Forget the 4% SWR, or Any SWR - It's All About Income

[foobar]

They are not 100% linked but that is the general problem with divs. It is the  33x portfolio size instead of 25x is what is making it safer. Not the payout in divs. The solution to market volatility isn't dividends. It is that bonds and cash portion of your portfolio that makes you sell  high and buy low at the cost of a some return.

If you're holding a non-dividend paying stock, and it's price gets cut in half, you're forced to sell at those low prices to pay your living expenses.

If you're holding a dividend-paying stock, and the dividend gets cut in half, you're forced to sell some of the stock to make up the difference - and the price when you sell will be low because the dividend just got cut in half.

If you're paid dividends then you don't sell shares at the bottom and at the end of the day the volatility doesn't matter.

Functionally, you do. Collecting a dividend is equivalent to selling that amount of the stock, because your holdings in that stock go down by (the market's expected value of) the dividend received. If the stock price is low that day, then your dividend is equivalent to "selling low".

If you sell 4% of a stock every year, it's totally equivalent to the stock paying you a 4% dividend every year (differing tax treatments of dividends and capital gains aside).

[Eudo]

If you're holding a non-dividend paying stock, and it's price gets cut in half, you're forced to sell at those low prices to pay your living expenses.

If you're holding a dividend-paying stock, and the dividend gets cut in half, you're forced to sell some of the stock to make up the difference - and the price when you sell will be low because the dividend just got cut in half.

True, which is why I said it only "roughly" solves it. But it does reduce it quite a bit. Dividends are far more stable than stock prices, and you have to sell less stock if you also received a dividend. I don't have the numbers in front of me, but if memory serves, during the last financial crisis, dividends on the S&P got cut by about 20% and recovered quickly after a couple years. During the dreaded 1970s, inflation eroded dividends by about 20% of their inflation-adjusted peak and stayed low for probably a decade. That sucks, but if you're living off dividends and you tighten your belt a bit, you can survive without going to plan B (going back to work). Once the market recovers, you're back in business and didn't take a permanent hit to your wealth, unlike what happens when you're forced to sell stock at fire sale prices to live.

But that brings us to your second point...

If you're paid dividends then you don't sell shares at the bottom and at the end of the day the volatility doesn't matter.

Functionally, you do. Collecting a dividend is equivalent to selling that amount of the stock, because your holdings in that stock go down by (the market's expected value of) the dividend received. If the stock price is low that day, then your dividend is equivalent to "selling low".

If you sell 4% of a stock every year, it's totally equivalent to the stock paying you a 4% dividend every year (differing tax treatments of dividends and capital gains aside).

I have to disagree, but you're using a pretty clever argument that forces me to nerd out on finance theory below.

Take a thought experiment with two different scenarios:

(1) You own 1 million shares of $1 stock that doesn't pay dividends. In other words $1 million dollars. You need $40k to survive. The stock takes a drastic 80% drop. Now it's only $0.20 per share. You need $40k to survive, so you are forced to sell  200,000 shares to survive. You only have 800k shares left. Then the stock market recovers to its old mood and the stock is trading for $1 again. You only have $800k, even though the "average return" was 0% and you only took out $40k. This is how portfolios often fail in firecalc simulations, btw.

(2) You own 1 million shares of $1 stock that pays $0.04 in dividends (4%). Again, $1 million. And again, you need $40k to survive. The stock takes a drastic 80% drop. But luckily, since it's paying a dividend of $0.04 per share, you get paid $40k, and you didn't need to sell anything. The market recovers to its old mood, and the stock is trading for $1 again (I know you object to this. See below.) You now have $1 million and you got your $40k living expenses paid for.

Ok, nerds only from here on.

Ok, I think your point is that if the stock paid a dividend back when it was $0.20 per share, then its price will drop by $0.04, so it will only be trading at $0.16. Then when the market recovers, all the stocks see a five-fold increase, so it's now trading at $0.80 and you only have $800k again.

Very clever, but let me explain why I think it's not quite right. It's true that stocks should drop by their dividend amount when it's paid, but there's more to the story than that. They also rise (slowly) before the dividend is paid.

Take an idealized company that is perfectly stable and profitable forever. Its profits never rise and never fall. Constant forever. It sells for $1 a share, and pays a $0.04 dividend. It's true that when a dividend is paid, the share price drops to $0.96. But clearly it can't keep doing this forever or in 25 years the stock price will be zero, even though it's a perfectly profitable company paying a $0.04 dividend! In this case, this idealized company stock is essentially a bond, and its price (in a rational, efficient market) would exhibit the same saw-tooth behavior that the dirty price of a bond has. It falls to 0.96, then slowly rises back to $1.00 and so on forever.

So, to get all theoretical, what should happen in our scenario (2) is this: When the non-dividend-paying stock falls to $0.20, the dividend-paying stock doesn't fall as much because part of the price is the imminent dividend payment. Instead it falls to roughly 0.232. Then it pays its dividend, and the new price is 0.192. Then the market increases 5-fold, and the new price is 0.96. Then throughout the rest of the year the price slowly recovers to $1.00 as before. Ugly, I know. But if you think about it, it makes sense. The market totally recovered to where it was before, with the same expectations that it had before the crash. The market knew the stock was going to pay a dividend. Why would people value the company lower just because it paid its dividend when the price was low?

[warfreak2]

Take a thought experiment with two different scenarios:

(1) You own 1 million shares of $1 stock that doesn't pay dividends. In other words $1 million dollars. You need $40k to survive. The stock takes a drastic 80% drop. Now it's only $0.20 per share. You need $40k to survive, so you are forced to sell  200,000 shares to survive. You only have 800k shares left. Then the stock market recovers to its old mood and the stock is trading for $1 again. You only have $800k, even though the "average return" was 0% and you only took out $40k. This is how portfolios often fail in firecalc simulations, btw.

(2) You own 1 million shares of $1 stock that pays $0.04 in dividends (4%). Again, $1 million. And again, you need $40k to survive. The stock takes a drastic 80% drop. But luckily, since it's paying a dividend of $0.04 per share, you get paid $40k, and you didn't need to sell anything. The market recovers to its old mood, and the stock is trading for $1 again (I know you object to this. See below.) You now have $1 million and you got your $40k living expenses paid for.

You're assuming that after an 80% drop, stock (2) still pays a $0.04 dividend. Obviously, if the assumption is that dividends never go down even in drastic circumstances, then the conclusion will be that dividends are safe income in perpetuity.

[Eudo]

Take a thought experiment with two different scenarios:

(1) You own 1 million shares of $1 stock that doesn't pay dividends. In other words $1 million dollars. You need $40k to survive. The stock takes a drastic 80% drop. Now it's only $0.20 per share. You need $40k to survive, so you are forced to sell  200,000 shares to survive. You only have 800k shares left. Then the stock market recovers to its old mood and the stock is trading for $1 again. You only have $800k, even though the "average return" was 0% and you only took out $40k. This is how portfolios often fail in firecalc simulations, btw.

(2) You own 1 million shares of $1 stock that pays $0.04 in dividends (4%). Again, $1 million. And again, you need $40k to survive. The stock takes a drastic 80% drop. But luckily, since it's paying a dividend of $0.04 per share, you get paid $40k, and you didn't need to sell anything. The market recovers to its old mood, and the stock is trading for $1 again (I know you object to this. See below.) You now have $1 million and you got your $40k living expenses paid for.

You're assuming that after an 80% drop, stock (2) still pays a $0.04 dividend. Obviously, if the assumption is that dividends never go down even in drastic circumstances, then the conclusion will be that dividends are safe income in perpetuity.

Fair enough. But even if the dividend gets cut in half (which is worse than what happened to the S&P during the Great Depression), you'd still only need to make up for $20k instead of $40k, which would mean selling 100k shares, which once the market recovered would be worth $900k instead of $800k, which is how much you'd be left with from a non-dividend-paying stock.

