# Will Your Indexed Universal Life Insurance Policy Produce an 8% Average Return?

Last week we discussed the probability of achieving various returns on a passive investment in the S&P 500.  This week I want to take a similar analysis and direct it to indexed universal life insurance.  Pursuing this information comes from two primary motivations.  On the one hand, I'm intrigued by the answer.  On the other, we, of course, received questions about how indexed universal life insurance looks when running through a similar analysis.

So let's feed our intellectual curiosity and review the results.

## The Design

The design in this analysis is similar (but not identical) to the one I used for last weeks blog post.  Using rolling monthly average data I took S&P500 data spanning the 1920's through 2017 to generate several 10, 20, and 30-year tranches (over 60 for each category) used to compute the compound annual growth rate of a hypothetical investment for each observable timespan.

In other words, I used real historical data to create a boatload of 10, 20, and 30 year periods of hypothetical investments and calculated the growth rate for each one to arrive at the probability of achieving various rates of return with the cash value of an indexed universal life insurance policy.

I chose an indexed universal life insurance policy with a one-year point-to-point index option where the cap rate was 11%, the floor 1%, and the participation rate 100%.

I then used the expense report generated by the indexed universal life insurance product to plug in expense deduction to accurately capture the cost of owning the life insurance policy.  All expenses are accounted for in all scenarios.

## The Results

Before I discuss the actual results, I'll note my assumptions coming into this analysis.

I hypothesized that the 10-year scenarios would all reflect uninspiring results in the sense that rate of returns would be low by most standards.  Since the index universal life insurance policy must overcome a set of early expenses, returns are never stellar for the first decade.

I also assumed that 5% returns would exceed 50% probability in the 20 and 30-year scenarios since that level of return has always been our target for success with indexed universal life insurance.

### 10 Year Scenario

As predicted 10-year results were not great for sure.  But, they were all positive.

This caught me a little by surprise.  I figured surely a scenario or two would exist where market returns were bad for such a span that insurance expenses would cause the cash value to see a loss. Not the case.

Further, there was one observation within the 10 year tranches, and only one, where the return fell below 1%.  And this sub 1% result was 0.77% annual growth per year.  In other words, extremely close to 1%.

More fascinating was the fact that 57% of the 10-year tranches resulted in annual returns of 3% or better.  This was not at all something I anticipated.  I assumed much lower results.

Most fascinating of all was the 12 incidences (nearly 19% of the time) where the return came out to 4% or better–there was even one observation that came out above 5%.

 CAGR 10 Years Probability of CAGR 0 100% 1 98% 2 86% 3 57% 4 18% 5 2% 6 0% 7 0%

### 20 Year Scenario

The 20-year scenario validated my prediction that 5% would occur 50% of the time (62% to be exact).  It also had no observations under 3%.  Very interesting if you remember that last weeks direct investment in the market had observations the fell below a 3% return over 20 years about 12% of the time.

Additionally, a return of 4% over the 20 year period comes in nearly at 100% of the time falling just short at 97% of the observations.

It's also noteworthy how dramatically the upper band falls off with indexed universal life insurance.  Just 6% of the observations came in over 6% return over the 20 year period. We've always known indexed universal life insurance as a strategy that gives up higher returns to guarantee against losses, and this is certainly supported by this analysis.

 CAGR 20 Years Probability of CAGR 0 100% 1 100% 2 100% 3 100% 4 97% 5 62% 6 6% 7 0%

### 30 Year Scenario

In the 30 year scenario, no observations fall below 4%. The 5% probability shoots up to 97% which is far higher than I predicted for the 5% return.

The 6% return probability improves, in fact, it doubles.  But it still remains at a rather low 13%.

I'm fascinated by the fact that indexed universal life insurance appears almost entirely unaffected by time and the risk time poses to higher returns.

Last week we showed that time can place higher than average returns at risk of evaporating because corrections become more likely the longer one remains invested in the market.  Index universal life insurance doesn't appear to suffer this same reality and instead sees improvement in higher band returns if only slight.

 30 Year CAGR Probability of CAGR 0 100% 1 100% 2 100% 3 100% 4 100% 5 97% 6 13% 7 0%

## My Thoughts on the Data

It's interesting to see validation of the so-called “downside protection” offered by indexed universal life insurance that comes with the sacrifice of “higher returns.”  I'd like to propose, however, that we must make a small yet important adjustment to this paradigm.

Indexed universal life insurance doesn't just protect against a loss, it affords protection with the assurance of moderate gains.  The tradeoff of this feature is the understanding that a 7% or higher year-over-year return on money in an indexed universal life insurance policy is highly unlikely.

But there are other considerations to weigh here…

Indexed universal life insurance cash values do accumulate tax-deferred and can be used without tax implication when accessed correctly.  The indexed universal life insurance policy also carries a death benefit that has a host of potential benefits to the policyholder.

There are additional points of intrigue that comes from this data.  You'd think that the likely lowest return from the 30-year scenario would be a time period wrapping up around 2008, but you'd be wrong.

The worst 30-year span actually took place from 1949 through 1978.  The 30 year period that ended with 2008, by the way, was better than the bottom 3rd of all observations.  This plays out because of the floor feature to indexed universal life insurance policies.  A market correction doesn't create a loss and only results in a small return for the year.

I also find some additional validation in this data for us at the Insurance Pro Blog.

We've warned against using unrealistically high index rate assumptions for a long time. We even took this head on years ago in a blog post about what assumptions seemed most prudent.  The fact that 6% returns come out with such low probabilities supports our claim that any projection resulting in an internal rate of return report above 6% consistently at a 20 or 30 year period is likely overshooting the index credit assumption.

