An Assiduous Reader writes in and says:
My background is in insurance and mathematics {now retired at 56}. I am intrigued by the perpetual preferred shares offered by quality institutions rated Pfd 1 or 2 {about 12 to 15 issuers}. With the higher yield and the Canadian dividend tax credit, I am comparing this investment strategy to purchasing a joint life-time annuity with a guaranteed period of 25 years. On the surface, the rate of return is about 6% annually. I recognize the annuity has a return of capital component.
So here’s my riddle…why wouldn’t I buy the quality perpetual preferreds for the ongoing income whereby I can transfer the shares to my spouse [or vice versa] and, likewise, to our children in perpetuity – unlike the annuity that would cease upon on our last-to-die joint death and/or 25 years. The fluctuation in the capital value of my preferreds would not concern me – anymore than the annuity where the entire capital is ‘lost’ upon the initial purchase.
Lastly, if I wanted to ‘boost’ my yield on my preferred portfolio, I would re-invest the dividends for a few years to a level of income that I would like [eg 10%/year} and then that would be the rate of return in perpetuity. If this 10%/year is approximately the long-term equity stock market return, why would I even bother taking on the extra risk with equities and the aggravation of trying to select the appropriate securities for my portfolio.
If this makes sense, then my entire investment strategy should be with perpetual preferreds. {I have other significant assets}
Maybe I am missing something so, thus, my email to you. This ‘riddle’ is starting to keep me up at night!
I am not particularly comfortable in this field of investment decision-making, so take everything below with a grain of salt and do your own research. But I’ll give it a stab! Any corrections or elaborations will be gratefully received.
But first, let me say that the Inquiring Reader is on the right track. Preservation of Income – as opposed to Preservation of Capital – is what preferred shares are all about.
Morningstar publishes life annuity rates; a sixty-year-old male is looking at an annual payment of about 7.2% of principal; at 70 it’s about 9%; and at 75 it’s about 10.8%.
According to Standard Life, the income is taxed as regular income:
Registered
The annuity payment is fully taxable
Non-registered
Only the interest portion is taxable
The taxable portion can be reported on a “prescribed” or “non-prescribed” basis
- Prescribed: level taxable portion each year
- Non-prescribed: taxable portion changes each year (interest reported each year reduces)
The prescribed taxation basis is attractive to taxpayers as it allows for the deferral of taxes. It is regulated and can only be used with specific types of annuities. All other annuities must be on a non-prescribed taxation basis.
To estimate how much of the annuity payment is return of capital, I used the Canadian Business Life Expectancy Calculator with the following data:
| Life Expectance Calculations | |||
| Data Required | Age 60 | Age 70 | Age 75 |
| Age | 60 | 70 | 75 |
| Sex | Male | ||
| Height | 6’0″ | ||
| Weight | 190 lbs | ||
| Frame | Small | ||
| Physically Active? | Somewhat | ||
| Level of Stwess | Low | ||
| Smoking | Less than 2 packs a day | ||
| Drinking Habits | Never more than three drinks | ||
| Eat Saturated Fats | How the hell should I know? I’m a GUY, ferchrissake. Call it once or twice a week | ||
| Elevated cholesterol | No | ||
| Normal blood pressure | Um … when talking about regulators and politicians? Or other times? Call it yes | ||
| Two subquestions skipped | |||
| Parents lived long? | Yes | ||
| Siblings with bum tickers? | No | ||
| Accidents or speeding tickets? | No | ||
| Post-Secondary | Yes | ||
| Not poor? | yes | ||
| Safety belt? | Even in bed! | ||
| Estimated Lifespan | 80 | 83 | 85 |
So at age 60, we’ll say 20-years to go; 13 years to go at age 70; and 10 years left at age 75.
Some quick work with MS-Excel, with the assumption that the capital is all gone at the end of the annuity results in required yields of 3.6%, 2.3% and 1.4%. We’ll summarize this in another table:
| Annuity Rates and Required Return | |||
| Age | Years Left | Annuity Rate | IRR |
| 60 | 20 | 7.2% | 3.6% |
| 70 | 13 | 9.0% | 2.3% |
| 75 | 10 | 10.8% | 1.4% |
Holy smokes! I’ve definitely made a mistake somewhere … could be the assumptions, the math, or the fact that I didn’t go into the insurance business.
However, before we leap wholeheartedly into PerpetualDiscounts as life-annuity substitutes, let’s take a look at the risks:
Credit Risk: Annuities are a far more senior claim on the insurers than preferred shares, especialy since – as far as the insurers are concerned – an annuity is a claim on the operating companies assets, while a preferred share is a claim on the parent. It is entirely possible that in times of trouble, a preferred shareholder could get nothing while an annuity holder could get paid in full … even in the absence of a government bail-out.
