Asset Location 

I agree you need to diversify a bit more.Transfer some to me 😏

I'm more of a "hollowed out book" man myself but I do like your style.

Agreed!

Thanks! I also did a new paper on asset location last year www.pwlcapital.com/wp-content/uploads/2019/11/2019-11-14_PWL_WP_B-Felix_Optimal-Asset-Location.pdf In my opinion it's one of the least useful things that people worry the most about.

I did a much deeper dive in this paper www.pwlcapital.com/wp-content/uploads/2019/11/2019-11-14_PWL_WP_B-Felix_Optimal-Asset-Location.pdf

Ben Felix - Thanks so much for the helpful resource!

Thanks!

Haha! The cost of waiting for the right time to invest. There’s a good video idea.

@@BenFelixCSI Yes, thank you! That's a wonderfully good video idea. :) Cheers, Doug

I agree, they do a great job!

Agreed!

Sure, but you also pay tax on the income and gains in the taxable account.

@@BenFelixCSI comparing it to RRSP, with RRSP you save withholding tax and you get tax deduction while on a non-registered account you pay tax on dividends (and possibly when you rebalance) but you get tax deduction on the cost of investment. On both cases if you cash your inversement out in retirement , you pay lower income tax . I remember I did a quick calculation on the difference of cost of inversement versus tax deduction on RRSP and the former made sense but I never did a full evaluation . Happy Easter by the way !

Great questions, Quinton! I agree that with a 58% success rate in the model there appears to be a positive expected outcome, but that could easily change if the realized returns end up being different from those used in my example. I arbitrarily tested the model using two sets of assumptions, but those assumptions do not come close to capturing reality. The point is not so much that value is added in 58% of trials. It is more that when we change the assumptions, the probability of value being added changes substantially. Further to that point, when the magnitude (0.07% in the example) and frequency (58% in the example) of positive outcomes are weighed against the potential drawbacks of an optimized location strategy (cost, complexity, regulatory risk, model risk) I do not think that this is an obvious value-add. In fact, I think that it introduces a new risk which, unlike the dimensions of returns, does not have a positive expected outcome. Taking on the risk that future returns and tax rates will be what you expect them to be is not a priced risk. It is very different from taking the priced risks of small cap and value stocks. Asset location ends up being a bet on your ability to predict the future. As the differences are likely to be small, I see no problem with professionals implementing asset location strategies for their clients if it makes the client happy. I similarly take no issue with any DIY or professional who chooses to have the same allocation across all accounts. As far as I know there is zero evidence that current tax rates are predictive of future tax rates. On a macro level tax rates change with governments. At the individual level tax rates change based their individual circumstances. I would be very hesitant to bet on future tax rates, or the magnitude of future returns, which are both requirements of asset location optimization.

I agree that we don't have good data on predictability of future tax rates (perhaps we should go halfers on a Master's/PhD thesis). However, in every financial planning software we use, we make the assumption that future tax rates will stay the same. I'm simply advocating for the null hypothesis in this instance. Since I have no data to support the predictability of tax rates, I can only assume that the future will be like the present. Perhaps it's not reasonable to assign a specific rate to capital gains/income, but I think it is reasonable to assume that capital gains will be more favorably taxed than regular income, and/or if that changes, such an event would not be predictable in advance. As far as cost goes, are we assuming that complexity adds cost to the end investor? (I would agree that it becomes less scalable, which is why for accounts under 1M, we generally don't fuss too much unless the client specifies). For clarification, specific to regulatory risk, is an investor not already exposed to that risk? Could you elaborate? Good stuff as always.

Maybe. It's an unknown. Even if we knew tax rates, we do not know future returns. The point is that because it is all unknown, it is very hard to concretely say that asset location will add value. Yes I am assuming cost is added for the end investor, which I think is true based on the scaling challenges. Assuming that leads to higher fees it gets that much harder to justify the value-added. Let's say it becomes relatively tax inefficient to hold Canadian stocks in a taxable account. If you have the same mix in all accounts you are only exposed on that portion and probably don't need to change. If you have all of your Canadian equity in your taxable account, you may need to change which would have tax implications.

I know that the Loonie Doctor has argued that asset location is especially important for CCPCs to defer the passive income limit for as long as possible. I am still skeptical that we can predict future returns well enough to make intelligent decisions about location. If you’re interested I just released a more detailed discussion of this topic, although not about CCPC: www.pwlcapital.com/resources/optimal-asset-location/

@@BenFelixCSI Thanks!

