Responsible/ESG investing

Let me preface this post by saying that I have nothing against social responsibility or having social responsibility reflected in an investment portfolio.

I read an article in Investment Executive recently discussing the evidence that Responsible Investing (RI) leads to as-good or better returns compared to traditional investing. The article compares the performance of well-known indexes to demonstrate the long-term benefits of RI. In the article, long-term seems to be defined as the trailing 10-years. In my opinion, and the opinion of most researchers, a 10-year period is insufficient to draw any conclusions.

It is true, as the article states, that the MSCI ACWI ESG Leaders Index has beaten the MSCI ACWI index since 2007, when the indexes became available. The performance difference has been 0.22% per year on average. However, this difference is not due to the responsible nature of the companies in the index.

The difference in returns between diversified portfolios is almost completely explained by exposure to certain types of securities. The characteristics that define those certain types of securities are called factors. Factors explain the differences in returns between diversified portfolios extremely well. A factor is a long-short portfolio where the portfolio is long one side of a characteristic and short the other. For example, the size factor is determined by subtracting the returns of large stocks from the returns of small stocks. If small beats large, the factor exhibits a performance premium.

Currently there is a five-factor model that has been published by Eugene Fama and Ken French, the men who pioneered factor research in the 1990s. The factors in the model are market (market minus the risk-free asset), size (small stocks minus large stocks), relative price (value stocks minus growth stocks), profitability (stocks with robust profitability minus stocks with weak profitability), and investment (stocks that invest conservatively minus stocks that invest aggressively).

We can use statistical analysis to see how much a portfolio’s return is explained by these known factors. It is probable that an ESG (environmental, social, governance) index is offering naïve factor exposure. In other words, by targeting ESG companies the index is likely getting exposure, by accident, to known factors.

To put to rest the idea that an ESG index is inherently better I have run five-factor regressions on an ESG index compared to a total market index using Ken French’s data. What we should expect is that exposure to the factors will explain the return differences between the ESG index and the total market index.

In the following table, the coefficients tell us how much loading the portfolio (index) has to each factor, and the t stat helps us to understand if the coefficient is statistically significant. A t stat over 2 indicates statistical significance.

ESG.png

In this case we see that most of the factor exposure is statistically insignificant and small, which we would expect for a total market index, except for RMW, which is a little bit larger and statistically significant. RMW is the profitability factor. This is indicating that the ESG index has more exposure to stocks with robust profitability than the market.

The research shows that more profitable stocks tend to outperform less profitable stocks over the long-term. The RMW premium over the time period in question (12/1/2007 – 7/31/2018) was 3.34%. That is, more profitable stocks beat less profitable stocks by 3.34% per year on average over the period.

We can estimate the amount of additional return that we would expect from RMW exposure for each index by multiplying the regression coefficient by the premium. Multiplying the difference in coefficients (0.23 for ESG minus 0.11 for total market) by the factor premium we get an expected performance difference of 0.40% per year on average.

Put simply, when we adjust for factor exposure, the ESG index has actually done a little worse than we would expect it to compared to the total market index.

The article that I am responding to also offers that the Jantzi Social Index, an index of socially responsible Canadian stocks, has beaten the S&P/TSX 60 since inception of the index in 2000. That is true.

Ken French does not publish factor data for Canadian stocks. AQR does have Canadian factor data, but they look at slightly different factors. They do not look at Profitability (RMW) or Investment (CMA). They add in QMJ (quality minus junk) which does include profitability among other characteristics. We will look at a four-factor regression including market, size, relative price, and quality. I was only able to obtain total return data for the Jantzi Social index back to May 2009, but the regression results should still be informative.

Jantzi.png

In this case we see similar negative loading to SMB for both indexes, which we would expect as these are both large cap indexes. We see slight loading to value, statistically significant for the Jantzi and insignificant for the TSX 60. The big statistically significant difference comes from QMJ, which for the sake of discussion is very similar to RMW. A relatively large and statistically significant loading to highly profitable stocks would easily explain the higher returns of the Jantzi index.

So far, we have seen that the higher returns of an ESG index can be attributed to factor loading as opposed to the inherently better returns of ESG companies.

The article cites an academic study showing that ESG mutual funds beat their benchmark 63% of the time. That study looked at a handful of Canadian mutual funds included in the Responsible Investment Association listings. The study, as far as I can tell, does not address survivorship. ESG investing aside, this study conflicts with the massive body of evidence that active mutual funds typically fail to beat their benchmark index.

