SWRs, RRBs, and taxes

[scomac]

[Split out of Real Return Bonds by ModeratorA]

I am currently at about 10% RRBs of FI. (This amount happily coincides with the about 10% of FI represented by bonds with terms to maturity exceeding 5 years). If 10% RRBs is too low, what is the appropriate percentage and what is the argument for that percentage

This should be dependant upon the amount of long term liability that needs to be defeased via the fixed income portion of the portfolio. The greater the total portfolio allocation to fixed income investments then the greater the need (IMO) for RRBs. If we assume that 4% is a safe withdrawal rate then any asset mix that has <60% in equities should include RRBs to defease the long term liabilities (beyond 10 yr. that can be effectively defeased through a nominal bond allocation). If an investor has only 30% allocated to equities then it would make sense to allocate up to 30% to RRBs to help defease the long term liabilties

[ockham]

I am currently at about 10% RRBs of FI. (This amount happily coincides with the about 10% of FI represented by bonds with terms to maturity exceeding 5 years). If 10% RRBs is too low, what is the appropriate percentage and what is the argument for that percentage?

This should be dependant upon the amount of long term liability that needs to be defeased via the fixed income portion of the portfolio. The greater the total portfolio allocation to fixed income investments then the greater the need (IMO) for RRBs. If we assume that 4% is a safe withdrawal rate then any asset mix that has <60% in equities should include RRBs to defease the long term liabilities (beyond 10 yr. that can be effectively defeased through a nominal bond allocation). If an investor has only 30% allocated to equities then it would make sense to allocate up to 30% to RRBs to help defease the long term liabilties

I confess to ignoring all the "safe withdrawal rate" discussions. Is the idea here that under a range of reasonable assumptions, a 4% withdrawal rate is safe only with an equity allocation of ~60%, this equity allocation being needed to defend this withdrawal rate against the predations of inflation? And the next step is to observe that RRBs afford an even more reliable protection. Therefore, your equity + RRB allocation should approximate (not less than) 60%?

[scomac]

I confess to ignoring all the "safe withdrawal rate" discussions. Is the idea here that under a range of reasonable assumptions, a 4% withdrawal rate is safe only with an equity allocation of ~60%, this equity allocation being needed to defend this withdrawal rate against the predations of inflation?

In my interpretation, the 4% safe withdrawal rate provides for portfolio sustainability under a variety of assumptions and a variety of asset allocations. It is simply more-or-less a general rule of thumb.

And the next step is to observe that RRBs afford an even more reliable protection. Therefore, your equity + RRB allocation should approximate (not less than) 60%?

If you accept the 4% rule then we can assume that short term and mid term liabilities -- the next 10 years worth -- should be adequately covered, on a real basis, with fixed income/nominal bonds at a present value of 40% of the portfolio. The ravages of inflation which can greatly impact your long term purchasing power need to be hedged with investments that respond favourably to inflation. Traditionally this has been equities, real-estate and gold. More recently, the advent of inflation linked bonds have given us another lower risk method of off-setting that risk. Therefore the remainder of the portfolio needs to be made up of these sorts of inflation sensitive assets. For the risk averse investor who has a low tolerance to volatility and wishes to restrict their equity allocation to 30%, for example, he/she would need a healthy exposure to RRBs of up to 30% of assets. In essence, what I am advocating is a sliding scale approach to RRB allocation (as a percentage of total portfolio assets) that is dependant upon exposures to other inflation sensitive assets. The investor that is in accumulation phase and has 70% in equities won't need any whereas the investor in withdrawal phase that has 40% in equities and real-estate would require about a 20% weighting.

[Shakespeare]

If you accept the 4% rule then we can assume that short term and mid term liabilities -- the next 10 years worth -- should be adequately covered, on a real basis, with fixed income/nominal bonds at a present value of 40% of the portfolio

The 4% is inflation-adjusted withdrawal. To cover it with normal bonds and cash requires somewhat more than 40%, depending on what future inflation rate is assumed. If you use the BoC target rate of 2%, 10 years will cause a total inflation of 22%, which is too large to ignore.