If you own a foreign stock that pays a dividend and have foreign taxes withheld on the dividend, you can use that foreign withholding tax to reduce your Canadian taxes. So assuming you have enough taxable income in Canada to use the credit, you end up recouping the foreign tax. While that sounds good, you would prefer to lose the withholding tax and _not pay any tax on the dividends,_ which are fully taxable at your marginal rate. Foreign stocks have higher _yields._ Dividend yields: XEF (international stocks) = 2.41%, XUU (US stocks) = 1.37%, VCN (Canadian stocks) = 2.37% taxed at a favorable rate. Did that all make sense?

that's what is sounded like to me...

I appreciate the discussion on the topic, but this one is not a point of opinion. It is mathematics. The only thing that matters to your after-tax net worth is your after-tax asset allocation. What you see in your accounts (pre-tax) means nothing to your ultimate result. The 60/40 will only act differently than the 75/25 *pre-tax*, which is a meaningless comparison because you cannot access your pre-tax dollars. Let's say that you have $1m 60/40 pre-tax allocation like my example in the podcast. You put $400k in the RRSP and $600k in taxable. Assume a 50% tax rate, which means that your after-tax allocation is 75/25. Say we have a 50% stock decline while bonds stay flat. Your portfolio could fund a $500,000 liability at that moment ($300k from stocks in the taxable account + $200k in after-tax bonds from the RRSP). Let's instead say we had a 75/25 mix in both the RRSP and taxable accounts: $300k in stocks + $100k in bonds in the RRSP, and $450k in stocks and $150k in bonds in the taxable account. Both the pre and after-tax allocations are 75/25. Stocks drop by 50%, now you have $150k in stocks + $100k in bonds in the RRSP, and $225k in stocks and $150k in bonds in the taxable account. If you liquidate, you again have $500k after-tax. $75k in stocks + $50k in bonds from the RRSP, and $225k in stocks and $150k in bonds from the taxable account. So, in a 2008 scenario, you are indifferent between the two portfolios with the same after-tax allocation. Your ability to fund your liabilities is impacted in exactly the same way. Do you agree now, after seeing the math? The point on basing location decisions on what is most tax efficient at a point in time ignores the (potentially substantial) tax cost of reshuffling the locations when circumstances change. Interestingly, bonds are not the most efficient asset class to hold in the RRSP at this time - Internationals stocks are. If we control for the higher after-tax allocation to stocks, there is no advantage to bonds in the RRSP. In fact, there is a disadvantage. That was one of my last points in episode 36 if I remember correctly. An accidental 75/25 is less efficient than an intentional 75/25 with all International stocks in the RRSP. Thank you for listening to the podcast, watching the videos, and especially for commenting!

@@BenFelixCSI I agree with your math, the problem in your example is that bonds usually don't stay flat in a recession, they usually get boost in positive return. I prefer to have 40% of bonds in an RRSP (pre-tax), than only having 25% bonds in a taxable account when that asset is getting a boost in return. 400k in an RRSP with a return of 6% on bonds (2008) will give me a return of 24k, take that return out of the RRSP and it's at 12k tax free. With your 75/25 portfolio, if you only have 250k of bonds with a 6% return, that gives you 15k and you still have to pay your taxes, which means a net return of 7.5k. The point I am trying to make is that you still get the boost of 6% return on all the 40% of bonds(pre-tax) you have in your RRSP. I'm sure I'm missing something, you have a bit more time to think about it than I do since it's your job :)

@@jonathanborg9274 If I re-run my previous example + a 6% return for bonds, in both cases the end result after-tax is $512k. There is no difference. Honestly, this one (pre vs. after-tax allocation and its interaction with asset location) has taken a whole lot of thinking and re-thinking. It is not at all intuitive.

This asset location stuff is nuanced. You lost unrecoverable foreign withholding tax in your registered accounts when you hold ETFs of international stock there, but you will pay more tax overall by holding international stocks in a taxable account. International developed and emerging markets dividend yields are high, and the dividends are fully taxable. Being able to claim foreign withholding tax is a small benefit relative to a big tax bill. Canadian dividends are taxed more favorably in a taxable account. More here www.pwlcapital.com/wp-content/uploads/2019/11/2019-11-14_PWL_WP_B-Felix_Optimal-Asset-Location.pdf

Thanks Rick! Domodoran is good company to keep. Not so sure about being in a list with Shkreli, but I'll take it!