I won’t recreate the study with a correction for survivorship for this post, but consider that over the trailing 10-year period about 50% of Canadian mutual funds have closed down, likely due to poor performance. Basing a study on funds currently in existence ignores all of the funds that have closed. The surviving funds are likely to have done better by nature of having survived, but that may well have been due to luck. I suspect that a survivorship correction would drastically change the results of this study.

The article also cites a report from MSCI which found that high ESG-rated companies tend to be more profitable with higher dividend yields and lower idiosyncratic tail risks. We have proven this to be true, at least the profitability part, with regression analysis. The problem for investors is that if you want to maximize risk adjusted returns through exposure to factors, including profitability, then ESG investing is a very inefficient and potentially inconsistent way of getting it.

Not to mention that the fees on ESG funds tend to be high. In the case of index funds, an ESG index fund will typically carry a higher fee than a total market fund. XEN, the Jantzi Social Index ETF has an MER of 0.55% while VCN, the Vanguard FTSE Canada All Cap Index ETF has an MER of 0.06%. Most ESG funds are actively managed, and carry fees well over 2%. The notion that an ESG screen will magically allow actively managed funds with high fees to beat their benchmark is not sensible.

Finally, investors need to understand that buying the securities of a company on the secondary market does very little to impact that company’s future. It is perfectly reasonable to feel guilty profiting from a company that engages in business that you do not agree with, but it is important to understand that owning shares in a company does not benefit the company.

As I mentioned at the beginning, I have no problem at all with ESG investing as long as it is done for the right reasons. The reason to be an ESG investor should be that it makes you feel happy. There should be no expectation that an ESG investment will outperform a non-ESG investment with similar exposure to the factors that explain returns.

As opposed to buying ESG investment products it might make sense to optimize a portfolio for low costs, diversification, and factor exposure – all things that an ESG portfolio typically gives up – and donate money directly to causes that are important to you.

Money you can afford to lose

If you go to any online community related to money you will find people talking about their desire to learn about investing by researching and selecting individual stocks. They understand that this is risky, they will say, but they will only try it out with money that they can afford to lose.

This statement implies a lot more than people realize when they write it. What puts you in a position to have money that you can afford to lose? At the very least, you need be financially independent. That is, have enough financial capital that you do not need to rely on your human capital to meet your desired lifestyle.

Let me give you some context to understand that: to spend $5,000 per month, adjusted for inflation, for 30 years, you would need to have about $1.7M in an investment portfolio. Most people, especially people wanting to get their feet wet with stocks for the first time, are not in this position. Until you have attained financial independence you cannot afford to lose any money.

Not only are most people not financially independent, but they are probably not saving enough to eventually become financially independent with the lifestyle that they want. They are also likely to have debt, a lack of emergency savings, and a lack of life insurance coverage. Where is this money that you can afford to lose coming from?

It’s not just losing money that you have to worry about either. You might pick a stock that drops a little bit and then sputters along for a while before you sell it. Great learning experience, right? Even if you do not sell it at a loss, you may have missed an opportunity if the market has gained over the holding period. This opportunity cost is as good as a loss in terms of achieving your financial goals.

It would be one thing if you had a reasonable chance at picking winning stocks, but you most certainly do not. For example, if you had picked one of the S&P 500 constituents at random in January 2017, you had a 53% chance of underperforming the index as a whole by the end of December. For anyone keeping score, that is worse than flipping a coin.

I know that the premise of this conversation is based on the idea that you are able to learn about investing by picking stocks, and maybe you can even get better at picking them as you learn. Unfortunately this is not how the stock market works. You may be able to learn a little bit about how to open a brokerage account, how stocks trade, and even how to analyze companies. Those things might be empowering to know, but there is no reason to believe that they will make you better at investing.

The stock market is a pricing machine. For the most part, it instantly prices stocks based on the aggregate information of market participants. No matter how much you may know about a company’s earnings, growth prospects, or business model, you have to know more than everyone else in order to have a performance edge. Put another way, it is not your absolute information and skill that matters in picking stocks, it is your relative information and skill.