I'm not saying that living only off of dividends, or switching your whole portfolio to high-yield dividend paying stocks is the best move, I'm only trying to show that there is a real effect here. Dividends really can get rid of much of the bad effects of stock volatility. Reaching for yield and buying cruddy companies just to get a high dividend, though, will probably lead to sorrow...

[foobar]

I have seen various numbers for the great depression ranging from 20% to 70%. I think it depends on what subset of the market (20 best div stocks, s&p 500, everything) you are measuring.

The unmentioned thing is what is the rate of return of dividend stocks versus non dividend. I have seen a ton of studies on it and it comes down to it depends. Against the general market, div stocks slightly outperform. BUT the studies also suggest that is strictly because of a value bias. You normalize the value bias and high dividend stocks underperform by a couple of percent.

Take a thought experiment with two different scenarios:

(1) You own 1 million shares of $1 stock that doesn't pay dividends. In other words $1 million dollars. You need $40k to survive. The stock takes a drastic 80% drop. Now it's only $0.20 per share. You need $40k to survive, so you are forced to sell  200,000 shares to survive. You only have 800k shares left. Then the stock market recovers to its old mood and the stock is trading for $1 again. You only have $800k, even though the "average return" was 0% and you only took out $40k. This is how portfolios often fail in firecalc simulations, btw.

(2) You own 1 million shares of $1 stock that pays $0.04 in dividends (4%). Again, $1 million. And again, you need $40k to survive. The stock takes a drastic 80% drop. But luckily, since it's paying a dividend of $0.04 per share, you get paid $40k, and you didn't need to sell anything. The market recovers to its old mood, and the stock is trading for $1 again (I know you object to this. See below.) You now have $1 million and you got your $40k living expenses paid for.

You're assuming that after an 80% drop, stock (2) still pays a $0.04 dividend. Obviously, if the assumption is that dividends never go down even in drastic circumstances, then the conclusion will be that dividends are safe income in perpetuity.

Fair enough. But even if the dividend gets cut in half (which is worse than what happened to the S&P during the Great Depression), you'd still only need to make up for $20k instead of $40k, which would mean selling 100k shares, which once the market recovered would be worth $900k instead of $800k, which is how much you'd be left with from a non-dividend-paying stock.

I'm not saying that living only off of dividends, or switching your whole portfolio to high-yield dividend paying stocks is the best move, I'm only trying to show that there is a real effect here. Dividends really can get rid of much of the bad effects of stock volatility. Reaching for yield and buying cruddy companies just to get a high dividend, though, will probably lead to sorrow...

[warfreak2]

I'm only trying to show that there is a real effect here.

There only appears to be a real effect when you make the assumption that profits paid out as dividends are more reliable than profits used to buy back stock. A dividend payout reduces your holding in the stock by the amount of the dividend (well, the market's expectation of the dividend, but this is more uncertainty, not less) and credits that much to your cash account. This is the same effect as if you had chosen to sell that amount of stock; in effect, the company is buying some stock back from you at the spot price. If the company just used the same money to buy stock back from the market, then the only difference really is that it's up to you how much you will sell them for cash. Reinvesting the dividend is equivalent to choosing to sell none.

Any "imagine two stocks both crash 80%, let's see what happens" hypotheticals which show a difference, are either miscalculated, or hiding an additional assumption. In this case you assume stock (2) is just about to pay out its dividend and then crashes 80%; the dividend paid will be $0.02, but the market expected the dividend to be $0.04, so the price will drop by $0.04, not by $0.02, and your holdings in the company will be smaller because of that. Also, you gave stock (2) an extra year to go from $0.96 to $1 again, but left stock (1) at it's immediate recovery price of $1.

[Nords]

Then I think that the "how do I FIRE" question has not been concisely and sufficiently answered here, at least not since I've been around this site. If it was so straightforward, we wouldn't see so many questions about it.

Buckets of money, for example, I'm sure you're familiar with?  That's a semi-popular one that helps explain and simplify the part you think people are confused on.  People like Brewer probably don't see the point.

I wonder where Ray Lucia is these days.  What an idiot.

An interesting and controversial side effect of the "buckets of money" strategy was that it led to a rising equity asset allocation in retirement, because people were busy emptying buckets before they refilled them.  Then more rebalancing techniques were piled on top of the original buckets idea in order to avoid a high-equity allocation for (hypothetically) aged retirees of declining cognition.  It seemed to get awfully complicated.

Now Kitces & Pfau are beginning to show that a "rising equity glide slope" in retirement might be a good thing.  In their case, however, I'm pretty sure that it's based on years of peer-reviewed research & analysis.  Maybe Lucia could license it from them...

[Another Reader]

Ol' Ray still has a radio show.  He's promoting gold, self help legal forms, and a bunch of questionable stuff on the hour that's broadcast locally on the "business" radio station.  His son is doing wealth management and is big on annuities.  He's on during the next hour, along with a couple of Ray's pals.  Ray's handing off the torch I guess.

Of course, we are practicing the BOM strategy by keeping enough cash to fund a couple of years of expenses beyond the pensions....

[Eudo]

I'm only trying to show that there is a real effect here.

A dividend payout reduces your holding in the stock by the amount of the dividend (well, the market's expectation of the dividend, but this is more uncertainty, not less) and credits that much to your cash account. This is the same effect as if you had chosen to sell that amount of stock;

Except that if you were paid a dividend, you own the same fraction of the company that you owned before, but if you sell the stock, you get cash, but own less of the company.

in effect, the company is buying some stock back from you at the spot price. If the company just used the same money to buy stock back from the market, then the only difference really is that it's up to you how much you will sell them for cash. Reinvesting the dividend is equivalent to choosing to sell none.

Interesting way to look at it, and I think I agree here. If you sell 4% of your stock, and the company simultaneously buys back 4% of its stock, I think I agree that's equivalent to getting a 4% dividend. In that case, you maintain the same fractional ownership of the company, you have 4% of your holdings in cash now, and the company has 4% of its market cap less in cash. Same as a dividend. 

Any "imagine two stocks both crash 80%, let's see what happens" hypotheticals which show a difference, are either miscalculated, or hiding an additional assumption. In this case you assume stock (2) is just about to pay out its dividend and then crashes 80%; the dividend paid will be $0.02, but the market expected the dividend to be $0.04, so the price will drop by $0.04, not by $0.02, and your holdings in the company will be smaller because of that.
 Also, you gave stock (2) an extra year to go from $0.96 to $1 again, but left stock (1) at it's immediate recovery price of $1.

I'll agree that if they cut their dividend from 0.04 to 0.02, then when they make that announcement, the price will drop by 0.02, then on the ex-dividend date it will drop by another 0.02. So a total of 0.04.

And yes, I simplified things by giving the dividend stock a year to recover. It gets a little wishy-washy to get into this kind argument about what the stock price should be theoretically in a year, but if we want an apples-to-apples comparison, we have to make an assumption about the average growth of the two stocks. Say they both give total returns of 7%, so that the first one grows 7% a year and the second grows 3% a year (and pays 4%). Then in a year the first stock is 1.07, but you only have 800k shares. The second is 1.03 and you have all 1 million shares. The second one still ends up better.

But all this talk of exact prices obscures the main point, which is: In case (1), your results depend on the path that prices take during your retirement, because you're constantly selling stock at those prices, but in case (2), your results are path-independent, as long as the dividend doesn't get cut (or at least as long as your living expenses are less than the newly cut dividend).

