To Put Money Down on a House...Or To Fill a Very Empty RRSP?

[Bullseye]

The key part of that article;

'Since about half of GIS recipients pay income tax, the combined effect of paying tax on the RRSP payments and the reduction of GIS benefits results in an unexpectedly horrendous situation – an effective marginal tax rate of 75 per cent on those RRSP benefits. Call it a clawback by another name.

Little wonder that a 2003 study by the C.D. Howe Institute called RRSPs a "poor investment" for those at the lowest end of the income spectrum. Many groups are calling for Ottawa to bring in Tax Prepaid Savings Plans (TPSPs) as one solution to the GIS/RRSP clawback. Under TPSPs, contributions are not tax deductible. But when money is eventually withdrawn from these plans in retirement, it is tax-free.

People should also be aware that once the GIS disappears, so do some other benefits like drug and rent subsidies and provincial aid programs. For example, the one-time federal Energy Cost Benefit of $125 was sent to eligible seniors in early 2006 – but only to those who qualified for the GIS.'

So in effect, a modest RRSP can be worse than no RRSP at all (because your effective marginal tax rate could be much higher than it was when you were working). This is why I recommend the OP's mother pay down the mortgage, then put money into non-registered investments after that. She can draw on the investments for income, but only pay taxes on the gains.

RRSP's are only good if you can be sure your tax rate will be lower in retirement. Depending on how much this woman could potentially accumulate in savings, her effective rate may indeed be higher. Especially as she doesn't appear to be a saver by nature.

[pitz]

So, is it the case that if you ONLY have income of a few thousand you are hurt the most? I don't want to retire and try to only earn a few thousand just so I can receive full government benefits... I'd rather retire on $60k/year including CPP/OAS.

Ideally you'd manipulate your finances in such a way that you gain access to wealth, but not have to actually call it 'income' for tax purposes. Focussing solely on 'income' , rather than wealth creation/building, can be downright dangerous (just ask income trust unitholders recently).

For instance, the imputed rent from owning a home with a paid-off mortgage provides a substantial return that cannot be taxed by the government because technically its not 'income' per se. The same home could also support a secured line of credit if additional augmentation is required -- also not called 'income' per se.

Non-registered investments can be pledged as collateral and borrowed against, again, not income, but spendable money nonetheless. Can't do that with a RRSP. In fact, the government dictates the schedule under which you can use your RRSP, and you are penalized if you do not use enough of it.

Take Mr. Warren Buffet, for example. His personal holdings consist primarily of a house, a car, and billions of dollars of Berkshire Hathaway stock (a stock that does not pay dividends). If he was a Canadian, and decided to retire as the manager of BRKA, Mr. Buffet could collect GIS/OAS/CPP/GST rebates, receive subsidized prescription drugs, and all the other perks. He could borrow literally billions of dollars against his holdings of Berkshire Hathaway, to indulge in a rich lifestyle, still receive Canadian welfare, and not pay much income tax. Pretty good scheme for a billionaire, eh?

Ideally you want to be Mr. Buffet, not some misguided person who is trying to 'retire' with some meagre RRSP savings and still paying off a mortgage. It is much more efficient, in the long run, to accumulate unencumbered equity, than it is to be shackled by the restraints of taxes and their very regressive nature.

[Xtreme2001]

Pitz, I always enjoy reading your posts. Very informative.

I have to complete my own research on this, but could you briefly tell me how borrowing against one's holdings works? Where would this be most beneficial? Are rates lower if you borrow against holdings? In your buffet example, how does he get away with everything, even though he has billions in one particular stock?

Thanks

[don242]

Great information everyone.

Obviously financial planning depends on each individual person's situation which makes it difficult to make any proper recommendations, but there has been lots to think about.

I do however, stand by my view that you should not plan to rely on social assistance programs (and the fact that millionaires can, shows not only the great planning by these people but also the flaws in a system that is meant to help those with less).

[Odysseus_Maximus]

Go to a casino

[majesus]

It is much more efficient, in the long run, to accumulate unencumbered equity, than it is to be shackled by the restraints of taxes and their very regressive nature.

