We are continuing to examine a few foundational sub-topics before getting into the meat of the active versus passive debate for fixed-income. Last week, we looked at the composition and tax implications of fixed-income returns. Today we will be looking at liquidity in bond markets.

I find that the layperson is familiar with the stock market and understands that equities are traded electronically on centralized exchanges. Bonds are quite different however. Consider a few of the below points:

• Many bonds are traded over the phone, rather than electronically. There is a big push towards electronic trading and a majority of US Treasuries may trade electronically now, but most corporate, mortgage, and muni issues still trade via phone.
• There is no centralized exchange (electronic or otherwise), but a network of brokers and dealers who request quotes from one another.
• Very few bonds trade on a given day. While nearly all stocks trade every day and many times over, very few bonds trade every day. Many US Treasury and Agency bonds trade every day, but a large portion of investment grade corporate bonds do not. A minority of high yield bonds trade on any given day and just a fraction of muni bonds trade on most days. For instance, both my clients and I own many bonds that have not traded in months.

Each of the above points contribute to wide bid-ask spreads. That is, the difference between what buyers are willing to pay and sellers are willing to accept is wide. When bid-ask spreads are tight and liquidity is deep, it does not much matter whether you’re a buyer or a seller because either will get close to the same price. But if bid-ask spreads are wide, then buyers and sellers will get quoted very different prices. Buyers will be quoted higher prices and sellers will be quoted lower prices. This type of market can be either good or bad, depending on the investor.

Investors who are forced to buy or sell are at a disadvantage. If an investor has some discretion over trading, he/she can exploit this illiquidity by selling when people want to buy and buying when they want to sell. The more illiquid the asset class, the more you want to be in control of buying and selling decisions. This dynamic certainly impacts how allocations evaluate manager mandates, fund structure, product wrappers, as well as management style (active vs passive) which we’ll get into more in coming weeks).

Before diving into an analysis of active versus passive management in fixed-income, it may be helpful to cover a few foundational topics. First up is taxation.

Tax-efficiency is major tailwind for passive management in equities, but not so much in bonds.

First, let’s look at some reasons why equity index funds are tax-efficient:

• Equity returns are composed of both appreciation and dividends. Under the current US tax code, short-term capital gains are taxed at ordinary income rates and long-term capital gains are taxed at 15-20%. Furthermore, “qualified dividends” also receive favorable tax treatment. Thus, there are a lot of tax benefits to reducing turnover in an equity portfolio.
• The ETF structure allows for in-kind redemptions (and creations, but that’s beside the point) and many equity index fund managers have been able to minimize or avoid any capital gains exposure since their inception.

Now let’s examine how these factors contrast with fixed-income:

• The majority of a bond’s return is from interest. Bond prices do fluctuate and investors can capture gains and experience losses, but a bond will only return a fixed-amount over its life. Hence the term fixed-income. Under the current US tax code, interest income is taxed at ordinary income rates. Even if an investor captures some appreciation, he/she will likely need to attribute a portion of the gain as income. Thus, there is not as much of a tax incentive to hold on to bonds for longer than need be, as investors will largely be taxed at ordinary income rates regardless of holding period.
• The bond universe is much larger than the stock universe, but also less liquid. Thus, bond ETFs utilize in-kind redemptions less than their equity fund counterparts, which means that there is often capital gains exposure.

Obviously the tax laws can change, but passively-managed fixed-income does not currently enjoy the same tax benefits as passively-managed equities. This does not mean that active management is necessarily better in fixed-income (that depends on many other factors which we’ll explore in coming weeks), but the hurdle is not quite as high.

Just came across this chart from Russell Investments, which corroborates much of the data from last week’s post. Thus, the conclusion remains the same: it is very, very difficult to beat the indices in domestic large-cap equities.

This post will look at US large caps through the lens of the active vs passive debate. As one the largest and best known asset classes, it is often the battleground of the debate.

There is no shortage of analysis and reporting purporting to shed light on the active vs passive debate, although most of it is noise. SPIVA publishes data quarterly, although knowing what percent of managers out/underperformed each quarter or year is meaningless, as is knowing the persistence of top ranked funds (see here). So, we’ll skip all the noise and red herrings and skip right to a good starting point: rolling returns over multiple years.