If you do not have better information and more skill than other market participants, then they may be taking advantage of you when you trade. When you buy a stock, you are buying it from somebody. When you sell, you are selling it to somebody. Who are these people on the other side of trades? They are mostly large institutions whose core business is buying and selling stocks. I think it is a bit of a stretch for any individual person, especially a person trying to learn how stocks work, to believe that they have an edge in terms of information or skill.

The odds are massively stacked against you. But don’t worry, it’s not just you that the odds are stacked against. Even those institutions have trouble beating the index. For the 10-year period ending December 2017, only 1.67% of mutual funds in Canada that invest in US stocks were able to beat the S&P 500 index.

If you get to a point in your life where you are financially independent and you still want to trade stocks, the next question that I would be asking is what causes are important to you? Is it more fulfilling to make risky bets with a negative expected outcome by picking stocks, or to donate all of that extra money that you can afford to lose to support something that is important to you?

Stock picking is well documented as a losing game. It is, by all accounts, akin to gambling. Experience trading stocks does not make you a better investor. I do not see how the average Canadian has any money to lose in an effort to learn about something that will not benefit them in any way.

I’ll leave you with a quote from the 1967 classic book The Money Game: “If you don’t know who you are, [the stock market] is an expensive place to find out.”

Figuring Out Factor Investing

Everyone likes to believe they’re smart consumers. That’s probably why the term “smart beta,” also known as “factor investing,” is so hot right now. “Stupid beta” probably wouldn’t attract many takers. 

So what is factor investing, and are you smart to use it? That’s what today’s post and related video are all about

A simple way to think about factors is as quantitative characteristics shared across a set of securities. If you look past their current, flashy popularity, there really is some substance there. For decades, evidence-based investors have been structuring their investment portfolios to tilt toward factors that are expected to drive investment returns, without the need to rely on random stock-picking or market-timing.

In this respect, factors are pretty smart. Unfortunately, lately, I’ve seen them being used in some pretty dumb ways, to market products that may be factor-based in name, but no longer in evidence-based substance.

Factors: Bringing Order to Evidence-Based Investing

Why does one investment portfolio do better than another? Unraveling this mystery, of course, is at the heart of what evidence-based investing is all about. Before we understood factors, most of the performance difference was usually attributed to the skill of the portfolio manager. In the absence of understanding, it was an easy, if erroneous assumption to make.

As factor research emerged, it became clear that the bigger determinant of different outcomes among different diversified portfolios was, by far, not the prowess of the manager (their “alpha”) but rather certain characteristics, or factors (the portfolio’s “beta”). No wonder we started caring about what those factors were, and which ones in what combinations were expected to deliver the most powerful, positive performance differences. Isolating and incorporating positive return differences exhibited by certain types of stocks has an obvious benefit to investors.

Fast forward to today. Currently, we have factor models that can explain over 95% of the return differences among diversified portfolios. It doesn’t take a guru to incorporate these models; you just need to seek the most efficient, cost-effective funds for doing so. 

This is problematic for active fund managers who are still chasing after “alpha.” In the past, their ability to “beat the market” was assumed to be due to their skill. We now know it’s far more likely to be a result of the factor exposures they’ve chosen, whether deliberately or as a random byproduct of their other activities – no prognostication skills required or desired. 

In other words, if you can manage your expected returns and related risks through one or a few simple factor-based funds, who needs an active manager or the additional costs they’re expected to incur? 

If you’re not quite convinced, check out my related video. [hyperlink] In it, I describe a classic 2015 blog post, “Active Funds Exposed,” in which my PWL colleague Justin Bender runs some real-life numbers for us.  

Figuring Out the Factors

Research on factors emerged in a landmark 1992 Journal of Finance paper by Nobel Laureate Eugene Fama and Kenneth French, entitled The Cross-Section of Expected Stock Returns. In the paper, they observed that, over time and in aggregate, small-company stocks outperformed large-company stocks, and value stocks outperformed growth stocks. The explanation for the return differences is that stocks with these characteristics, (small and value), were riskier – more volatile. Investors required higher expected returns before they were willing to take on these riskier assets,

In 1997, Mark Carhart added the momentum factor to the body of research; in 2012 Robert Novy-Marx added the profitability factor. This gave us five factors, which come together to explain over 95% of the return differences among diversified portfolios, as touched on above. In 2014, Fama and French came out with their own five-factor model. Their model combined market (i.e., investing in anystocks, versus risk-free assets), size, relative price (value), profitability, and investment. They left out momentum. 

