How I Set My Retirement Number

Last week I posted I Set My Retirement Number and Boy It's a Doozy! and I had this comment from a reader:

If you don't mind sharing, I'd like a little more information as to the five methods that you used in calculating the retirement number. I'm particularly interested in the two methods of calculation that you devised yourself.

So here we go -- I'm going to detail how I set my number.

First of all, I googled around a bit and found three retirement calculators that were to my liking. Each of them asked different questions (and thus made different assumptions), so the answers varied more than a bit. However, I recorded the answers provided by Monroe Bank, Money Central and Dollar Times as three data points in setting my number. Then I set off on developing my own estimations in two different ways.

The first estimation was the simple "4% method." It's developed on the premise that you will be able to take out 4% of your nest egg in year one of retirement -- and if you can do this (and live on it) -- you'll be set throughout retirement. So here's what I did:

• I took the annual salary I would need today to retire. (Note: This was not figured as a percentage of my current salary or our living expenses. I simply determined what we would need to live on today based on actual expenses. I did not adjust downward -- as many people recommend -- since I like to be conservative and I assume any decrease in living costs at retirement will be eaten up by increased health care costs.)

• I adjusted for inflation (assuming 3.5% inflation per year) to get the amount I'd need to withdraw the first year I'm retired. In other words, I increased the number I got in step #1 by 3.5% per year for 23 years.

• I then divided this number by .04 (4% -- you can also multiply by 25 -- it gives the same result) and that was my retirement number estimate #4.

For further details on this method, see Will My Savings Run Out in Retirement? where they comment:

If you withdraw 4% of your personal savings during your first year in retirement and adjust subsequent withdrawals to compensate for inflation, you're virtually assured of never outliving your money over a 30-year retirement.

Remember, in my original post, I said I wasn't counting on anything from Social Security, so I have to save this entire amount myself. If I wanted, I could subtract what I already have saved (plus growth of that amount) to see how much MORE I need to save before retirement. 

By the way, this method gave me the highest estimate of all the methods I used.

As an example, here's what this method would look like for a person planning retirement:

• Let's say the person decides he needs $50,000 a year in today's dollars to retire.
• In 23 years at 3.5%, that number balloons to $110,306 per year.
• Divided by 4%, $110,306 means the person needs $2.8 million at retirement.

From this amount, the person could subtract what he expects to get from Social Security as well as what he thinks his current savings will be worth at retirement to get how much he needs to save in the next 23 years. For instance, someone counting on nothing from Social Security but who thinks his current portfolio will be worth $1 million at retirement only needs to save an additional $1.8 million (principal and growth) in the next 23 years.

My last method took all of the same steps I employed in the example above, but I just laid it out year-by-year. For every year (from now until I'm 92), I listed the following:

• My income needs (increased by 3.5% annually).
• The value of my current savings. I estimated annual appreciation at 6% to be conservative, but my spreadsheet allows me to change this rate to do "what-if" analysis.
• The value of additional savings -- the amounts I contribute annually to my retirement savings for the next 23 years. These numbers include both the amounts I save as well as 6% annual appreciation.
• A running balance for my retirement funds (the value of my current savings plus the value of additional savings less my income needs)

I then took all five methods, and averaged them to get "my" retirement number. This number was a bit higher than the one I got in method #5, so I felt that it was one that would allow me to reach the savings goals I needed.

I end up with lots more money when I die at 92 than I had at 65 when I retire -- my needs are just not as great as the growth of the amount I'll save. This could be looked at two ways:

1. I'm saving too much and thus don't need to put as much away.

2. This protects me against potential snags in my assumptions (what if it costs more to live than I estimate, what if I can't get the annual returns I have estimated, etc.?).

Finally, this number is not set in stone. I will re-visit it in 3-5 years, see how I'm doing and make any needed adjustments. I hope to put in more than planned in the next few years to really boost my savings, so hopefully that review down the road will show I've made good progress towards my goal.

Comments

[Jay]

Are you talking about your savings or is this for you and your wife?

[FMF]

This is for both of us.

[Keith]

I am frequently confused by these retirement calculators because I don't know if I'm meant to enter my desired retirement income in todays dollars or inflation-adjusted future dollars.

For example, say I'm planning to retire in 30 years and I expect to need $70,000/year in today's dollars. If I expect inflation to be 2.5%, then I'll need almost $140,000 in 2036 dollars to have the same buying power that $70,000 give me today. So in this case do I enter $70,000 or $140,000 in the "Desired Income" field?

Obviously, it makes a HUGE difference in the results, and it annoys me that it is not explained better on these sites.

[FMF]

Keith --

That's why I did my own methods as well -- so I knew what number I was getting.

[retirementbuff]

I've hear the "4% method" quoted by numerous sources, but I've always had two questions/reservations about this method:

1. Does it take into account inflation over the course of your retirement?

2. I've seen this method applied to retirement at various ages (both earlier and later than 65 years). Wouldn't the exact percent depend significantly on expected length of retirement?

[Joachim]

1. Yes
2. No, not really, for a typical US age-based risk behavior

Consider 2 extreme cases:
A) age 40, life expectation 80 or more, you live longterm on only what you make good in real value year after year after tax and inflation, therefore all assets in stocks:
R=8.5 % expected annual return, T=25 % tax, I=2.5 % inflation, Retirement Yield Y = R * (1-T) - I = 3.9%
As long as your consumption stays under that, your real
value increases year after year
B) age 67, life expectation smaller than 92 (=67+25) or so, most of your assets in fixed interest with todays 4.5 % (= +1.2% real after tax and inflation) As long as annual return - taxes - inflation is better than 0,
your 1 / 4 % = 25 years expenditure can be consumed

In reality most the times your are somewhat inbetween those 2 extreme cases, both for age and asset allocation.

Of course, if you invest age 40 in government bonds with 4 %, have a tax rate of 30 % and inflation of 3 %, you will run out of money quickly

The main uncertainties in this game are predicting tax rates and inflation 20 or more years into the future

One last thing to consider is, to not take stock values at the peak of the bubbles, like 2000, but when they are in line with long term yield expectations, like 12/31/2005

Hope that helps you

[beef]

Example and Question:

Let's say I want to retire when I am 60 or 30 years from now and I expect to live to 90.

** Need $50,000 a year in today's dollars to retire.
** In 30 years at 3.5%, that number balloons to $140,340 per year.
** Divided by 4%, $140,340 means I need $3.51 million at retirement

Okay, how exactly do I figure out when I will run out or if I will have enough then when I am 90?

Year 31 - 3.51mil - 140k = 3.37 mil
Year 32 - 3.37mil (1.05) - 140k (1.035) = 3.39 mil
Year 33 - what formula?

What about taxes?

[FMF]

Beef --

Check out this post for some help:

http://www.freemoneyfinance.com/2007/08/some-good-finan.html