The below shows that active managers really underperformed in 1990s, then enjoyed a period of outperformance during the recession of the early 2000s, but have not added much value since.

The picture does not change much when we go from 3-year rolling returns to 5-year rolling returns. The median manager added modest value for a brief time in the mid-2000s and again around 2010.

Vanguard put out this beautiful chart of 10-year rolling returns that includes data on all percentiles of active managers (instead of just the mean or median manager). On average, most active managers underperform over most 10-year time periods. Again, this is before taxes, so the real after-tax numbers are even worse for active managers.

There is some evidence that active managers perform better in severe down markets, which you can see in the above charts as well as in the scatter plot below. It’s is debatable whether this is simply due to lower equity allocations or whether active managers are successful risk managers. Distinguishing between those two factors is really just splitting hairs, so I’ll move on and wrap up.

Conclusions:

The data largely shows what the passive camp already knows: large cap indices are hard to beat. Over 3-, 5-, and 10-year periods, most active managers underperform most of the time.

However, the active camp will rightfully point out that there have been some time periods when more than half of managers outperform. This is true, although it is a minority of the time and does not include tax drag.

Of course, these numbers are in aggregate. Some managers may pitch themselves with a superior 10-year return and my hope is that readers will respond by saying, “Show me a history of your rolling returns.”

My final conclusion is that outperforming a US large-cap index is very, very difficult to do. It is even more difficult in taxable accounts. I don’t think its impossible and I own an actively-managed US large cap fund at the moment, but investors need a very compelling case because the odds favor passive management.

My general view (explained here) is that investors should use passive strategies in asset classes where it is difficult to beat the averages and active strategies where it is probable that the averages can be beat. Unfortunately, the active vs passive debate in many asset classes hinges on biases, dogma, and platitudes. Anecdotes and misused statistics abound, while rigorous questioning and examination are much harder to find.

Below are some questions that are commonly asked when evaluating whether active management can succeed in a particular asset class. Each is problematic in some way and also used (and the answers overplayed) by both the passive and active camps. Hopefully, the below is helpful in cutting through the BS and developing a framework for evaluating management style.

1. How many managers beat their benchmark in any given year?

The problem with the above question is that it is impossible to separate luck from skill. The number is likely to be higher in markets with higher volatility or more randomness. Therefore, we might ask:

2. How many of this year’s outperforming managers beat the market in consecutive years?

On the surface, this is a good question. However, it completely ignores the magnitude of outperformance or underperformance. Is it better for someone to gain 5% three years in a row or to get two years of 12% and one year of -2%?

3. We can look at active manager alpha over rolling 3-, 5-, 10-year periods, which should account for magnitude of gains of losses and focus on total return.

Yet, like any of the previous metrics, rolling returns vary over time. A particular manager or strategy might do well in one 10-year period and terrible in another. However, the data may show that it is consistently easy or difficult to outperform in a given asset class consistently through various regimes.

This is also a good place to pause and point out that questions #1-3 may lead to an aggregated answer, but can miss pockets of persistent outperformance. For instance, it is possible for subset of managers to consistently outperform in markets even if most underperform. That’s why #4 is important.

4. Shifting from asset classes to individual managers, it’s best to dig into the portfolio’s historical data and understand what drove performance in past years. This process is called “return attribution.”

Reviewing past performance and attribution is helpful, but certainly not prescriptive. Evaluating portfolio construction, strategy, historical returns and attribution provides a baseline understanding, but portfolios evolve over time and the future will inevitably unfold differently than the past. However, investors can understand a manager’s process and judge whether the returns were due to skill or luck and, ultimately, whether past success is repeatable or not.

The above is just a brief summary of some commonly asked questions and why they are problematic. As I mentioned, hopefully it is helpful in calling someone’s BS. We’ll dig deeper into specific asset classes over the next few weeks.

This week, we have a reader question from Jasmin:

How is the stock market index of any significance for individuals?

The significance of a stock market index (or any other index for that matter) is that it is a benchmark. In other words, investors can use indices as a yardstick to measure and/or compare slices of their portfolio to.

For most investors, indices can be used to evaluate the performance of the individual asset classes in their portfolio. Investors can compare the risk and return metrics of an allocation to those of an appropriate index. When utilizing an actively managed strategy, an appropriate index or index fund can be viewed as the opportunity cost of using the active strategy.