There seem to be two main objections:

a) Dividends tend to get cut in crashes. Of course, if your dividend is cut, you're worse off. And if the dividend is stopped altogether, you're in the same boat as the non-dividend holding guy. But dividends don't get cut all that often. Only in extreme cases, and even then in a diversified portfolio they don't get stopped completely. The cuts are typically smaller than the price drops. And the rest of the time, the the stock price doesn't matter at all.

b) Non-dividend payers could buy back stock, and if you simultaneously sell an equal percentage of stock, you can simulate dividends. True, and if all your companies always opportunistically buy back stock at rock bottom prices, you'll have great returns. But most don't, so you can't really bank on that. I mean, if you own the S&P index, how are you going to know how much stock you need to sell to balance out share repurchases in order to maintain the same proportionate ownership?

Again, I'm not advocating anything for your portfolio. I personally am more or less indifferent to dividends when I'm buying stock for now. (If it pays large dividends I put it in my tax-advantaged account.) But the question was what is the point of living off dividends, and I think people dismiss it too easily without realizing that the strategy really can mellow out volatility.

Let me put it more starkly. If you did a simulation of early retirees where you had two different scenarios, one with an average 7% annual price appreciation and the other with 3% price appreciation and 4% dividend yield. And both had the same volatility (say 15% annualized volatility). And both had the same withdrawal rate, the first one would have a higher failure rate. There would be some scenarios where the market tanks and stays down for 10 years and the retiree just sells off all his stock.

[MDM]

Quote
in effect, the company is buying some stock back from you at the spot price. If the company just used the same money to buy stock back from the market, then the only difference really is that it's up to you how much you will sell them for cash. Reinvesting the dividend is equivalent to choosing to sell none.

Interesting way to look at it, and I think I agree here. If you sell 4% of your stock, and the company simultaneously buys back 4% of its stock, I think I agree that's equivalent to getting a 4% dividend. In that case, you maintain the same fractional ownership of the company, you have 4% of your holdings in cash now, and the company has 4% of its market cap less in cash. Same as a dividend. 

It's not quite equivalent.  E.g., assume you own 100 shares and the rest of the world owns 100 shares, so you own 50%.  You sell 4 shares that the company buys.  Now you own 96/196 = 49%.  Don't know how that affects the rest of the discussion.

[Eudo]

Quote
in effect, the company is buying some stock back from you at the spot price. If the company just used the same money to buy stock back from the market, then the only difference really is that it's up to you how much you will sell them for cash. Reinvesting the dividend is equivalent to choosing to sell none.

Interesting way to look at it, and I think I agree here. If you sell 4% of your stock, and the company simultaneously buys back 4% of its stock, I think I agree that's equivalent to getting a 4% dividend. In that case, you maintain the same fractional ownership of the company, you have 4% of your holdings in cash now, and the company has 4% of its market cap less in cash. Same as a dividend. 

It's not quite equivalent.  E.g., assume you own 100 shares and the rest of the world owns 100 shares, so you own 50%.  You sell 4 shares that the company buys.  Now you own 96/196 = 49%.  Don't know how that affects the rest of the discussion.

Yeah, I was assuming the company bought back 4% of its entire market cap. So they buy back 8 shares (and you sell 4, so they buy another 4 from the world). Then you own 96/192. This just emphasizes how hard this trick is. You don't have the freedom to just sell as much as you want and call it a "dividend". The percentage of your holdings that you sell has to equal the percentage of the market cap that the company is repurchasing. And you have to do it at the same average price as them. It's interesting, but not really practical.

[warfreak2]

I think the "percentage of the shares owned by you" is a red herring, since you don't really have control over that anyway. If other people sell their shares back to the company, your percentage goes up, and if the company issues more shares, your percentage goes down. Stock (1), which pays no dividend, has the company using their profits to buy back shares at the spot price (instead of using the same profits to pay a dividend), so if you sell the same amount as what the dividend would be, then you meet this criteria:

The percentage of your holdings that you sell has to equal the percentage of the market cap that the company is repurchasing. And you have to do it at the same average price as them.

[steveo]

Most people saving for ER are not very interested in continuing to work, so they might find the 4% SWR much more compelling.

I prefer the 4% SWR based on this point. I'm interested in quitting work not working till I get to some state where I 100% confident the money will never run out.

[arebelspy]

Most people saving for ER are not very interested in continuing to work, so they might find the 4% SWR much more compelling.

I prefer the 4% SWR based on this point. I'm interested in quitting work not working till I get to some state where I 100% confident the money will never run out.

Doubly true, since there's no such thing as 100% sure.  :)

[steveo]

Most people saving for ER are not very interested in continuing to work, so they might find the 4% SWR much more compelling.

I prefer the 4% SWR based on this point. I'm interested in quitting work not working till I get to some state where I 100% confident the money will never run out.

Doubly true, since there's no such thing as 100% sure.  :)

Definitely. To me the 4% rule is a reasonable rule to base being FI on and therefore being able to retire. I think we all will utilise different asset allocations that each of us will feel safe with. Some people will probably be 100% stocks whereas others 100% bonds. Some will own their house or have rental properties.

I will personally own my own house, not count that as part of my total assets and have a high equity allocation excluding my own house. I also have 3 kids so I figure my spending will decrease not increase over the course of my retirement so 25x my current spending to me is fine. I consider myself having FU level of savings when my assets are at less than current spending 25x level but at myself and my wife's expected 25x spending level. Once I reach FU level I'm ready to start looking towards actually retiring.

[foobar]

If they have the same return, they will have the same failure rate if you take the same money out (there is a bit of hand waving as dividends tend to trail stock prices).  That stock paying out 2% in divs is either going to have share price that is dropping 2% more than the 0 div stock OR it is is outperforming.

Personally I expect a lot of traditional high dividend stocks to outperform the market during bear markets. It has what they have done in the past. The problem of course is they underperform in during the bull markets.  Having 150k cut to 75k during a bear market is bad. Having 100k cut to 75k leaves you in the same position.


People like to focus on the great depression but the 70s were worse (from a pure stock market point of view not a total economy point of view). Between Jan 1929 and 1939, you had  a real return  (after divs and inflation) of 5.33%.  And you were making a killing on your bonds. Between 1973 and 1983, you had a return of .16% and at the same time your bonds were losing money.  The good news is that we have products available today to help deal with that situation.

Let me put it more starkly. If you did a simulation of early retirees where you had two different scenarios, one with an average 7% annual price appreciation and the other with 3% price appreciation and 4% dividend yield. And both had the same volatility (say 15% annualized volatility). And both had the same withdrawal rate, the first one would have a higher failure rate. There would be some scenarios where the market tanks and stays down for 10 years and the retiree just sells off all his stock.

[Eudo]

Let me put it more starkly. If you did a simulation of early retirees where you had two different scenarios, one with an average 7% annual price appreciation and the other with 3% price appreciation and 4% dividend yield. And both had the same volatility (say 15% annualized volatility). And both had the same withdrawal rate, the first one would have a higher failure rate. There would be some scenarios where the market tanks and stays down for 10 years and the retiree just sells off all his stock.

If they have the same return, they will have the same failure rate if you take the same money out (there is a bit of hand waving as dividends tend to trail stock prices).  That stock paying out 2% in divs is either going to have share price that is dropping 2% more than the 0 div stock OR it is is outperforming.

I'm talking about a simulation where the second stock never has a dividend cut. For example, a $1.00 stock paying a $0.04 dividend. If the stock price falls to $0.50, then it still pays its $0.04 dividend (which is now 8%).

 A lot of people might think that's unrealistic, but the realistic case isn't as bad as a lot seem to think. To use your example, in 1973 the S&P was at 118.40 and paid a dividend of 3.16 (or 2.7%). In 1983, it was at 144.30 and paid a dividend of 6.88 (or 4.8%). Adjusting for inflation, the new dividend was 3.00. So, yeah, after adjusting for inflation the dividend got cut by about 5%. Not the end of the world, though. (I'm using data from Schiller's excellent excel spreadsheet for 1/1973 to 1/1983, so our numbers might not agree exactly depending on which dates you used.)