Which factors form THE ideal model that explains 100% of the return differences among diversified portfolios? This is unknown. Frankly, it’s unlikely we’ll ever arrive at a universal answer. But researchers continue to test new factor models aimed at inching us ever closer to the elusive nirvana of a perfect model.

And therein lies the challenge. 

Faux Factor Investing 

Factor research has become not only important to our understanding of finance and investing, but a way for academic researchers to make a name for themselves … and for fund companies hungry for a fresh marketing hook to differentiate themselves by injecting “new & improved” factors into the mix. 

Duke University’s Campbell Harvey, Texas A&M’s Yan Liu, and University of Oklahoma’s Heqing Zhu have identified over 300 factors in academic literature. This is problematic for investors. Cost-effectively targeting five factors in a portfolio is hard enough. What do you do if there are 300 of them? Unfortunately for the researchers (and fortunately for investors), many of these factors do not pan out. In many cases they turn out to be a re-packaging of the original factors.

In my video, I cover the sniff tests you can use to help decide when a new factor is really all that new or worth pursuing … and, conversely, when common sense tells us it’s safe to ignore it. To be taken seriously, I would suggest a factor should be persistent, pervasive, robust to alternative specifications, investable, and (my personal favorite) sensible. Check out my video to dig into each of these characteristics in more detail. 

Now, about those fund families and their marketing programs. As we know from our experience with other retail products, “new and improved” is usually just the same old you-know-what, stuffed into a fancy new box. Same thing with factor or smart-beta investments. With some 300 “flavors” to choose from, countless new factor products have emerged, but very few of the companies creating them have impressed me. 

Maybe that’s because, these days, factor research has become a commodity that any fund manager can access. The difference between implementing the evidence well or poorly comes down to how well the company vets the research, who does the vetting, how they interpret the data, and how effectively they manage the inherent limitations of factor models. 

So far, Dimensional Fund Advisors – the company that introduced factor investing (with Fama and French on their board) – continues to stand apart in the field. That said, Dimensional’s funds can only be accessed through vetted advisor firms like PWL Capital. Where does that leave the DIY investor? For now, I think you’re better off focusing on simplicity rather than trying to sort out all these factors on your own. 

The Canadian Couch Potato model portfolios used to pursue the size and value factors, but my colleague Dan Bortolotti changed the models in 2015 to ignore factors entirely. Part of his explanation was that “many DIYers make costly mistakes when they try to juggle too many funds. Meanwhile, there are exactly zero investors in the universe who failed to meet their financial goals because they did not hold global REITs or small-cap value stocks.” I agree with him in full. 

Have you tried to implement a factor portfolio? Tell me how it went in the video’s comments.

Original post at pwlcapital.com.

When GICs beat bonds

The fixed income component of the Dimensional Fund Advisors global portfolios has been low to-date in 2018, and also for the past 12 months. It is down 0.51% year-to-date at July 2018, and down 0.27% for the 12 months ending July 2018.

Fixed income is typically the safe portion of a portfolio of risky assets. Fixed income is not held for the primary objective of maximizing returns, but for minimizing volatility through diversification. Fixed income assets tend to have imperfect correlation with stocks. This offers a buoy effect if stocks are falling. The opposite is also true; when stocks are doing well bonds do not tend to do as well. Similar to stocks, there are many sub asset classes within fixed income. These sub asset classes have different risk and return characteristics.

Fixed income factors

Some characteristics of fixed income have been identified as responsible for most of the asset class returns. The two factors that explain the majority of fixed income returns are term and credit; longer-term bonds and bonds with lower credit tend to have higher long-term returns. This is intuitive as each of these characteristics, term and credit, result in an asset that is riskier to hold. As always, risk and return are related.

Assuming the CDIC limits are followed in portfolio construction, a GIC is close to risk-free. In other words, it has no exposure to the credit factor. GIC terms longer than 5-years are not covered by CDIC, so an investor would typically limit themselves to 5-year or shorter GICs. This limits exposure to the term factor. 

With this in mind we would expect two things: 

  1. GICs will outperform bonds in some years
  2. Bonds will have higher long-term returns than GICs

Similar to owning stocks in lieu of bonds, owning bonds in lieu of GICs will inevitably result in years of underperformance. This is one of the risks that investors endure to access the higher expected returns of an asset class. If bonds were guaranteed to outperform GICs at all times then bond prices would increase, decreasing their yields, and we would not even be having this conversation.