Return Order Risk: An annuity withdraws principal on a steadily increasing basis – even if that basis has to be calculated on a post hoc basis. Thus, if we are performing a direct comparison, we also have to withdraw principal from our preferred share portfolio on a steady basis. This means we are exposed to Order of Returns Risk. And that’s even before we consider:
Principal Evaporation Risk: With an annuity, the insurance company takes the risk that you will last longer than expected, and covers it with their chances that other clients will make up for it. With a preferred share portfolio – or any investment portfolio – you’re the one stuck with that risk.
Call Risk: Say preferred share yields fall dramatically and your shares get called. This will definitely foul up your long-term returns because your returns after the call date will reflect the coupon of your PerpetualDiscount, and not the initial yield – and that’s even before you account for frictional costs of the process.
As noted in the comments, this is overstated. Your yield will go down but, to at least some extent, your capital will have increased on a call. However:
- There will be capital gains tax to pay
- There is no guarantee that suitable replacements will be available
- A call will normally take place when the issuer can refinance cheaper, so there will be a yield hit to reflect “cheaper”.
Tax Risk: The tax regime for dividends could change, eliminating at least some of the dividend advantage
Inflation Risk: This will be about the same for both strategies, but you do have the option to buy an indexed annuity, whereas there are no indexed preferred shares at present. At some point, a deeply discounted FixedReset with a microscopic spread against five-year Canadas might be functionally equivalent, but we don’t have any of those yet. Other floating rate perpetuals (Ratchet, FixedFloater, Floater) might be considered equivalent, but then you have basis risk (either prime or five-year Canadas vs. inflation) and extant non-FixedReset Floating Rate issues don’t have sterling credit quality.
All in all, the risks are significant, but the returns are certainly juicy. I would advise that annuities are good for the bare-bones-beans-on-a-hotplate portion of retirement income, while preferred shares – and other investments – provide the income that you spend in Florida.
There are probabily mistakes in the above – this is not a topic I spend a lot of time on. My job is to take the investor’s allocation to preferreds and do a better job with it than he could himself – not to decide on the allocation. Any commentary will be be appreciated.
Update: See also Lifetime Financial Advice: Human Capital, Asset Allocation, and Insurance, Ibbotson, Milevsky, Chen & Zhu, ISBN 978-0-943205-94-6
Update, 2010-3-17: See also the So you are going to buy an annuity. With what? discussion on Financial Webring Forum.
Update, 2010-3-17: Another good article is Annuity Analytics: How Much to Allocate to Annuities? by Moshe A. Milevsky.
Update, 2010-3-19: There’s a good table in Milevsky’s Annuitization: If Not Now, When?:
| Value of Unisex Mortality Credits: Assuming 40m/60f (static) Annuity 2000 Table at 6% net interest. |
|
| Age of Annuitant |
Spread Above Pricing Interest Rate (in Basis Points = 1/100 %) |
| 55 | 35 |
| 60 | 52 |
| 65 | 83 |
| 70 | 138 |
| 75 | 237 |
| 80 | 414 |
| 85 | 725 |
| 90 | 1256 |
| 95 | 2004 |
| 100 | 2978 |
| Source: The IFID Centre calculations | |
Update, 2010-3-25: Interesting conclusions and charts in Kaplan’s Asset Allocation with Annuities for Retirement Income Management
Good summary …. among the risks I would include Liquidity Risk against the annuity. Once you buy that’s it, you are locked in whereas the preferred shares can always be sold in case of change of plans – emergencies, to fund long term care etc.
I don’t get the impression that indexed annuities are very available. None of the insurance comp websites talk about them. Based on fact that inflation protection for other products such as LTC and Critical illness insurance offer such protection in the form of a choice of an arbitrary 2 or 3% as opposed to matching unknown CPI increases and the cost is a lot more, I doubt there is much benefit to an indexed annuity. Unlike the basic spreading of the risk of when you die amongst the annuitants, which the insurance company is uniquely able to do and the individual cannot, the insurance company has no special advantage for inflation.
You may want to change the external link URL for “Lifetime Financial Advice: Human Capital, Asset Allocation, and Insurance” and remove the last part, leaving it as http://www.cfapubs.org/doi/pdf/10.2470/rf.v2007.n1.4580 — otherwise, on a browser which has not visited the URL before, one gets a message similar to http://www.cfapubs.org/cookieabsent.html, i.e. “An error occurred during attempt to set your user cookie”
Adrian
I would add two points
1. The reader hints that the funds were in a taxable account, in which case the pref approach with, say 5.5% yields from top quality issuers is quite stellar. In a RRSP/RRIF, one can get some 5+% very long term yields on provincial and excellent quality corporate bonds with considerably less risk than prefs.
