Long term investment analysis

Disclaimer: Not financial/legal/tax advice.

There are many debates and analyses regarding staying invested in the stock market. In this article, I take a deeper dive into some of the data regarding two sides of the argument, to stay invested during volatile times or not. A series of articles debunking common myths about investment for the long-run, is a series called The Myths of Stocks for The Long Run, written by Lance Roberts, and Michael Lebowitz from Real Investment Advice (RIA). The other side is represented by the article The case to always stay invested posted by JPMorgan.

In The Myths of Stocks for The Long Run, the authors were analysing some of the common myths about investing, namely "time in the market is better than timing the market". This statement is usually true in bull markets, however, at other times, the effects of being in the market can be devastating. Starting an investment in the NASDAQ100 before the dot-com crash would have resulted in a loss that would take around 12 years to breakeven. Similar effects were shown for the S&P500 index in the series of articles. The reason simply is that, major crashes destroy the compounding effect of wealth growth, for example, a 40% crash needs 5 years of 10.8% growth, 3 years of 18.6% growth, or 2 years of 29% growth, just to recover the loss. They managed to show a simple technical strategy that can reasonably avoid the most devastating moves in a bear markets (hence timing the market somewhat), by simply staying in the market as long as the price is above the 12 months moving average.

In The case to always stay invested, the main premise behind the article is that, trying to "time the market" by getting out of the market after volatile events has high chance (of 70%) of timing errors and missing the best performing days, and missing the 10 or 20 best performing days significantly decreases the growth of the portfolio. Missing the best 10 days will cut the annual growth rate from 9.4% to 5.2%, missing the best 20 days will cut it to 2.5%, and missing the best 30 days will cut it to 0.32% (nearly no growth at all).

The main thing I agree with JPMorgan's article about is one should not make emotional decisions in investing. There are many corrections in the market that recover fast enough, taking an emotional decision to sell after a volatile event can lead to near worst-case scenario (selling near the bottom, then the market recovers shortly after). However, the argument laid by the article of JPMorgan is mostly misleading in my opinion, even when the data shown is correct. The reason I find it misleading is that, throughout the history of the S&P500, the best days mostly come in extremely volatile bear markets, almost always after one of the top worst days, or series of days (check the table below, conducted on SPY, which is an ETF tracking the S&P500 since 1993). Hence, one does not magically wrongly time missing the good day without timing missing some of the bad days preceding it. For instance, if we look at the list of top 20 days, according to the argument of JPMorgan, October 2008 and March 2020 look like two of the greatest months to invest, however, they have -16.52% and -13.00% returns, respectively. They are actually the worst and third-worst months of the entire history of SPY. On the other hand, part IV of the first article discussed missing the worst 10 or 20 days, which is also misleading for similar reasons. Perhaps, a better analysis would include missing both best and worst days, or in other words, the most volatile days.

Presented Analysis

Missing the best 10 days or worst 10 days are theoretical setups that will rarely occur; probably a good practical alternative is missing the best and worst days altogether. As a result, one should analyse practical strategies that can be applied, and not theoretical scenarios that will rarely manifest, hence I present a modified version of the analysis made by JPMorgan, based on variants of the technical strategy presented by RIA.

The strategies are staying the market when the price is above the 50-days moving average (50 DMA), 100-days, 150-days, 200-days (9 months period), and 250-days (12 months period). The 250-days moving average is a smoothened variant of the 12 months moving average, since it also contains a lot of samples from within the months, not only the closing prices. The benchmark strategy is the classical buy and hold. Furthermore, I extend the theoretical strategies to missing the best, worst, and both 10, 20, or 50 days. I also investigate staying only intraday (only during American exchange market hours, buying at open, selling at close), or the opposite of staying outside the open hours (selling at open, buying at close).

The capital is invested in two fashions, starting from the beginning of 2002 until the end 2021 (20 years in total), either lump sum or Dollar Cost Averaging (DCA). In both cases, dividends are reinvested. The lump sum invests a capital of 1 million dollars at the beginning of the investing period. DCA assumes a monthly gain of 1,000$, that are invested at the beginning of the month.

In this table, we can see the annualised return for the lump sum investment and the dollar cost averaging. The numbers seem to be slightly different than the study by JPMorgan, since they conducted the study on S&P500 index and not on an ETF tracking it.

As hypothesised, avoiding both the best and worst days is generally better than just the buy and hold, as seen by missing the best 50 and worst 50 days, and similarly for 20 or 10. This is probably due to the distribution of negative days in volatile times of the markets; additionally, there is a natural skew in the math of compounding loss, favouring the downside, for instance, a 10% loss followed by a 10% gain is in 1% net loss, a 25% loss followed by a 25% gain is in 6.25% net loss, and a 50% loss followed by 50% gain is in 25% net loss.

Interestingly, the moving averages strategies perform better than missing the worst 10 or 20 days (but not the worst 50). This shows that the moving averages are good indicators of volatile markets, hence they are good filters for the good times to invest. The shorter the period of the moving average, the less it is affected by the lagging effects of the moving average, hence the earlier it will find the beginning and end of bear markets, but also the more likely it will give false-positive signals; it is very frequent that the market falls below a 10 DMA or 20 DMA. Another advantage of the higher period DMAs, due to the lagging effect, if they give false-positive signals, the false-positive signals tend not to have big effects. For example, following the strategy in 2015, the strategy would exit SPY on 20 Aug at 204 and reenter on 28 Oct at 208, and again would exit on 31 Dec at 204 then reenter on 16 Mar at 203, this is a missing 1.4% gains for skipping a 50-months volatile period. As a result, even when the 50 DMA strategy is more profitable, it can also be more volatile, compared to the 200 DMA or 250 DMA.

Portfolios growth

Here I show the portfolios' growth for some of the strategies.

Important remarks

1. The strategies above are not considering the tax situation, where exits are subject to capital gains tax, which will eat a lot of the compounding gains, especially when the strategy gives false positive signals.
2. The analysis assuming that not being in the market simply means staying in cash, which might be an extreme measure for some investors. RIA suggested that it doesn’t need to be purely in cash, but it can be possibly a period of de-risking, by diversifying in bonds for example, or hedging with options.
3. The given strategy is a technical strategy, that does not consider any fundamental aspects of the economy and so on, so it is more beneficial to be used as diversifying or augmenting other strategies, rather than using it alone.

The data is acquired from Yahoo Finance.

Created on 2022-09-15 at 10:39