Constructing ETF portfolios

Following the global financial crisis, I set out on a path to offer my clients the best possible investment solution. I came to the realization that portfolios made up of ETFs would offer the best outcome over time.

Also read: Making the paradigm shift to ETFs

As I did not yet have my discretionary portfolio manager’s licence at that time, I decided to build my clients’ portfolios so that I could easily convert their assets to a discretionary business model once everything was in place (licencing, investment policy, compliance approval, etc.).

Although building and managing ETF-based portfolios doesn’t require a discretionary portfolio manager’s licence, the argument for making the leap beyond fee-based management to a discretionary model is quite convincing. This was part of my business development plan from the moment I began building these portfolios.

The first thing I had to decide was how many portfolios I would need to create to offer a variety that was sufficient to meet the needs of my clients’ diverse risk profiles. I decided to build four core portfolio models: conservative, balanced, growth and dynamic growth.

Some advisors will work with two portfolios: one that’s all stocks and another all bonds, then combine them to create different portfolio profiles. For example, 100% stocks, 80% stocks and 20% bonds, 60%/40% and 20%/80%.

I didn’t want to go that route as I thought it was important to have the flexibility to use different investments in each portfolio, such as more low-volatility ETFs and fewer sector-based ETFs in the equities portion of the conservative portfolio compared with the growth portfolio, for which the risk profile is higher and there’s more of a focus on equities.

Next, I needed to benchmark each of the portfolios. Once again, to earn my discretionary portfolio manager’s licence, I would need to have benchmarks, along with an investment policy that outlined asset-allocation parameters and risk factors.

The choice of a benchmark is an extremely important factor in a portfolio’s returns over time. It will dictate how you run your portfolios and how you explain performance attribution to clients. You cannot change a benchmark from one year to the next on a whim, so the amount of time and effort put into deciding what the benchmark will be for each portfolio is critical. Remember the adage: asset allocation accounts for upwards of 80% of returns. Some might say it’s even say more.

I had several discussions internally with our research team and reviewed the asset allocation of many of the bigger mutual fund portfolios (fund of funds) in order to create benchmarks for each of my four portfolios.

Once that was done, I selected the ETFs for each portfolio, working with core and satellite positions for asset classes and geographical allocations. Core ETFs are chosen primarily based on their ability to offer inexpensive broad-based geographical and asset allocation. Core positions that meet these requirements often are the traditional, market capitalization ETFs.

I then use satellite ETF positions to overweight sectors such as technology or financial services. I only use satellite positions if there’s compelling research showing that these sectors are poised to outperform the broad market or core index position.

In each of my portfolios, the core equity positions are a combination of ETFs that represent the S&P/TSX composite index for Canadian equities, the S&P 500 composite index for U.S. equities and the MSCI EAFE index for international equities. On the fixed-income side, the core ETF holding replicates the FTSE Canada universe bond index.

From those core positions, I branch out with sector or specialty ETFs based on market research and vary the asset allocation and geographic allocation, within the parameters of the portfolios investment policy.

Since launching my portfolios more than six years ago, there have been many developments in the ETF space. The ability to hedge international currency exposure is one of them. With important swings in the value of the Canadian dollar (C$) vs the U.S. dollar, I had to make the decision as to whether hedging was a performance criterion. I decided that the answer was no. So, the international exposure in my portfolios is usually not hedged back to the C$.

In addition, the race to enter the ETF space in Canada has lowered the management expense ratios on even the most inexpensive plain-vanilla ETFs, and we’ve an explosion of smart beta or rules-based ETFs. The proliferation of products has been the catalyst for changing many ETFs within my portfolios. This has been done not only to reflect changing market conditions, but also because better and cheaper investment products have become available.