But back to the simulation I proposed. Let me put it even more starkly, if the dividend does not get cut in dollar terms, then if you have a 4% withdrawal rate, failure is mathematically impossible in the second case, because the money our retiree is living off of is exactly equal to the dividend in dollars, and that dollar dividend is never cut. But in the first case clearly you still can have failure, since the results of our robotic 4%-withdrawing retiree depend on what stock prices are.

[warfreak2]

As I said, obviously if you assume that dividends never go down then the conclusion will be that dividends are a secure income stream. That's a tautology, but a silly assumption. You might as well assume that stock (1) always grows at least 4% per year, it's the same assumption.

[Ziggurat]

I'd like to thank everyone for their posts here, very helpful for me in getting my head around all these issues.

I was leaning toward dividend growth style investing, but a couple of "aha" moments reading here have turned me back:

1. As mentioned, living off dividends requires a bigger stash to begin with.  That bigger stash would also give more safety to a fixed SWR route.

2. Dividend growth seems so safe because over the last X years, these big solid companies are still around and have not cut (or only cut a little) their dividends. Over the longer term, however, there is no guarantee those companies will still even exist, much less continue to thrive.  It requires the investor to spend a lot of time "picking stocks", and as we know, the vast majority of experts have not picked stocks successfully, certainly not enough to surpass the indexes.

This discussion also prompted me to read more, and I found a great series on canadian couch potato debunking dividend myths too, starting here (http://canadiancouchpotato.com/2011/01/18/debunking-dividend-myths-part-1/). I particularly like this part "Investors categorize $1,000 in dividends as income that they will happily spend, but the idea of selling $1,000 worth of stock is “dipping into capital,” which causes them great anxiety" (true of me, I'll have to fight it).  Also, in part 6, they debunk the yield-on-cost idea, showing that you can sell your stock and immediately buy it back and get the exact same dividend income from it. 

Having said that, I will probably go for a bit of a hybrid approach, as I will likely need a somewhat stable income source to bridge for 10 years or so until taking pension.  Over short periods starting with a partial stash, if you will, some portion in dividend-solid companies would help smooth the bumps.  Over a small period, the risk of their demise is quite small.  Or, is even that kind of thinking not correct?

[Eudo]

As I said, obviously if you assume that dividends never go down then the conclusion will be that dividends are a secure income stream. That's a tautology, but a silly assumption. You might as well assume that stock (1) always grows at least 4% per year, it's the same assumption.

So I think we agree that my argument hinges on dividends being far more stable than stock prices. You think it's silly to assume that. I think it might be silly to assume they never go down at all, but it's not silly to assume they are far more stable than prices. In the '73-'83 period (which was picked by someone else because it was one of the worst in history for retirees), inflation-adjusted dividends dropped by only 5%.

What we really need as a test is to have data for two stock indices during that period. Say, the S&P non-dividend payers and the S&P dividend payers. Assume a retiree robotically withdraws 4% a year the first year and increases it by inflation. See where each one stands in 1983. I don't have the data, but I'm pretty sure I know what we'll see. If I have time I can do a similar test. I know XLU (utilities dividend paying ETF) data goes back to 1999. We could do the test on the S&P, which paid pathetic dividends back then vs XLU. I haven't done the test, but again I'm pretty sure I know the outcome.

[warfreak2]

So you don't have actual evidence to back you up, but you're nonetheless sure what the outcome will be?

Businesses have three ways to use their profits: they can (A) pay a dividend, (B) buy back stock, or (C) reinvest to grow the business. (I'll ignore (C) for now since it's harder to analyse and so far we've been discussing the difference between (A) and (B) anyway.)

If you're a stockholder, then when the business pays out its profits to shareholders, you receive a share of those profits in proportion to the stock you hold. You can receive that same amount of money either (A) in cash, or (B) in your stock balance. The amount of money is decided by the profits of the business, and the proportion of stock that you hold - but not by the method it's paid out (taxes aside).

The actual profits a business makes, and the proportion of them that you receive, are not affected by the method that it pays them out. The only difference is that in (B), it's up to you how much you convert into cash.

[MDM]

I don't have the data, but I'm pretty sure I know what we'll see. If I have time I can do a similar test. I know XLU (utilities dividend paying ETF) data goes back to 1999. We could do the test on the S&P, which paid pathetic dividends back then vs XLU. I haven't done the test, but again I'm pretty sure I know the outcome.

So you don't have actual evidence to back you up, but you're nonetheless sure what the outcome will be?

If forced, I might bet that warfreak2 is correct.  But I also have to give Eudo kudos for intellectual honesty: he has put forth a hypothesis, but also a test that could falsify it.  From that link: "
1) Devising alternative hypotheses;
2) Devising a crucial experiment (or several of them), with alternative possible outcomes, each of which
will, as nearly as possible, exclude one or more of the hypotheses;
3) Carrying out the experiment so as to get a clean result;
1') Recycling the procedure, making subhypotheses or sequential hypotheses
to refine the possibilities that remain;
and so on."

[warfreak2]

I'm actually not at all sure I would win that bet! It's possible that dividend-paying stocks (A) are more secure than stocks from businesses which use their profits to buy back shares (B). My point is that there's no mathematical reason for them to be different, and that the arguments presented so far don't establish a difference. However, an invalid argument does not an incorrect conclusion make - for example, perhaps reliably profitable businesses have some other reason to prefer paying out dividends than buying back stock?

[beltim]

Category A beats category B for the same reason that most experts trail stock indices: company management is more likely to buy back stock when times are good are their share price is high than when the share price is low (when they would get the best deal).

https://doc.research-and-analytics.csfb.com/docView?sourceid=em&document_id=x454556&serialid=e45HdNbTyfzxqJiuVkQbC5cPZZDE1SdAOSK1XWRCbxs=

Edit:  Actually, I guess it's closer to the reason that most individual investors trail broad stock-market indices: poor timing.

[Nords]

Having said that, I will probably go for a bit of a hybrid approach, as I will likely need a somewhat stable income source to bridge for 10 years or so until taking pension.  Over short periods starting with a partial stash, if you will, some portion in dividend-solid companies would help smooth the bumps.  Over a small period, the risk of their demise is quite small.  Or, is even that kind of thinking not correct?

That hybrid approach is also known as "diversification".

If you want to get into picking dividend stocks (which is a lot of work when done correctly) then you could start with the Dividend Growth Investor's blog (http://www.dividendgrowthinvestor.com/) and with a library copy of Josh Peters' "The Ultimate Dividend Playbook". 

DGI is a blogging animal who puts out several posts per week, so it's probably better to subscribe via e-mail or RSS and then dig deeper when one of his posts catches your eye.  His key is focusing on the 300 or so stocks which have paid rising dividends for at least a decade, and diversifying among sectors.  If a holding cuts its dividend then he immediately sells, otherwise he's likely to hold for decades.

Personally I think it's easiest to buy an index fund or ETF like the Dow Dividend ETF (DVY).  It came through the 2008-09 recession with a 10%-15% dip in dividends (which has since returned), but if you kept DCA'ing into the fund then today you'd be looking at a double-digit yield on your original cost. 

[foobar]

I would bet on you losing the bet. Obviously to some extent it is going to come down to how you pick the div stocks but the traditional divs stocks (think Utils) tend to outperform in  bear markets. But the world doesn't start in 1973 or end in 1983. The better performance in the bear market might have followed a decade of underperformance on either side.  If it was a no brainer that you could have the same return and less variance by an investment strategy, everyone would do it.

I don't want to come across as anti divs. They are nice. But you don't want to chase them. I would rather retire with an S&P 500 portfolio with a 2% div than some subsector that you pick out that pays 4%.

I'm actually not at all sure I would win that bet! It's possible that dividend-paying stocks (A) are more secure than stocks from businesses which use their profits to buy back shares (B). My point is that there's no mathematical reason for them to be different, and that the arguments presented so far don't establish a difference. However, an invalid argument does not an incorrect conclusion make - for example, perhaps reliably profitable businesses have some other reason to prefer paying out dividends than buying back stock?

