An appropriate benchmark is probably one of the most important aspects of long term investing. Yet, most individual investors have never thought much about benchmarks.

A benchmark is important from a couple of perspectives; 1) it provides a basis to compare how you are doing against a passive alternative and 2) it provides a gauge to measure the performance of the portfolio, not individual securities within the portfolio. 

The key to a good portfolio is that it reduces risk and enhances return. Risk reduction can only come through diversification, and diversification is only valid if the securities in the portfolio have low correlations with each other. Put another way, in a properly constructed portfolio, there should always be something that is not doing well.

If you and / or the Advisor become fixated on the individual securities in the portfolio, we guarantee two things will happen: the fist fifteen minutes in the meeting will be spent talking about the securities that went up in value, and next hour there will be a discussion around the securities that went down in value. Nothing of value will get done, and both sides end the meeting in frustration. 

Of course saying you need a benchmark and actually designing one are two very different issues. To be useful, a benchmark must have certain properties. Which is to say, for a benchmark to have any value it must reflect what it is you are trying to accomplish. It does not good to use a GIC portfolio as a benchmark, if the return from a GIC portfolio do not provide sufficient growth to meet your objectives. So, if we are to use a benchmark it has to be a good benchmark. 

If you are using an advisor, then the right benchmark must be one that both of you agree with in advance. It is no good picking a benchmark after the results are already in, which, by the way, is how it is often done.

Advisors talk to clients about their investment returns, and add a qualifier like “At least it’s better than inflation.” And then the client fires back that “Yeah, but it’s not as good as what the TSX did.” And next year it won’t be inflation and the TSX... it will be the money market and average housing prices. Or the GDP and the average baseball salary increase. Or whatever.

Further the benchmark has to be consistent and be something that is useful year after year after year. If it was a good benchmark to begin with, it will be a good benchmark for years after that.

A good benchmark also has to be challenging. It is one thing to say, for example, “My goal is to outperform the Consumer Price Index.” It is quite another to say that “we intend to beat the CPI by 3% annually.” 

On the other hand, the problem with the CPI or something like it, is that it is not very useful on its own. It would be like Barry Bonds having a target of 10 home runs in a single season, which like beating inflation, is nothing to write home about. In fact, you could succeed in beating a CPI benchmark year after year, yet after tax, you are no further ahead. 

The point is, beating the CPI by a certain percentage might be a good starting point, but it still does not tell you whether the manager’s performance was good, bad, or indifferent. That’s because a good benchmark also has to be relevant, and the CPI is as relevant to investors as the number of days the sun shines in the summer. It’s not relevant at all. So we can say that a good benchmark must be both challenging and relevant. 

Some managers use the return on an index such as the S&P/TSX Composite index, the S&P 500 composite index (a broadly based basked of 500 US stocks), or the Scotia Capital Universe Bond Universe Index as their benchmark. These indexes are more challenging to beat. And more relevant as well, because at least they are investment-based. In fact, many Advisors do use the S&P/TSX or the S&P 500 as a target to be beat.

But a good benchmark should go even a step further. The TSX and the S&P are both all-equity indexes - one Canadian, the other American. The ScotiaCapital Bond Universe Index is an all Canadian bond index. We do not believe that measuring your portfolio’s performance against a single-country, or a single-asset-class, is appropriate. 

Here’s why. First, the typical investor, and what we have been talking about all along, holds a portfolio made up of several asset classes. Rarely does someone hold an asset mix of 100% Canadian equities. Nor would they typically hold 100% U.S. equities or 100% Canadian bonds either. They will hold some cash, a range of bonds, and various Canadian, U.S., and international equities. The benchmark should reflect that diversity. A single-country, single-asset-class index simply does not reflect a passive version of the client’s active holdings. 

For serious investors, it simply does not cut the muster. Remember, what a benchmark is supposed to do is provide a basis for comparison. You compare what your portfolio made with what you could have made. A portfolio benchmark tells you what you could have made if you had simply stayed invested in the benchmark portfolio all along.

But if the benchmark consists of 100% equities, or 100% bonds, it’s not a reasonable comparison, because you would never have invested that way in the first place. So if you would never have invested in the benchmark, it’s not a reasonable alternative to what you did invest in.

We also know that investors have different risk tolerances. And these risk tolerances span a very wide range. Measuring the performance of an ultra-conservative and an ultra-aggressive investor against the same benchmark also does not make sense. The benchmark should have the same risk as the portfolio the client invests in. By contrast, the S&P/TSX as a benchmark reflects a single risk level, and as such, cannot reflect the entire spectrum of risk tolerances in the marketplace.

In fact, risk adds a whole new dimension when dealing with advisors. Say your portfolio matched the performance of your benchmark, but upon further analysis, did so with less risk. That’s called positive risk-adjusted performance, or positive alpha. Just as equal performance with less risk, under-performance with substantially less risk, or over-performance with equal risk, are examples of positive risk adjusted returns. These are good things, but unless you have a portfolio benchmark to compare with, there is no way to have a good discussion with your advisor as to the merits of his work over a period of time. 

One more thing: In an ideal world, the benchmark should be investable. If you cannot invest in the benchmark, it’s not really an alternative. At one point, the ScotiaCapital Bond Universe Index was not something you could buy. Today, you can buy it with the iShares Broad Bond Index (symbol XBB, listed TSX). 

There is real value in having something that is investable. If you are able to buy the benchmark instead of the portfolio, then you have a valid alternative that can form part of the discussion with your Advisor. If your portfolio does not beat the benchmark, then simply buy the benchmark. 

So, a good benchmark has to be agreed on, in advance. It has to be challenging, relevant in that it has to be diversified by asset class and geographical region. It has to be risk-adjusted to match the client’s risk tolerance, and ideally, it should be investable. That’s the only way it can be presented as a valid alternative to what was actually included in your portfolio.

Next week, we will look at the Real World of investment management, and look at building benchmarks that take into account the costs associated with financial planning work that an advisor may bring to the table.