WOW... VERY VERY solid point.
That's PITZ

[Bullseye]

I have to complete my own research on this, but could you briefly tell me how borrowing against one's holdings works? Where would this be most beneficial? Are rates lower if you borrow against holdings? In your buffet example, how does he get away with everything, even though he has billions in one particular stock?

Thanks

Non-registered investments are an asset, so they can be borrowed against, the lender simply puts a lien on them. This gets you a better interest rate. Some people use it to leverage even more investments, but in Pitz's theoretical example, an investor could also borrow against it to generate a non-taxable income stream.

Our tax system is flawed in the sense that it taxes income, not wealth, and that is what Pitz is highlighting. A similar real life example would be that guy who retired at 34 and wrote a book about it (Derek Foster - Stop Working). He has a half million in stocks and a paid off house, but he only lives on the dividend payments from those investments. Because dividends are taxed very favourably, he is able to live decently on fairly little. He also qualifies for many social benefits, because he is essentially 'low income', which adds to his standard of living (at taxpayers expense, mind you! But the ethics of that is for another post). So he's wealthy in assets, but 'poor' in income according to our tax system, because he's manipulated that system to work to his best benefit.

[mork]

This thread has been an interesting read. Something that occurs to me is that there is some assumption that the retiree is going to draw on their RRSP at a steady 'equalized' pace.

Could a retiree not just simply withdraw their entire RRSP, pay all taxes owed on it, and do whatever with the money (some non-RRSP investment vehicle).

Then.. in subsequent years of retirement they may have the ability to demonstrate a very low income, and as a result receive some of the government benefits discussed here.

Are there some constraints associated with RRSPs to prevent this?

I realize withdrawing the large sum in a single year would guarantee that it will be taxed at a higher rate, but perhaps there is some balancing point with the benefits realized in future low-income years?

[pitz]

This thread has been an interesting read. Something that occurs to me is that there is some assumption that the retiree is going to draw on their RRSP at a steady 'equalized' pace.

Could a retiree not just simply withdraw their entire RRSP, pay all taxes owed on it, and do whatever with the money (some non-RRSP investment vehicle).

The problem that then occurs is that the effective incremental rate of tax on the RRSP is elevated over its entire lifespan.

We can model taxation on RRSPs as follows:

eff tax rate = (1 - j + {[(1+j)*(1+k)^n]*(1-m)}^(1/n))/k

For RRSP holdings for n years, with an annualized return of k, and incremental tax rate of j at time of contribution, and incremental rate of m at time of withdrawal.

Obviously the computed savings associated with a RRSP is at its maximum when n=0, and the effective tax rate drops to 0 when n = infinity.

A 10-year-holding period on a RRSP still generates an average tax rate on RRSP investment greater than what one could achieve using tax-deferred non-RRSP equity investment. After 10 years, the effective tax rate on a RRSP for someone in the 40% tax bracket is still 20%. The same rate immediately achieved by investing in equitiees that produce capital gains. Of course, deferred capital gains follow a very similar curve.

So to achieve any tax savings in a RRSP, you have to hold for a very long period of time, in the order of at least 20 years, if not longer, just to equal the treatment of non-registered capital gains in terms of average annual tax rates.

So if one plans on cashing their RRSP out early in its entirety, not only do they end up paying at the top marginal rates (because their income is spiked), but they never do realize effective tax rates below what would have been achieved using a simple buy and hold non-registered stock strategy. Plus their money is locked up and cannot be leveraged to boot.

You're a physicist, so the math shouldn't be difficult, but private message me and I can send you a spreadsheet with the above calcs if you want.

Then.. in subsequent years of retirement they may have the ability to demonstrate a very low income, and as a result receive some of the government benefits discussed here.

Sure, but the question is one of optimization of current savings, and which structure is the most appropriate if it is likely that such a scenario is contemplated.

Are there some constraints associated with RRSPs to prevent this?

Not really. Some RRSPs/RRIFs are locked-in, however, and are subject to minimum and maximum yearly withdrawals under the Income Tax Act and provincial legislation. These plans are usually associated with commutation of a defined benefit pension plan, or are otherwise, workplace sponsored pensions.

I realize withdrawing the large sum in a single year would guarantee that it will be taxed at a higher rate, but perhaps there is some balancing point with the benefits realized in future low-income years?

