Great article as always. The statement that any investor would have underperformed selling 1996 compared to the investor who held is not really correct (maybe for the S&P, but not for the Nasdaq). Also does not account for opportunity cost (an investor who would have bought treasury bonds would have outperformed the buy and hold investor).
Professor, doesn't the choice of endpoint influence the results? If you choose an endpoint where stocks are near all-time highs, as stipulated earlier in your essay, then buy-and-hold is likely to be superior to market timing. If you choose an endpoint closer to the average, then market timing using a filter rule based on the entire history is likely to be superior to buy-and-hold.
It also matters very much whether you are a retiree or a working investor still adding to one’s assets.
I have no doubt whatsoever that if many years from retirement, buy and hold - and in fact keeping 100% of one’s portfolio in equities - is the superior strategy for several reasons. Two of the biggest: the difficulty of market timing, and the fact that the future additions you will make to your portfolio are logically more like cash than equities.
In fact for someone not close to or in retirement, keeping any appreciable amount of assets in bonds or cash is itself a form of market timing!
OTOH, for a retiree consuming 3% or more annually of one’s portfolio, the answer is far less clear, because of sequence of return risk. Market timing is an alternative to a 60-40 portfolio (and at any rate in the modern world I’d argue bonds are a not very good hedge for equities and in fact gold is a much better hedge) or other similar, and can in fact have positive results over a fixed portfolio percentage, especially when stock prices are elevated.
The Safe Withdrawal Rate (SWR) series at earlyretirement.com covers these issues extremely well and is highly recommended.
Very helpful. I am in a situation where I will need to allocate a lot of money to new positions in the markets starting Q1. Do you have any suggestions on a method to build an allocation without jumping in all at once?
Decades of rolling 30 year ( 30 or more years being relevant to an "accumulation stage" investing experience ) periods of total returns produced by US small cap "value" stocks shows them having outperformed US large cap stocks in every period up until the most recent period ( 1995 - 2024 ).
An argument "for" the inclusion of small cap into a portfolio is one of "honoring" more of a "Sharpe / Markowitz" portfolio diversification ( using an ETF such as Vanguard Small cap Value adds 800+ holdings to the mix ), while still seeking portfolio "alpha".
A single 30 year period by which the small cap value premium underperforms large cap hardly seems enough to reject inclusion of it into a portfolio. And modern ETF products make it easy to add another source of alpha premium through the impact of the "capitalization weighting" structure - such as the addition of the QQQ ( which shares many of its top portfolio weightings with the S&P 500 index ). From there, a large cap value ETF can be added in order to add many more of the same companies that the S&P 500 holds - yet in a value weighted structure.
The inclusion of small & large value stocks has also added a downside volatility buffer to that produced by the QQQ.
Professor your article is very good , complete and transparent in the data that you use. I just want to ask how models to time market or conclude if its overvalued or under can be used when we don’t know how optimistic or pessimistic assumption are on the estimated future earnings growth. When optimism is high analysts are very confident of future earnings growth and when low its the opposite so wouldn’t this skew your models which rely on future cash flows. How do you manage this risk of over estimation or under estimation of future earnings.
What stands out here is how valuation models bend under institutional uncertainty. Once the policy map is unclear, markets stop pricing fundamentals and start pricing confidence in the model itself. That’s the part that makes ‘fairly highly valued’ less about numbers and more about narrative stability.
Excellent as usual from Prof. Damodaran.
Great article as always. The statement that any investor would have underperformed selling 1996 compared to the investor who held is not really correct (maybe for the S&P, but not for the Nasdaq). Also does not account for opportunity cost (an investor who would have bought treasury bonds would have outperformed the buy and hold investor).
Professor, doesn't the choice of endpoint influence the results? If you choose an endpoint where stocks are near all-time highs, as stipulated earlier in your essay, then buy-and-hold is likely to be superior to market timing. If you choose an endpoint closer to the average, then market timing using a filter rule based on the entire history is likely to be superior to buy-and-hold.
I would have to agree with you, but i think the periods used in the article are somewhat standard periods used by market and finance professionals.
It also matters very much whether you are a retiree or a working investor still adding to one’s assets.
I have no doubt whatsoever that if many years from retirement, buy and hold - and in fact keeping 100% of one’s portfolio in equities - is the superior strategy for several reasons. Two of the biggest: the difficulty of market timing, and the fact that the future additions you will make to your portfolio are logically more like cash than equities.
In fact for someone not close to or in retirement, keeping any appreciable amount of assets in bonds or cash is itself a form of market timing!
OTOH, for a retiree consuming 3% or more annually of one’s portfolio, the answer is far less clear, because of sequence of return risk. Market timing is an alternative to a 60-40 portfolio (and at any rate in the modern world I’d argue bonds are a not very good hedge for equities and in fact gold is a much better hedge) or other similar, and can in fact have positive results over a fixed portfolio percentage, especially when stock prices are elevated.
The Safe Withdrawal Rate (SWR) series at earlyretirement.com covers these issues extremely well and is highly recommended.
https://earlyretirementnow.com/safe-withdrawal-rate-series/
https://earlyretirementnow.com/2021/03/02/pre-retirement-glidepaths-swr-series-part-43/
Very insightful. Thank you.
Very helpful. I am in a situation where I will need to allocate a lot of money to new positions in the markets starting Q1. Do you have any suggestions on a method to build an allocation without jumping in all at once?
Decades of rolling 30 year ( 30 or more years being relevant to an "accumulation stage" investing experience ) periods of total returns produced by US small cap "value" stocks shows them having outperformed US large cap stocks in every period up until the most recent period ( 1995 - 2024 ).
An argument "for" the inclusion of small cap into a portfolio is one of "honoring" more of a "Sharpe / Markowitz" portfolio diversification ( using an ETF such as Vanguard Small cap Value adds 800+ holdings to the mix ), while still seeking portfolio "alpha".
A single 30 year period by which the small cap value premium underperforms large cap hardly seems enough to reject inclusion of it into a portfolio. And modern ETF products make it easy to add another source of alpha premium through the impact of the "capitalization weighting" structure - such as the addition of the QQQ ( which shares many of its top portfolio weightings with the S&P 500 index ). From there, a large cap value ETF can be added in order to add many more of the same companies that the S&P 500 holds - yet in a value weighted structure.
The inclusion of small & large value stocks has also added a downside volatility buffer to that produced by the QQQ.
Just superb work , thanks so much for this
https://open.substack.com/pub/pramodhmallipatna/p/ais-grand-entanglement-the-subprime
Professor your article is very good , complete and transparent in the data that you use. I just want to ask how models to time market or conclude if its overvalued or under can be used when we don’t know how optimistic or pessimistic assumption are on the estimated future earnings growth. When optimism is high analysts are very confident of future earnings growth and when low its the opposite so wouldn’t this skew your models which rely on future cash flows. How do you manage this risk of over estimation or under estimation of future earnings.
What stands out here is how valuation models bend under institutional uncertainty. Once the policy map is unclear, markets stop pricing fundamentals and start pricing confidence in the model itself. That’s the part that makes ‘fairly highly valued’ less about numbers and more about narrative stability.