As I am self-employed and paid in dividends rather than salary, the majority of my investments are in non-registered accounts. I’ve been struggling with the general question of expected returns vs. diversification on a couple of fronts. First, it is generally accepted and understandable that Canadian equities should be expected to outperform foreign equities on an after-tax basis due to the dividend tax credit. Therefore, it makes sense to have *some* over-weighting of Canadian equities. By doing so, you are adding geographical and sector concentration aka reducing diversification, so can expect higher volatility. Currency exchange risk also adds to that volatility even though it shouldn’t affect very long term expectation. In many cases, currency exchange costs would also have to be factored in, although in my personal situation they are not significant given the nature of my business. So for me, the question boils down to how much added volatility I am willing to accept in return for the added expected return of the Canadian dividends. (A tricky question to answer without being able to exactly quantify the volatility involved!)

Thinking about this though, I realized it basically mirrors the general equity/fixed income split issue. I currently have a 75/25 equity/income split, where the fixed income is split between preferreds and short term bonds (again trading off between safety/diversification and expected after-tax return). But if I’m contemplating over-weighting Canada in order to increase my expected return at the cost of higher volatility, what am I doing holding fixed income at all? Couldn’t I just as easily move to 100% equity, essentially making the same tradeoff, except with a larger boost in expected return?

Put another way, in a non-registered account, does it make any sense to start over-weighting Canada at all, while still holding fixed income investments? Even if the fixed income is preferred shares, you could expect to gain more on average by switching that into diversified equities than by over-weighting your equities in Canada. (On the other hand, perhaps the increase in volatility is greater with this change as well.. I don’t even know how to begin to quantify that.)

Finally, on the far end of the spectrum is the approach of over-weighting small caps. As above, small caps have theoretically higher expected returns but greater volatility. I don’t have exact numbers here, but I expect that the increase in volatility is greater, and the expected increase in return is lesser, than either of the options above. So then if this line of thinking makes sense, even if all you care about is expected return, it isn’t logical to start adding focus on small-cap stocks until you already have a 100% equity, largely Canadian portfolio. (And likely not even then, since then you would be sacrificing Canadian dividends to do so.)

What do you think? Again, all this only applies to non-registered accounts, and basically consists of my musings and assumptions. I haven’t thought about this over a long period of time or done any definitive research. And I’m certainly not suggesting just going 100% Canadian equities in registered accounts; just that if a person wants to trade added volatility for added expectation, the various methods for doing so (more equities vs fixed income, more Canadian equities, more small-cap) should be prioritized based on which method increases expectation the most and volatility the least. Hopefully someone can definitively tell me which that is!

Money in my pocket = Money invested + Gains – Taxes

So, assuming “Money Invested” and “Gains” are similar(*) between the different accounts (RRSP, TFSA, and non-sheltered), the primary goal should be to minimize taxes paid in order to get as much “Money in my pocket” as possible.

Here’s an example using a $2,000 dollar investment (50% TFSA, 50% non-sheltered) with a 50% GIC/50% stocks allocation. Assume the GIC has a 2% and the stocks 8% (capital gains only) annual return. After 3 years, what happens?

(assuming 50% marginal tax rate)

Scenario 1: $1000 GIC in TFSA, $1000 Stocks non-sheltered.
– end of year 1: TFSA $1020, non-sheltered $1080 (no tax on unrealized gains)
– end of year 2: TFSA $1040, non-sheltered $1166 (no tax on unrealized gains)
– end of year 3: TFSA $1061, non-sheltered $1194 [$1259 - 65$ (taxes)]
Total in my pocket after 3 years (after taxes): $2255

Scenario 2: $1000 stocks in TFSA, $1000 GIC non-sheltered.
– end of year 1: TFSA $1080, non-sheltered $1010 [$1020 - $10 (taxes)]
– end of year 2: TFSA $1166, non-sheltered $1020 [$1030 - $10 (taxes)]
– end of year 3: TFSA $1259, non-sheltered $1030 [$1040 - $10 (taxes)]
Total in my pocket after 3 years (after taxes): $2289

So, scenario 2 wins, mostly due to less taxes paid.

