Timing Market Dips

· 2 min read

With all the Coronavirus panic the DOW has been falling off a cliff the past few days:

DOW plummet
DOW plummetting to $21k based on Yahoo finance data. Click the image for an interactive full-size plot.

Which raises a valid question: if you're sitting on some cash should you buy a index fund like the DOW?

Let's backtest

This isn't the first time the DOW has dipped more than 5%. It's happened 18 times before 2020:

OccurrenceDateDip %
11987-10-1922.61
21987-10-268.03
31988-01-086.85
41989-10-136.90
51997-10-277.18
61998-08-316.36
72000-04-145.65
82001-09-177.12
92008-09-296.97
102008-10-075.10
112008-10-097.33
122008-10-157.87
132008-10-225.69
142008-11-055.04
152008-11-195.07
162008-11-205.56
172008-12-017.70
182011-08-085.54
192020-03-097.78
202020-03-115.85
212020-03-129.98
DOW dips
DOW dips >5%.

Buying the dip immediately

Buying the dip on the close yields the following percent gains:

Dip Date1 year5 year10 year
1987-10-1924.21%82.56%356.5%
1987-10-2621.15%78.80%337.4%
1988-01-0814.80%72.92%317.4%
1989-10-13-6.66%50.82%314.4%
1997-10-2716.82%17.91%90.9%
1998-08-3143.64%24.89%55.3%
2000-04-14-1.42%.95%6.7%
2001-09-17-7.99%29.59%28.1%
2008-09-29-6.01%47.20%155.0%
2008-10-072.94%58.10%179.9%
2008-10-0914.98%72.23%208.7%
2008-10-1517.31%78.37%195.4%
2008-10-2218.33%80.67%198.6%
2008-11-059.48%71.12%176.5%
2008-11-1929.19%99.76%217.7%
2008-11-2036.62%11.41%231.2%
2008-12-0128.50%97.40%210.9%

Waiting for less volatility

But wait, it might be safer to sit on cash or invest in something else during periods of volitility! A decent measure of volatility is the standard deviation over a 20 day period (1 month) divided by the close price. This volatility can be as high as 10% (like during 5% dips) and as low as .5% when not much is happening. If we wait for volatility to die down <3%, this yields the following percent gains:

Dip DateBuy date1 year5 year10 year
1987-10-191987-11-1811.3863.69309.39
1987-10-261987-11-1812.0765.41304.70
1988-01-081988-01-1112.8069.92310.20
1989-10-131989-10-16-9.7645.82300.70
1997-10-271997-11-1410.4711.5080.54
1998-08-311998-09-2332.8015.4643.66
2000-04-142000-04-17-4.00-1.684.00
2001-09-172001-10-19-10.8225.6024.21
2008-09-292008-09-30-10.2140.62143.67
2008-10-072008-12-1913.3674.10208.27
2008-10-092008-12-1914.9872.23208.73
2008-10-152008-12-1917.2978.35195.36
2008-10-222008-12-1917.5179.41196.58
2008-11-052008-12-1916.6382.29194.56
2008-11-192008-12-1920.4386.22196.22
2008-11-202008-12-1920.2786.11191.60
2008-12-012008-12-1922.0587.50195.35

When we use this measure we wait between a few days and a few weeks to buy. Interestingly all the dates after the housing crisis in October 2008 wait to buy until 2008-12-19 when volatility dies down.

Comparison

There are 3 cases when you could have gotten a better "sale" by waiting for even lower prices than a 5% dip:

DateNaive priceDelayed priceWorth waiting
1987-10-1917381939No
1987-10-2617931939No
1988-01-0819111945No
1989-10-1325692657No
1997-10-2771617572No
1998-08-3175398154No
2000-04-141030510582No
2001-09-1789209204No
2008-09-291036510850No
2008-10-0794478579Yes
2008-10-0985798579Yes (by 8 pennies)
2008-10-1585778579No
2008-10-2285198579No
2008-11-0591398579Yes
2008-11-1979978579No
2008-11-2075528579No
2008-12-0181498579No

But how much better off would you have been when using the Naive strategy? This is the Naive strategy's advantage over the delayed strategy:

Date1 year diff5 year diff10 year diff
1987-10-1912.8318.8647.19
1987-10-269.0713.3932.76
1988-01-081.993.007.25
1989-10-133.095.0013.74
1997-10-276.346.4010.36
1998-08-3110.839.4211.72
2000-04-142.572.642.79
2001-09-172.833.993.94
2008-09-294.206.5811.40
2008-10-07-10.41-15.99-28.32
2008-10-09-1.1e-5-1.6e-5-2.9e-5
2008-10-15.01.02.04
2008-10-22.821.262.08
2008-11-05-7.14-11.17-18.05
2008-11-198.7613.5421.55
2008-11-2016.3525.3039.64
2008-12-016.449.8915.58

As you can tell the Naive strategy wins a decent amount more than the delayed strategy. But when the delayed strategy works, it's really the better strategy to use by a 10% margin.

HorizonMean naiveMean delayed
1 year15.0511.01
5 year63.2257.80
10 year193.0182.8

As you can see, these differences are so miniscule that timing the market doesn't really matter for people looking to invest over long time horizons.

Conclusion

If you're focused on a 10+ year time horizon go ahead and take the Naive strategy of buying an index fund on the Coronavirus dip! Any reason is a good reason to start investing, and the statistics favor it in the majority of historical dips.

Python code available on Github.