It's better to invest when the market is at a record high.

[Cecil]

I know, provocative title. But I ran some numbers.

I analyzed the nominal closing values of the S&P 500 index back to Jan 2, 1957. Of the 14190 daily closes,

853 (6.0%) were a record high.
2289 (16.1%) were within 1% of the current record
5774 (40.7%) were within 5% of the current record.

$1000 invested on an average day turned out to be worth $1583 after 5 years (annualized return of 9.6%)
$1000 invested on an average day in which the index was at a record high turned out to be $1601 5 years later (9.9%)
$1000 invested on an average day in which the index was at least 20% below the current record, turned out to be worth $1480 after 5 years (8.2%).

I guess what I'm saying is market timing doesn't work.

[gdborton]

It's better to invest when the market is at a record high.

You can't say that it's "better" only that it is still a good idea.

[Joet]

I think we can only say that backtested... as far as backtesting means anything, suggests it doesnt make much difference.

The results do surprise me, to be honest. I would have thought that investing on 'record high' days would have underperformed. See? what do I know. Not much

[NumberCruncher]

It's better to invest when the market is at a record high.

You can't say that it's "better" only that it is still a good idea.

It does appear to be "better" according to the numbers - $1601 five years later on record highs vs $1480

Interesting to think about!

[Johnny Aloha]

How did you do this?  Is there a website for it?

Makes me happy - I just rebalanced today!

[Cecil]

It's better to invest when the market is at a record high.

You can't say that it's "better" only that it is still a good idea.

True. I've just heard a lot of people lately complain about how they're staying out of equities because the market is "overvalued" (aka it's setting a record high")

How did you do this?  Is there a website for it?

Excel. :)

[Bank]

Interesting.  Are you using the S&P Total Return or just the straight index?  I wonder if it would make a difference.

Thanks for sharing.

[Roland of Gilead]

Funny, I always heard it was better to invest within two weeks of a full moon...

[The_Dude]

I'm very fascinated by your analysis.

Care to expand it at all?  I wonder what the results would be compiled against the typical investment result categories of 1, 3, 5, and 10 years. 

[tomboalogo]

I'm fascinated as well.

Does the analysis just take into account the stock price appreciation or do you correct for the better yield of dividends? I've tried to buy at a lower price, so my locked in yield is better.

[arebelspy]

I love these counter intuitive maths.

[rjack]

Interesting.  Are you using the S&P Total Return or just the straight index?  I wonder if it would make a difference.

Thanks for sharing.

If he didn't include dividend reinvestment then the results don't mean much.

[gdborton]

I don't want to bash the idea, as I think it is always a good idea to invest (timing and all that).

The math here only shows three examples over a relatively short time frame, so it doesn't prove that the optimal time to invest is when the market is at it's highest point.  It could be that the "best time" to invest is when the market is at 37.5% of it's most recent high.  Which isn't shown.

[arebelspy]

I don't want to bash the idea, as I think it is always a good idea to invest (timing and all that).

The math here only shows three examples over a relatively short time frame, so it doesn't prove that the optimal time to invest is when the market is at it's highest point.  It could be that the "best time" to invest is when the market is at 37.5% of it's most recent high.  Which isn't shown.

The takeaway is to not that "investing at a market peak is the best time to invest."

It is: don't be one of the people sitting largely in cash waiting for a pullback because you feel the market is too high.  Execute your AA and don't worry about it being at a market high.

[gdborton]

Yea I got it, just want to be sure other people aren't reading into it that they should wait for record highs.

[mpbaker22]

I presume this is because a record high is often followed by another record high and another record high.  No one knows whether it will be followed by another record high or a massive pullback.

An example of a data point - Someone who invested in the S&P 500 on March 7, 1997 would have lost nearly 8% within a month.  But if they held on for a full year, they would have been up over 30%.

[GreenGuava]

First, very cool.  Just curious:  how did you define 'average day'?

$1000 invested on an average day turned out to be worth $1583 after 5 years (annualized return of 9.6%)
$1000 invested on an average day in which the index was at a record high turned out to be $1601 5 years later (9.9%)
$1000 invested on an average day in which the index was at least 20% below the current record, turned out to be worth $1480 after 5 years (8.2%).

I guess what I'm saying is market timing doesn't work.

[Bank]

If he didn't include dividend reinvestment then the results don't mean much.

I disagree.  It shows something different (total return vs. price appreciation), not something non-meaningful.  If one were living on dividends rather than re-investing them, you might be more interested in the latter.  I admit that I am much more interested in the former.