Ever since reading “The 3 Percent Signal” by Jason Kelley and debunking its claims in my post, I’ve been curious about whether buy and sell signals for stocks truly exist. For those who don’t know, the 3 Percent Signal suggests buying or selling shares of an ETF to secure a 3% gain each quarter. When I tried to validate this, I found it didn’t work as promised.
I believe there’s some logic to buying or selling stocks when their prices are particularly low or high. After all, it’s just math, with small gains adding up over time. However, knowing when a price is truly low is challenging because “low” is relative.
It’s easy to look at the price history of an investment and identify a low price that seems like a good buying opportunity. But is that enough to consistently beat the market? I decided to test this theory using a simple scenario: buying when the investment is 10% or more lower than it was 12 months ago. Let’s call this the “10% 12-month” signal.
I tested this with the Vanguard small-cap NAESX mutual fund. The process was straightforward: every month the fund dropped 10% or more below its price from 12 months ago, I’d buy $1,000 worth. If it dropped less than 10% or gained value, I’d do nothing. To compare, I created a benchmark approach where an investor would buy shares worth the same total investment on January 1st each year – the “No Signal” approach.
I analyzed data from 2000 to 2015. At first glance, there’s no profound difference between the two strategies. The Signal strategy does slightly better, ending with $80,492 compared to $73,078 from the No Signal strategy—a difference of $7,414. While not massive, it’s notable, leading to an annualized return of 11.3% for the Signal strategy versus 10.5% for No Signal.
Could we be on to something? To see, I tested the same process on the S&P 500 Index. Unfortunately, the results showed a decrease of $1,674 with the Signal strategy. Thus, there’s no clear benefit and sometimes a disadvantage.
Returning to the small-cap index, I wondered if the spread of nearly 1% between the two strategies would hold over 30 years. The results showed the Signal strategy ending with $13,765 more than No Signal, but it’s not significant over 30 years ($436,134 vs. $422,369), with nearly identical annualized returns of 9.7% and 9.6%.
I applied the same 30-year analysis to the S&P 500. Both strategies again produced nearly identical results: $128,614 vs. $125,185. Hence, there’s no meaningful difference.
From these four simulations, I conclude that trying to time investments based on buy and sell signals offers almost no long-term benefit. You’d do just as well by automating investments on January 1st rather than monitoring the markets monthly to outsmart the system.
To develop this further, I could run multiple sets of 15 and 30-year data to see if the strategy works in different periods. I could also investigate other investment types, as small-cap and S&P 500 indices are popular but not the only options available.
