Executive Summary
For the comparison, I assume a 500k portfolio over 30 years to be earning 6% annually, drawing down by 4% initially, then increasing the withdrawal by 3% per year for inflation and discounting future advisory fees by 6% annually. Here’s how an annuity stacks up to a managed account:
If an advisor put $500k in a single premium immediate annuity, commissions to all levels of the hierarchy are likely to be less than 5% of the initial premium, paid one time – for a total of $25,000 to the entire distribution hierarchy. The same $500k in a managed account, using the assumptions above with advisory fees assessed annually at a total fee of 1.25%, would be a present value of over $77k. If the advisor’s fee was .75%, lifetime fees would still be over $51k.
Let’s take another case – a variable or indexed annuity with an income rider. Typical commission across all levels of a hierarchy could be as much as 10% at all levels for a longer surrender product, for a total commission of 50k. If the product has an income rider, it needs to be monitored each year to determine whether to engage the income rider guarantee or to take free withdrawals while continuing to let the rider guarantee roll up. It takes specialized knowledge and time initially research and identify appropriate guarantee structures for a given client and specialized knowledge and time to monitor these guarantees. It requires an annual review process for the same reasons. When an advisor allocates to these products, he expects those funds to stay there for the rest of the client’s years. At 50k for the deferred annuity with income rider, compared to $51k-$77k, the present value of the expected AUM fees, the deferred products with income riders still represent less compensation than an AUM model.
I use annuities with my own clients and I work with other advisors on using annuities in their practices. I do the same with investment portfolios. The appropriate question is how much to each.
If you don't like the simplified nature of the mainstream dialogue on annuities, read the actuarial journals. You'll realize that most of the arguments against annuities are oversimplified and are not representative of reality. As an intro, take a look at Moshe Milevsky's Longevity Risk and Annuities.
The fact is there is no investment portfolio alone can hedge longevity risk. Period. The only way to protect against longevity risk with a portfolio is to have a ridiculously low withdrawal rate (ie 2-3%); to hedge it, you need to pool risk via the use of annuities. As advisors who are truly interested in the best interest of our clients, whether we came up through the insurance industry or we came up through the investment industry, we all need to recognize our own biases and actively combat them with processes designed to help us avoid dogma and make recommendations in the best interest of the client. Despite the fact that annuities carry commissions that need to be disclosed to clients, its not only appropriate, but imperative to use them when the fact pattern requires it.
Alan Moore says
For anyone that cares, I went through the excel spreadsheet and it is correct 🙂
As for the statement, “there is no investment portfolio alone can hedge longevity risk. Period.” I find this interesting. I feel that a 4%(ish) withdrawal rate, along with a multi-year cash hedge, monthly rebalancing, and frequent (at least annual) reviews does hedge against longevity risk.
The 4% withdrawal rate keeps the withdrawals to a reasonable level, the cash provides the client with the ability to not pull money out of the market when it has bottomed out, the rebalancing insures buy low, sell high throughout the life of the portfolio, and the annual reviews insure that the client is on track, or make adjustments as needed.
The market has always beaten the return on annuities, and will continue to do so. If it doesn’t, the annuity companies would go bankrupt so it wouldn’t really matter.
You are correct that annuities present a conflict of interest, in that those that charge on AUM or net-worth would lose money, however I feel it is more than that. I think that annuities provide a very valuable tool when dealing with psychological issues, such as over spending. However, just like they aren’t always wrong, they are not always right for a client.
Michael Kitces says
Alan,
The 4% withdrawal rate research is built on the assumption that you spend down the portfolio over the 30-year time horizon.
Quite literally, the reason why Bengen’s original research had the safe withdrawal rate at 4.1% is because if you used 4.2%, you ran out of money in year 29 instead of year 30. But similarly, it assumes you DO spend the money in 30 years. Under Bengen’s own research, if you lived to year 31, you had no money.
This is the point in Elsasser’s post as well – safe withdrawal rates may be a viable strategy for maintaining a 30-year spending plan, but it doesn’t inherently do ANYTHING to protect you if the bad returns occur and you have the “misfortune” to live to/beyond year 31.
– Michael
I agree with the statement that an investment portfolio alone cannot hedge longevity risk, and that some form of annuity should be utilized for some portion of a retirement lifetime depending on the client’s situation. SPIAs (or the more recently developed longevity SPDAs) should work as long as the costs are made clear. The author states “commissions to all levels of the hierarchy are _likely_ to be less than 5%” How do we and our clients find out for sure?
Note also that Michael himself analyzed GMWBs and GMIBs and concluded that holding stocks plus a ladder of treasuries should outperform in most market situations.