If you own US large-cap stocks, you might compare the volatility and performance of those stocks to the S&P 500. If you own international stocks, you might compare them to the MSCI EAFE. You can compare the bond portion of your portfolio to the Barclays Agg or a number of other indices. There are multiple indices for various sectors, regions, countries, and so on. In fact, today there are more indices than stocks!

Although the above is relatively straightforward, many investors misuse indices. Below are two of the more common errors that I see:

• Comparing a multi-asset class portfolio to a single asset class index. I meet many people who compare their portfolio of global stocks and fixed-income to the S&P 500. The S&P 500 could probably be used to benchmark the US large-cap equity portion of the portfolio, but certainly not the entire portfolio.
• Viewing an index as a goal. Each individual investor has unique goals and their portfolio should reflect this. Even the largest indices have some fairly substantial risks, which investors should be aware of and may want to avoid. For instance, the “China Region” (China, Hong Kong, and Taiwan) makes up 40+% of most broad-based emerging markets equity indices or the Barclays Agg is overwhelmingly government bonds, which offer much more duration than yield today. These are just two examples of where investor goals may be inconsistent with an index’s composition. Note: Money managers may benchmark to a single index, but this is because they typically run single asset class strategies and are concerned with relative performance, neither of which is typically true of individual investors.

So, the short answer is that individuals can use indices to evaluate single asset class strategies/managers and should not be used for much more than that.

If nearly all investors engage in active management through the development or selection of a portfolio’s asset allocation, then what is the active versus passive management debate all about?

The debate is whether individual asset classes should be accessed via active managers or via passive strategies. Thus, the debate over active versus passive management is at the asset class level, rather than portfolio level (because all investors are active at the portfolio level!).

Below is a brief summary of active and passive management:

Active:
The traditional form of investing, where an investor picks individual securities that he or she expects will result in the best risk or return metrics.

Passive:
At some point (I think in the 1950s) investors began to both deduce and notice that active managers (in the aggregate) could not outperform the average returns in various markets. If managers could not beat the averages, then investors would be better off accepting average returns and minimizing costs. Average returns could be nearly achieved by simply buying the entire market or a representative sampling of them.

Below is an example of where an investor must decide whether to utilize an active or passive strategy:

A hypothetical investor decides to allocate 10% of her portfolio to emerging markets bonds. The investor can do one of two things with the 10%:

1. Purchase individual emerging markets bonds (or hire a manger that selects the “best” bonds, based on research, analysis, etc).
2. Invest in a passive strategy that attempts to mirror the risk and return characteristics of the broad emerging market bond market.

If the investor is comfortable with the risk and returns of the broad emerging market bond market, she should consider a passive strategy. If not or she thinks that she can beat this market, active management may be a better choice.

It should be clear that neither active management nor passive management is inherently better. The decision to use one over the other should depend on both investor objectives, as well as the asset classes’ characteristics and market structure. Yet, the debate does carry on, which I believe is driven both by bias from investing product sponsors and by dogma from under-informed investors.

Looking ahead, we will examine various asset classes and how active and passive strategies fare in each.

Thus far in this series, we’ve looked at market efficiency and management styles in a general sense. Yet, investors must examine these factors and make decisions at multiple points, including when they select an asset allocation.

A completely passive asset allocation would mirror the market portfolio, which may look something like the below. Hypothetically, an investor could construct such a portfolio.

However, some investors will want a more aggressive portfolio while others will desire a less volatile portfolio. Consequently, they will adjust their allocation weights to produce different portfolios along the efficient frontier. 

Yet, the efficient frontier is not a constant thing. Does it represent a 3-, 5-, 7-, 10-year, or some other time horizon? The efficient frontier will look different for each horizon, as seen below:

Once settled on a time horizon, investors face a new decision: should historical data be used or future assumptions? If historical data is assumed, which timeframe should be used? 100 years of data? Or just the most recent 10 years? Or a decade that most resembled our current situation? Each will provide a different frontier.