[Eudo]

If I have time I can do a similar test. I know XLU (utilities dividend paying ETF) data goes back to 1999. We could do the test on the S&P, which paid pathetic dividends back then vs XLU. I haven't done the test, but again I'm pretty sure I know the outcome.

-----
EDIT: Originally, I used the wrong data for XLU and got much more favorable results for it. I've corrected the data and changed some of the conclusions.
----

Ok, perhaps having a little OCD in me, I did the exercise. I looked at the price and dividends of XLU (a high dividend paying ETF) each year since 1999 (the first full year I have data for). Then I compared it with the S&P and its dividends during the same period. I created two fictitious retiree robots who started with $1 million each at the beginning and started withdrawing $40k a year, and increased their withdrawals with inflation each subsequent year. The data is below for all to critique and find my mistakes in, but here are the results:

The retiree who put all his money in the S&P ended up with $680k at the end of 2013. He's currently drawing down $56k a year, which is almost 10% of his portfolio, so he will be broke soon without moving to plan B. His dividends are paying him $13k.

The retiree who put all his money in XLU ended up with $902k at the end of 2013. He's also drawing down $56k a year and his dividends are paying 35k of that. Unfortunately for him, though, he's still taking a big chunk of his portfolio each year, so he probably won't have great results.

So in the 1999-2013 period, the dividend method works better than the non-dividend method. But it's important to not learn the wrong lessons from this. It's important to understand why it worked. The price of XLU was probably more stable than the price of the S&P, but that's not the main reason this worked. Look at the "XLU Div" column and compare with the "Withdrawal" column. His dividends were a large part of his withdrawals. When your dividends are higher than your withdrawals, the price of the stock you own doesn't matter. You're never selling stock, so what do you care what the price is? The important thing is that for this to work, your dividend must be somewhat stable and higher than your living expenses. The reason it this didn't work perfectly for retiree #2 was that even though XLU pays higher dividends, it still started with a low rate of 1.85%, so he was still selling stocks. If someone has their favorite dividend ETF that yields closer to 4% in 1999, I'm pretty sure you'll see much better results. It was the dot com bubble after all. And I keep claiming that dividends are very stable compared to stock prices, and apparently a lot of people don't believe it. This example encompassed both the dot com crash and the '08 stock crash and you can see that dividends were fairly stable.

Meanwhile, the S&P investor sold all his shares at the bottom. If you've ever looked at portfolio simulations, you know that this is the way portfolios fail: you sell your shares in a bear market, then the market eventually recovers, but you don't have any shares left, so it doesn't matter. You still go broke.

Keep in mind that there's a cost to this out-performance. This was a bear market. If we instead looked at '83-'93, I'm pretty sure you'd see the opposite. The non-dividend paying portfolio would outperform the dividend-payer. (Again, I haven't done the test, but I'm pretty sure I know how it would turn out :) ) That's the whole point. The original question was, "Why live off dividends?" and my answer was, "Because it gets rid of the sequence of returns risk, roughly." It gives in bear markets and takes away in bull markets. It's not for everyone, but the conventional wisdom is that if you need the income, a dividend portfolio will be more stable for you, and in this case I think the conventional wisdom is correct.

-------
Ok, here's the data. I got the data for XLU from Yahoo finance, and asked for monthly data for price and all dividends. (That's why the dates aren't exactly the first trade date of the year in all cases). The S&P and CPI data come from the Schiller Excel spreadsheet I mentioned in a different reply.

Date   CPI   XLU Price   XLU Div   XLU Div Rate   Beginning Balance   Capital Gain   Dividend   Withdrawal   Ending Balance
1/4/1999   164.30   30.23   0.56   1.85%   1,000,000    -85,676   18,525   40,000    892,848 
1/3/2000   168.80   27.64   0.964   3.49%   892,848    118,874   31,140   41,096    1,001,767 
1/2/2001   175.10   31.32   0.667   2.13%   1,001,767    -137,535   21,334   42,629    842,936 
1/2/2002   177.10   27.02   0.92   3.40%   842,936    -259,557   28,701   43,116    568,964 
1/2/2003   181.70   18.7   0.797   4.26%   568,964    155,476   24,249   44,236    704,454 
1/2/2004   185.20   23.81   0.874   3.67%   704,454    136,985   25,859   45,088    822,209 
1/3/2005   190.70   28.44   1.01   3.55%   822,209    107,257   29,199   46,427    912,238 
1/3/2006   198.30   32.15   0.783   2.44%   912,238    125,983   22,217   48,278    1,012,161 
1/3/2007   202.416   36.59   1.095   2.99%   1,012,161    72,752   30,290   49,280    1,065,923 
1/2/2008   211.08   39.22   1.235   3.15%   1,065,923    -278,303   33,565   51,389    769,796 
1/2/2009   211.143   28.98   1.273   4.39%   769,796    14,610   33,815   51,404    766,816 
1/4/2010   216.687   29.53   1.004   3.40%   766,816    56,868   26,071   52,754    797,001 
1/4/2011   220.223   31.72   1.368   4.31%   797,001    74,373   34,373   53,615    852,133 
1/3/2012   226.665   34.68   1.123   3.24%   852,133    46,685   27,594   55,183    871,228 
1/2/2013   230.28   36.58   1.465   4.00%   871,228    52,398   34,892   56,063    902,455 

Date          CPI   S&P Price   S&P Div   S&P Div Rate   Beginning Balance   Capital Gain   Dividend   Withdrawal   Ending Balance
1/4/1999   164.30   1248.77   16.28   1.3%   1000000   141,595   13,039   40,000    1,114,635 
1/3/2000   168.80   1425.59   16.57   1.2%   1,114,635    -70,338   12,958   41,096    1,016,160 
1/2/2001   175.10   1335.63   16.17   1.2%   1,016,160    -148,677   12,302   42,629    837,156 
1/2/2002   177.10   1140.21   15.74   1.4%   837,156    -179,419   11,554   43,116    626,174 
1/2/2003   181.70   895.84   16.12   1.8%   626,174    165,435   11,268   44,236    758,640 
1/2/2004   185.20   1132.52   17.60   1.6%   758,640    32,750   11,790   45,088    758,091 
1/3/2005   190.70   1181.41   19.70   1.7%   758,091    62,449   12,643   46,427    786,756 
1/3/2006   198.30   1278.73   22.41   1.8%   786,756    89,478   13,788   48,278    841,744 
1/3/2007   202.416   1424.16   25.08   1.8%   841,744    -26,833   14,825   49,280    780,457 
1/2/2008   211.08   1378.76   27.92   2.0%   780,457    -290,489   15,804   51,389    454,383 
1/2/2009   211.143   865.58   28.01   3.2%   454,383    135,436   14,704   51,404    553,118 
1/4/2010   216.687   1123.58   22.24   2.0%   553,118    78,293   10,950   52,754    589,607 
1/4/2011   220.223   1282.62   22.96   1.8%   589,607    8,256   10,556   53,615    554,804 
1/3/2012   226.665   1300.58   26.74   2.1%   554,804    76,708   11,405   55,183    587,734 
1/2/2013   230.28   1480.4   31.54   2.1%   587,734    135,762   12,520   56,063    679,953 

[MDM]

...
I looked at the price and dividends of XLU (a high dividend paying ETF) each year since 1999 (the first full year I have data for).

Eudo, nice work and fairly presented with appropriate disclaimers. 

Went looking for more info on XLU (don't have that currently).  Found a few links with charts that look similar to this one:


The starting 2013 value shows ~35, similar to your table, but the starting 1999 value appears closer to 30 instead of the 17.87 in your table.  Is there more than one XLU or am I missing something...?