Historical data

Looking at annual returns going back to 1985 through 2017 we observe a few things that illustrate my previous comments. 5-year GICs have outperformed bonds in 5 of the 33 calendar years examined, or 15.15% of the time. Interestingly, 5-year GICs have outperformed stocks in 9 of the 33 years, or 27.27% of the time. In a year where GICs beat stocks we should not abandon stocks. The same is true for GICs beating bonds. Bonds are riskier than GICs, but they also have higher expected returns. This relationship shows up in the historical data.

Asset Class Annualized Return 1/1/1985 - 12/31/2017
Average 5-Yr GIC CAD* 4.97%
World Government 1-5 Yr (CAD Hedged) CAD 6.13%
Canada Universe Bond CAD 7.97%
US Stocks CAD 11.10%

Any asset class with a positive expected return in excess of the risk-free rate carries risk. The higher the risk, the higher the expected return. While it may not always feel good at the time, holding an asset class through periods of underperformance is a necessity when seeking higher returns.

Through all of this we should not lose the primary objective of holding bonds: reducing volatility through diversification. While bonds are riskier than GICs, they are still much safer than stocks in terms of volatility, and they tend to shine when stocks are crashing. For example, when the global stock markets dropped 31.21% in CAD between March 2008 and February 2009, the Bloomberg Barclays Global Aggregate Bond Index (hedged to CAD) returned +2.82% for the year.

Original post at pwlcapital.com.

All time highs are normal

We may have just witnessed the longest bull market in US market history. There is some disagreement on the definition of a bull market, but no matter how it is defined it is clear that the S&P 500 has been rising without too much interruption for a long time. When the market is rising, it tends to hit levels that have never before been seen.

Information like this makes a lot of people feel uncomfortable. What goes up must come down. I don’t want to invest at the top. It may be normal to feel this way, but it is probably not sensible. Fortunately, we have price data on the S&P 500 from Robert Shiller going back to 1871 to help us think sensibly.

Going back to 1871, and ending August 21, 2018, a dollar invested in the S&P 500 has hit an all time high in 16.3% of months; in other words, 289 of the 1,771 months going back to 1871 have seen all time highs. In hindsight we know that despite hitting all time highs the market has continued to rise. We also know that most of these all-time highs have been followed not by crashes, but by more all-time highs. Of the 289 all-time high months over the period, 194 of them were followed by another all time high. That’s 67% of the time.

Think about that. Investing in a month when the US market is hitting an all time high has historically been followed by another all time high 67% of the time. This strong post-all-time-high performance has not only been true for the US stock market. The MSCI EAFE and S&P/TSX Composite have shown similar results, with 63% and 66% of all-time highs being followed by another all-time high.

The green bars in the chart represent all-time highs.

ath.jpg

Are markets expensive?

Hitting all-time highs may not be out of the ordinary, and they are probably not a reason to get nervous, but as prices rise faster than earnings, the Shiller PE also rises. As of August 22, 2018, the Shiller PE is sitting at 32.89, compared to a historical mean average of 16.56. Based on that it seems that the US market is expensive. Whether or not that is meaningful information is another story entirely.

A 2012 white paper from Vanguard showed that the Shiller PE does have some ability to predict future returns, but probably not enough to base any investment decisions on, especially short-term decisions. The paper found that the Shiller PE explains about 40% of the variance in future 10-year stock returns. More importantly for anyone feeling nervous based on current valuations, the paper found almost no explanatory power over 1-year stock returns. Put simply, even if the Shiller PE is high, that tells us almost nothing about the returns for the following 12 months. Stay invested.

The S&P 500 is not the world

If normalizing the S&P 500 hitting all-time highs and debunking a high PE as an indicator of a coming crash is not enough to calm the nerves, it is always important to remember that the S&P 500 is not representative of the world. While the US market has been soaring since March 2009, International and Canadian stocks have been much more restrained.

Total CAD Return March 2009 – July 2018

S&P 500 Index

18.05%

S&P/TSX Composite Index

10.98%

MSCI EAFE Index (net div.)         

