Wednesday, January 8, 2014

Greatest Hits


This short post provides an updated and revised version of a March 2012 post I wrote about how low the stock market can go. 

We must remember that the stock market is risky and can experience extended downturns for long periods of time. To get a sense of this, I’ve tallied up all the cases of stock market drops greater than 50% in inflation-adjusted terms for the 20 countries included in the Dimson, Marsh, Staunton Global Returns Dataset. The final entry in the table is for a GDP-weighted “world” portfolio diversified across these 20 countries. These calculations are based on annual data, and the drawdowns from peak to trough may be even bigger with monthly data, had that data been available. The data provides total market returns, which includes reinvested dividends. Though World War I and II account for some of these significant market drops, there are still plenty of other examples from more peaceful times. 

The table shows the country name, years (beginning of the first listed year to end of the second listed year), and the percentage drop in real terms for the stock market over that period. I also provide the year that the real stock market value would again exceed the level prior to the market drop, as well as the number of years it took for this to happen. 


12 comments:

  1. Wade -

    Scary Chart.

    Would be even more alarming if the recent (Oct 2007 - March 2009) 53% crash of the DOW/S&P was included.

    ReplyDelete
  2. Yes, I can only see Jan. 2007-Dec. 2008 with this data, as so that period missed a 50% or greater drop.

    ReplyDelete
  3. Every time I have the urge to go 100% stocks, I'm going to take a look at this!

    ReplyDelete
  4. Great post Wade. Thanks for crunching the numbers. Remind me not to move to Finland.

    ReplyDelete
  5. I believe the S&P 500 with dividends reinvested is still below where it was in March of 2000 (using Vanguard's VFINX)- if adjusted for inflation. "Stocks For The Long Run" indeed professor Siegel, who has probably mislead more people than Dave Ramsey I suspect.

    ReplyDelete
  6. This is very sobering. But while it grabs my attention, it doesn't provide a prospective that aids in selecting an appropriate degree of risk. As a next step, I'd love to see how various portfolio mixes would have reacted over the same periods, perhaps color coded to distinguish between inflationary and deflationary periods. Adding in the effect of rebalancing would be a very helpful. I recall seeing scatter plots of various alocations about 5 years ago that showed rolling average returns over a range of periods (1-5 years?), as well as average returns and 1-3 std deviations for each period. It looked like a funnel with reduced volatility over time. I've tried to find something like it for some time.

    ReplyDelete
    Replies
    1. Yes, if you make a graph of the volatility of average annual returns it appears as though "risk" declines with time. This is known as The Fallacy Of Time Diversification because the volatility of the final outcome INCREASES over time. Retirement could be viewed as as succession of intermediate "final" outcomes as one makes withdrawals successively. If this seems counter-intuitive there are some nice graphs at www.norstad.org/finance/risk-and-time.html You can be sure you won't see this sort of presentation from stock floggers or mutual fund promoters either. Fred

      Delete
    2. Thanks for setting me straight. Having been in the company of Boggle et al is of little solace. Considering part 2 (thank you again) of your post, I'm at risk of going from being a conservative planner to a paralyzed one :)

      Delete
  7. I never realized how long the recovery periods have been for some of the countries listed. What is there (other than track record) to suggest that the U.S. could not also sustain a lengthy recovery period similar to the other countries listed?

    ReplyDelete
  8. Even this chart perpetuates the overoptimistic idea that the 1929-1931 crash in the United States only lasted seven years. This IS technically accurate, but it ignores the fact that the recovery at the end of 1936 presented no more than a brief three-month window of opportunity, during which a successful market timer could have recouped. It was followed immediately by a 50% crash at the start of 1937 that took until 1945 to recover.

    Benjamin Roth's “The Great Depression: A Diary" makes it clear how disheartening this was.

    So, there were back-to-back bear markets of 7 years each with only a 3 month rest between them. In most respects and for most people this would have been experienced as a 14-year bear market.

    My source for the determination of start and end points is Table 12-4 of the 2010 Ibbotson SBBI Handbook, which calls out the 1929 crash as recovered by Nov 1936, and the 1937 crash as beginning in February, 1937. The 1937 crash is shown as the fourth largest crash in U.S. history, and comparable in severity to 2008-2009.

    ReplyDelete
    Replies
    1. Thank you. I did notice cases like this going on as I worked through the data, but didn't want to make judgment calls about what to do. Things really were worse for those individuals.

      On the other hands, there are also cases where these big market drops were preceded by a substantial and quick run-up in stock prices, such that a buy-and-hold investor who had bought in a couple of years before the drop would have weathered it in a much better fashion than implied above.

      Thank you, as your point is very valid.

      Delete
  9. I was extremely pleased to find this website. I wanted to thank you for ones time just for this wonderful read!! I definitely loved every bit of it and I have you saved as a favorite to look at new things on your blog.
    ACA impact on payers

    ReplyDelete