(*)Note that the gains on GICs are reduced in scenario 2 (e.g. in year 3: 2% of $1020 [$20] is less than 2% of $1040 [$21]). This difference in the “Gains” part of the “Money in my pocket” formula should be taken into account in a full picture.

The long-term average annual gains of stocks and (some types of) bonds are much closer than stocks/GICs, making it preferable to hold the bonds (in preference to stocks) in TFSA/RRSP accounts in most cases.

Personally, I use an estimate of the difference in both taxes and gains over the life of an investment to guide my decisions.

My rule of thumb for putting money in RRSP/TFSA:
– the bonds part of my long-term goals goes in first
– if space remains, the stocks part of my long-term goals goes in second
– if space remains, the GICs of my short-term goals go in.
Any overflow goes in taxable accounts.

They take the position that the income you earn on your investments in a RRSP is actually tax free, because the tax that you pay when you remove the funds is just the amount the government loaned you in the form of the RRSP contribution credit plus interest. Consequently, they feel it is no different from a TFSA and that you should place the funds that will be taxed the most heavily in total dollars into the RRSP. In other words, if your going to make enough capital gains to pay more tax dollars than on the current low bond interest, then you should put your equity investments into the RRSP preferentially, even if they are Canadian, because you aren’t actually paying any tax on your portion of the RRSP, so it doesn’t matter that you can’t claim the capital gains tax credit.

In any case, they do make a fairly convincing case and have forced me to reconsider how I think about my RRSP allocation. I have a nagging feeling that there is a flaw in their logic somewhere, but I haven’t been able to find one. I’d be much obliged if you could have a read through and see if you agree with their conclusions, because if they are correct then all equity index ETFs that are being purchased with the intent to hold for the long-term should be held in RRSPs on the basis that stocks have consistently higher returns than bonds/GICs.

If you can only afford an RRSP then everything goes into that (assuming that your current tax rate > expected withdrawal tax rate otherwise TFSA is an option but has lower contribution limits) and allocation % will be based on your risk tolerance. You could be almost 100% in Bonds or 100% in equities or anything in between depending on your age and risk tolerance.

If you have additional funds that you want to invest then looking at the portfolio as a whole is the way to go. Determine your desired allocations and then administer across all the accounts. Do think about the tax implications but also think about time horizon.

For instance, you are getting close to sending your kid off to school, you may want to make that portion of the portfolio less risk adverse to ensure that you don’t see a major drop in the 2 years before you need the funds for instance. This would mean bumping up your equity portion in other parts of your portfolio so as to maintain your desired overall portfolio asset allocation.

Another example, would be if you had a large capital loss from previous years that you can use against future capital gains. You may want to have as much of your equity portfolio outside of the registered accounts to benefit from any capital gains that are triggered regardless if it is canadian or foreign. Yes with foreign outside of registered accounts you will have 15% withholding tax on dividends but most can claim that against taxes owed at tax time so I don’t see that being a big reason to not to move foreign equity outside especially if you are trying to generate lots of capital gains. Better to diversify the capital gain generation than depend on just Cdn equities for the gain generation. If your main reason is to generate capital gains, the small 15% tax on say 1/5 to 1/10 of the total annual return is minimal compared to the benefit of using the capital loss.

All to say, there are more factors involved in determining which assets go in which account other than tax implications. Risk tolerance, time horizon, current tax situation vs withdrawal tax situation, etc.. All must be taken into account but I do agree with the general rule of ” income producing assets in tax-deferred accounts and dividend / capitals gains producing assets in non-registered accounts”.

Knowing the tax implication of each type of asset in each type of account is the beginning step and then making it work for your situation is next.

I can see some comments coming on this one.