Read that last sentence again, there's a lot packed in there…be sure not to confuse internal rate of return with the index credit.

One last comment on this data addresses the claim that you cannot lose money with indexed universal life insurance.  This statement has met ridicule in the past (even by us), but I think it's fair to point out that timeframe is an important context to this notion.

Sure you could purchase an indexed universal life insurance policy and cancel within three years and very likely walk away with less money than paid in premiums.  But after the 10th year–assuming the policy was designed and implemented to accumulate cash values–it's pretty unlikely from this data that a loss would take place.

## Indexed Universal Life Insurance Versus Passive S&P500 Investment

I never set out to prove that one idea was better than the other and I caution everyone against making that mistake by drawing conclusions from this data or last week's data.

The intention is not to call winners, it's more to give everyone a more complete understanding of some basic functionality.

I think the biggest takeaway in both situations is that some practitioners have overstated expectations for both savings strategies and this hopefully will bring more reasonable guidance to either pursuit.

Keep in mind that both strategies have slightly different goals in mind.  It's true that both are a means to accumulate wealth.  They both seek out this goal with a different priority in mind.  One is much more speculative than the other.  This speculative nature hopefully brings greater returns over some longish time period, but temper that notion with the fact that time does not guarantee better gains when speculating.

The other holds more solid moderate returns from very early on with little to no expectations of any windfalls.  This more moderate approach, however, isn't as affected by the risks assumed as the timeline extends and instead appears to improve results, though moderately.

### 15 thoughts on “Will Your Indexed Universal Life Insurance Policy Produce an 8% Average Return?”

1. If the CAGR is lucky to even hit 6%, wouldn’t a whole life product that offers a dividend in the 5%+ range be a better offering? My first read of this post led me to that conclusion, but listening to your more recent podcasts is making me question if I understand the numbers correctly. (For instance, my 5%+ doesn’t include the premium charges but maybe your analysis in this post did.)

Thank you for so many great posts on this topic! All the hoopla around IBC brought me to your blog as I try to understand the details past the shiny sales pitches.

• Hi Thumper,

This will depend on a few factors. The quoted dividend rate references the amount attributable to investment performance on the reserve of the policy. There really isn’t a standard for this, so one company’s 5% is not another company’s 5%, but we can loosely anchor the concept to the policy reserve, which is somewhat closely tied to cash value.

Given this notion, 5% on cash value is very different from 5% on premiums paid.

This being said, there is reason to assume that some whole life policies might achieve an actual return on a compound annual basis of 5% or more. Assuming this does happen, there are those would would argue that this might be mighty competitive versus regular stock investing. It will come down to which camp you want to belong. If you find value in stock ownership, feel the return might be better, or have some other attraction to public company ownership that you think will be advantageous, you’d likely still take your chances on passive index investing. If on the other hand you find value in stable returns with no risk of a retraction due to economic declines, whole life insurance certainly will look attractive.

2. I really liked your S&P and IUL analysis. Any chance you can do one for whole life polices? (preferably from mutual companies)

• Hi James,

As much as we’d love to, the problem is we don’t possess the data to do such a thing. Index universal life was easy because we could take market data against the indexing feature and use company disclosed fees to estimate results. In truth, there are factors that might change these results. We are assuming current market conditions, so options pricing leaves us with current cap rates, but that’s a variable I’m willing to accept at current value.

Whole life insurance figures would require we know way more than any life insurer would ever disclose to us in terms of actuarial calculations to arrive at policy values from past performance.

We’ve also been asked to perform a similar analysis on the bond market. There is also a problem with adequate data points on that front. Bond pricing is very variable and holds no standard reporting like stock pricing. We could work through various analogues such as various bond indices, but the only real clear data that exists comes from government issued debt. That’s an okay basis, but corporate debt (where most people would invest) varies incredibly from government pricing. And government debt in this case only refers to Federal Government Debt. Municipal debt (another high source of investor activity) would also be nearly impossible to capture any meaningful data.

3. How is the policy in this example funded? Level or increasing death benefit? Switch? Health of insured? Premium relative to target?

• The premium is guideline level premium on an increasing death benefit (i.e. minimum db solve). I assumed premium payments in all years so there was never a death benefit switch. Preferred risk class.

4. I’m curious to know what you think of GE executives getting IUL’s in compensation plans. These premiums are in the \$100,000 range a year. Highest was \$445,000. That’s alot of money going toward a IUL that’s being used as a retirement plan.

• Hi Joshua,

Life insurance used as part of executive compensation is a long established practice. These premiums might appear large, but keep in mind that they are funding a benefit for someone with a likely seven figure income from the company, with a lot of offers from all sort of places to bring their talents elsewhere. It’s a great incentive to keep top talent in place. It’s also a strong vote of confidence when very large companies choose life insurance as the place to save money for these benefits.

5. Could you explain what you mean by a “tranche”? Is it a 10-yr point to point CAGR? If so, how did you pick the 60 periods out of the many hundreds available?

• A tranche was one data set or X year observation created with rolling monthly averages from 1928 through 2017, which resulted in over 60 observations for each year span category. I elaborated more on the design in this blog post made a week earlier.

• Thanks. I looked at the previous post and I’m still confused. I can generate 948 rolling 10-year, 828 20-year, and 708 30-year monthly CAGRs for 1928 to 2017. 60 just seems like a small number on that basis.

• Feel free to run your own analysis and let me know if you end up with markedly different results.