11.10%

Source: S&P Dow Jones Indices, MSCI, Dimensional Returns Web

A globally diversified investor has less to worry about in terms of an extended period of abnormally high returns. Even an investor who owns the total US market, or better yet a US market portfolio tilted toward small and value stocks, would have much less to worry about as those asset classes have not followed the S&P 500’s trajectory. That is, if all-time highs were anything to worry about, which they are not.

The MSCI EAFE hit an all-time high in January 2018, but has not surpassed that level since; it also finished its last bear market in early 2016. The S&P/TSX Composite index has reached all-time highs through 2018. While not quite a bear market, the Canadian market did see a drop of over 17% between 2014 and 2016.

Original post at pwlcapital.com.

Asset Location: Promising Theories vs. Practical Applications

The nonsense of trying to come out ahead by picking stocks or timing the market has long been one of my recurring themes. So, if you’ve stuck with me so far, you probably already know to avoid these sorts of active pursuits. There are other, more sensible ways to seek more from your investments – such as effective tax planning to help you keep more of your returns as your own. Asset location is one form of tax planning. 

So far, so good. But now, I’m going to shake things up a bit. While asset location sounds great in theory, once we work the actual numbers, we often discover it may not be all it’s cracked up to be. At the very least, you don’t want to get lost in the weeds when trying to locate your assets. 

What Is Asset Location?

First things first: What is asset location? I am not talking about asset allocation, which is deciding how much of each asset class you should hold overall. Asset location is the practice of holding certain asset classes in certain account types. Typically, you follow a set of rules for holding the assets with the highest expected tax costs in your non-taxable accounts. For example, you may deliberately hold less-tax-efficient bonds in your RRSP and more-tax-efficient Canadian stocks in your taxable account. 

Easy right? Well, it is easy in theory, but there’s a catch. To know where the highest expected tax costs will occur, it helps to know in advance what your future taxes and returns are going to be. Without a crystal ball to help you peer into the future, it can be really, really hard. 

Can Asset Location Add Value?

On the plus side, much has been written about asset location and its potential contribution to after-tax returns. Here are several analyses that compare an asset location strategy versus holding the same asset mix in each of your accounts (such as the same 60/40 stock/bond allocation in every account you own): [Can you hyperlink to these papers?]

  • In a 2013 paper from Morningstar, David Blanchett and Paul Kaplan determined that asset location might add up to 23 basis points (bps) of value per year to after-tax returns. 
  • In a 2014 paper, my PWL colleagues Dan Bortolotti and Justin Bender found that optimal asset location would have added 30 bps per year to the after-tax returns in an ETF portfolio held from 2003–2012.
  • In a 2017 paper, I used statistical analysis to test an asset location model and found that optimal asset location could ideally add an average of 23 bps per year to after-tax returns. My ideal situation included an investor taxed at Ontario’s highest marginal rate, with enough room in their RRSP to hold most of their bonds while staying in line with their target asset allocation.

But, again, we cannot know in advance what returns a portfolio will actually earn. So, in my own paper, I used Monte Carlo analysis to stress test my optimal asset location strategy across a range of about 1,000 potential outcomes. I found that asset location did in fact add value about 80% of the time. 

While this was an interesting finding, it was still based on forward-looking assumptions. Any asset location optimization decision is based on the expectation of future returns … which, of course, we cannot know except in hindsight.

To test the robustness of my optimal asset location decision against my inability to know the future, I next ran the Monte Carlo model against actual returns instead of the expected returns that had been used to make the optimal asset location decision. Under these circumstances, the average value-added dropped from 23 bps to 7 bps. More importantly, the optimal asset location strategy only outperformed holding an identical asset allocation in all accounts 58% of the time.

Should You Pursue Asset Location?

When we realize that asset location does not guarantee higher after-tax returns, we should start to wonder about other issues that may arise. I explore these possibilities in more detail in my related video, but to name a few, there are: 

  • Regulatory risks – What if tax rates or other tax laws change? 
  • Room for error – Given all the future unknowns, even financial professionals and academics often heatedly debate just what qualifies as “optimal” asset location. 
  • Added complexities – Obviously, it takes a lot more time and energy to engage in asset location than to simply duplicate the same asset allocation in each account. Is the potential value-added worth it?
  • Debilitating distractions – Asset location may cause more harm than good if it distracts you from other investment best practices, such as remaining fully invested and engaging in periodic rebalancing. 

