4 ms·

Great call-out. Here's a non-paywall version to that article: https://www.berkshirehathaway.com/1999ar/FortuneMagazine.pdf https://www.berkshirehathaway.com/199

Great call-out. Here's a non-paywall version to that article: https://www.berkshirehathaway.com/1999ar/FortuneMagazine.pdf https://www.berkshirehathaway.com/1999ar/FortuneMagazine.pdf Interesting to note that if you'd bought and held the S&P500 index for the 20-year period of 1998 to 2018, you would have invested through the dotcom bubble and the great recession -- but still ended with an average real total return of 3.3% per year over that timeframe (adjusted for dividends and inflation)! Edit: typo

roflyear 4y ago
I never thought this was being too honest, as no one just buys the SPY once and forgets about it. Most people invest at some interval (through their 401k, IRA, or whatever) and it is much better to run some kind of test that mimics this to some degree. If you bought the SPY once a year, once a month, once a XXX from 1998 to 2018, what are your returns looking like?
- DMell 4y ago
  These backtests used to be fun when I had access to a Bloomberg terminal. But, yes.. this discussion shows the importance of dollar cost averaging.
  - roflyear 4y ago
    Yup! Also diversification. It is rare to have ONLY equities (in my experience, though I only have equity & cash positions!!) and having even 10-20% of bonds would make a MASSIVE difference, especially through the 70s, 80s, 90s!!! (when yields were an average, EYEBALLING, like 8-9% or something!) see: https://www.macrotrends.net/2016/10-year-treasury-bond-rate-yield-chart https://www.macrotrends.net/2016/10-year-treasury-bond-rate-...
- scottLobster 4y ago
  Paul Merriman has a fun "lifetime investment calculator" that can get you those numbers, it assumes investing Jan 1 of every year (It's tuned to their recommended allocations but includes the S&P 500 as a baseline). https://paulmerriman.com/lifetime-investment-calculator/ https://paulmerriman.com/lifetime-investment-calculator/ If you'd invested 100% in the S&P 500, say 10,000 a year on Jan 1 every year from 1998 to 2018, you'd have $522,135 from a $200,000 total principal, or roughly 13% annualized return (not including inflation)
- teraflop 4y ago
  That question is also relatively easy to answer: https://www.portfoliovisualizer.com/backtest-portfolio?s=y&timePeriod=4&startYear=1998&firstMonth=1&endYear=2018&lastMonth=12&calendarAligned=true&includeYTD=false&initialAmount=10000&annualOperation=1&annualAdjustment=10000&inflationAdjusted=true&annualPercentage=0.0&frequency=4&rebalanceType=1&absoluteDeviation=5.0&relativeDeviation=25.0&leverageType=0&leverageRatio=0.0&debtAmount=0&debtInterest=0.0&maintenanceMargin=25.0&leveragedBenchmark=false&reinvestDividends=true&showYield=false&showFactors=false&factorModel=3&portfolioNames=false&portfolioName1=Portfolio+1&portfolioName2=Portfolio+2&portfolioName3=Portfolio+3&symbol1=SPY&allocation1_1=100 https://www.portfoliovisualizer.com/backtest-portfolio?s=y&t... The "rate of return" isn't quite as simple to define when you're talking about multiple contributions over time. The most straightforward approach is to look at it as a weighted average, and the exact value depends on how you do the weighting. But in this case, it's in the ballpark of 7%/year nominal, which is probably 4-5%/year after inflation. As you can see, you get a big boost (at least in this cherry-picked example) by continuing to invest through the downturn.
- getToTheChopin 4y ago
  Agreed that an example of buying once per year (or quarter / month) over time and calculating the return of that scenario is more applicable to the experience of a regular investor. This can be eye-balled on this chart: https://themeasureofaplan.com/wp-content/uploads/2023/01/Rolling-Returns-1074.png https://themeasureofaplan.com/wp-content/uploads/2023/01/Rol... The bottom graph shows the annualized return for each 20-year period in market history. 20-year holding periods have this behaviour: - Max: 13.2% - Min: 0.5% - Average: 6.6% - Std. Dev.: 3% So, if you dollar cost average over time and hold each of those investment cohorts for ~20 years, you should expect an average annual return of roughly 6% (+/- 3%).