2. From my experience, the back of the envelope calculation for annuities seems more consistent if you add a couple of years to the expected life (as an insurance profit — they will say adverse selection) and use the average Canada bond yield over the period (e.g. a 10-year yield for a 20-year payout). That might get you a better explanation of your IRRs. [My guess is the reader knows more about this than us!]
Given that the insurance company deliberately assumes we will live longer and this reduces IRR, the only people who really need an annuity are those who would run out of money if they lived longer than expected. This can be simulated by amortizing with your spreadsheet to age 91 or even 100. What you find is that, like a mortgage, the annuity payments are not sensitive to changes in long time term.
To answer the original reader’s question, I agree that all James’ noted risks are relevant, but the two largest (in comparison to an annuity) are credit risk (including 100% concentration in financials with the pref strategy) and call risk (if perpetual pref rates fall, 6% yields will disappear and be replaced with something else). These are both systemic risks undiversified by this approach. They may be worth taking to some degree in a taxable account, but bonds might be better in an RRSP or RRIF. Indeed, if the reader has both taxable and registered accounts, he can do both and achieve James’ ideal of covering beans with bonds and Florida with prefs.
Finally, I have to say, as I always do, that although the historical long run return of the stock market may have been 10%, the future expected return is 6%. The difference is that the historical record includes 1% for growth in the market P/E ratio (which can’t continue forever), 2.5% for dividends higher than they are now, and inflation a touch higher. In the past and in the future, index EPS grow at only about 1.7% in real terms (after inflation)
Looking forward, equities (before costs like MERs etc) should return:
2.3% from Dividends
1.7% from real EPS growth
2.0% from inflation
= 6%
Look out DB pension plans like the CPP and (especially mutual funds), but this has been going on for a decade! Actuaries don’t get it. They just look at the historical record without seeing where the returns came from.
However, equities do have a degree of inflation protection that prefs and long term bonds lack, so having some is good diversification in a taxable account (prefs might do better in a lower inflation environment, until they were called).
among the risks I would include Liquidity Risk against the annuity
Quite right.
From my experience, the back of the envelope calculation for annuities seems more consistent if you add a couple of years to the expected life (as an insurance profit — they will say adverse selection)
I’d prefer to use a spread. Also, note that my figures are for a specific person and don’t use mortality tables; on the basis that I don’t really care how much profit the insurance company makes, I just want to price the annuity against other instruments for comparison purposes
call risk (if perpetual pref rates fall, 6% yields will disappear and be replaced with something else).
I over-stressed call risk in the original post; If you buy a pref paying $1 at $20 and it later gets called then sure, you’re reinvesting at 4% rather than your original 5%, but you’ve got more capital.
However, there are very significant frictional costs in the transaction (capital gains tax and (essentially) the issuer’s cost of a new issue) so it’s still a big risk.
equities do have a degree of inflation protection that prefs and long term bonds lack
I always recommend that inflation protection be sought in the equity market, rather than trying to buy fixed income with bells and whistles.
“equities do have a degree of inflation protection that prefs and long term bonds lack”
Is the extra inflation protection the protection against “unexpected inflation” as presumably the “expected inflation” premium is already priced in the yield on the preferred? Or is there some other meaning that I am missing?
Is the extra inflation protection the protection against “unexpected inflation”
Yes. The rationale is that if inflation spikes, then commodities – oil, sugar, etc., – will maintain their real value and that the equity in producers of these commodities will accordingly also maintain their value (at least!).
Naturally, not all equities offer the same amount of protection – it depends on the pricing power that each company can reasonably be expected to have during periods of unexpected inflation, with resource firms expected to have lots of it.
I may be too late to add this, but the inflation protection offered by equities is in the EPS growth, which I had split into an inflationary and real parts when looking at the history (and the future). Thus, I view equities as providing complete inflation protection (in the long run). If inflation goes down, so, likely will nominal equity returns in the long run. Indeed, if real GDP growth slows it is likely that real EPS growth will slow too. Note that 1.7% real EPS growth is roughly equal to long run real GDP growth per person.
In the short run, rising inflation may be bad for equities if the P/E ratio falls, too — as it did in the 1970s (to 8 vs about 20 today). However, Brazil had an average 8% inflation per MONTH for 25 years and traded at about the same 13.5 P/E as developed markets have in the 20th century.
Very high inflation would be bad for the long run real returns from discount prefs, which presumably would never be called and so wallow at very low levels compared with today. On the other hand, low inflation (or continued return to normalcy in the bond-pref spread) could cause today’s discount prefs to be called and re-issued at lower rates (albeit with some ofsetting potential capital gain, as James notes). Thus, a pref is a non-symmetrical bet on relatively stable inflation that should be diversified.
[…] this 2010 essay (which builds upon the PrefBlog post Preferred Shares & Annuities, I looked at annuities as a component of a retirement portfolio and concluded in […]