In response to a comment on one of his blog posts, my PWL colleague Justin Bender echoed my own thoughts perfectly: “There are many thoughts on the asset location decision and in the end, it probably doesn’t matter much.”

Put simply, since we cannot predict the future, I often question the value of trying to optimize for asset location. It may even make an investor worse off if they’re struggling with the complexity or, worse, delaying the implementation of their portfolio due to asset location concerns. 

Do you have an opinion on the asset location decision? Leave me a comment about it in my asset location video.

Original post at pwlcapital.com.

Mortgage debt and asset allocation

Mortgage debt is a normal thing to have in Canada. Based on 2016 census data, 61.7% of Canadian families owned a principal residence, and 57.3% of those families had a mortgage.

If you have both a mortgage and an investment portfolio you might have wondered if it makes sense to use your investments to pay off your mortgage or keep investing while following a normal mortgage repayment schedule.

I think that the most common advice out there, and probably the most common thinking, is that it makes sense to keep the portfolio in tact to benefit from the higher expected returns of financial market investments compared to the relatively low cost of mortgage debt.

This might seem logical from the perspective of maximizing expected returns, and for some people it might be logical. However, if the investment portfolio is not allocated 100% to stocks, paying off the mortgage and increasing the portfolio’s equity allocation may be a more sensible approach to increasing expected returns.

To illustrate this point I will use the PWL Capital financial planning assumptions for expected portfolio returns and a mortgage rate of 3.00%.

I will imagine someone with a $500,000 home, a $900,000 investment portfolio currently invested in 50% stocks and 50% bonds, and a $400,000 mortgage. They are deciding between continuing to pay off the mortgage using their income over time or selling their investments to pay the mortgage off immediately.

Here are the two options up for consideration. Keeping the mortgage and a 50/50 portfolio or paying off the mortgage and keeping the remaining capital a 50/50 portfolio.

Keep the mortgage + 50/50 Pay off the mortgage + 50/50
Home 500,000 500,000
Portfolio 900,000 500,000
Mortgage (400,000) -
Net Worth T+0 1,000,000 1,000,000

Initially there is no impact on net worth. There could be some consideration for taxes payable on selling $400,000 of the portfolio, or penalties to pay off the mortgage in a lump sum, but I will ignore them to keep things simple.

I will also ignore the impact of mortgage payments on the mortgage balance. In reality the payments would decrease the mortgage balance, but they could also be assumed to be added to the portfolio in the no-mortgage scenario. It is easiest to ignore them on both sides.

If we assume that the 50/50 portfolio earns 4.68% per year and the mortgage costs 3.00% per year, here is how it looks after one year:

Keep the mortgage + 50/50 Pay off the mortgage + 50/50
Home 500,000 500,000
Portfolio 942,120 523,400
Mortgage + 3.00% interest (412,000) -
Net Worth T+1 1,030,120 1,023,400

Clearly keeping the mortgage and growing the larger portfolio was better, but we are not comparing apples to apples. Keeping the mortgage results in a substantial amount of leverage. Leverage increases risk.

Comparing apples to apples

Based on this, it might be more reasonable to compare keeping the mortgage and a 50/50 portfolio to paying the mortgage off and increasing the portfolio risk level to 95/5, that is 95% stocks and 5% bonds, with a higher expected return of 6.09%.

Keep the mortgage + 50/50 Pay off the mortgage + 95/5
Home 500,000 500,000
Portfolio 942,120 530,450
Mortgage + 3.00% interest (412,000) -
Net Worth T+1 1,030,120 1,030,450

With the more aggressive portfolio you get a nearly identical expected outcome whether you keep the mortgage or not. Despite the appearance of taking more equity risk, arguably your overall risk has decreased.

If you do not think you could handle a 95% equity portfolio, it might help to think about this another way. In either case you have a $500,000 home. Keeping the mortgage and investing $900,000 effectively means investing $500,000 of your own money and borrowing the remaining $400,000 as a mortgage. This is a leveraged investment.

In the worst 12 months of the financial crisis, a 50/50 portfolio lost 20.58%, so a $900,000 portfolio lost $185,220. In our example only $500,000 of that was your own money. A $185,220 loss on $500,000 equates to a 37% loss, but you also paid 3% in mortgage interest on your $400,000 mortgage that year for a total loss of 39.44%.