There are always optimizations possible, but also keep in mind that tax policy is constantly shifting in Canada. Just a little more than a decade ago, capital gains were taxxed at the same rates as income. Now the Tories are talking about further reductions in capital gains taxes, and have implemented an enhanced tax credit associated with Canadian-sourced dividends. Also planning optimizations that might work in one set of market conditions, may not work in another (such as leverage).

[83_gemini]

Actually I think lower taxes on dividends and capital gains are good-they encourage investment. What's needed are effective consumption taxes combined with credits for those with low incomes, so people who are wealthy, live off dividens and consume rather then invest pay appropriate levels of tax.

Canada's tax system is not particularly progressive in any case; average rates are about 33% across the board, and marginal rates can be very wonky depending on income thresholds.

[cannon_fodder]

We can model taxation on RRSPs as follows:

eff tax rate = (1 - j + {[(1+j)*(1+k)^n]*(1-m)}^(1/n))/k

For RRSP holdings for n years, with an annualized return of k, and incremental tax rate of j at time of contribution, and incremental rate of m at time of withdrawal.

Obviously the computed savings associated with a RRSP is at its maximum when n=0, and the effective tax rate drops to 0 when n = infinity.

A 10-year-holding period on a RRSP still generates an average tax rate on RRSP investment greater than what one could achieve using tax-deferred non-RRSP equity investment. After 10 years, the effective tax rate on a RRSP for someone in the 40% tax bracket is still 20%. The same rate immediately achieved by investing in equitiees that produce capital gains. Of course, deferred capital gains follow a very similar curve.

So to achieve any tax savings in a RRSP, you have to hold for a very long period of time, in the order of at least 20 years, if not longer, just to equal the treatment of non-registered capital gains in terms of average annual tax rates.

Where did you find this formula? Do you know how it was justified? Does this factor in the tax refunds being reinvested?

Based on the above statement, it suggests that one might be better, as one nears retirement, to not invest in an RRSP. On the other hand, if you withdraw from your RRSP/RRIF for more than 20 years, one could always think that the last dollar in is the last dollar out - thus giving a rolling 20+ year window.

[pitz]

Where did you find this formula? Do you know how it was justified? Does this factor in the tax refunds being reinvested?

I developed the formula. Its pretty simple:

1) The amount of a RRSP contribution, on a pre-tax basis, is equal to (1+j)
2) That RRSP contribution compounds at the rate of k%/year, for n years, thus (1+j)*(1+k)^n
3) A withdrawal from the RRSP is made, and is taxed at the incremental rate of m%, in other words, leaving (1-m) of the value of the RRSP remaining.
4) Therefore, we can find the annualized rate by taking the result of (3), and finding the n-th root (ie: ^(1/n)).
5) This result can be expressed as a ratio of the rate of return, to thus determine a tax rate, by subtracting 1, and then dividing by the pre-tax rate of return, k.

Not quite grade 9 math, but not very difficult once you work through it.

Based on the above statement, it suggests that one might be better, as one nears retirement, to not invest in an RRSP. On the other hand, if you withdraw from your RRSP/RRIF for more than 20 years, one could always think that the last dollar in is the last dollar out - thus giving a rolling 20+ year window.

Indeed. RRSPs are a terrible investment if you buy, and only hold for 10 (or die within) 10 years, compared to tax-deferred equities.

Basically you need to optimize for conflicting goals. For example, young investors ultimately benefit the most from the long-term tax-deferred compounding of RRSPs, but they also tend to have the lowest tax rates. Whereas older investors often do not have enough time in order to make RRSPs worthwhile. Undoubtely there is a sweet spot, probably occurring in one's 30s and early 40s.

[pitz]

Here's a graphical view of RRSPs versus ideally tax efficient non-RRSPs for an individual who contributes at a 35% incremental tax rate, and withdraws at a 35% incremental tax rate:



Note that one is clearly better off, return wise, with the non-RRSP investment until approximately 13 years has passed. Cashing out RRSPs earlier than 13 years is financially stupid.

