Any evidence that returns are greater after sudden drop?
Any evidence that returns are greater after sudden drop?
There are suddenly a bunch of posts about how one can buy stock "at a discount" given the recent drops in the market. That if you have a bunch of stock sitting around, now is a great time to buy. This can only be true if the (long-term?) returns following a sudden dip are predictably greater than otherwise, even if only by a small amount.
It's been a while, but the last time I played around with the data I seem to remember that there was *virtually no evidence at all* that this was true. But I might not be remembering it right, or maybe I missed something, and so on.
So, can anyone show me any reasonable empirical evidence that returns following a dip are greater? Take the opportunity here to school myself and other Bogleheads. If there isn't any good evidence, can we, I don't know, provide some wiki page or sticky thread to link to whenever someone suggests it?
(Please do not confuse this with the question "are returns greater if you buy at a market low"?)
It's been a while, but the last time I played around with the data I seem to remember that there was *virtually no evidence at all* that this was true. But I might not be remembering it right, or maybe I missed something, and so on.
So, can anyone show me any reasonable empirical evidence that returns following a dip are greater? Take the opportunity here to school myself and other Bogleheads. If there isn't any good evidence, can we, I don't know, provide some wiki page or sticky thread to link to whenever someone suggests it?
(Please do not confuse this with the question "are returns greater if you buy at a market low"?)
Re: Any evidence that returns are greater after sudden drop?
https://www.fool.com/investing/2019/04/ ... he-st.aspx
"Many of the best days in the market come right after the worst days. According to the J.P. Morgan study, six of the 10 best days occurred within two weeks of the 10 worst days. One example was in 2015: The best day was Aug. 26, just two days after the worst day in the stock market that year.
"The lesson here is that investors are rewarded for sticking to their investment plan and riding out the bad days in the market over time. It may seem harmless to wait out the bad days with your money, but this also means missing the up days that should boost performance over time."
"Many of the best days in the market come right after the worst days. According to the J.P. Morgan study, six of the 10 best days occurred within two weeks of the 10 worst days. One example was in 2015: The best day was Aug. 26, just two days after the worst day in the stock market that year.
"The lesson here is that investors are rewarded for sticking to their investment plan and riding out the bad days in the market over time. It may seem harmless to wait out the bad days with your money, but this also means missing the up days that should boost performance over time."
Re: Any evidence that returns are greater after sudden drop?
Thanks. Interesting, but would like to see long-term results. How do they look after 5/10/20 years? Not particularly impressed with "6 out of 10" sample sizes.Sandi_k wrote: ↑Mon Mar 09, 2020 3:54 pm https://www.fool.com/investing/2019/04/ ... he-st.aspx
"Many of the best days in the market come right after the worst days. According to the J.P. Morgan study, six of the 10 best days occurred within two weeks of the 10 worst days. One example was in 2015: The best day was Aug. 26, just two days after the worst day in the stock market that year.
"The lesson here is that investors are rewarded for sticking to their investment plan and riding out the bad days in the market over time. It may seem harmless to wait out the bad days with your money, but this also means missing the up days that should boost performance over time."
Data on market returns are out there. Surely someone besides myself has combed through the data and looked for such trends?
Re: Any evidence that returns are greater after sudden drop?
Nothing?
To get us started, here's a repeat of a post that finds evidence to the contrary:
I just pulled monthly SP 500 prices, going back to Jan 1 1927. I then looked at 10 year returns from every first of the month. I asked the question, "what are the 10-year returns following a month that drops 1%, 2%, 5%, 8% and 10% in price?" Basically, do we see better future long term returns after various kinds of bad months? Are they better than average?
The answers: average 10 year returns after a drop are lower than the overall average in every case. This certainly doesn’t support the claim that returns are greater. Did not bother to check significance for that reason.
Data from here: https://finance.yahoo.com/quote/%5EGSPC ... quency=1mo
To get us started, here's a repeat of a post that finds evidence to the contrary:
I just pulled monthly SP 500 prices, going back to Jan 1 1927. I then looked at 10 year returns from every first of the month. I asked the question, "what are the 10-year returns following a month that drops 1%, 2%, 5%, 8% and 10% in price?" Basically, do we see better future long term returns after various kinds of bad months? Are they better than average?
The answers: average 10 year returns after a drop are lower than the overall average in every case. This certainly doesn’t support the claim that returns are greater. Did not bother to check significance for that reason.
Data from here: https://finance.yahoo.com/quote/%5EGSPC ... quency=1mo
Re: Any evidence that returns are greater after sudden drop?
