We all want the perfect portfolio, the portfolio that achieves the highest amount of return for the lowest degree of risk. But one of the inconveniences of a system as dynamic as a financial market is that it’s impossible to consistently maintain the perfect portfolio. This pursuit, unfortunately, causes more damage than good since it leads to increased activity, higher fees, higher taxes and usually lower returns. I have argued in my new paper, Understanding Modern Portfolio Construction , that this pursuit of alpha is misguided and that we should seek the appropriate portfolio as opposed to the optimal portfolio. Here’s my basic thinking: There is an abundance of data supporting the fact that more active investors do not consistently generate alpha or excess return.¹ Alpha is elusive because it doesn’t exist in the aggregate and because we all generate the after tax and fee return of the aggregate financial markets. So, the diversified low fee indexer must ask themselves – if I want to be properly diversified and alpha is impossible to achieve in the aggregate, then is this a pursuit I should bother engaging in? For most people, the answer should be no. For most people, the generation of “alpha” is not a necessary financial goal. Asset allocators should be concerned with generating the appropriate return as opposed to the optimal return. This means building a portfolio that is consistent with your risk profile and managing it across time so that you maintain that profile while maintaining an appropriately low fee, tax efficient and diversified approach. The pursuit of alpha generation not only reduces returns by increasing taxes and fees, but also misaligns the way the portfolio manager perceives risk with the way the client sees risk. Since the portfolio manager is benchmarked to a passive portfolio they likely cannot outperform they will often exacerbate many of the frictions that degrade portfolio returns all the while increasing the risk that the client will not achieve their financial goals. Of course, the “optimal” portfolio might not seem so different from the “appropriate” portfolio, but I would argue that there’s a substantive difference. For instance, let’s look at an example of a 40-year-old man with $500,000 to allocate. Let’s assume he uses the simple “age in bonds” approach and comes to a 60/40 stock/bond portfolio. Every year this asset allocator should rebalance his portfolio so that he owns approximately 1% more in bonds. In all likelihood, the stock piece of the portfolio will outperform the bonds over long periods of time so he will consistently be tilting further away from stocks and into bonds. But why does he rebalance? He rebalances to maintain an appropriate risk profile, not to optimize returns and generate alpha. He is accepting the high probability of a good return and foregoing the risks associated with pursuing the perfect return. This should be the approach taken by most asset allocators seeking to build a proper savings portfolio. Countercyclical Indexing takes this process of risk profile based rebalancing a step further.¹ Since a 60/40 portfolio derives 85%+ of its risk from the equity market piece (and even more late in a market cycle) it is prudent to try to achieve some degree of risk parity across the market cycle. But we should be clear about the process of this rebalancing – we are not rebalancing to achieve alpha. We are rebalancing to better balance our exposure to asset risk across time. Said differently, we don’t implement this rebalancing to capture the best portfolio, but to capture an appropriate portfolio. In doing so, we are accepting that our portfolio might merely be “good,” but by pursuing the appropriate portfolio we are avoiding many of the pitfalls involved in pursuing the perfect portfolio. If more asset allocators abandoned the false pursuit of the optimal portfolio, I suspect they would perform better. Instead, they’ve let perfect become the enemy of the good. ¹ – See the annual SPIVA reports. ² See, What is Countercyclical Indexing ?

Robo Advisors Won’t Die As Fast As High Fee Human Advisors

Michael Kitces has a very good post up discussing some of the big trends in the Robo Advisor space. Michael notes that the robo growth is falling off fast and that this could be a sign that the trend here is beginning to dry up. He ultimately concludes that the biggest winners here are the companies augmenting human advisory services with the benefits of the robo technology: “advisor platforms are quickly seeking to build or acquire it to provide it to them, and the tech-augmented humans are increasingly pulling ahead of both their robo and human counterparts.” I think this is pretty much dead on. When I wrote my original piece on the Robo Advisors I said that the endgame here was clearly some combination of human advisors and technology.