Can someone please explain MMM's 7% rule for me

[Siecje]

This is a post from June 6, 2011. I am wondering how -37% in one year is not the same in subsequent years as it is a percent.

1-year holding periods: Worst: -37% (and it was in 2008!) Best: +52.62%
5-year periods: Worst: -2.35% Best: +28.55%
10-year periods: Worst: -1.38% Best: +19.35%
15-year periods: Worst: +4.31% Best: +18.93%
20-year periods: Worst: +6.53% Best: 17.87%
25-year periods: Worst: +7.94% Best: +17.24%

[matchewed]

It's in the paragraphs below what you quoted.

The gist of it is that the market returns over a long enough period equal to about 10%. Subtract inflation ( about 3%). And you get 7%.

Quoted below for reference -

The average for all periods is of course the average annualized stock market return for the overall period, which is about 11.1%

After adjusting for inflation, this 11.1% annualized number would become 7.16% over the 1950-2011 period.  (Inflation was very high during the 1980s,  so it averages to 4% in this example instead of the 2-3% currently forecast for our own future).

[arebelspy]

I'm not sure I fully understand the question, but basically what that part you are questioning is saying is that if you invested money in the market for one calendar year, and then took it out a year later, the worst you'd have done historically is lose 37% of your money.

If you invested it one year and took it out five years later, the worst historically you'd have done is lost 2.35% of your money (and often done much better than that, and had a positive return...say 94-99).

And so on and so forth.

Do those numbers make more sense now?

And the 7% comes from long term 10% market returns minus 3% inflation.  Both, obviously, highly debatable numbers.

[Siecje]

I understand what the numbers mean, I don't understand where they come from, if I lose 37% one year why wouldn't I lose that again the next year

[arebelspy]

I understand what the numbers mean, I don't understand where they come from, if I lose 37% one year why wouldn't I lose that again the next year

Um, because the market doesn't perform the exact same year after year.

If you lost 37% in the worst year (2008), why would that necessarily mean you would the next year (2009)?

This post only added to my confusion of what you are asking.  :D

I'm not quite sure why you think you would lose that again the next year... please explain what you're thinking so maybe we can help guide you.

[Sekk]

I understand what the numbers mean, I don't understand where they come from, if I lose 37% one year why wouldn't I lose that again the next year

The numbers are historical.  They are not a rule or strange mathematical quirk of investing.

[k9]

You could loose 37% (or a little less, or a little more) twice in a row, but that never happened (at least in the US). That would result in a 60% loss after the 2 years, which is very big but happened in many places around the world, though.

As for the 7%, that's the same idea : you can probably expect that kind of number over the time, as it already happened in the past, but really we don't know : it could be much bigger, or much smaller. If you aim for 7% after several decades and don't care about volatility, the stock market is your best bet, but you can't expect that for sure.

[shadowmoss]

So, I lose 37% of my investment, I have to make back a bit over 37% just to break even.  Do those number show that?  Or is the + starting from 0 again?

[JohnGalt]

The numbers are saying that over every historical period of X years (insert 1, 5, 10, etc years for whatever row you are looking at in the OP numbers), those are the minimum and maximum returns that were seen.  So, for the 5 year period - that means that there is a 5 year period where the average annual returns over that period were -2.35% and that is the lowest of all 5 year periods (1900 - 1905, 1901 - 1906, 1902-1907,...,2007 - 2012). 

MMM's 7% rule comes from the assumption that you will be investing for the long term (25+ years) and the lowest average annual return that has ever actually happened is +7.94%.  So, if you assume that future returns will, at worst, be just slightly worse than the worst historical 25 year period, 7% is a conservative growth rate to expect. 

Obviously, past results are not guaranteed for the future, so you'll ultimately have to decide for yourself what you are comfortable going with - but the historical extremes can provide a good starting point. 

[tooqk4u22]

So, I lose 37% of my investment, I have to make back a bit over 37% just to break even.  Do those number show that?  Or is the + starting from 0 again?

That is not the point of the post but the answer is no - if you lose 37% then you need to get a 59% return to breakeven. 

The ten year data is pretty compelling especially for those that are worried about getting in the market at the wrong time as it says historically that if you invested all of your money in the worse possible time ever that ten years later you would be down only 1% nominally - although you would lose roughly 25% to inflation so in real terms you would be down 26% (although I don't know if the data was inflation adjusted or factored in dividends).

[Mrs MM]

I've put together a simple example below to help (maybe this is too simplistic, but it seems that real numbers were needed to explain better).

Let's say you invest $10,000 in 2012.  You put your money in and you leave it there.  You don't take it out or add any more.

Imagine you bought 1,000 shares each worth $10 (that's $10,000 = 1,000 shares x $10) of some index stock.

At the end of the year, the share price went down 37% (worst case, of course), so the share price is now $6.30.  You still have 1,000 shares, but now your initial investment of $10,000 is down to $6,300.  You don't do anything.  You keep your money in here.

At the end of the next 5 years (of course there are ups and downs throughout the years), the share price has gone down 2.35% from your ORIGINAL amount (as JohnGalt explained below).  So, now your shares are worth $9.77.  You now have $9,770 of your original $10,000.

At the end of 25 years (worst case historically), your shares have gone up +7.94%.  Now your shares are worth $10.79.  Again, you still have 1,000 shares, so now your money is worth $10,790.

If you panic and start to take money out when the market is low and buy when it's high, then things don't look so good.  The secret is just to buy and hold.