The guest article was about expenses, not appropriateness of use. Given such a narrow discussion based on assumption, how can anyone comment intelligently? “About 5%” is surely less rigorous than the 1.25% and the -0.75% investment advisory fee examples. Furthermore, doesn’t the insurance company charge a maintenance fee against the account periodically? Frankly, my preferred strategy is to buy an annuity in tranches at market and interest rate peaks, if I were to buy an annuity at all. Thus, I am getting more annuity $$ in the higher rates and I am using a profitable account to buy the annuity tranch. How much in a tranch? Maybe10%to 20% of invested capital each time. I can appreciate the effort that went into the expense comparison, but I would have preferred real life billings/examples.
Jim,
The article is talking about an immediate annuity, not a variable annuity. The typical M&E expense, investment advisory fee, maintenance fee, etc., don’t apply. You can’t really “see” the costs in an immediate annuity, you can only estimate it indirectly by netting out an estimate of the insurance company’s expenses/costs/profit margins to arrive at it. That’s what the 5% is intended to estimate. As far as I’m aware, it’s a reasonable approximation in the single premium fixed immediate annuity context, although I’ll grant I’m not spending a lot of time hanging out with annuity company actuaries these days. 🙂
But the bottom line is that I’ve never seen an annuity company at any point in history even publish the number you’re asking for here. It can only be estimated indirectly by looking at the business model and product line profitability models that the actuaries use, and that will still only provide you with a general industry estimate, not a product-specific one.
Respectfully,
– Michael
I actually used 5% to be on the very high side for immediate annuities. Having been part of a large national marketing organization, with top contracts in that space, I can tell you that few are even that high, and if they are, the payouts to the client aren’t competitive.
On the IA side, I didn’t include the “value added services” that are disclosed in all of our ADVs that are funded by platform or shelf space fees paid by fund firms. So the analysis isn’t perfect and can’t be until our entire industry is considerably more transparent. That said, the goal was to be as fair as possible.
For immediate annuities, there are no maintenance fees per se. You get a series of firm quotes from different carriers and the difference in payout represents the difference in how the insurer views the risk and their expectation for expenses across a block of business. when a client buys a contract, they simply exchange the lump sum for a lifetime payout.
I actually agree with a laddering strategy and would love to see someone put together some metrics that would suggest appropriate timing for annuity purchases using some combination of technical indicators.
a few quick thoughts:
The 4% “safe withdrawal rate” also assumes inflated distributions as noted; SPIA’s generally do not inflate payouts. And the ones that do have horribly low starting payouts.
If you want to know determine the SPIA rate of return, run a 20 and 25-year period certain and you can back into a “rate of return” with your 12c. At present, returns are paltry.
Everyone needs to remember the big building theory. Insurance companies, like casinos, or other large gambling institutions (Wall Street), have large buildings. There is no such thing as a free lunch. A market based portfolio gives an investor a chance to earn that 6,7,8% or whatever. A 30-year SPIA will probably not ever get much above 4% with no COLA.
Thanks for the oportunity to hang out on here and learn some stuff I never heard of http://www.kitces.com before.
Joe, I find it an interesting calculation, but I’m not sure I’m seeing the usefulness. We need to consider all costs, not simply compensation.
As others mention, with the immediate annuity you basically are accepting purchasing power risk. Your immediate annuity today in 30 years will buy little compared to today. Entering an unknown era where the chances of inflating are great, that is too significant a risk to ignore, and simply buy ‘what’s cheapest.’
On the comment of the variable and index annuity, I won’t go into more but to my mind the important aspect to the client is the amount of income their savings buy. When I’ve run these numbers in the past, you could go into an immediate annuity for well over a 30% discount, meaning you could go into retirement with maybe 1/2 to 1/3 of a variable annuity and still have more income. That’s the cost (the biggest one anyhow) in my mind.
I guess the takeaway to me is you get what you pay for. Ongoing management to cover inflation risk, and to manage money better than an annuity (which I’ve never found to be a valid investment since you don’t know what the investments will be in the future) has a cost.
The comparison between the two scenarios is not applicable because if an RIA is giving ongoing advice that is an extra service.
To fairly structure the debate some adjustments need to be made: When buying an annuity to compare with spending down a portfolio using a 4% withdrawal rate plus CPI for a total payout of about 6% the annuity would need to include the purchase of a life insurance policy to hedge against the risks that the annuitant would die early and leave nothing to heirs. Further, in avoiding an annuity and investing in appreciating assets then the capital gains tax is waived at death due to basis step-up; buy contrast the annuity is ordinary income. Also an annuity lacks diversification as it is like a loan to the insurance company. No one would tell a retire to own more than 5% or even 2% of their assets in any one company yet when people buy an annuity they may put a huge percent of their net worth into a non-cancellable immediate annuity into just one company without realizing that if the insurance company goes bankrupt then they would lose their nestegg.
Please keep in mind that this is a pretty narrow article. I make no argument that an annuity should or should not be used for any specific fact pattern, only that annuities are absolutely unique in their ability to hedge longevity risk and that they shouldn’t be dismissed solely due to the argument that I hear most often – “annuities are bad because of the high commissions”. Thanks for the comments!