Lately, a lot of our analysis and writing hovers around retirement and RRSPs. We started with a brief post about Not Blindly Contributing to your RRSP a few months back and continued in March with more elaborate one explaining how we were Dealing with Complex Retirement Considerations. Today we persist with some profound thoughts on the never-ending RRSP versus non-registered debate. You’ll still note our research is not exhaustive as it particularly refers to our personal situation.
Some people feel RRSPs are a government scam to take more of their money thru taxes. Many angry retirees almost consider RRSPs as evil when they realize how much tax they owe at withdrawal. They just forgot all about the juicy tax refunds they received when they deducted their RRSP contributions in the first place. Over the years, these RRSP deductions helped fund a great portion of their retirement stash.
We can view RRSP tax refunds as a loan the government allows you to make to yourself. You only have to remember that you’ll have to reimburse it with interest some day (at withdrawal). If you are in the same tax bracket, the interest rate of that artificial loan will be equivalent to your investment return. If you now fall in a lower tax bracket, good for you, you’ll pay less «interest». Similarly, if your tax bracket is higher, tough luck, you’ll end up paying more «interest» via income taxes.
RRSPs are Better with Equivalent or Lower Withdrawal Tax Rate
RRSPs are still the uncontested winner if you expect a lower withdrawal tax rate. Paying less tax is obviously better. After that, the next best thing is paying zero tax. TFSAs can give you that. You get the same with RRSPs under equivalent deduction/contribution versus withdrawal tax rates. But if you expect steeper tax rates because of higher income in retirement, non-registered may come into the game. Knowing a related tax bill is inevitable, the objective here will be to reduce it as much as possible.
Again, let’s make it clear, there no question RRSPs are better if your withdrawal marginal tax rate is lower than your deduction (same as contribution most of the time but not always) marginal tax rate.
Furthermore, RRSPs are still a better choice if your withdrawal and deduction tax rates are equal. In fact, the tax treatment in that situation is equivalent to investing in TFSAs. It’s like paying no tax at all.
We are insisting on those points to make sure our position is clear, especially to anti-RRSP protestors. From a mathematical standpoint, RRSPs are better or at least equivalent in most situations. Then again, we know tax implications are not all that simple. We’ll even talk about some of them further down.
With those clarifications out of the way, the only interesting remaining question is how much higher your withdrawal tax rate has to be to tilt the advantage in favor of non-registered.
The ensuing puzzle is quite complicated as, among other things, the solution varies both with time and investment return. An equation with more than one variable is always difficult but maybe more fulfilling to solve.
Before we franticly dive into that, let’s talk about some other considerations.
Why Even Consider Non-Registered
In usual cases, retirement income will be lower than job or career income. So, most people will be subject to a lower marginal tax rate in retirement. In those circumstances, RRSP offer a preferential tax treatment for your investment.
Even so, in practice, one of you will eventually die. We know it’s not pleasant to talk about it, but it’s the sad truth. At that point, there’s a good chance the surviving spouse will have to deal with more taxable income even if the RRSP balance of the deceased can be smoothly rolled over to him/her. Without income splitting possibilities accessible to couples, the individual taxable income of the survivor will probably significantly rise and may tip over the OAS clawback annual limit ($77580 in 2019). Even though it’s not technically a tax, the OAS clawback may artificially boost your marginal tax rate by 15%. That’s not negligible.
On top of all this, the lone survivor will inevitably also pass away. This will result in the entire RRSP balance being taxable all at once at death at a very high tax rate. So, from a fiscal standpoint, RRSP may not necessarily be the optimal option.
Maxing Out TFSAs First
When we first got on the job market and for a lot of years, TFSAs were not even around. So, we got used to RRSPs being the sacred tax saving program. Young workers now have additional options like the TFSA, that is much more suited for their lower income level as they begin their career. TFSAs also give them much more withdrawal flexibility as it will never impact their taxable income. In that context, maxing out TFSAs first seems like a great idea.
Unless their employer matches up their RRSP contributions, young people should be better off waiting till their income reaches higher levels before to deduct their RRSP contributions. Some experts even suggest to still contribute to RRSPs now if you have accessible funds to do so but only to use related income deductions in higher income years. The RRSP program allows to carry forward deductions to any subsequent year. That way, your investments could still grow tax-sheltered right away. Tax-sheltered growth is considered the second main advantage of RRSPs after income deduction that is the first.
We have to confess that we did not understand or consider all tax liabilities that came with using RRSPs at a young age making less money. We already talked about our Reserve concerning The Home Buyers Plan. With retrospect, early contributions (or deductions should we specify) to our RRSPs were a mistake. They will inevitably result in withdrawals at a higher tax rate.
One minor argument persist is favor of RRSPs versus TFSAs, TFSAs won’t allow you to avoid the 15% withholding tax on US dividends like RRSPs do. For that matter, non-registered will also have that advantage over TFSAs. In non-registered accounts, you will first pay US dividend withholding tax but will get it credited back when you file your taxes. This is not possible with TFSAs. That withholding tax is still not that expensive. For instance, it will effectively only cost you 0.60% on a 4% annual dividend (4% x 15% = 0.60%).
Getting back to our original deliberation, TFSAs are generally better than non-registered. No tax or just a little tax if you consider US dividend withholding tax is always better than more tax, no matter how low it is.
So today, TFSAs are our first incline, particularly for our Canadian holdings and in lower income years. The tax treatment of TFSAs in retirement is also much simpler, especially if, like us, you plan on having higher income when you retire. We won’t necessarily need more money, but our taxable income will certainly be higher.
Higher Taxable Income in Retirement
In our expected retirement situation, because we’ll have plenty, it won’t be about how much will we spend but rather how much of our income will be taxable and which tax rate will apply.