¹ After all, a robo “advisor” isn’t really an advisor at all. It can’t risk profile you correctly, it can’t know your intricate financial details, it won’t help you stick with a plan when it looks like the world is falling apart, it can’t provide the appropriate financial planning needs, etc. As I’ve stated before, these services are just Robo Allocators. And if you ask me, they aren’t doing anything all that sophisticated so the value proposition is limited from the start. Let me explain why I believe this realization is driving the slowdown in growth we’ve seen. 1 – Robos Build Fancy Looking and Expensive Versions of the Vanguard Three Fund Portfolio. One of the points I highlighted last year was the fact that the Robo portfolios all mimic some version of the Vanguard Three Fund Portfolio . That is, they take 6-20 positions and build something that almost perfectly resembles the specific allocation of the much cheaper Vanguard portfolio. The result is that you’re adding the management fees of the Robo on top of the expense ratios of the underlying funds, resulting in a high fee version of something very simple: (WealthFront Moderate Profile vs. Vanguard 3 Fund Portfolio) (Betterment Moderate Profile vs. Vanguard 3 Fund Portfolio)¹ As you can see, the performance is nearly identical. And if you can rebalance once a year and harvest some losses on your own it becomes difficult justifying a management fee for this. This is ultimately the biggest impediment to Robo Advisor growth. As more people use these services they’re realizing that what they’re getting is little more than a really simple passive portfolio that they can easily build on their own without the Robo middleman. There are some side benefits like tax loss harvesting and the automation of the rebalancing/reinvestment, but it’s not going to be worth it for anyone who has the time to look at their portfolio a few times a year (which is everyone). 2 – The Stock Market Scare Exposed the Same Old Flaws in Traditional Portfolio Theory. The Robo advisors all pride themselves on using “Nobel Prize winning” approaches to investing. But this approach has also exposed the same old flaws we saw during the financial crisis – most portfolios constructed using Modern Portfolio Theory will have highly correlated equity heavy allocations. Even more “balanced portfolios” constructed using Modern Portfolio Theory are not really very “Balanced” at all from the perspective of drawdown risk. One of the key points I highlight in my new paper on portfolio construction is the need for balanced risk in a portfolio. And when you measure risk by the academic notion of volatility, you tend to arrive at portfolios that are always equity heavy which is the case with the Robos. This creates a temporal conundrum and behavioral problem for most asset allocators – are you willing to go through potential periods of substantial unrealized losses in exchange for the potential that you will make it all back “in the long-run”? The cause of this is the fact that stock heavy portfolios are always overweight purchasing power protection (reaching for gains) at the expense of permanent loss protection (protecting against downside exposure). This is because bear markets will expose a traditionally “balanced” portfolio like a 50/50 stock/bond portfolio to excessive permanent loss risk since 85%+ of the downside comes from the stock component. That is, even a 50/50 stock/bond portfolio is not balanced at all as the majority of the negative volatility comes from the 50% stock piece. The two portfolios mentioned above are the moderate profile portfolios for two of the dominant Robo firms and these portfolios will undergo 40%+ declines in a bear market like the 2008 crisis. (Betterment Moderate Portfolio Drawdowns) These are extraordinary drawdowns for a moderate risk profile. To put this in context, a moderate Robo portfolio is designed in such a way that it will take on almost 85% of the downside risk of the S&P 500 during bear markets. That’s quite the rollercoaster ride for most people and not the level of certainty they want from their savings. I’ve referred to this as a major flaw in Modern Portfolio Theory, but I suspect that the deficient risk profiling process in the Robos is compounding the problem by placing the vast majority of their clients in portfolios that are exposed to substantial negative volatility. The cause of this is simple – they’re trying to be fiduciaries who serve the best interests of their clients when the reality is that they are asking 4-5 insufficient questions in the process of risk profiling and then placing you in a very general allocation that could be wildly incorrect because they don’t actually know their clients. The result in many cases is an overly aggressive and insufficiently customized portfolio. I suspect that these are the two primary drivers of the slowdown in Robo growth. But despite the flaws in these approaches, I have to disagree with Michael to some degree. I don’t think the recent weakness in asset flows are the death of the Robos. I suspect something bigger is happening and these firms are merely pushing the human advisory space in its logical direction – towards a much lower fee platform. After all, while I am here criticizing a portfolio that costs 0.28%-0.38% (Betterment’s and Wealthfront’s all-in costs on portfolios over $100K) I would be remiss if I didn’t also clarify that I think it’s absurd that most advisors still charge 1% for constructing something that is usually the same. So no, this isn’t the end of the line for the Robos. In fact, it’s all just the beginning of a long-term decline in human advisory fees. And the combination of these new technologies with lower human advisory fees will create a nice blend of real advisory services with low-cost investing. But we’re not there yet. Human advisory fees have a long way to fall and I suspect that the human advisory space will contract substantially more (in relative size to the Robo space) before all is said and done. ¹ – See, Should You Use an Automated Investment Service? ² – In order to perform these longer backtests I used the JPM GBI for the global bond piece that Betterment uses.

How Scared Should We Be About Future Returns?