Does that make sense?  (hopefully that's all correct - please correct me if I made a mistake!)

When holding for the long term, historically things look pretty good.  One year fluctuations can be pretty big (and certainly day to day can be), but once you're looking at 5+ years, the numbers even out and the graph starts to look like a nice steady upward slope.

[projekt]

Those are annualized growth rates so in your example, you lose $3,700 the first year, but then you lose an average of 2.35% each year for the next five years. So you now have $10,000 x ((1-0.0235)^5) = $8,879. Assuming the worst 25 year period, you would have $10,000 x ((1+0.0794)^25)) = $67,539 after 25 years. Assuming inflation eats 3% of your return each year, that'd be something like $26,277 in today's dollars.

Risks: the biggest gains are often at times of highest inflation, past performance ≠ future results, you might not get the same return as the index, fees may eat into your profits, and many more things that better investors than me can think of. Not to be a complainypants though, as I will be planning to get pretty good returns as I invest, and I will also be diversifying into other investments like MMM has.

[Scottma]

At the end of 25 years (worst case historically), your shares have gone up +7.94%.  Now your shares are worth $10.79.  Again, you still have 1,000 shares, so now your money is worth $10,790.

If you panic and start to take money out when the market is low and buy when it's high, then things don't look so good.  The secret is just to buy and hold.

Does that make sense?  (hopefully that's all correct - please correct me if I made a mistake!)

When holding for the long term, historically things look pretty good.  One year fluctuations can be pretty big (and certainly day to day can be), but once you're looking at 5+ years, the numbers even out and the graph starts to look like a nice steady upward slope.

Thanks for the example Mrs MM! I have a follow up question to your example with regards to the safe withdrawal rate. Say for those 25 years that you were financially independent and withdrawing at the safe withdrawal rate. It seems to me that you would have started to cut deeply into your money and would end up with quite a loss after 25 years, even though the market did go up 7.94%. Is it possible that you could end up using so much of your money up that you will eventually run out? I mean, I guess that's the whole point of having a "safe withdrawal rate", I'm just confused as to how it works...

[JohnGalt]

Thanks for the example Mrs MM! I have a follow up question to your example with regards to the safe withdrawal rate. Say for those 25 years that you were financially independent and withdrawing at the safe withdrawal rate. It seems to me that you would have started to cut deeply into your money and would end up with quite a loss after 25 years, even though the market did go up 7.94%. Is it possible that you could end up using so much of your money up that you will eventually run out? I mean, I guess that's the whole point of having a "safe withdrawal rate", I'm just confused as to how it works...

If you want to add safe withdrawal rates, I think http://www.firecalc.com/ is the best tool.  It will show you how your withdrawal/portfolio would have fared historically assuming you retired in 1901, 1902, 1903,...,1987.  Just enter in the number of years you want to measure, a starting portfolio amount and a spending amount (spending / starting portfolio is the withdrawal rate you are testing).  It uses actual inflation rates to increase spending each year and actual market returns to track how often you would have run out of money. 

Withdrawing only 3% of your starting portfolio each year (increasing for inflation every year) would never have run out of money over any 25 year period for which data is available.
4% would only have run out <2% of the time. 

This is where the safe withdrawal rate numbers come from. 

Of course, if you increase your spending faster than inflation, it falls apart, but firecalc has options to let you run different scenarios like that. 

[grantmeaname]

It seems to me that you would have started to cut deeply into your money and would end up with quite a loss after 25 years, even though the market did go up 7.94%. Is it possible that you could end up using so much of your money up that you will eventually run out? I mean, I guess that's the whole point of having a "safe withdrawal rate", I'm just confused as to how it works...

The 7.94% figure is the annual return. If your money grows by 7.94% and you use it at less than 7.94%, the total assets invested will grow each year. If you're using like 4%, as many early retirees aim to, you'll have your money run away from you and your net worth will increase exponentially as the difference between your spending and investment income grows each year.

[tooqk4u22]

It seems to me that you would have started to cut deeply into your money and would end up with quite a loss after 25 years, even though the market did go up 7.94%. Is it possible that you could end up using so much of your money up that you will eventually run out? I mean, I guess that's the whole point of having a "safe withdrawal rate", I'm just confused as to how it works...

The 7.94% figure is the annual return. If your money grows by 7.94% and you use it at less than 7.94%, the total assets invested will grow each year. If you're using like 4%, as many early retirees aim to, you'll have your money run away from you and your net worth will increase exponentially as the difference between your spending and investment income grows each year.

Technically this is correct but the 4% SWR vs. the 7% ROI is needed to compensate for inflation.  So your net worth will grow in nominal terms but will be flat in real terms as you withdraw.

[grantmeaname]

Right. I was leaving that out for simplicity.

[k9]

To be even more precise, don't forget also that an average 7.94% means there will be good years (with +20%) and bad years (with -10%). Depending on the year you start withdrawing your money, you may have to consume a part of the capital (if the first year your stash loses 10% and you consume 4% more, you have lost 1/7 of it on the very first year). Money lost the very first years will be hard to recover even on good subsequent years.

That and inflation are why you have to choose a conservative figure for the SWR, even if, on average, you could withdraw a little more.

[grantmeaname]

Or, alternately, why you find other work or tighten your belt if a big recession hits in the first few years of your retirement. You only need to be more conservative if you also need to be inflexible (like, if you're living really cheaply with kids and have no room to adjust).