Again, too much income is a good problem to have. It’s a pre-requisite that justifies our present quarrel between RRSPs versus other options. We also realize that those specific circumstances probably won’t apply to most people.
We’ve even explored the new thing courtesy of the latest federal budget: Advanced Life Deferred Annuities (ALDA) With an ALDA, however, Canadians should now be able to defer payouts until the end of the year their turn 85.
After looking at it, we see two problems with ADLAs in our situation. First, annuities nowadays equate to lower returns and we simply like higher returns better. Second, although it could help to control taxable income early in retirement, after 85, ALDAs would mean more imposed income, which we already have too much.
We still think ADLAs could be another interesting option to consider in different circumstances. For instance, they could potentially allow to defer more taxes after 85 and provide a safe alternative to deal with longevity risk.
After all that research, we essentially have circled back to non-registered as the best option.
Before we take it further, we will make a nuance about our calculations. We know that because of inflation, money spent now is not equal to the same amount spent in the future. We still have decided to ignore it to somewhat simplify calculations. In the end, what matters to us is how much more money can we give or leave our daughter and grandchildren.
When examining the question with that perspective, inflation relatively loses relevance. Only what will remain is important. We chose to focus on taxes as the main factor we can influence.
RRSPs vs Capital Gains
With probable higher income in retirement for us, we will try to calculate to tipping point (or points as we will see later) where the fiscal advantage begins to shift to non-registered versus RRSPs.
Generally, non-registered stock investing will generate two forms of significant taxable income: capital gains and dividends. The tax treatment of dividends is much more complicated and elaborate. It also varies quite a lot depending on your province of residence. So, for the sake of our analysis, we will assume only capital gains.
This will allow us to obtain the minimal spreads between tax rates (deducted vs withdrawal) necessary to favor non-registered. The fiscal impact of dividends should be similar but would make things even a little worse for non-registered. Remember that dividends are grossed-up by 1.38 for income tax purposes before any dividend tax credits kick in and that their biggest flaw in that regard is that they will make it more difficult to avoid OAS clawback.
I have a somewhat twisted nature, ambiguously lazy and hard-working at the same time. The relentless worker in me has been well served with this problem. To obtain meaningful results, a multiple-variable equation had to be solved by iteration. For those who are not familiar with it, this solving method consist of guessing an answer, looking at the related result and then adjusting your guess accordingly. You repeat that process again and again till you get a precise answer. I usually would have written a program to deal with all those iterations but decided to do it manually (still with a little help from Excel) for this equation. The same excruciating process had to be repeated about 50 times to produce results for the various times and returns. As I said, I sometimes can get crazy persistent.
Results vary with both time and expected return. To simplify calculations a tad, we also fixed the initial deducted tax rate at 38%. It corresponds (for 2018) to the middle-class combined marginal tax rate for taxable income between about 46K$ and 93K$ in our province. Most of my RRSP contributions were deducted at that level. For a while, my wife’s contributions have been deducted at a superior level for both her own RRSP and our spousal RRSP that will eventually be withdrawn under my name.
Even though the related chart doesn’t seem that elaborate, it still required some amount of work. So, with sweaty fingers, we quickly concluded two things: the spread increase with time but also with return.
Just to mention it, it wouldn’t be realistic for anybody to consider capital gains within a period of less than at least 5 years because that type of short-term investing would be too risky.
After all, it does not bode well for capital gains or non-registered as both these variables are normally quite high in our situation. Our time horizon is long, and our expected return will easily fall in the 6% to 10% range. Our tax-level spread is still just about 8.50% (46.5% vs 38%) with taxable income over 93K$. Only individual revenues over 144K$ would result in a spread over 12% (51% vs 38%).
The only factor left in favor of non-registered would be OAS clawback issues. But these complications won’t apply every year and certainly won’t even be considered if both spouses are still alive.
The only factor left in favor of non-registered would be OAS clawback issues. But these complications won’t apply every year and certainly won’t even be considered if both spouses are still alive.
Even a big chunk taxable at death all at once won’t significantly tilt the balance towards non-registered because it will take a very long time to get to that point and as our research shows, time only widens the spread to the benefit of RRSPs.
Non-Registered Only as a Last Resort
Maybe we did all this for nothing but at least, we got to the bottom of it. It did not amount to much for now, but we will persist and continue to look for ways to reduce our taxes in retirement and beyond.
Even if it’s not black or white because of more complex OAS clawback considerations, until proven otherwise, our initial incline to use non-registered only as a last resort will prevail.
Again, let’s mention some people have a false sense that capital gains are better than RRSPs because only 50% of capital gains are taxable. But half of some tax is still worse than no tax at all. In practice, most people will pay less tax with RRSPs. Even with higher income at withdrawal, provided some time and decent returns, RRSPs should still stay ahead of capital gains or non-registered in general.
For us, TFSAs will play a more prominent role as we will put more emphasis on maximizing them first. To some extent, we will even transfer some funds out of RRSPs to achieve this. Yet, with the perspective of this recent research, we will probably keep it more balanced and avoid using non-registered if possible. Hence, we will keep almost all the rest of our holdings registered in RRSPs. They are still a very interesting and viable option to minimize taxes.
We will continue to apply basic fiscal planning strategies like deducting (not necessarily in the contributing year) RRSP contributions in higher income years, and only withdrawing in lower income years. My expected decreased work schedule should provide plenty of opportunities to utilize reduced income years.
In that sense, keeping my taxable income just under the lower tax bracket upper threshold ($46605 in 2018) will remain a key in our broad fiscal strategy.
Photo Credit
No comments:
Post a Comment