Purely on the merits of the content of videos. Shkreli seems to be a scumbag IRL, it seems you aren't! Have you had a chance to watch his investing/finance lessons... curious what you think of them...?

In all seriousness Shkreli does have some good content out there. Thanks again Rick!

Rick Chen I have not spent a ton of time watching them, but am familiar with his content. I think he talks a lot about valuation which is something that does not align with what I suggest is a sensible approach to investing. If you’re using low cost index funds to invest then the ability to value a company is not useful.

That would be a good future topic. Thanks Dave! I did touch on asset allocation a bit in the video How Much Risk Should You Take? ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-t0EYr8Aex1I.html and also in the video Why Not All Index Funds Are Created Equal ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-8dkq4EwIawg.html but it would be good to go into detail on how much of each asset class it makes sense to hold and why. I will add that to the list!

Why does the internet love asset location so much?! This video was supposed to tell the world that asset location is probably not worth your time. Anyway, I think REITs would be best held in a TFSA *if you are so inclined to attempt to optimize.*

Thanks Ben, I think the internet ''love'' asset location so much because all those financial / tax things are confusing to average joe and we don't want to screw up too much.

Right, that makes sense. As long as you are using low-cost total market index funds I think you are winning the battle of not screwing up. All of the other stuff like asset location, factors (small cap and value stocks), US listed ETFs in RRSP accounts etc. adds a bit of value on the margin, but the trade off between complexity and vale added on those things is not always obvious.

@@BenFelixCSI how many basis points would you say those things add: Seperate funds insead of one asset allocation, factor investing, asset locationI, and US. Listed in RRSP? Personally im hoping to get the xeqt and also the 2 small cap and value etfs u suggested in ur paper, maybe even some REITs, almost like a balance between complexity and comfort. More complex than one asset allocation but not so much so that im pulling my hair out.

I agree. The asset location decision is independent of the "which account should I use" question.

Always max out registered accounts first. Canadian dividends are more tax efficient than foreign dividends once you get to the taxable account.

Assuming no regulatory changes, an argument could be made to do as you have described.

I was wondering if maybe that since RRSPs are taxed as income when they are withdrawn and index funds since they have low turnover are tax efficient anyway, that long term tax efficient growth in the non-registered account with eventual capital gains as you draw down might outweigh that tax free growth in the RRSP that will get nailed eventually, because you would be growing relatively tax efficiently anyway in that non-registered account. I guess expected dividend yield would be the main factor to consider?

@@romiguma good question. It's not accurate to think of the RRSP as converting capital gains and dividends to regular income. It appears that way, but it is not the case. When we compare an RRSP to a taxable account we have to account for the tax deferral on the RRSP contribution. For example, at a 50% tax rate, $10,000 invested in an RRSP should be compared to $5,000 in a taxable account. The extra $5,000 in the RRSP belongs to the government, but the other $5,000 is yours and it grows tax-free.

So the argument goes. The problem that you run into is that the tax liability is not only a function of the tax treatment of the asset, but the returns of the asset. Comparing VTI in a TFSA to VTI in an RRSP misses this point. We should be comparing VTI in an RRSP to IEFA in an RRSP. All else equal VTI looks better, but this ignores the fact that all else is not equal - the relative tax savings are not guaranteed because the relative returns are not guaranteed. This would be a non-issue if attempts to optimize asset location did not add complexity and risk. The added complexity comes mostly from the need to take after-tax asset allocation into account in the portfolio management process (video on that is forthcoming). The risks come from the potential for tax changes and the uncertainty of future returns. Trump decides to take away the RRSP withholding tax exemption, and instead adds an additional tax. That optimal asset location is no longer so attractive. US underperforms for 10 years, same story. Favoring bonds in the RRSP is also generally flawed. It is a requirement to take the initial tax savings into account when comparing the RRSP to a taxable investment account. For example, at a 20% tax rate investing $10,000 in the RRSP must be compared to investing $8,000 in a TFSA. When this is accounted for properly it becomes clear that there is really no difference between holding bonds in the RRSP vs. TFSA. There are other issues, like premium bonds, which make bonds relatively tax inefficient in taxable accounts, but that should not last forever. In any case, I think most DIY investors will find some happy medium that works for them, or at least feel "good enough". I don't think there is an optimal solution. I do agree that once the registered accounts are full, it's a good problem to have!

