## Lies, damned lies and investment illustrations

### By David Lewis

#### Twin Tier Financial

Savings. Your clients want it. One of the best ways for them to get it is to invest in good companies and ride the returns to financial freedom. There's just one small problem: Your clients need to predict something that's not easily predictable. Oh, of course we come up with a mountain of data, Monte Carlo simulators and hypothetical formulas that attempt to give us a glimpse of the future, but those forecasts often fall far short of what we hoped. Evidence?

**Personal savings looks pretty dismal**

According to the Employee Benefit Research Institute, the average retirement savings of people between the ages of 65 and 75 is a measly $56,212. And younger generations don't fare any better. Here's how average savings breaks down for different age groups:

- Ages 55–64: $69,127

- Ages 45–54: $43,797

- Ages 35–44: $22,460
- Under 35: $6,306

**There's a lot of drag on investment performance**

DALBAR's famous Quantitative Analysis of Investor Behavior consistently shows that investors are often their own worst enemy, buying and selling investments at the wrong time. Some financial professionals chalk this up to overly-emotional investing or ignorance of how to invest and there is evidence to support this idea. However, there's another explanation for why some investors end up with sub-par returns: real life happens.

New research by Hello Wallet shows that 1 in 4 Americans raid their 401(k) plan for non-retirement needs. Out of those that do withdraw, 75 percent report that they use their savings because of basic money management problems. Penalties on these types of withdrawals have increased from $36 billion to about $60 billion between 2004 and 2010. Ouch. Once you get past the fees and penalties, it's no wonder investment returns (and total savings) is so low. But there's more to this problem.

**Illustrations don't predict the future**

When used to predict variable equity returns, Monte Carlo simulators, and every other mathematical model that is predictive, suffer from the same basic problem that no one likes to talk about. The software, and the very method itself, is entirely deductive. What does that mean? Well, you've probably heard of deductive reasoning before, right? A well-known deductive argument goes something like this:

All men are mortal;

Socrates is a man;

Therefore, Socrates is mortal.

All deductive arguments have two premises and a conclusion that is deduced from the first two premises. This is important in understanding the flaw in Monte Carlo simulators. On the surface, all deductive arguments make sense, but what you have to ask yourself is: How do I know the premises on which the deduction is based are true? In other words, how do you know that "all men are mortal" and that "Socrates is a man?" If the only thing you're relying on is the deductive argument presented, you don't know. Not really. The deductive argument doesn't prove that "all men are mortal" and that "Socrates is a man." All it really tells you is that if all men are mortal and Socrates is a man, then Socrates is mortal. So, how does all of this relate to Monte Carlo simulators?

Well, the simulator's math is perfect, but it relies on inputs that you provide. You, as the advisor, must specify an assumed rate of inflation, an average rate of return, your client's life expectancy and a few other assumptions. How do you know any of the assumptions are valid? You don't. Like the deductive argument above, all the simulator will tell you is that if inflation is a certain percentage, and if you receive certain investment returns, and if you live to a certain age, then your retirement savings has a certain percent chance of supporting you. That's a lot of assumptions.

Here is a simplified version of what Monte Carlo would do for you and your clients. Let's say your client invests $100,000 and earns an average of 12 percent annually. An average of 12 percent could look like this:

Year 1 = 20% = $120,000

Yearr 2 = 4% = $124,800

Year 3 = -10% = $112,320

Year 4 = 24% = $139,276

Year 5 = 22% = $169,916

Now, let's repeat this example, but use a straight 12 percent rate of return:

Yearr 1 = 12% = $112,000

Year 2 = 12% = $125,440

Year 3 = 12% = $140,492

Year 4 = 12% = $157,351

Year 5 = 12% = $176,234

Now, let's compare that with another scenario involving one down year in which your client lost a catastrophic 40 percent (similar to what many people went through in 2008):

Year 1 = 25% = $125,000

Year 2 = 30% = $162,500

Year 3 = 20% = $195,000

Year 4 = 25% = $243,750

Year 5 = -40% = $146,250

The problem you will always run into is that you're trying to deduce investment performance, inflation, and life expectancy for one individual from statistical data. Statistics are fine in the context of the law of large numbers, but they can't be applied to individual circumstances with any kind of meaningful accuracy.

This same problem exists in the insurance industry, too. How many times have you seen a fellow advisor tout the importance of life insurance policy illustrations? These illustrations state right on them that they are not predictive. Non-guaranteed columns in the illustration often use a simple straight-line assumption (i.e. a straight 6 percent equivalent compounded return over 30, 40, 50, etc. years.) Indexed life is notorious for this, but it's also common in participating whole life assumptions. This isn't realistic, even under the best of scenarios.

**A possible solution**

No amount of hypothetical returns or data mining of historical returns will give advisors an advanced warning of future equity performance. However, that doesn't mean that investing is a futile endeavor. We can help clients in one of several ways (or a combination of all three):

- Encourage a stronger focus on saving money and let fixed rate investments make up the difference, or;

- Teach clients how to apply qualitative securities analysis (the Philip Fisher method), which would largely eliminate the need for hypothetical performance assumptions, or;

- Do the analysis for the client.

What's needed is a simple solution that's more predictable than what we're using now. One that doesn't rely on lies, damned lies and investment illustrations.