And the two curves pretty much remain fairly converged after 13 years, which means you are not really extracting a premium for locking up your equity within the RRSP. Whereas, if you kept the equity outside of the RRSP, you could be using it to lower your cost of credit insofar as borrowing is concerned, and you could be applying leverage to your investments.

Of course, one is far worse off with the RRSP if they are paying a higher effective tax rate in the withdrawal phase than they were paying in the accumulation phase.

So aside from the short-term pyschological effect of saying 'up yours' to the taxman, there is little reason to accumulate anything but fixed income investments in a RRSP. And even then, if one has a mortgage, they shouldn't be buying fixed income anyways -- they should be merely paying down their mortgage, as it doesn't make sense to be simultaneously a lender, and a borrower, when borrowing costs significantly more than the returns on lending (due to varying credit spreads, or a flat/inverted yield curve).

edit: the graph also proves my other point, that only people in their 30s and early 40s should consider investing in RRSPs at all, over and above amounts that can be withdrawn under the HBP/LLP. If you invest in a RRSP when too young, you've locked up your money for all those years for no additional return (as the curves completely converge after 40 years anyways). If you invest in a RRSP when too close to retirement, you never do achieve the advantage of the tax deferral inside the RRSP.

a further note: non-RRSP savings can be accumulated in tranches, and removed on a LILO basis, highest cost-base first, so as to maximize long-term tax deferral, whereas no such optimization is possible on a RRSP. Borrowing can also be used to defer tax until death. This is not possible with RRSPs/RRIFs in the withdrawal phase.

[cannon_fodder]

I've checked twice, and I seem to have copied your formula exactly but it is giving answers which don't make sense so I must not be understanding the results. I would have expected an answer less than j but I'm getting answers in the 20's (i.e. I put in j=.463, k=0.08, m=0.32 and n=1 to 25).

Regardless, I would like to understand the following:

1) The amount of a RRSP contribution, on a pre-tax basis, is equal to (1+j)

Assuming incremental tax rate is marginal tax rate, why not equal to 1/(1-j)? The grossed up amount?

It is March 1st and I have $1k in my bank account. I can invest that in a non-registered account or I could borrow $666 and add that to my $1k and contribute it to my RRSP. I efile immediately and get my return about the 2nd week of March (except for the borrowing, this is what I do - if I had to pay taxes, I'd wait until April 30th to submit). My tax refund comes in, and because I'm in the 40% tax bracket, I get my $666 back which I use to pay down my loan. (Whether you actually get a loan or not I think is irrelevant, especially considering the interest on such a short lived loan would be about $1.67). So, I have $1,666 working for me in my RRSP as opposed to $1,400 or (1+j).

2) That RRSP contribution compounds at the rate of k%/year, for n years, thus (1+j)*(1+k)^n

In my scenario it would be (1/(1-j))*(1+k)^n

3) A withdrawal from the RRSP is made, and is taxed at the incremental rate of m%, in other words, leaving (1-m) of the value of the RRSP remaining.

I misinterpreted what you said here. In case anyone else does, in pitz assumptions, you are withdrawing the entire value of whatever the contribution grew to, e.g. if you contributed $1 and it grew to $3.45 you are withdrawing the entire amount. By saying "the value of the RRSP remaining" I thought pitz meant that there was some $ left in the RRSP after the withdrawal.

4) Therefore, we can find the annualized rate by taking the result of (3), and finding the n-th root (ie: ^(1/n)).

It would be helpful if you would be very specific. Remember, we don't all think like you and understand where you are coming from. You mean you get the annualized growth rate net of taxes do you not?

5) This result can be expressed as a ratio of the rate of return, to thus determine a tax rate, by subtracting 1, and then dividing by the pre-tax rate of return, k.

I'm tired so perhaps it is my error, but your description doesn't match the formula. Your formula adds 1 and subtracts j to/from the result after step 4 and then divides by k. Hopefully you can address what you are trying to achieve because at this point I'm not sure it matters.

The reason I say that is simple - if someone came to you and said they could put you in two investment solutions that would grow and your choice would be to pay no tax, or pay your marginal tax rate what would you want to know? After taxes, how much will I have in either scenario?

