Thursday, 5 November 2015

How Much do Interest Rates Need to Rise for Deferral of Annuity Purchase to Make Sense?

When interest rates are higher, annuity payouts are higher. The big question is, exactly how much would interest rates need to rise in order to justify deferring the purchase of an annuity? Not that much, is the answer, but there are some crucial differences according to age, sex and assumptions about portfolio returns. And there is a significant risk if interest rates do not rise but merely stay the same.

A 1.2% increase in interest rates over five years (i.e. 0.24% per year) could suffice to make it worthwhile for a man, taking money out of an RRSP or RRIF, to defer buying an annuity for five years from age 60 to 65.

The table below shows the results of our calculations. The base case of the assumptions we think most logical is shown in the left most column, along with other possible assumptions for male annuities and the base case for women in the right-most column.

In all cases our approach is simply to make the ultimate lifetime income the same, whether the annuity is bought immediately or deferred five years. While the deferral is under way, we assume the same amount is withdrawn every year from the investor's portfolio as the immediate annuity pays out. The later deferred purchase is then priced on whatever amount would remain after withdrawals and each year's portfolio return and, crucially, at the higher annuity payout rate the five year older person would get. The older you are, the more mortality credits boost the annuity payout, as we explained in Is it Worthwhile to Wait for Higher Interest Rates to Buy an Annuity?.
(click on image to enlarge table)

Here are some other observations on the results:
1) The younger you are, the more worthwhile it is to defer - At age 60, the required rise for a male annuity buyer is only 1.2%, but deferring from 70 to 75 would require a 2.4% rise. The same age pattern exists for women, it's only the numbers are different. The reason is that mortality credits (the money that comes from other annuitants dying) are much more important the older you are, so missing out by deferring makes less and less sense.

2) It takes less of an interest rise to make deferral worthwhile for a woman than a man - The right-most column for female annuity pricing shows that deferring from age 60 to 65 requires only a 0.95% rise in interest rates to gain higher annuity income. The reason again is mortality credits. Because women live longer than men, it is as if the whole age scale for mortality credits is shifted youngwards - women get fewer mortality credits than men of the same age. Note that single-sex, men vs women, annuity pricing, where the differing lifespans influence payouts, applies only to non-locked in registered retirement accounts, TFSAs and non-registered accounts. Locked-in retirement accounts are subject to unisex pricing rules that create payouts in between those of the single-sex prices, (Our table does not show those payouts)

3) A high return in the investment portfolio during deferral can significantly reduce the required interest rise - When we set the investment portfolio's annual return at 2.7%, matching the return embedded in the annuity, the required interest rise was only 0.3% for the 60 year old investor, versus the base case 1.2%. The required interest rise was much lower for the other two age groups as well. ... however ...

This raises a crucial point - what rate of return on the interim investment portfolio is most logical to assume. Our base case uses 1.0% per year, which is about the net return (yield to maturity of 1.26% minus the fund's MER of 0.28%) currently on the iShares Short Term Bond Index ETF (TSX symbol: XSB). This ETF has very high credit quality and its duration of 2.79 years means that it will have limited exposure to capital value losses that would result from the very interest rate rises the investor is trying to take advantage of. As we discussed in What Happens to a Bond ETF When Interest Rates Rise? duration also tells us how long it will take for the ETF to recover from capital losses as newer higher yielding bonds replace outgoing issues in the ETF portfolio. At the end of the deferral period, the investor wants to be sure to have enough money to buy the annuity. Investing in much more volatile longer duration higher yielding bond ETFs, or a stock bond portfolio, is taking on considerable exposure to volatility risk.

One thing that does not make much difference is how much of the interest rate increase is captured by the interim investment portfolio. The second column from the right shows the results when only half of the interest rise gets reflected in the interim portfolio. This might well be the case for the short term bond fund when the interest rise, which is the rise in the long term rate used to price the annuity, is not matched at the short term end.

4) If interest rates stay the same, there is a big hit to the deferred annuity income - The 60 year old man would be forever stuck with about 13.1% less yearly annuity income, the 65 year old 15% less and the 70 year old 18.7% less. There will be less of a hit to a woman but it will still be very noticeable as our table shows.

Bottom Line: Is it worth taking the chance? What are the chances of a rise in interest rates and by how much? We'll explore that question in the next post.

How we came up with the calculations - Note that this blogger is not an actuary or annuity pricing expert. Our calculations are rough (and rounded to the nearest 0.05%) but hopefully reasonable. First we followed the method for roughly pricing an annuity laid out by professional retirement researcher Wade Pfau in Income Annuity 101. Then we applied recent mortality tables from the Canadian Institute of Actuaries - Canadian Pensioners' Mortality, published in February 2014 - that contain data on pensioners of private plans (apparently public plan pensioners live longer!). Next we compared resulting actual current prices of annuities as quoted in the Globe and Mail to find what implicit interest rate is embedded in the current quotes. That's how we got the 2.7% rate to discount annuity payouts. The 2.7% discount rate also lines up pretty well with what a portfolio of highly secure investment grade long term bonds would yield nowadays i.e what the insurance companies would invest in to back up the annuities. For instance, a mix of federal ( holdings like BMO's Long Federal Bond ETF (TSX: ZFL) with a yield of 2.25%) and provincial (holdings like BMO's Long Provincial Bond ETF (TSX: ZPL) with a yield of 3.28% could average out to 2.7%.

There are other signs we are close to the mark in the calculations. First, retirement researcher Wade Pfau in his post Annuity Pricing Sensitivity also calculated the effect of interest rate changes on annuities based on annuity pricing principles. Historical Canadian data in this post came up with 0.58. The two approaches give us reassuringly close estimates that every 1% rise in interest rates will cause annuity payout rates to rise about 0.6% (Pfau gets 0.63 to 0.65% and my Canadian data shows 0.58). On our simple annuity model the figure is 0.61%.

Second, the TIAA-CREF Institute, a research offshoot of one of the largest retirement income providers in the world, looked at the general issues around possible deferral of an annuity purchase in Annuities: Now, Later, Never? At unchanging interest rates they found that annuities always look better than deferring indefinitely and living off drawdowns from an investment portfolio.

However, they also showed that deferring in period of rising interest rates could be worthwhile. For instance, they calculated that if interest rates increased 0.25% a year for five years (i.e. 1.25% in total after five years), deferring the annuity purchase at age 65 by five years (and living off drawdowns from the portfolio in the meantime) would enable the purchase of an annuity with more than 7% higher income. The benefit tails off with a longer deferral. A 10-year deferral would produce only 3% higher income. A big caveat that they note is this result depends on the interest rate embedded in the annuity pricing (i.e. the interest rate the insurance companies receive from investing the lump sum premiums) being the same as the interest rate / return the portfolio could obtain. The interest rates won't be the same, if one is to construct a realistic apples to apples comparison. With the annuity, the insurance companies will invest in long term highly rated bonds while someone intending to defer five years would logically opt for much shorter term bonds as we argued above. So our results are more cautious than TIAA-CREF's though with the same patterns.

Disclaimer: This post is my opinion only and should not be construed as investment advice. Readers should be aware that the above comparisons are not an investment recommendation. They rest on other sources, whose accuracy is not guaranteed and the article may not interpret such results correctly. Do your homework before making any decisions and consider consulting a professional advisor.

Copyright 2015 Jean Lespérance All Rights Reserved