Of course, one's personal inflation rate may be lower or higher than the CPI rate.

[marty123]

I confess to ignoring all the "safe withdrawal rate" discussions.

The 4% rule was never really fully accepted by everyone and "officially" defined, but the people that seem to push it seem to believe that without great market fluctuations (aka 1930s or 2007-20xx), the return on investment should cover both the 4% withdrawals and the inflation increase. For example, a constant 7% return combined with 3% inflation was going to preserve the inflation-adjusted capital until death. Worst case, to account for fluctuations and recessions, the 4% was generally supported by testing the withdrawal rate against historical market data to determine that the distribution was safe in "the real world". For example, a 55-year-old would enter numbers in an electronic calculator (Google "Firecalc") and his proposed distribution would be tested against historical market data from the 20th century. Until the current market crisis, a figure of 4% from a balanced portfolio was generally seen and back-tested as safe for a typical retiree (55-65) seeking to preserve the distribution until his/her 90s (possibly depleting all the capital by then though). I don't know what would happen when we consider back-test data that include 2008 and 2009, or whether anyone knows how long the markets will take before getting back to normal after the crisis.

Is the idea here that under a range of reasonable assumptions, a 4% withdrawal rate is safe only with an equity allocation of ~60%, this equity allocation being needed to defend this withdrawal rate against the predations of inflation?

If you subscribe to the 4% rule of thumb, then you'll realize that a 100% bond portfolio at the current rates cannot yield enough to cover both a safe 4% withdrawal and to allow for some of the interest to be reinvested to inflation-protect the capital. A $100 bond yielding 3.5% would be worth $99.50 after 4% withdrawal and $96-97 after the effect of inflation. In a matter of a few years, you'd lose quite a bit of inflation-adjusted capital. Under that same rule of thumb, the only way to achieve an annual return which is high enough to provide a "safe" 4% inflation adjusted withdrawal for several decades is to have a yield of 6-8% per year (inflation + 3-5%) so that you preserve enough of your annual growth to preserve inflation. With the bond returns we've seen in the past decade, you'd need to supplement your portfolio with a large percentage of equities (assuming you don't hit a crisis like the current one too often). It's possible that the few websites that back-test safe withdrawal rates against historical data have found that 60% equities was optimal in the few decades leading up to 2007.

which is better strip RRB or coupon RRB? (I think for coupon bonds, i get a payment 2x a year, stripped i don;t)

To go back to the recent question about strip vs no-strip: assuming both of them can be found, you'd use a strip if you perceive the yield to be high and :

1. You are young and want the entire invested amount and its yield to be inflation-protected until 20XX when you retire, annuitize or need the inflation-protected funds;
2. You may be older, but want to a certain amount fully kept and inflation protected for a certain number of years;
3. You don't want to bother with taking the small coupon payment, and reinvesting it in a new RBB every year.

You'd want to pick the no-strip bond if:

4. You are retiring and want or can use the cashflow;
5. You don't mind or want to periodically reinvest the proceeds in new RRBs or in a different investment.
6. The yield is so low that it doesn't matter: the invested amount will be inflation adjusted, and the cash it throws will be used elsewhere

A full bond is broken into a large strip that follows the original maturity (2021, 2026, 2031 & 2036 in the case of RRBs) and smaller strips that are the coupon payments. It would be easier to find strips with that are the original principal with an original maturity of 2021, 2026, 2031 or 2036, than it would be finding strips that represent the original coupons.

When I called TDW the yield where different (i forgot exactly)
where the stripped RRB has a higher yield.

The strip will be more expensive (have lower yield) if the real return of the bond was high (it was the case when real returns where 2.4-2.8% recently). You know that your principal will continuously be reinvested at today's high yield.

In theory, the strip would be less expensive (have higher yield) if the real return of the bond was very low because you can then reinvest the coupon in something else every year. In practice, it may not be the case because of the convenience, and the cheap cost of reinvesting a very small annual amount.