If after the 1st, 2nd, 3rd, nth years I have more money, after taxes, by investing in my RRSP versus a non-registered plan, isn't that a better measurement of investment success rather than which requires me to pay more taxes?

Based on breaking down your formula, it is clear that you end up with more money after taxes with the RRSP scenario unless your marginal tax rate when you deposit is significantly less than your marginal tax rate when you withdraw.

I believe whenever I've seen arguments in favour of non registered investing over RRSP's there are two (sometimes not explictly stated) premises:
1. The subject doesn't gross up their RRSP contribution or equivalently they don't reinvest the tax refund. In fact it is never used to do anything useful like pay down credit card debt, the mortgage, etc. It somehow "disappears".
2. The marginal tax rate differential between when the money goes in and comes out is nominal at best and, at worst, higher when coming out.

I hadn't taken the time to actually run some numbers as I always assumed RRSP's were the best first choice. The recent articles/posts warning about government benefit clawbacks and the tax advantage of non RRSP investing over RRSP investing had caused me concern. Now I feel more assured.

The unassailable advantage of an RRSP (as long as your marginal tax rates don't skyrocket from working to retirement) is the tax refund. Compounding tax free is a benefit on a downward sliding scale as one moves from interest producing investments to straight equities even turning negative if you have to sell at a loss.

I really appreciate (as do many others I'm sure) the time you take to post in this forum. It is good to have one's ideas challenged, get the mind thinking and not just accepting.

[cannon_fodder]

Here's a graphical view of RRSPs versus ideally tax efficient non-RRSPs for an individual who contributes at a 35% incremental tax rate, and withdraws at a 35% incremental tax rate:

Note that one is clearly better off, return wise, with the non-RRSP investment until approximately 13 years has passed. Cashing out RRSPs earlier than 13 years is financially stupid.

And the two curves pretty much remain fairly converged after 13 years, which means you are not really extracting a premium for locking up your equity within the RRSP. Whereas, if you kept the equity outside of the RRSP, you could be using it to lower your cost of credit insofar as borrowing is concerned, and you could be applying leverage to your investments.

Of course, one is far worse off with the RRSP if they are paying a higher effective tax rate in the withdrawal phase than they were paying in the accumulation phase.

So aside from the short-term pyschological effect of saying 'up yours' to the taxman, there is little reason to accumulate anything but fixed income investments in a RRSP. And even then, if one has a mortgage, they shouldn't be buying fixed income anyways -- they should be merely paying down their mortgage, as it doesn't make sense to be simultaneously a lender, and a borrower, when borrowing costs significantly more than the returns on lending (due to varying credit spreads, or a flat/inverted yield curve).

edit: the graph also proves my other point, that only people in their 30s and early 40s should consider investing in RRSPs at all, over and above amounts that can be withdrawn under the HBP/LLP. If you invest in a RRSP when too young, you've locked up your money for all those years for no additional return (as the curves completely converge after 40 years anyways). If you invest in a RRSP when too close to retirement, you never do achieve the advantage of the tax deferral inside the RRSP.

a further note: non-RRSP savings can be accumulated in tranches, and removed on a LILO basis, highest cost-base first, so as to maximize long-term tax deferral, whereas no such optimization is possible on a RRSP. Borrowing can also be used to defer tax until death. This is not possible with RRSPs/RRIFs in the withdrawal phase.

Beyond stating what I did in the previous post, your conclusion is based on a marginal tax rate that stays the same, i.e. does not get lower in retirement. I'd say that this is more likely the lower the tax bracket in which you are today, but less likely the higher the tax bracket in which you are today.

I like your point about the best fixed income investment whether in/out of an RRSP... it is called paying down your mortgage.

Based on the varied sophistication of the forum readers, and the good deal of credibility that is accorded you, it may be wise to be specific, and not general, when stating conclusions. Someone without a critical eye might look at your post and conclude that they should only invest in an RRSP if all of the following criteria are met (with possible relaxing of these 'rules' for HBP/LLP scenarios):

- you are in your 30's or 40's
- you leave it in for at least 13 years
- you will not have a higher marginal tax rate when retired as compared to working
- you have no mortgage
- you only buy interest income investments for your RRSP

If this is your recommendation then I'm sure it will be challenged without promise of consensus.