3/4/2022
People tend to be overly focused on avoiding losses. Much of this is due to a cognitive bias known as loss aversion. This is no surprise since being hyper focused on preventing/avoiding negative consequences was a very useful evolutionary trait during our hunter gatherer times. Those early humans that didn’t focus huge amounts of energy preparing to defend against an attack from a foreign tribe would have probably lost all their valuable resources very quickly.
With that being said, the modern world of investing is quite different to the world of hunter-gatherer tribes. So, if an investor is using dollar cost averaging (DCA), is it worth trying to avoid the paper losses that market corrections cause by trying to sell at the peak and buy back in at the trough?
To get some insight into this question let’s look at the price movements, excluding dividends, of VAS.AX as a proxy for Australian All Ordinaries from 1/1/2018 through to the 2/1/2020 (Yahoo finance 2022). The simulations will compare 4 scenarios where an investor receives $5,000 per month to invest over a 25-month period. Transaction fees will be fixed at $20 per transaction and, for taxation purposes, their marginal tax rate will be 32.5%. A correction is defined as a peak to trough fall in VAS.AX pricing that is greater than 10%. In this simulation a correction occurs once, between the 30/8/2018 ($81.30) to the 21/12/2018 ($69.63), representing a loss of 14.35%. Across all scenarios the total amount invested will be equal to:
Scenario 1: dollar cost averaging
The investor will purchase $5,000 worth of VAS.AX ETF on the first trading day of each month starting on the 1/1/2018 and making their last purchase on the 2/1/2020 (first trading day of 2020).
Scenario 2: optimal sell and buy
In this scenario, the investor has timing that only God can achieve. They will dollar cost average until they sell their entire holding at the peak of the market on the 31/8/2018 for $81.30 and buy in again at the lowest point on 21/12/2018 at $69.63. During the time they are out of the market (31/8/2018 – 21/12/2018) they will continue to accrue cash at a rate of $5,000 per month – this will be added to the cash balance for repurchasing. After this point they will continue to dollar cost average on the first trading day of each month.
Scenario 3: sell at the peak but buy back in late
This investor will get the timing perfect by selling at the peak of the market but buy back in 28 trading days late on the 5/2/2019 at $75.90. Dollar cost averaging occurs in the same manner as scenario 2.
Scenario 4: sell late but buy back in at the trough
The investor will sell their entire holding 28 trading days after the peak on the 9/10/2018. However, they manage to buy back in at the lowest point on the 21/12/2018. Dollar cost averaging occurs in the same manner as scenario 2.
Scenario 5: sell late and buy late
In this scenario the investor is 28 trading days late to sell and 28 trading days late to buy back in. Personally, I think this scenario may be closer to how most retail investors would behave, but more on this later. Dollar cost averaging occurs in the same manner as scenario 2.
Results
Buy/sell points
Figure 1 below shows the buy and sell points for all 5 scenarios
Summary data
| Scenario 1: DCA | Scenario 2: Optimal | Scenario 3: Buy 28 days late | Scenario 4: Sell 28 days late | Scenario 5: Buy and Sell 28 days late | |
| Total number of units owned at 2/1/2020 | 1579.991021 | 1693.044982 | 1612.456108 | 1653.603327 | 1576.272679 |
| Market value at 2/1/2020 | $133,309.44 | $142,923.98 | $136,138.91 | $139,543.67 | $133,034.54 |
| Total number of transactions | 25 | 23 | 21 | 25 | 23 |
| Total transaction cost | $500.00 | $460.00 | $420.00 | $500.00 | $460.00 |
| Tax | $0 | $706.89 | $706.89 | -$185.67 | -$185.67 |
| Total amount invested (Note 1) | $125,500.00 | $126,166.89 | $126,126.89 | $125,500.00 | $125,460.00 |
| Raw gain (%) | 6.22% | 13.28% | 7.94% | 6.22% | 6.04% |
What the results show
DCA produces an overall, non-annualized, return of 6.22% over the circa 2-year period that this simulation was run. For an investor with God power (scenario 2), their return was more than double that of doing DCA.
Scenarios 3 and 4 performed similarly to scenario 1 with 7.94% and 6.22% returns respectively; and scenario 5 slightly underperformed (6.04%) when compared to scenario 1. It should be noted that total return for scenarios 2 and 3 were reduced by the need to pay $706.89 in taxes given that the entire holding was sold for a $2,175.04 profit. If taxes were not included then the return would have risen to 13.92% and 8.55% respectively.
Concluding thoughts
As expected, if you managed to sell at the peak and buy back in at the absolute lowest point of the market you would completely crush the returns produced by any other portfolio. What is quite sobering though, is even with demi-god like powers i.e., selling the at the peak buy buying back in late; or selling late and buying back in at the bottom, the returns are comparable to DCA.
In my opinion, most retail investors are highly unlikely to conduct their sell transactions near the peak and instead will sell much too early or too late. I believe this to be also true on the buy side. This implies that scenario 5, would be the most representative of many retail investors. Having said that, after several office ‘watercooler conversations’ about completely avoiding investing through all of 2020 due to fear of the markets falling further, scenario 5 could be overestimating the willingness of active retail investors’ to return to the markets after corrections.
This simulation implies that there is little to be gained by trying to avoid corrections for broad based index funds – being late to the sell and re-buy actually produces worse results for greatly increased effort compared to DCA. However, since corrections play out in many different ways, I suspect that trying to avoid corrections by selling out and rebuying may produce better returns in some cases. As an example, if a recovery doesn’t occur during a set sample period, then the negative consequences of selling early and buying back in late may be reduced. Additional simulations would be required to demonstrate these results.
Overall, since short term price movements are almost impossible to predict, outperformance of a DCA strategy using market timing is likely the result of luck rather than skill. Further to this, timing the market requires an investor to be correct 2 twice in a row which is significantly more effort than periodically investing a set amount. To me, success at avoiding corrections is not repeatable and therefore should not be relied upon to produce positive investor outcomes. Hence at the FFE house, we’ll be sticking to DCA.
Engineer your freedom
Notes
- Total amount invested is calculated by adding the total funds available + transaction fee x number of transactions + tax payable if greater than $0. Tax losses can only be written off against future capital gains so are not subtracted from total amount invested
References
Yahoo finance, 2022, Vanguard Australian Shares Index ETF (VAS.AX) 31/12/2017 – 31/12/2020 (Daily), Yahoo finance, available from: <https://au.finance.yahoo.com/quote/VAS.AX/history?period1=1514678400&period2=1609372800&interval=1d&filter=history&frequency=1d&includeAdjustedClose=true>