[scomac]

If you accept the 4% rule then we can assume that short term and mid term liabilities -- the next 10 years worth -- should be adequately covered, on a real basis, with fixed income/nominal bonds at a present value of 40% of the portfolio

The 4% is inflation-adjusted withdrawal. To cover it with normal bonds and cash requires somewhat more than 40%, depending on what future inflation rate is assumed. If you use the BoC target rate of 2%, 10 years will cause a total inflation of 22%, which is too large to ignore.

You are assuming that the coupons aren't being reinvested. A nominal ladder should cover off expected inflation when the coupon is reinvested over time. Another way of looking at it is to construct a stripped ladder with each wrung having a PV of 4% of assets thus representing the nut plus inflation upon maturity and the ladder is sustained by selling other assets to fund the purchase of a new bond each year as part of the rebalancing function.

[Shakespeare]

You are assuming that the coupons aren't being reinvested.

In withdrawal phase they often won't be unless the holder has a very large portfolio.

If you are running a roughly balanced portfolio, it's unlikely that the equity/REIT side will itself produce sufficient cash flow by itself unless the portfolio is sizeable. The 4% rule-of-thumb also is before expenses and assumes the portfolio can run to zero at age 95 in a worst-case scenario.

[scomac]

You are assuming that the coupons aren't being reinvested.

In withdrawal phase they often won't be unless the holder has a very large portfolio.

I realize that from a practicality perspective it will simply be easier to spend the coupons upon withdrawal. That said, if the investor understands that the bond coupons must be reinvested to preserve purchasing power then this can be addressed at rebalancing with an additional infusion of capital to renew the bond ladder. We are both correct, it's just that we are looking at the problem from different perspectives.

[Bylo Selhi]

The 4% rule was never really fully accepted by everyone and "officially" defined, but the people that seem to push it seem to believe that without great market fluctuations (aka 1930s or 2007-20xx), the return on investment should cover both the 4% withdrawals and the inflation increase.

The seminal studies, including Trinity, Bengen and Harvard endowment, all assume ~4% real annual withdrawals.

That said, the 4% is only a starting point, not some immutable law. It's useful for estimating how much you need to save in order to retire with a high probability that your portfolio will outlast you. But there are no guarantees. You still have to monitor your portfolio throughout your retirement, making minor adjustments, especially during trying times such as these.

[adrian2]

The 4% rule was never really fully accepted by everyone and "officially" defined, but the people that seem to push it seem to believe that without great market fluctuations (aka 1930s or 2007-20xx), the return on investment should cover both the 4% withdrawals and the inflation increase.

The seminal studies, including Trinity, Bengen and Harvard endowment, all assume ~4% real annual withdrawals.

Actually, the 4% rule does include all the US financial history in the past century or so, Great Depression et al.

That said, the 4% is only a starting point, not some immutable law. It's useful for estimating how much you need to save in order to retire with a high probability that your portfolio will outlast you. But there are no guarantees. You still have to monitor your portfolio throughout your retirement, making minor adjustments, especially during trying times such as these.

Ditto. I would add that, IMO, 4% was historically safe even starting at market tops (IIRC, even with 100% equity) and there was some debate if one can adjust it after market declines (i.e. if your portfolio has declined 25% from the top but you're starting withdrawals now, after the decline, you can increase the 4% by 1/1.25 (+ inflation adjustment) as if you had started withdrawals at the top. I'm not sure if i would go with a much higher adjustment, but IMO one could make an allowance for current market conditions.

[marty123]

The 4% rule was never really fully accepted by everyone and "officially" defined, but the people that seem to push it seem to believe that without great market fluctuations (aka 1930s or 2007-20xx), the return on investment should cover both the 4% withdrawals and the inflation increase.

The seminal studies, including Trinity, Bengen and Harvard endowment, all assume ~4% real annual withdrawals.

Actually, the 4% rule does include all the US financial history in the past century or so, Great Depression et al.

I didn't say it didn't. I said that except for 2-3 big crises, one would probably find that the principal is protected while, in the next paragraph which you did not quote, I said that one we consider the worst downturns, the likelihood of maintaining withdrawals is good, but the principal will get depleted.

