Life After Full-time Work Blog

I’ve been asked to say more about future inflation and its effects

You’ll remember that blog posts #185 and #186 came to the surprising conclusion that the rise in interest rates in 2022 was more than enough to compensate current workers for the loss in value of their retirement assets. For my hypothetical 45-year-old, at the end of 2022 the projected annual retirement income from age 66 was 31% of current pay, whereas at the end of 2021 it was only 18% of current pay – a very big improvement in retirement security, despite the big fall in asset values.

Some readers noticed that those percentages were calculated in terms of annual income that stayed constant in nominal dollars, after retirement. The implication is: wait, that bond yield of 3.88% that you used has a certain amount of expected inflation built into it, and you’re taking credit for it; but surely what the worker can buy after retirement will be reduced every year by whatever inflation actually results. In other words, the 31% is misleading: that’s only what can be purchased in the first year after retirement, and it will fall with inflation every year after that.

True.

I was taking the standard (and easy) way out, by not projecting an inflation-indexed lifetime income stream after retirement. Most defined benefit pension plans in North America, in the private sector, don’t (and didn’t) provide inflation-indexed benefits: those are far too expensive. (Yes, it’s very often the case that public sector workers enjoy inflation-indexed benefits, despite the cost. That’s a story others can get into.)

But that’s an excuse, on my part. I just wanted to make the point that the replacement ratio (like its sibling, the Personal Funded Ratio) is a superior and more informative measure than simply reporting asset values. Now I’ll get to grips with projected inflation. I’ll aim to satisfy those who questioned my approach.

I’m going to do this in two stages. The first will deal with projected inflation after retirement. The second will deal with inflation between now and retirement.

***

I used 10-year US Treasury bonds to represent the current interest rate prevailing at the end of each of the last 15 years, for my 45-year-old who has been in this pension plan for those 15 years. Well, the US Treasury also issues TIPS (short for Treasury Inflation-Protected Securities), in which there’s a “real” interest rate to which actual inflation is added each year. So, if I use the “real” yield on 10-year TIPS to calculate the hypothetical annuity purchased at age 66, that will ensure that the projected payments will in fact keep pace with inflation after retirement.

The difference in yields between the 10-year Treasuries and the 10-year TIPS tells you what amount of (estimated) future inflation is built into those TIPS prices.

At the end of 2022, we’ve seen that 10-year Treasuries yielded 3.88% per annum. The 10-year TIPS yielded 1.58% per annum. The difference of 2.30% per annum represents the annual inflation estimate built into the TIPS price. Of course actual inflation may turn out to be higher or lower; but it’s an average 2.30% inflation that bond buyers are prepared to pay for.

At the end of 2021, in contrast, 10-yeat Treasuries yielded 1.52% per annum, and 10-year TIPS yielded (-)0.97% per annum. (Yes, that’s right, bond buyers were prepared to receive a negative yield … as long as those payments were proofed against inflation.) Again, the difference of 2.49% per annum represents the inflation estimate built into the TIPS price.

OK, with that as illustrative background, let’s see what the reported income replacement ratios would have been for our current 45-year-old over the past 15 years, making the assumption that the retirement income stream would be generated by annuities that index each year’s payments to inflation.

Obviously, lower than I’ve mentioned in the two previous blog posts. I won’t bother with giving you all 15 years, because (trust me) there are no surprises. You get (as before) a generally increasing but volatile series of numbers. I’ll just give you the last two numbers.

At the end of 2021, instead of an 18% nominal replacement ratio, it would reflect a 14% inflation-indexed ratio.

At the end of 2022, instead of a 31% nominal replacement ratio, it would reflect a 25% inflation-indexed ratio.

Two quick comments. First, to sustain inflation-proofing the starting post-retirement income has to be lower than a non-inflation-proofed income. Second, there’s still a big improvement from the end of 2021 to the end of 2022, so (to confirm my original point) 2022 was in fact a very good year for our 45-year-old’s retirement prospects.

***

In fact, even though I can calculate them, in practice inflation-indexed annuities have always been very difficult to buy, and in most countries are simply not available. There are two main reasons.

One is that an insurance company’s obvious assets to back the annuities of this type that it sells are government-issued inflation-indexed bonds, like the TIPS I mentioned; and in most countries these are simply not issued. (For example, in Canada the government announced in November 2022 that it would no longer issue its equivalent, called “real return bonds.”)

The other is that people are short-sighted, and don’t like the lower starting payments. For example, compare the end-of-2022 results I mentioned: 31% replacement in nominal terms and 25% replacement in real terms. What this means is that, for every $31,000 a year that some given amount of capital would purchase as a nominal annuity from a hypothetical insurance company, the same company would offer an inflation-indexed income stream that would start at $25,000 a year. And people overwhelmingly go for the $31,000 stream.

It’s true that over their lifetime, if inflation averages the built-in 2.30% a year, the two streams would have the same value. And if inflation exceeds the average 2.30% a year, the inflation-indexed $25,000 stream would be more valuable. But people typically just see $31,000 versus $25,000, and choose the $31,000.

By the way, what if you would actually like to buy the $25,000 stream – but you can’t find an insurance company willing to sell it to you? What you do, then, is to buy the $31,000 stream, spend the $25,000 stream, and invest the excess each year in TIPS so that it becomes available for spending when inflation takes your spending needs above $31,000. It’s an imperfect solution, but it’s probably the closest you can come. Talk to an annuity expert about it. (I’m obviously not giving you advice in my blog posts.)

***

I said earlier that I would do this in two stages, and the second would deal with inflation between now and retirement. Where’s the protection against inflation in that period?

Answer: that’s up to you. You need your pay to keep pace with (or, of course, preferably to exceed) inflation.

Then your annual series of replacement ratio projections will, one year at a time, adjust for inflation indirectly, because it will be based on each year’s pay, and you can see if (as you hope and expect) the replacement ratio shows a gradual (though probably volatile) increase over time. You’ll notice that it did indeed do so, in our hypothetical example.

Would you want that volatility to decrease, as you approach retirement? Most people definitely want that. And that’s what glide paths are designed to achieve, in the accumulation phase.. My blog posts on glide paths and target date funds explain this.

***

One final point. Aaron Minney wrote to me with an angle I had never considered, as I calculated how price changes affect people of different ages. And I think his angle is a fascinating and important insight, so I’ll draw it to your attention.

It’s clear that a fall in prices is good for buyers (think: young people saving for the future) and bad for sellers (think: retirees cashing out savings to generate money to live on). Aaron takes it further, and points out that, over the age spectrum, there must be a cross-over point, an age at which the impact is neutral: the benefit from higher interest rates in the future exactly offsets the loss in value from the fall in prices. He says, succinctly: “This age will vary with different market adjustments, but it is always there.”

That never occurred to me. In my fumbling way I noticed (in blog post #186) that the benefit of the increased interest rates got lower as the age of the worker increased, and it was still marginally positive for a 66-year-old, the oldest age I considered in my hypothetical work force. But the extension to a cross-over point escaped me; and the notion that there’s always a cross-over point is, I now think, a fundamental principle that Aaron has brought to my attention.

Thank you, Aaron. Your students are lucky to have you.

***

Takeaway

Income streams in the after-retirement payout phase can, in principle, be adjusted to try to keep pace with inflation. Such streams start out lower than nominally-constant payout streams.

Leave a Comment