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