The 4% rule was never really fully accepted by everyone and "officially" defined, but the people that seem to push it seem to believe that without great market fluctuations (aka 1930s or 2007-20xx), the return on investment should cover both the 4% withdrawals and the inflation increase. For example, a constant 7% return combined with 3% inflation was going to preserve the inflation-adjusted capital until death. Worst case, to account for fluctuations and recessions, the 4% was generally supported by testing the withdrawal rate against historical market data to determine that the distribution was safe in "the real world". ... (possibly depleting all the capital by then though).

It's also important to understand that these stats are based on history repeating itself. It would only take a Japanese-style bear market or a few years of stagflation to mess up these odds. None of these scenarios are far fetched.

[adrian2]

I didn't say it didn't. I said that except for 2-3 big crises, one would probably find that the principal is protected while, in the next paragraph which you did not quote, I said that one we consider the worst downturns, the likelihood of maintaining withdrawals is good, but the principal will get depleted.

The rule is targetting a Systematic Withdrawal Rate (i.e. for retirement) which ensures a portfolio survival for several (3?) decades, based on past US history, including the worst case in the past century. The limit condition is that you die broke. In the majority of cases, the portfolio end value will be significantly above zero, including, as you note, the principal surviving intact.

It's also important to understand that these stats are based on history repeating itself. It would only take a Japanese-style bear market or a few years of stagflation to mess up these odds. None of these scenarios are far fetched.

100% agreed.

[ghariton]

You can get an almost perfectly safe 4% withdrawal rate, after inflation, using 100% RRB strips.

I say almost perfectly safe because, as Bylo says, there is always some risk that the Government of Canada might default. That opens the possibility of including some TIPS, OATS, etc. If all of those governments default as well, I guess the only diversification worth while is the proverbial hundred acres far from civilization, a machine gun and lots of ammunition, etc.

If you can't get RRB strips and have to make do with ordinary RRBs, as I do, then there is some risk that, for my withdrawals, I might have to sell some RRBs below (inflated) face value. But the bonds I hold have a coupon of 4.25%. So I won't be selling very much in any given year, at the start. I will have to sell more as time goes on, but then time to maturity shortens as well, and so deviations from face value will decrease as effective maturity shortens.

Again, I am not advocating 100% investment in anything. However, I personally have a two-thirds allocation to RRBs. I think that unexpected inflation in the 5% - 7% range is a significant risk. That kind of inflation would significantly erode the value of a 5-year nominal bond, never mind a 10-year one. Long periods of negative returns on nominal bonds, and even T-bills, have happened before (most recently the 1970s).

George

[Bylo Selhi]

It's also important to understand that these stats are based on history repeating itself. It would only take a Japanese-style bear market or a few years of stagflation to mess up these odds. None of these scenarios are far fetched.

Of course. Ya [s]pays[/s]invests yer money and ya takes yer chances.

All investment strategies incur some risk. That's why 4% SWR is only a rule-of-thumb starting point for thinking about the issues. Not only do you have to consider some catastrophic event that requires George's "proverbial hundred acres far from civilization, a machine gun and lots of ammunition, etc." there's a small risk of some personal catastrophe like contracting a disease that may require 10s or 100s of $1,000 for treatment. How many 4% SWR portfolios can deal with a hit like that? Yet I'd argue that the risk of the latter is orders of magnitude larger than the former. (Yes you can buy insurance, but at certain ages or with certain pre-existing conditions the cost of the premium might consume most, all or even more than 4% of your portfolio.)

[Springbok]

When you bandy about the 4% withdrawal rate, shouldn't you be considering taxes?

Interest income gets taxed at a high marginal rate - lets say 35%. Even if the bonds are held in registered accounts and just the income is withdrawn. Our 4% becomes 2.6%. About same as inflation rate.

Are you considering this in the above discussion?

[Bylo Selhi]

Are you considering this in the above discussion?

Before retirement the bulk of your income comes from employment so it's taxed the same as interest. Once retired it depends. Pensions are ordinary income, as are RRSP/RRIF withdrawals.