McKinsey had a really nice piece this week on the future of financial market returns. The basic conclusion – lower your expectations and hunker down for some lean years in the financial markets. McKinsey says that equities have benefited from unusually favorable conditions in the last 30 years such as low valuations, falling inflation, falling interest rates, strong demographic growth, high productivity gains and strong corporate profits. Specifically, they say: ” Despite repeated market turbulence, real total returns for equities investors between 1985 and 2014 averaged 7.9 percent in both the United States and Western Europe. These were 140 and 300 basis points (1.4 and 3.0 percentage points), respectively, above the 100-year average. Real bond returns in the same period averaged 5.0 percent in the United States, 330 basis points above the 100-year average, and 5.9 percent in Europe, 420 basis points above the average .” That’s a nice clean view of the future relative to long-term returns. I think McKinsey is dead right – the last 30 years were unusual and something closer to the 100-year average is probably reasonable. I’ve stated in the past that the math here isn’t terribly controversial (or shouldn’t be). If a 50/50 stock/bond portfolio has generated 30-year average returns of 9.5%, then we should expect the future returns to be lower or more volatile. In other words, you can, with near certainty, expect that the high risk adjusted returns of the last 30 years are gone. Why is this a certainty? Well, it’s a simple function of the current interest rate environment. Because the post-1980 era involved a huge bond bull market, the risk adjusted returns of a balanced portfolio were unusually high. For instance, from 1985-2015 a 50/50 stock/bond portfolio posted returns of about 9.5% with a Sharpe ratio of 0.7 and a Sortino ratio of 1.5. That’s because the bond piece, which is inherently more stable, generated average annual returns of 7% with a Sharpe ratio of 0.76 and an eye popping Sortino ratio of 2.12, while the stock piece generated annual returns of 12.5% with a Sharpe ratio of 0.5 and a Sortino of just 0.92. In other words, bond investors have done extraordinarily well over the last 30 years thanks to the favorable tailwind of falling inflation and falling interest rates. And those outsized bond returns had a hugely positive impact on diversified investors. We also know that the best predictor of future bond returns is current yields so, do the math on the 1985 starting overnight interest rate of 7.5% versus today’s rates of 0%. A bond aggregate held for the next 10 years is unlikely to outpace the current yield of 2.25% by much. So, we know for a fact that the bond piece won’t generate anything close to the types of returns it did in the last 30 years. But there’s also good historical precedent here. In the 1940s, rates were as low as they are today. So, how did the bond market do? It did okay, but it certainly wasn’t anything like the post-1980 period. From 1940-1980, bonds posted annual returns of 2.75%, but were very stable (much more stable than is commonly believed in a rising interest rate environment). The stock piece, however, performed very similarly to the post-1980 period, with rates of returns from 1940-1980 at 12.4% vs. 12.5% for the 1985-2015 period. As a result of this, a balanced portfolio from 1940-1980 generated an average 8% return with a Sharpe ratio of 0.58, significantly lower than the average 10% return with Sharpe of 0.7 that we experienced in the last 30 years. In other words, in the only reasonable historical precedent a balanced portfolio generated lower nominal and risk adjusted returns than the post-1985 period. Now, I think backtests and historical references are a bit dangerous and overused by the financial community, but I also don’t think we need these historical precedents to establish a reasonable probability of future returns. All we need is a little common sense when comparing the next 30 years to the last 30 years. After all, we have empirical proof that most of those tailwinds are in fact waning. For instance: Current interest rates are the best predictor of future returns in the bond market, and this period is certain to be a low return period for future bond holders. Valuations, which have a strong tendency to correlate with future equity returns, are high historically. Demographic trends have shifted substantially in the last few decades from a world of higher growth to a much more modest pace of growth. High productivity gains have waned and have now become an area of great concern for economists. Corporate profits, as a share of national income, have never been higher as they rode the back of the liberalization of tax rates and regulation and could come under pressure given the anti-corporate climate we are entering. I don’t think any of this should be terribly controversial, and you don’t have to be an expert forecaster to see what’s coming. At the same time, we shouldn’t panic as some people have implied . If the aggregate stock and bond markets generate anything close to that 8% return of the 1940-1980 period, then most investors will still generate positive real returns. However, there are a few key takeaways here: It is crucial to understand the most important principles of portfolio construction so you can grow comfortable with a process and a plan. See Understanding Modern Portfolio Construction . It’s time to temper expectations in the markets. The future is likely to be an era of lower returns and potentially bumpier returns; however, it doesn’t mean returns are going to be catastrophic. It’s time to hunker down on your taxes and fees in your portfolio. As a % of assets, these frictions will become increasingly important in a lower return environment. See, Understanding your Real, Real Returns . Be patient! Find a good plan and learn to stick with it. The lower and bumpier returns will create periods of frustration for most investors. The grass will always look greener somewhere else. Switching in and out of plans and chasing the next hot guru will very likely result in higher taxes and fees, leading to lower average returns. See, How To Avoid the Problem of Short-Termism . Invest in yourself, continue to save and pour that savings into your portfolio. You might not get world beating returns from your portfolio in the coming 30 years, but we know cash will be the riskiest asset in the future as it will guarantee a negative real return in such a low interest rate environment. See, Saving is not the Key to Financial Success . Be careful reaching for yield. All safe assets aren’t created equal and reaching for yield in the wrong places could create more volatility without the guarantee of stable income. See, Reaching for Yield or Reaching for Risk? Don’t let the scaremongers get to you. If the future is one of lower returns and bumpier returns, there will be lines of people trying to sell you something in exchange for your fear. These people should not be trusted. The world of the future might not be the gangbusters growth period of the 80s and 90s, but it also won’t be the end of times either.