[Eudo]

The starting 2013 value shows ~35, similar to your table, but the starting 1999 value appears closer to 30 instead of the 17.87 in your table.  Is there more than one XLU or am I missing something...?

No, it was my mistake. Yahoo finance provides data for the "adjusted stock close". I thought it was adjusted for splits only, but it was adjusted for splits and dividends. So I was effectively double counting dividends. I've corrected the original post with the correct data and changed some of the conclusion. The bottom line is that XLU still out-performed the S&P for the retiree over that time period, but not by nearly as much. (You end up with $902k, not $2.3 million. This is compared with the S&P which gives $680k). Big difference quantitatively, but same basic conclusion. The dividends even out the steep drops.

[foobar]

XLU was at 30.02 . The 17.87 is an adjusted price that accounts for dividends and splits. I am guessing XLU never split and that dividends were basically double counted:). Since dec 1998 (inception), XLU has returned 5.83% with divs reinvested. Your money will not last 30 years with that type of return. Of course the s&p 500 returned 4.7% so your not living large off that either:)
 
You can look at DJU to see how bad this index underperformed from 1985 to 2000.  Looking at single 10-15 year periods isn't overly useful. You need to look at a ton of them and figure out what the best and worst cases are.

...
I looked at the price and dividends of XLU (a high dividend paying ETF) each year since 1999 (the first full year I have data for).

Eudo, nice work and fairly presented with appropriate disclaimers. 

Went looking for more info on XLU (don't have that currently).  Found a few links with charts that look similar to this one:


The starting 2013 value shows ~35, similar to your table, but the starting 1999 value appears closer to 30 instead of the 17.87 in your table.  Is there more than one XLU or am I missing something...?

[Eudo]

XLU was at 30.02 . The 17.87 is an adjusted price that accounts for dividends and splits. I am guessing XLU never split and that dividends were basically double counted:). Since dec 1998 (inception), XLU has returned 5.83% with divs reinvested. Your money will not last 30 years with that type of return. Of course the s&p 500 returned 4.7% so your not living large off that either:)
 

Yes, that's exactly what happened. (It's been corrected in the original post.) And my conclusions exactly. (XLU out-performed, but neither retiree would be sleeping well at night these days.)

[warfreak2]

The price of XLU was probably more stable than the price of the S&P, but that's not the main reason this worked. Look at the "XLU Div" column and compare with the "Withdrawal" column. His dividends were a large part of his withdrawals. When your dividends are higher than your withdrawals, the price of the stock you own doesn't matter.

This is a fundamental misunderstanding. The price of a stock reflects how reliably profitable the underlying business is, and the dividends are your share of the profits. The price of the stock obviously does matter, because a low price tends to mean that the business isn't doing so well, and that means it can't pay out its usual dividends, so to make up the difference you are going to have to sell some of your stock, and when you do so is going to be when the stock price is low.

Obviously the price of XLU was more stable because the dividends were reliable.

You're never selling stock, so what do you care what the price is?

As I said twice already, if you (ridiculously) assume "you're never selling stock", you are assuming that dividends are a reliable source of income, which is exactly what you are trying to prove in the first place.

-----

You managed to pick an ETF which (over this time period) paid reliable dividends, which means the businesses profited reliably. Moreover, over the 15 year period, if you include dividends, XLU grew by 130%, while VFINX (Vanguard's S&P 500 index fund) grew by 90%. Annualised, XLU grew at 5.6% and VFINX at 4.3% - no shit, it was safer to withdraw from XLU. What you have shown is that if you pick an ETF that beats the S&P 500, then you can beat the S&P 500.

If you want to do this test in a fair way, you need to compare a dividend stock/ETF which, with dividends reinvested, would have performed pretty much equally to the S&P 500. If you want to compare the safety of your withdrawal, you need to compare between investments that do equally well with no withdrawal.

[arebelspy]

The price of XLU was probably more stable than the price of the S&P, but that's not the main reason this worked. Look at the "XLU Div" column and compare with the "Withdrawal" column. His dividends were a large part of his withdrawals. When your dividends are higher than your withdrawals, the price of the stock you own doesn't matter.

This is a fundamental misunderstanding. The price of a stock reflects how reliably profitable the underlying business is, and the dividends are your share of the profits. The price of the stock obviously does matter, because a low price tends to mean that the business isn't doing so well, and that means it can't pay out its usual dividends, so to make up the difference you are going to have to sell some of your stock, and when you do so is going to be when the stock price is low.

Except for the fact that during a market crash, a solid business that is doing fine canand will have its stock price decrease simply due to the panic in the markets.  Big banks failing, for example, can cause unrelated stock price drops.  This doesn't mean that the unrelated business is not doing so well (aside from a "well the economy as a whole is worse, so it will affect this company as well ... maybe, somewhat, yes).

Yes, what I'm saying is during those (irrational) times, you may be smarter than the market.  Not able to time it, per say, but the market could be undervaluing a business that has no need to cut its dividends, or not by as much as the share price would indicate it should.

In other words, the share price reflects how profitable the business is, but it doesn't always accurately reflect that.

[foobar]

You could go looking for another product maybe SDY ( the "aristocratic" div index) or REITs (they did ok in 2000-2002 which might be enough to  survive an a really brutal 2007-2008 stretch) to try an make the "income" strategy work but eventually you will find  that they have a 15 year period really doesn't look good when they are out of favor and if that period of time happens to start at the same time as your retirement, your screwed.

The big worry for me isn't surviving the bad years at the start sequence risk.  In those cases it is easy to go back to work. It is investing in underperforming assets that result in you having a declining standard of living and you being 65 and not in a great place to work.

XLU was at 30.02 . The 17.87 is an adjusted price that accounts for dividends and splits. I am guessing XLU never split and that dividends were basically double counted:). Since dec 1998 (inception), XLU has returned 5.83% with divs reinvested. Your money will not last 30 years with that type of return. Of course the s&p 500 returned 4.7% so your not living large off that either:)
 

Yes, that's exactly what happened. (It's been corrected in the original post.) And my conclusions exactly. (XLU out-performed, but neither retiree would be sleeping well at night these days.)

[warfreak2]

Except for the fact that during a market crash, a solid business that is doing fine canand will have its stock price decrease simply due to the panic in the markets.  Big banks failing, for example, can cause unrelated stock price drops.  This doesn't mean that the unrelated business is not doing so well (aside from a "well the economy as a whole is worse, so it will affect this company as well ... maybe, somewhat, yes).

Yes, what I'm saying is during those (irrational) times, you may be smarter than the market.  Not able to time it, per say, but the market could be undervaluing a business that has no need to cut its dividends, or not by as much as the share price would indicate it should.

In other words, the share price reflects how profitable the business is, but it doesn't always accurately reflect that.

Yes, you're correct that I'm assuming efficient markets. I readily admit that there are various differences between dividend-paying stocks and non-dividend-paying stocks.

E.g., when a dividend is paid, the stock price goes down by what the market "expected the dividend to be". When the market underestimates dividends, the buy-and-hold investor receives more, and conversely if the market overestimates the dividends then the buy-and-hold investor receives less. What you propose is a psychological reason that the market might systematically underestimate the dividends.

I only argue against the alleged mathematical difference, which wasn't made with reference to market inefficiency.

[arebelspy]

Agreed, there should be no difference in theory.

That just may be one reason why there could be a difference in practice.

I don't have enough data to determine either way, and I tend towards total market for other reasons (and cap gains in general over dividends for tax reasons).  But I can see why one would want to say it works that way.

[totoro]

FWIW I don't use the 4% rule as my retirement calculator, although I do think it is a revelation.