If one chooses to use future assumptions, what assumptions should be used? Rarely do people agree on what the future will look like and nobody can consistently predict it accurately. Below is an example of possible efficient frontiers, based on differing assumptions:

It’s not just average investors that cannot agree on what the future looks like. There is debate amongst the largest asset managers who have vast resources, huge budgets, and large teams of economists, analysts, and researchers. Below are the 2017 capital markets assumptions from three of the largest asset managers:

• Vanguard
• Blackrock
• JP Morgan

Regardless of whether an investor allocates to actively managed funds or passive index funds, selecting an asset allocation is an exercise in active management. In fact, allocation decisions will likely be more determinative to a portfolio’s risk and return than whether active or passive managers are used within each asset class.

Before delving deeper into the topic of market efficiency and active vs passive, it may be helpful to briefly review the definitions of beta and alpha.

Beta

At it’s simplest, beta can be defined as the volatility of an asset relative to a benchmark. The benchmark is often an index of a specific sector or an entire asset class. For example, a US large-cap stock that is 20% more volatile than the S&P 500 index has a beta of 1.2. Or if the volatility is 20% less than the index , it’s beta is .8. Beta is a measure of relative volatility.

Alpha

Within the Capital Asset Pricing Model (CAPM), this same beta is used to forecast returns. Under CAPM, higher beta leads to higher returns and lower beta results in lower returns. In other words, a fund with more volatility should have higher returns and a fund with lower volatility should have lower returns. Reality rarely unfolds as modeled, so an ex-post CAPM equation was created. It adds a variable called alpha to account for the difference between an asset’s actual return and it’s expected return (which CAPM predicts to be the benchmark return, adjusted for volatility). Below are a couple of examples on calculating alpha:

• An investor invests in large-cap domestic stocks and reports a 10% return, while the Russell 1000 returns 7%. The investor has produced 3% alpha.
• An investor invests in global equities and reports a 9% return, while the MSCI All Country World Index returns 13%. The investor has (negative) alpha of -4%.

Beta, Colloquially

The term beta is often used to refer to the risk and return characteristics of a benchmark. Some examples: US large-cap beta may refer to the Dow Jones Industrial Average or the S&P 500, investment grade bond beta is often synonymous with the Barclays Agg index, or tech beta may simply refer to a tech-sector ETF. I often refer to index ETFs as cheap, liquid beta. Some may compare large-cap beta to small-cap beta or consumer staple beta to consumer discretionary beta, when discussing risk and returns. Below are some examples of how someone might refer to beta:

• Utility sector beta is sensitive to interest rates.
• An index fund may be called beta exposure.
• An investor holds investment grade bonds and reports a -1% return while the Barclays Agg is down 3%. Some might say that the beta return was -1% (and alpha was positive 2%).

To recap:

• Beta is an asset’s volatility, relative to a benchmark.
• Alpha is an asset’s return above it’s benchmark, after controlling for differences in volatility.
• Technical definitions notwithstanding, the term beta is often shorthand for a benchmark or may refer to the benchmark’s risk and return characteristics.

Does it matter whether markets are efficient or not?

Not really and here’s why:

• As I mentioned last week, nobody knows how efficient or inefficient markets are (due to the joint hypothesis problem, eloquently summarized here)
• Asking whether a market is efficient is often a proxy for asking whether one can outperform. However, the questions are different, as we’ll see below.

Last week’s post mentioned Eugene Fama, who won a Nobel prize for developing the Efficient Markets Hypothesis (EMH). However, Jack Bogle (who founded Vanguard) also has a hypothesis called the Costs Matter Hypothesis (CMH), which he introduced here in 2003. Consider the following, which both Fama and Bogle agree on:

• In any market, aggregate investor performance will equal the overall market performance minus costs. This statement holds whether a market is efficient or inefficient. Regardless of a market’s efficiency, investors as a whole will underperform because they bear costs. To outperform the market, you must beat the market by more than your costs.
• Since total investor performance is limited to market performance minus costs, any investor’s outperformance comes from another investor’s underperformance. So to beat the market, you must also beat other participants.

Thus, the important question for investors is not whether markets are efficient or not, but whether they can reliably and consistently beat both the market and other investors net-of-fees. If the answer is yes, investors might benefit from using active management. If the answer is no, it is likely better to use a lower-cost vehicle, such as an index fund.

Of course, there is more than one market. There are many asset classes and markets, so the question of whether you can beat the market or not must be asked over and over. Each market is unique and changes throughout different environments, so the question needs to be continually asked: is it possible for me to beat this market net-of-costs?