Wednesday, February 17, 2016

How Long Can I Afford to Live in Retirement? —Part 2

No sooner had I finished my most recent post, when I ran across this article in US News Money Personal Finance section that serves to illustrate some of the concerns expressed by Moshe Milesky in his article that I discussed in my last post.  The US News article asks the question for a specific hypothetical client situation of whether the client could satisfy his goals of having adequate retirement income over an expected 40-year retirement, leave a significant legacy to his children and still maintain a conservative investment strategy.  Much to the hypothetical client’s presumed delight, his financial advisor determined that there was a 96% probability of achieving the client’s goals with a small tweak in the client’s investment strategy.  The article implies that the client used this good news to go ahead and actually retire, assured that he could indeed “have it all.”

Now I don’t want to pour cold water on the client’s rosy view of his retirement.  But, I am an actuary and  prone to being perhaps a little more conservative than the next guy.  So, let’s look at the facts presented in the article and see if we reach the same conclusions for this hypothetical client.

The hypothetical client is age 66 and still working.  He has accumulated $1,200,000 in pre-tax savings (401(k) and IRA assets).  He is eligible to receive Social Security and an “old pension” totaling $3,500 per month, which we are told is “net of taxes.”  It isn’t really clear to us whether the client’s annual income goal of $6,500 per month is net of taxes or before taxes.  I’m going to assume that this income goal is before taxes and that his Social Security benefit is $2,000 per month ($24,000 per year increasing with inflation) and his “old pension” benefit is a fixed dollar amount of $1,500 per month ($18,000 per year). 

There is no discussion in the article about other assets that the hypothetical client may have, such as long-term care insurance or even home equity.  So, I’m going to assume that the client and the financial advisor decided in a separate discussion that the client’s home equity, if any, would cover his future long-term care costs or perhaps even some of his desire to leave money to his children.

There is also no discussion in the article about setting aside assets for future emergency expenses or about how future medical expenses may increase at a faster rate than expected inflation.  We are simply told that the client believes “he could live very comfortably and do all the things he wants on $6,500 per month.”  Fair enough. 

So if we input the hypothetical assumed data into the Actuarial Budget Calculator V 1.0 spreadsheet from this website with the recommended assumptions and we further assume that the hypothetical client desires to have constant real dollar spending in retirement for his 40-year desired retirement period and no amounts left to heirs at the end of this period, we come up with a first year spending budget $78,965, or $6,580 per month.  If he wants to leave $1,000,000 at death ($372,431 in real dollars under these assumptions), his monthly income would be $6,071 according to the calculator on an expected basis.  So, under these assumptions, the Actuarial Budget Calculator would not deliver the good news that the client wanted.  On the other hand, if the client used the recommended retirement period of 29 years (rather than a 40-year period), the Actuarial Budget Calculator would indicate that his assets would be expected (under the recommended assumptions) to be able to support a spending budget of $6,500 per month and a legacy bequest of $1,000,000 at death.

Of course the most important aspect of the Actuarial Approach is that it is a dynamic strategy and not a static “set and forget” strategy.  There is no probability of ruin (or success).  There is only this year’s actuarially determined budget that balances the present value of the retiree’s assets with the present value of his liabilities.  Next year, the actuarially determined budget would be based on next year’s data, assumptions and goals.  I realize that this process may not be as comforting as having someone tell you that your plan is 96% bullet-proof and you don’t have to worry about it ever again.  But remember that Monte Carlo modeling (generally) uses historical experience (which may or may not be adjusted for future expectations) to model future experience.  It is no better (and depending on assumptions employed, may be worse) at predicting actual future experience than using best-estimate deterministic assumptions.   And while MC modeling may be great at giving some clients comfort, in many ways it is just a black box that requires a high level of faith and does very little to educate the client on the reasonableness of the assumptions being made on his behalf. 

But, I’m not saying that Monte Carlo modeling is useless.  It can be very useful for certain purposes, such as modeling different asset allocations or kicking the tires on alternative strategies.  I am, however, saying that just as some experts have concerns about using deterministic models, these same concerns also apply to Monte Carlo simulations. 

Perhaps a few less esoteric items can be taken away from this article:

  • It is probably a good idea to consider all relevant assets and liabilities (such as Long-Term Care costs and unexpected expenses) when developing a spending plan. 
  • It is probably a good idea to determine whether the comparison of assets and liabilities is going to be on a pre-tax or a post-tax basis.  While conceptually the Actuarial Approach can match a retiree’s assets and liabilities on an after-tax basis, it is generally anticipated that this comparison will be on a pre-tax basis and that the retiree’s spending budget will include taxes.  
  • It is probably a good idea to quantify specific bequest motives.   
  • It makes a difference whether future benefits received from an “old” pension or annuity are fixed dollars or indexed.