roflyear 4y ago

I never thought this was being too honest, as no one just buys the SPY once and forgets about it. Most people invest at some interval (through their 401k, IRA, or whatever) and it is much better to run some kind of test that mimics this to some degree. If you bought the SPY once a year, once a month, once a XXX from 1998 to 2018, what are your returns looking like?

DMell 4y ago
These backtests used to be fun when I had access to a Bloomberg terminal. But, yes.. this discussion shows the importance of dollar cost averaging.
- roflyear 4y ago
  Yup! Also diversification. It is rare to have ONLY equities (in my experience, though I only have equity & cash positions!!) and having even 10-20% of bonds would make a MASSIVE difference, especially through the 70s, 80s, 90s!!! (when yields were an average, EYEBALLING, like 8-9% or something!) see: https://www.macrotrends.net/2016/10-year-treasury-bond-rate-yield-chart https://www.macrotrends.net/2016/10-year-treasury-bond-rate-...
scottLobster 4y ago
Paul Merriman has a fun "lifetime investment calculator" that can get you those numbers, it assumes investing Jan 1 of every year (It's tuned to their recommended allocations but includes the S&P 500 as a baseline). https://paulmerriman.com/lifetime-investment-calculator/ https://paulmerriman.com/lifetime-investment-calculator/ If you'd invested 100% in the S&P 500, say 10,000 a year on Jan 1 every year from 1998 to 2018, you'd have $522,135 from a $200,000 total principal, or roughly 13% annualized return (not including inflation)
teraflop 4y ago
That question is also relatively easy to answer: https://www.portfoliovisualizer.com/backtest-portfolio?s=y&timePeriod=4&startYear=1998&firstMonth=1&endYear=2018&lastMonth=12&calendarAligned=true&includeYTD=false&initialAmount=10000&annualOperation=1&annualAdjustment=10000&inflationAdjusted=true&annualPercentage=0.0&frequency=4&rebalanceType=1&absoluteDeviation=5.0&relativeDeviation=25.0&leverageType=0&leverageRatio=0.0&debtAmount=0&debtInterest=0.0&maintenanceMargin=25.0&leveragedBenchmark=false&reinvestDividends=true&showYield=false&showFactors=false&factorModel=3&portfolioNames=false&portfolioName1=Portfolio+1&portfolioName2=Portfolio+2&portfolioName3=Portfolio+3&symbol1=SPY&allocation1_1=100 https://www.portfoliovisualizer.com/backtest-portfolio?s=y&t... The "rate of return" isn't quite as simple to define when you're talking about multiple contributions over time. The most straightforward approach is to look at it as a weighted average, and the exact value depends on how you do the weighting. But in this case, it's in the ballpark of 7%/year nominal, which is probably 4-5%/year after inflation. As you can see, you get a big boost (at least in this cherry-picked example) by continuing to invest through the downturn.
getToTheChopin 4y ago
Agreed that an example of buying once per year (or quarter / month) over time and calculating the return of that scenario is more applicable to the experience of a regular investor. This can be eye-balled on this chart: https://themeasureofaplan.com/wp-content/uploads/2023/01/Rolling-Returns-1074.png https://themeasureofaplan.com/wp-content/uploads/2023/01/Rol... The bottom graph shows the annualized return for each 20-year period in market history. 20-year holding periods have this behaviour: - Max: 13.2% - Min: 0.5% - Average: 6.6% - Std. Dev.: 3% So, if you dollar cost average over time and hold each of those investment cohorts for ~20 years, you should expect an average annual return of roughly 6% (+/- 3%).