But capital gains from the sale of securities as well as stock dividends are taxed more advantageously. Better still if you have bonds or GICs in a taxable account and you dip into capital, that's not taxed at all. So if you have a lot of this stuff, then for the same level of nominal income as before retirement, your after-tax net may be greater. It's certainly worth considering. (I'm sure steves will provide more info )

Our 4% becomes 2.6%. About same as inflation rate.

That doesn't account for investment fees like MERs or an advisor's AUM fees, say 2%. So in the worst case it's 4% minus 2.6% minus 2.6% minus 2%, er..., wait a minute!

[Shakespeare]

Interest income gets taxed at a high marginal rate - lets say 35%.

When the loonie is in your pocket and you pass it over to the store clerk, he or she doesn't care where it comes from. Therefore, although the various marginal rates can be used to calculate the average rate for a given asset mix and age, it is that average rate that should be used IMO. And that will be much less than 35%.

[scomac]

Our 4% becomes 2.6%. About same as inflation rate.

That doesn't account for investment fees like MERs or an advisor's AUM fees, say 2%. So in the worst case it's 4% minus 2.6% minus 2.6% minus 2%, er..., wait a minute!

...In the worst case it maybe preferable to opt for a prescribed annuity rather than continually line the advisor's pockets.

[Springbok]

Interest income gets taxed at a high marginal rate - lets say 35%.

When the loonie is in your pocket and you pass it over to the store clerk, he or she doesn't care where it comes from. Therefore, although the various marginal rates can be used to calculate the average rate for a given asset mix and age, it is that average rate that should be used IMO. And that will be much less than 35%.

My concern is that retirees might use the 4% to determine how much of their investment income will be available for spending. If this is the safe withdrawal rate, then the tax must be deducted. They need to be aware that the 4% will be taxed and they might only get 3% or so after tax and who knows what after inflation.

I am not so sure that using the average tax rate is correct. Retirees get a base income that would include CPP and OAS. Investment income withdrawal rate is discretionary and would be on top of the base income and would be taxed at the applicable marginal rate.

The other problem in using the 4% rule, is what is it a % of?

Is it 4% of the original capital.
Is it 4% of original plus inflation.
Or 4% of current portfolio value?

I always favored the latter - until the markets tanked.

I am assuming that anyone here would not be foolish enough to further reduce their potential withdrawals by paying an advisor to help them lose money.

[Jaunty]

Springbok wrote:

The other problem in using the 4% rule, is what is it a % of?

The one I've seen most is x% of the portfolio on a chosen annual date - ie. January 1st - adjusted for inflation.

[Bylo Selhi]

My concern is that retirees might use the 4% to determine how much of their investment income will be available for spending. If this is the safe withdrawal rate, then the tax must be deducted.

Why? Do you negotiate your salary and/or raises based on the top or bottom line of your pay stub? For most people their retirement income will be taxed at a lower rate than their employment income was when they were working. So if they used to live on a before-tax income of $xxk, then it doesn't make sense that they plan their retirement needs based on after-tax income of $yyk.

They need to be aware that the 4% will be taxed and they might only get 3% or so after tax and who knows what after inflation.

The 4% traditionally includes provision for indexing. It excludes investment expenses like MERs.

The other problem in using the 4% rule, is what is it a % of?
1. Is it 4% of the original capital.
2. Is it 4% of original plus inflation.
3. Or 4% of current portfolio value?

I'm not aware of any studies that use 1.
The studies I referenced above used 2.
Bill Bernstein has written about the pros and cons of thinking in terms of 3. He concludes:

One point cannot be made often enough -- when you retire, are you going to be withdrawing a fixed inflation adjusted amount on a regular basis, or are you going to be withdrawing a fixed percentage of your portfolio? This is not a semantic fine point. If you need a fixed amount, plan on withdrawing no more than about 4% of your starting amount in inflation adjusted terms. A fair dollop of bonds won't hurt in this situation.