I don't because:

1. I don't feel secure relying on the market for all time periods and needs.  I'm not a statistician/mathematician/engineer.  I do have experience with investments that did not perform well and I did need to cash out some of them at a loss.  I know that this was my lack of knowledge at the time, but I still am wary of relying on an index fund and I'm not naturally bent in a way that I can relax on the basis of probabilities and past performance.
2. I won't have the same financial needs each year due to paying for kids' education or receiving a pension.
3. I believe you need to save less if you use leverage plus tax planning to create an income stream through the use of rental properties and business dividends or eventual sale (yes, it is a bit more work).  Both rentals and businesses are responsive to my strategic planning/efforts/creative thinking - unlike an index fund.
4. I want my assets to continue to grow so that I can leave a charitable legacy or provide for the unexpected like a grandchild with a disability.

[foobar]

1,2, and 4 are not really issues. They just requiring planning. 3 is the interesting one to think about. In real estate is gets hard to the same numbers that we get for the stock market so it is hard to evaluate risk. For example how did the landlord who owned 10 units due during the great depression? Or in a more recent case, what if you owned your units in 80s right as the bust started or if you loaded up on Detroit property in 1990? There used to be some really cool real estate empire blogs in the mid 2000 as people assembled empires. They pretty much have all gone silent when those 3 million in gains disappeared over the next 24 months and those cash flow positive properties turned bad.  Investing in the peaks of anything is always bad.

That being said, if I am in the right area in during my retirement, I expect to go back to owning rentals. Diversification of risk and all that. And no I don't think REITs perform the exact same role.

FWIW I don't use the 4% rule as my retirement calculator, although I do think it is a revelation.

I don't because:

1. I don't feel secure relying on the market for all time periods and needs.  I'm not a statistician/mathematician/engineer.  I do have experience with investments that did not perform well and I did need to cash out some of them at a loss.  I know that this was my lack of knowledge at the time, but I still am wary of relying on an index fund and I'm not naturally bent in a way that I can relax on the basis of probabilities and past performance.
2. I won't have the same financial needs each year due to paying for kids' education or receiving a pension.
3. I believe you need to save less if you use leverage plus tax planning to create an income stream through the use of rental properties and business dividends or eventual sale (yes, it is a bit more work).  Both rentals and businesses are responsive to my strategic planning/efforts/creative thinking - unlike an index fund.
4. I want my assets to continue to grow so that I can leave a charitable legacy or provide for the unexpected like a grandchild with a disability.

[arebelspy]

There used to be some really cool real estate empire blogs in the mid 2000 as people assembled empires. They pretty much have all gone silent when those 3 million in gains disappeared over the next 24 months and those cash flow positive properties turned bad.  Investing in the peaks of anything is always bad.

I'd wager the vast majority of them were investing for appreciation, not cash flow, and were massively overleveraged.

If you've done it right, a recession and corresponding 20% drop in income and increase in vacancies will hurt, sure, but won't make you cash flow negative or cause you to lose any properties.

Do you have any links?  I'd love to read some, whether to laugh or learn, even if they haven't had new content for years.

[totoro]

Investing for appreciation seems too risky to me and where I live in Canada I don't believe we will see appreciation for a while.  My view is that you want to look at cash flow and not care about appreciation/depreciation because you are never forced to sell. You choose the sale and time it to market conditions.

In our case, a nice four-bed/two-bath triplex unit in our neighbourhood is at least $500 000 per unit and often much more when you can find one.  If we wanted to stay in this neighbourhood we would have to pay a house off and invest about one million additionally to use the 4% rule and retire.  We want to stay in this neighhbourhood. 

We ended up buying a run-down multi-family with three units that is not strata (units can't be sold individually so valuation/unit is lower) for $750,000 with $150 000 down. We put an additional $50,000 in on repairs and renovations.   Total investment of $200 000. The rents from the other two units cover our mortgage and expenses.  It took us over a year of looking to find a place that would work this way.

In the end, instead of paying the principal and interest from our after-tax wages for a similar house, we have used only $200 000 and we own three units.  In addition, when the mortgage is paid off by rental income, we will have $800,000 in today's dollars in equity plus $3000/month in income from rent ongoing and a place to live mortgage-free or rent out if we want to travel.

In addition, this is our primary residence the capital gains on the percent of the structure we occupied will be tax exempt if we sell.  The rental income and expenses are deductible against interest creating a disincentive to pay off early as when it is paid off we will be in a lower tax bracket.

Rather than getting to one million invested and the amount required to pay off a house on after-tax wages, this strategy could allow someone to retire on substantially less invested capital over a shorter period of time.  It also lets them invest the portion they would have paid for a mortgage into other income-producing assets, like an index fund, along the way.

While I know this works for us, I also know that this is not for everyone.  We have relatives and friends that think it is quite odd that we live in a carriage house.  I think we get put into the "not very successful" box.

[foobar]

I will have to go through biggerpockets and see if I can dredge them up. The general pattern they were in hot places like Vegas. They would buy and hold a bunch of units and would constantly be refinancing out (units were still cash flow positive in general but often times the margin was reduced). The the crash hit and vacancies doubled (or worse the tenant stopped paying but had to be evicted)  and rents dropped and then went from cash flow positive to negative. And of course that 2 million dollars in real estate was now worth 1 million but they owed the bank 1.5 million.  Mortgage payments stop getting made and the blogs go dead.


It is definitely easy to see the mistakes now (investing in bubble locations) and some even at the time (they were far to aggressive with leverage). But how much of that is like saying don't invest in tech stocks in 1999 and you should be 100% in stocks?:)

There used to be some really cool real estate empire blogs in the mid 2000 as people assembled empires. They pretty much have all gone silent when those 3 million in gains disappeared over the next 24 months and those cash flow positive properties turned bad.  Investing in the peaks of anything is always bad.

I'd wager the vast majority of them were investing for appreciation, not cash flow, and were massively overleveraged.

If you've done it right, a recession and corresponding 20% drop in income and increase in vacancies will hurt, sure, but won't make you cash flow negative or cause you to lose any properties.

Do you have any links?  I'd love to read some, whether to laugh or learn, even if they haven't had new content for years.

[arebelspy]

I will have to go through biggerpockets and see if I can dredge them up. The general pattern they were in hot places like Vegas. They would buy and hold a bunch of units and would constantly be refinancing out (units were still cash flow positive in general but often times the margin was reduced). The the crash hit and vacancies doubled (or worse the tenant stopped paying but had to be evicted)  and rents dropped and then went from cash flow positive to negative. And of course that 2 million dollars in real estate was now worth 1 million but they owed the bank 1.5 million.  Mortgage payments stop getting made and the blogs go dead.


It is definitely easy to see the mistakes now (investing in bubble locations) and some even at the time (they were far to aggressive with leverage). But how much of that is like saying don't invest in tech stocks in 1999 and you should be 100% in stocks?:)

I lived through the Vegas bubble burst.  Vacancies went up a bit, but doubling .. maybe from like 4% to 8%?  And rents stagnated, but didn't drop (all the foreclosed homeowners suddenly needed a place to rent)

Like I said:

I'd wager the vast majority of them were investing for appreciation, not cash flow, and were massively overleveraged.

If you've done it right, a recession and corresponding 20% drop in income and increase in vacancies will hurt, sure, but won't make you cash flow negative or cause you to lose any properties.

I'm sure they were cash flow negative (in fact I know it, because our prices from 05-07 couldn't have been cash flow positive based on our rents, unless you put down something like a 70% down payment) and speculating on appreciation.  Whoops.  :)

[Eudo]

This is a fundamental misunderstanding. The price of a stock reflects how reliably profitable the underlying business is, and the dividends are your share of the profits. The price of the stock obviously does matter, because a low price tends to mean that the business isn't doing so well, and that means it can't pay out its usual dividends, so to make up the difference you are going to have to sell some of your stock, and when you do so is going to be when the stock price is low.

Ok, so I just want to make sure we’re on the same page here, so I know where the disagreement is. Do you agree that

(a)   If the dividend of a stock is not cut, and you are living on less than the size of the dividend, then it is mathematically impossible for your portfolio to fail, no matter what price the stock trades at?