I agree that for tax efficiency with bonds, TFSA versus RSP is a wash. I guess I prefer bonds in the RSP because I expect them to underperform stocks, and while I ultimately want my TFSA to grow to infinity, I kinda want my RSP depleted to zero through low-bracket withdrawals long before government-mandated retirement age. Sheltered compounding of uncertain market returns is a nice perk of the RSP, but I can approximate sheltered compounding without the shelter using e.g. HXS. I am much more interested in capturing the 20%~30% guaranteed swing between marginal income tax rates at RSP deposit and withdrawal times. If that advantage is not being realized, I think the RSP is barely worth using (thus I also don't agree with the general idea that 10k in the RSP equals 8k in the TFSA - if you've allowed that parity to hold, you've actually blown your RSP advantage). RSP contribution room is generally much larger than TFSA room for the high-income earners who would bother to use the RSP in the first place. My most golden goose of all, my TFSA room, is quite scarce and precious compared to my spacious but treacherous RSP. So I want my TFSA packed with fast-growing stocks which I will ideally never touch, while comparatively liquid slow-growing bonds can keep my RSP from getting too big (or too upside-down) to efficiently close out. It will take 20 years to withdraw a 200k RSP tax free. As a proponent of FIRE, that's getting close to a kind of optimum, for me.

That all makes sense, and it is definitely a preference. Unless someone has a very low income, they would have to have a very large RRSP (assuming no pension) to end up upside down in terms of the tax rate spread between contribution and withdrawal. I agree in full that the tax rate spread is the best attribute of the RRSP.

Assuming that the taxation of investments is somewhat similar, you might be able to follow similar rules. In terms of account types, here is the translation: RRSP ~= 401k TFSA ~= Roth IRA

Ben Felix thanks. Why put not put bonds in a personal taxable account if they have lower expected gains than stocks?

Because the bond return is mostly interest, which is taxed annually at a relatively high rate, while the stock return is largely long-term capital gains, which is deferred until sale and taxed at a relatively low rate. The tax cost of owning bonds is higher despite them having lower returns.

Ben Felix and one last question... If one currently has same asset location in all accounts but wanted to change up the location to be more optimal, would it be reasonable to just adjust their Investment direction going forward rather than selling and re buying to change the whole current portfolio? Thanks!

Well, my view is that you probably shouldn't be trying to optimize your asset location because it increases complexity and is not a guaranteed way to increase your after-tax returns. If you feel compelled to optimize, then I would do it over time as you add new assets as opposed to doing it all at once which might have tax implications. But again, I don't see a ton of value in trying to optimize this.

Thanks Justin! Good to know our previous comment conversation was helpful to you.

@@BenFelixCSI I have a kind of interesting question to add. so the company i work for started giving employee share grants a couple years ago on a drip plan. its a nice perk but being a utility stock its not really a growth investment it does pay over a 4% dividend however. my question is if i hold each block of shares i get (roughly $2500 a year) for two years i get a 50% income tax deduction if i sell that block. my thought process is i can see them and take about a $500 tax hit but reinvest that money in my RRSP and get about a $800 tax refund on it which i would put into my RRSP as well. and then invest that money into something with better long term growth like the sp500. does this idea make sense or am I missing something. or would you think its better to just hold on and let the drip build wealth that way. im only 32 so I do think theres a better potential return in something with better growth by the time i retire when im 55.

i make on average about $110,000 a year so im in the 40%ish marginal tax bracket i believe.

@@BenFelixCSI also if you dont want to answer the question thats totally fine to i understand people pay you for your experise and i wouldent expect you to give your knowledge out for free. :)

I need to understand this share purchase plan a bit better. So you receive $2,500 of shares each year, and then you are required to hold them for two years? Where does the 50% deduction come from? Is that just the capital gain inclusion rate?

Interesting question. If all of the returns had the same characteristics (deferred capital gains) then you would want to hold the asset classes with the highest expected gains in the tax preferred accounts. You might end up with HBB in your taxable account and the equity ETFs in your tax advantaged accounts.

Ben Felix thanks for the reply! 👍

Depends on how you account for your allocation: if you don't make the complex pre-tax adjustment for RRSP allocations, then you'll want the lowest expected return in the RRSP (i.e., bonds in RRSP) and the highest expected return in your TFSA. ;)

John Potato very true. I have a video on pre tax vs after tax allocation coming out in a few weeks.