If you can be more flexible and spend a fixed percentage of your nest egg each year, then you can indeed keep you entire retirement stash in stocks and spend 5% annually. Just remember that your stipend will likely fluctuate wildly over the decades of your retirement. Keep a few cans of Alpo in the cupboard if you decide to go this route.

I am assuming that anyone here would not be foolish enough to further reduce their potential withdrawals by paying an advisor to help them lose money.

That would be a valid assumption for people who have studied SWR. I suspect the vast majority of people who pay advisors via MERs have not. (I recall that in the late 1990s Templeton published SWR charts that went as high as 8% and 10% -- using a portfolio that consisted solely of Templeton Growth Fund. I wonder how investors who bought into that are doing now.)

[Shakespeare]

I am not so sure that using the average tax rate is correct. Retirees get a base income that would include CPP and OAS. Investment income withdrawal rate is discretionary and would be on top of the base income and would be taxed at the applicable marginal rate.

In that case the correct rate to use is the average rate (not the marginal rate) on the income above the CPP+OAS+DB (if any) funds.

Added: the differences can be considerable. From my 2008 return:

Marginal tax rate - 25%
Average tax rate - 7.4%
Average tax rate on investment income after subtraction of DB and CPP benefits - 10.4%.

[Springbok]

My concern is that retirees might use the 4% to determine how much of their investment income will be available for spending. If this is the safe withdrawal rate, then the tax must be deducted.

Why? Do you negotiate your salary and/or raises based on the top or bottom line of your pay stub?

I think the original was clear, but have illuminated it a bit .

I have not had a salary for, maybe, 30 years? Maybe longer. But, I will give it a shot.

When you receive your first paycheck, you suddenly find that it is less than you expected. After that you know what you can spend.

When you are planning retirement, you are just working with numbers and without running a trial tax return, don't know how much of that 4% you will actually receive. Just like your paycheck, you find out after the fact once everything else is dialed in.

Or you can run Steve's program!

[couponstrip]

When you receive your first paycheck, you suddenly find that it is less than you expected. After that you know what you can spend.

When you are planning retirement, you are just working with numbers and without running a trial tax return, don't know how much of that 4% you will actually receive. Just like your paycheck, you find out after the fact once everything else is dialed in.

Indeed, I ran into this problem when I decided to put together my own spreadsheet and run some numbers. Once I got into the detail of the analysis, I started to realize how complicated this could get based on both allocation and the taxation of that allocation.

Consider two colleagues who both want to retire and and have determined that their retirement needs will require $2.5 million. They both arrive at 2.5 million in the same month, 500k in RRSP, and 2 million in taxable accounts, and both retire. However, one investor started buying 4 ETF's 30 years ago (XIU, VTI, VEA, XBB), and has never rebalanced or sold anything. Investor B has jumped around over the years moving from advisor to advisor with moderate success and has mutual funds all of which were bought in the last two years plus some bonds. They both may have 2.5 million in the kitty, but Investor A has a much larger tax liability on withdrawal than investor B.

Trying to consider what tax might look at 2 or 3 (or 5-6) decades down the road, and the picture gets even murkier. As long as you guess that the taxation will be something different than what is now, then at least you will have some chance of being right.

Finally, in retirement, as you have highlighted, the taxation of your gains should be considered since this can vary significantly from a 100% bond portfolio to a bond-preferred-equity portfolio to a 100% equity portfolio.

FWIW, I posed this question sometime over a year ago on this forum, and the fairly unified response I received was "to do that type of complicated determination, you need advanced software like RRIFmetic". The permutations and combinations do make it a fairly complicated bit of DIY spreadsheet work. But I do agree that your tax concerns are not trivial. IIRC, after a rough calculation, I figured that at the extremes, the amount of disposable income available in retirement could vary by as much as 100% if allocation and taxation are ignored.

[queerasmoi]

How about, for a rough calculation, to include in your portfolio a tax liability for existing capital gains and RRIF/RRSP assets that will be taxed upon removal? Overestimate it slightly to be on the safe side. Take your 4% of *that* and consider it your safe withdrawal.