6 Mos market strategy

[dc_publius]

If you look at random numbers, you will eventually find patterns and correlations. Especially in such a short period of about 100 test intervals. While interesting, and possibly correlated, the deviation is so huge and correlation so small that it would be hard to profit from this.

Not surprisingly, there are also other seemingly correlated variables, like month, day of the week, weather, etc.

[Jeff W]

[ QUOTE ]

If you look at random numbers, you will eventually find patterns and correlations. Especially in such a short period of about 100 test intervals. While interesting, and possibly correlated, the deviation is so huge and correlation so small that it would be hard to profit from this.

[/ QUOTE ]

I'm a skeptic, but there seems to be a lot of supporting evidence for the strategy.

Can anyone access this paper:

Jacobsen, Ben and Bouman, Sven: "The Halloween Indicator, 'Sell in May and Go Away': Another Puzzle" (July 2001)

[ QUOTE ]
Abstract:
We document the existence of a strong seasonal effect in stock returns based on the popular market saying 'Sell in May and go away', also known as the 'Halloween indicator'. According to these words of market wisdom, stock market returns should be higher in the November-April period than those in the May-October period. Surprisingly, we find this inherited wisdom to be true in 36 of the 37 developed and emerging markets studied in our sample. The 'Sell in May' effect tends to be particularly strong in European countries and is robust over time. Sample evidence, for instance, shows that in the UK the effect has been noticeable since 1694. While we have examined a number of possible explanations, none of these appears to convincingly explain the puzzle.

[/ QUOTE ]

[technologic]

[ QUOTE ]
[ QUOTE ]

Instead of finishing with $24,608,711 as in the original post, the strategy now is at $88,037,236. This increases the annual return from 6.76% to a whopping 9.57%.

This only opened a can of worms for me, since I am leaving out stock dividends.

[/ QUOTE ]

Dividends & taxes make a huge difference in results. My guess is this approach gets creamed once you calculate those in.

Specifically, if your funds with this strategy actually compound at the after tax return rate. Todays short term cap gains tax rates are around 35-40% depending upon your tax bracket and state. So your compounding rate on your 9.57% drops to somewhere around 5.5-6%.

The ability to defer capital gains and pay only 15-25% (federal+state) at termination is a very compelling advantage for buy and hold over long periods.

[/ QUOTE ]

maybe develop a 1.5 year strategy? ie getting that half year in with a full year. it would be half as effective, but still i assume would yield better returns if this strategy is statistically significant.

[Jeff W]

[ QUOTE ]
maybe develop a 1.5 year strategy? ie getting that half year in with a full year. it would be half as effective, but still i assume would yield better returns if this strategy is statistically significant.

[/ QUOTE ]

Also, this is a moot point for a tax-deferred account. Consider someone with an allocation of 80% Equities/20% Bonds in a tax-deferred account. It seems like a no-brainer to put 40% bonds during Apr. 15-Oct. 15 and then switch to 0% bonds for Oct. 16-Apr. 14., even if market timing is a myth--risk is prob. somewhat higher as I don't think asset allocation and risk are linearly correlated.

I know most buy-and-hold investors see market timing as the dark side of the force, but I think it's worth evaluating the merits of the strategy. I can't contribute much yet, but I'm going to buy a book on timing and decide for myself.

[Jeff W]

I do think Timing is prob. not worthwhile for taxable accounts(where most of my money will be).

[DesertCat]

[ QUOTE ]


Also, this is a moot point for a tax-deferred account.

[/ QUOTE ]

A tax deferred account takes care of the tax issues, but no one has addressed the dividend issue yet. I believe the current S&P 500 dividend yield is about 1.8% per year. So by pursuing this strategy you lose .9% a year in dividends over buy and hold.

The average dividend yield from 1970 to now was probaby around over 3%. So it appears OP's analysis is under-estimating buy and hold returns by approximately 1.5% per year.

{ Edited historical data to match OP's time period }

[Jeff W]

I don't know why most analyses exclude dividends. The Wiki article on "Halloween Indicator" claims that even after dividends and trading costs, the 6 mo. strategy beat buy and hold(references the paper I mentioned earlier).

Edit: Got the paper. Reading it now.

[DesertCat]

[ QUOTE ]

If you bought $1M worth of the Dow in 1958, it would be worth $23,103,238 today, using 540.10 and 12481.49 quotes. No dividends or taxes.

[/ QUOTE ]

$1M to $23.1M in 49 years is only 6.6% per year, clearly that's the market return without dividends. Adding in a conservative guess at dividends gets you to 9.6% per year (I suspect the real return was a little higher). 49 years at 9.6% ends up at $89.3M.

[DesertCat]

[ QUOTE ]
I don't know why most analyses exclude dividends. The Wiki article on "Halloween Indicator" claims that even after dividends and trading costs, the 6 mo. strategy beat buy and hold(references the paper I mentioned earlier).

Edit: Got the paper. Reading it now.

[/ QUOTE ]

It's possible. I can't do the comparison with the OP's numbers because OP assumed zero percent returns for half the year. In theory, the strategy should earn enough interest during that 6 months to offset the dividend loss.

But I still think it's kind of a screwy strategy. Essentially you are counting on the market being close to break even that six months, since even a small level of capital appreciation + dividends should be larger than the interest you earn. Right now, you are assuming that the market's appreciation is less than 3.2% during those months, if your annual dividend rate is 1.8%, and the risk free interest rate for six months is 5%.

[Jeff W]

Link to the Paper

Tables and Figures at the bottom.