(b)   If you are living on the same amount as (a) but own non-dividend paying stocks it is possible for your portfolio to fail if the stock price falls low enough for long enough?

It sounds to me like you also believe these statements, but are arguing that it doesn’t matter because whenever stock prices fall, dividends are cut. I think this is just false, and it has nothing to do with inefficient markets or irrational traders. Stock prices can fall quite rationally, even without an expectation of a dividend cut. (They can fall irrationally, too, of course.) Think of US treasury bonds. They pay a coupon analogous to stock dividends. The market places roughly a 0% chance of the US government defaulting on its interest payments. Yet the price of bonds (aka interest rates) swing up and down all the time.

In fact speaking of bonds, this is a perfect analogy. There are regular coupon-bearing bonds which pay some amount (say, 4% a year), and there are also zero coupon bonds that pay no coupon. They are sold at a discount to their face value, and then at maturity you redeem them for face value. So you get your return entirely through price appreciation. This is very analogous to dividend-paying and non-dividend-paying stocks. If your argument is correct, it seems to me you would argue that there’s no real reason for coupon bearing bonds to exist. Why don’t we all just buy zero coupon bonds? Isn’t it a lot easier to just have one big payment at the end? And then if you retire and need cash flows, why don’t you just sell off little chunks of your zero coupon holdings? The reason is that the prices of zero coupon bonds are very volatile. You don’t know what bond prices (aka interest rates) will be when you need to sell, so you could end up getting very bad results, depending on the path of bond prices over the next several years. This is a well-known phenomenon and is the flip side of reinvestment risk. (If you don't need the cash flows any time soon, it’s better to buy a zero coupon bond because otherwise you’re going to get coupons that you will have to reinvest at some unknown future interest rate). In the bond world, this is all very well-understood. If you need regular cash flows a regular coupon bond gives more predictable, less volatile results. The only difference is that I’m applying the same reasoning to stocks and dividends. The math is the same, though.

The argument that there’s no mathematical difference between a company paying dividends and a company buying back shares sounds very similar to the argument Buffett gave in his 2012 shareholders letter for why he doesn’t pay dividends (pg. 19). But his argument works. He gives a mathematical example showing that his shareholders will be better off if he doesn’t pay dividends, and they just sell off parts of their holdings if they need money. But he makes it clear that one of his assumptions is that the price of BRK never goes too low! In his case, it makes sense because he then goes on to say that he has a policy of always buying back shares if they ever drop below 1.2 times book value. He’s putting a floor on the stock price.

[Another Reader]

My experience in Phoenix during the crash was that rents stagnated, then went up.  Vacancy and collection loss went up as jobs were lost but went down later when people lost their houses in foreclosure or short sales and had to rent.  Values went down by as much as 2/3, but as long as you weren't overleveraged or had negative cash flow, you were just fine.  The same was true of my dividend paying stocks, except for the banks.  Overall, the loss of income was small, and the income was independent of how markets viewed the value of the underlying real estate or businesses on any given day.

If I was relying on selling 4 percent of my portfolio to pay my expenses during that time, I would have been hurt.  If I was getting an average of two percent in dividends and selling to make up for the deficit, the blow would have been softened.  Mostly I deposited the checks, worried unnecessarily about structural changes in the markets that largely did not occur, and went about my business.  I also added to my income producing real estate portfolio at very favorable prices at the market bottom.

And that's why it really is all about the income.

[warfreak2]

(a)   If the dividend of a stock is not cut, and you are living on less than the size of the dividend, then it is mathematically impossible for your portfolio to fail, no matter what price the stock trades at?

(b)   If you are living on the same amount as (a) but own non-dividend paying stocks it is possible for your portfolio to fail if the stock price falls low enough for long enough?

This is a bogus comparison, because you're assuming total stability for the first stock, but you have no similar assumption for the second stock. The analogous assumption for (b) would be: the company has totally stable profits, it regularly spends the same chunk of its profits buying back shares, and your withdrawal is always less than your proportion of those profits (an analogous assumption), then it would likewise be mathematically impossible for your portfolio to fail.

There just isn't any magic in dividends. None. If you reinvest all your dividends, you have stock in the same company, with the same profits; it's equivalent to a non-dividend-paying stock, yet it's just as stable and reliable as a dividend-paying stock. Weird! Dividends exist because it's simpler and easier for the company to pay its shareholders in cash rather than by increasing the share price; not because they magically allow the company to pay the investors more (or more reliably) using the same damn profits.

[beltim]

(a)   If the dividend of a stock is not cut, and you are living on less than the size of the dividend, then it is mathematically impossible for your portfolio to fail, no matter what price the stock trades at?

(b)   If you are living on the same amount as (a) but own non-dividend paying stocks it is possible for your portfolio to fail if the stock price falls low enough for long enough?

This is a bogus comparison, because you're assuming total stability for the first stock, but you have no similar assumption for the second stock. The analogous assumption for (b) would be: the company has totally stable profits, it regularly spends the same chunk of its profits buying back shares, and your withdrawal is always less than your proportion of those profits (an analogous assumption), then it would likewise be mathematically impossible for your portfolio to fail.

This is an interesting comparison, and I think  you've identified a legitimate difference, but I still think you're incorrect.  If you assume constant profits for both a and b, and live only on the dividends for a, you clearly can't have your portfolio fail.  Now, the question is whether company b, which bought back stock with an equivalent percentage of profits as company a, could ever fail.  And the answer to that theoretical question is obviously yes.  Take a company like Tata motors.  From 2008 to 2009, their sales doubled, their profits also went up, but the stock went down by about 2/3.  Why?  Because of fears about the global economy.  Because someone selling stock is exposed to the marketplace, they are dependent on the share price - specifically, what others are willing to pay for their shares.  That variability introduces the way for that portfolio to fail that is not present in the company a example.

[beltim]

Also, I'd encourage you to read, the link I posted above about how company management times stock buy back decisions very poorly.  This negatively affects stock value compared to dividends.  It wouldn't if companies had constant buy back policies like most dividend-paying companies do with dividends, but very few companies have such a steady buyback policy. And the ones that do tend to have the best results - that's also shown in that report.

[Eudo]

Also, I'd encourage you to read, the link I posted above about how company management times stock buy back decisions very poorly.  This negatively affects stock value compared to dividends.  It wouldn't if companies had constant buy back policies like most dividend-paying companies do with dividends, but very few companies have such a steady buyback policy. And the ones that do tend to have the best results - that's also shown in that report.

By the way, I don't know about anyone else, but the link was broken for me (required a Credit Suisse login)

[beltim]

Also, I'd encourage you to read, the link I posted above about how company management times stock buy back decisions very poorly.  This negatively affects stock value compared to dividends.  It wouldn't if companies had constant buy back policies like most dividend-paying companies do with dividends, but very few companies have such a steady buyback policy. And the ones that do tend to have the best results - that's also shown in that report.

By the way, I don't know about anyone else, but the link was broken for me (required a Credit Suisse login)

That would explain why no one addressed it.  Weird, because I have no connection or login for Credit Suisse.  Let's try: http://www.google.com/url?sa=t&rct=j&q=credit%20suisse%20adding%20value%20or%20destroying%20value%3F&source=web&cd=1&ved=0CCgQFjAA&url=https%3A%2F%2Fdoc.research-and-analytics.csfb.com%2FdocView%3Fsourceid%3Dem%26document_id%3Dx454556%26serialid%3De45HdNbTyfzxqJiuVkQbC5cPZZDE1SdAOSK1XWRCbxs%3D&ei=iuJFU4S3KISdyQHvvoHwAQ&usg=AFQjCNFCUHmca-Qnzwc9RH5K1jP42Cb_eQ&bvm=bv.64507335,d.aWc

[arebelspy]

That link works, beltim.