Alternatively, owning a 95/5 portfolio through the financial crisis resulted in a portfolio loss of 38.92% for the year, with no interest costs. You should be almost indifferent with a slight preference for the riskier portfolio with no mortgage debt.

If a 95/5 portfolio still seems way too risky, then you may have purchased a house that is outside of your risk-appropriate price range! Whether you actually have the 95/5 portfolio, or you have a 50/50 portfolio with the mortgage, you effectively have the 95/5 portfolio.

In the case of a starting equity mix that is more aggressive than 50/50, say 70/30, it would not have been possible pay off the full mortgage balance without affecting expected returns. However it would still be possible to pay off a substantial portion of the mortgage while increasing equity exposure to maintain expected returns.

Keep the mortgage + 70/30 Pay some of the mortgage + 100/0
Home 500,000 500,000
Portfolio 900,000 644,444
Mortgage (400,000) (144,444)
Net Worth T+0 1,000,000 1,000,000

And one year later:

Keep the mortgage + 70/30 Pay some of the mortgage + 100/0
Home 500,000 500,000
Portfolio 947,880 684,658
Mortgage + 3.00% interest (412,000) (148,778)
Net Worth T+1 1,035,880 1,035,880

In either case, having a mortgage and a conservative portfolio, or having no mortgage and an aggressive portfolio, the risk and return characteristics are very similar. Paying off the mortgage has the added benefit of eliminating the guaranteed mortgage interest cost. Where this does not work is if the portfolio is already at 100% equity. In that case there is no way to increase expected returns other than adding leverage.

Mortgage debt should be considered when evaluating the risk and return characteristics of a portfolio. Where possible, paying down mortgage debt while increasing the portfolio’s equity allocation can result in equivalent risk and return characteristics without the added risk and cost of leverage.

Then, and now

Charley Ellis was on the Capital Allocators podcast with Ted Seides last week. Charley has been in the investment management business since the early 1960s. Today he is one of the most vocal proponents of index funds as the most sensible investment for most people.

I write often about why index investing makes sense, and why it is challenging for active investment management to generate consistent outperformance. While that has always been true, there are some good arguments for why it is harder now than ever for active managers.

On the podcast, Charley offers some reminders about how the world and the stock market have changed since the 1960s.

Then

According to Charley 10% of trading, at most, was done by institutions, and 90% was done by individuals. Most of those individuals were, as Charley describes, “nice people” who were investing their savings in stocks without doing much, if any, research. In other words, there were lots of easy targets for an active manager to exploit.

In the past, it was possible for an analyst to set up a private meeting with a company’s management in order to get an information edge.

Charley estimates that there were less than 5,000 people in the active investment management business.

Securities firms had 10 to 12 analysts who were only searching for interesting investments for the firm’s partners, and were not publishing their research.

About 3 million shares traded each day.

Now

Charley says that 99% of all trading today is done by computers. Most investors involved are highly skilled and have instant access to information. In other words, there is no easy target left to exploit.

Today, under Regulation FD, the SEC requires that any material information that is disclosed must be disclosed publicly. Under Reg. FD It is not possible, legally, to get a competitive information advantage.

Charley estimates that there are over 1,000,000 people in the active investment management business. There are around 154,000 CFA Charterholders in the world today.

Securities firms today have hundreds of analysts with diverse expertise in every major financial hub constantly publishing their research.

More than 5 billion shares trade each day.

The Paradox of Skill

Piling an increasing number of skilled professionals into the investment management business might seem like it would benefit investors, unfortunately it does not. Michael Mauboussin and Dan Callahan explained the paradox of skill in a 2013 Credit Suisse white paper:

In investing, as in many other activities, the skill of investors is improving on an absolute basis but shrinking on a relative basis. As a consequence, the variance of excess returns has declined over time and luck has become more important than ever. Still, differential skill continues to exist. This process is called the paradox of skill.

Put simply, if equally skilled and informed investors are competing with each other, the winner will be defined by luck rather than skill. It is not absolute skill that matters, but relative skill.

If there has ever been a time for active money management to flourish, that time is not now. This shows up consistently in the results of active fund managers; the vast majority of them underperform the index over any given time period.

There is little question that simple low-cost index funds are the most sensible investment for most people. Paradoxically, the case for index funds grows stronger as the investment management industry gets more skilled.