Wouldn't data that compares returns with rebalancing vs without rebalancing essentially answer this? Maybe look at 50/50 s&p/cash rebalance vs 50/50 s&p/cash no rebalance?rbaldini wrote: ↑Mon Mar 09, 2020 3:34 pm There are suddenly a bunch of posts about how one can buy stock "at a discount" given the recent drops in the market. That if you have a bunch of stock sitting around, now is a great time to buy. This can only be true if the (long-term?) returns following a sudden dip are predictably greater than otherwise, even if only by a small amount.
It's been a while, but the last time I played around with the data I seem to remember that there was *virtually no evidence at all* that this was true. But I might not be remembering it right, or maybe I missed something, and so on.
So, can anyone show me any reasonable empirical evidence that returns following a dip are greater? Take the opportunity here to school myself and other Bogleheads. If there isn't any good evidence, can we, I don't know, provide some wiki page or sticky thread to link to whenever someone suggests it?
(Please do not confuse this with the question "are returns greater if you buy at a market low"?)
- firebirdparts
- Posts: 4411
- Joined: Thu Jun 13, 2019 4:21 pm
- Location: Southern Appalachia
Re: Any evidence that returns are greater after sudden drop?
I've been looking at index charts for 30 years. It's pretty easy to see the evidence, isn't it?
If there's any confusion, I guess it would be in the area of "compared to what?". Returns after the drop are greater than the average including the drop. If I'm wrong, just use a shorter average.
If there's any confusion, I guess it would be in the area of "compared to what?". Returns after the drop are greater than the average including the drop. If I'm wrong, just use a shorter average.
This time is the same
Re: Any evidence that returns are greater after sudden drop?
Market level is presently at the level last registered in mid January 2019. Use the search function to look for posting dated back then. A good fraction of them stated that market level was "expensive", or at least "fair".
Yield curve at that time was fairly flat around 2.5%, with the T-Bond at just a smidge over 3%. Today we are about 2% lower across all maturities.
On the other hand, in January 2019 there was no inkling of a possible substantial disruption in the World's supply chain and the drop in the demand for many sectors.
Earnings for the S&P 500 were also a little higher.
You draw your own conclusions, but it would not be prudent to ingnore the possibility that the present market level might turn our to be expensive.
Yield curve at that time was fairly flat around 2.5%, with the T-Bond at just a smidge over 3%. Today we are about 2% lower across all maturities.
On the other hand, in January 2019 there was no inkling of a possible substantial disruption in the World's supply chain and the drop in the demand for many sectors.
Earnings for the S&P 500 were also a little higher.
You draw your own conclusions, but it would not be prudent to ingnore the possibility that the present market level might turn our to be expensive.
Re: Any evidence that returns are greater after sudden drop?
Can't say I am surprised by your findings. And if you're reacting to people claiming that tomorrow will be a great buying opportunity after the hiccup we've experienced in the past 2 weeks (a bigger 'dip' than what you tested though), then yes, it is very likely they will be disappointed. Heck, the big dip of the past few days essentially nullified the climb of the past few months. I am not quite sure to see where is the buying opportunity.rbaldini wrote: ↑Mon Mar 09, 2020 7:07 pm Nothing?
To get us started, here's a repeat of a post that finds evidence to the contrary:
I just pulled monthly SP 500 prices, going back to Jan 1 1927. I then looked at 10 year returns from every first of the month. I asked the question, "what are the 10-year returns following a month that drops 1%, 2%, 5%, 8% and 10% in price?" Basically, do we see better future long term returns after various kinds of bad months? Are they better than average?
The answers: average 10 year returns after a drop are lower than the overall average in every case. This certainly doesn’t support the claim that returns are greater. Did not bother to check significance for that reason.
Data from here: https://finance.yahoo.com/quote/%5EGSPC ... quency=1mo
Now if you were to test the actual returns over 10 years after a drawdown of 30% or more (hence a bit more than a 'dip'), then I suspect you'll find more evidence of some form of mean reversion... Even though drawdowns are terribly misleading, by virtue of being relative.
- Brianmcg321
- Posts: 1875
- Joined: Mon Jul 15, 2019 8:23 am
Re: Any evidence that returns are greater after sudden drop?
Rules to investing: |
1. Don't lose money. |
2. Don't forget rule number 1.
- CyclingDuo
- Posts: 6006
- Joined: Fri Jan 06, 2017 8:07 am
Re: Any evidence that returns are greater after sudden drop?
Not sure Ben Carson's post last week in any way qualifies that returns following a dip are greater, but it does highlight what returns were after a big down month 1 year, 3 years, and 5 years hence since 1926...rbaldini wrote: ↑Mon Mar 09, 2020 3:34 pm There are suddenly a bunch of posts about how one can buy stock "at a discount" given the recent drops in the market. That if you have a bunch of stock sitting around, now is a great time to buy. This can only be true if the (long-term?) returns following a sudden dip are predictably greater than otherwise, even if only by a small amount.
It's been a while, but the last time I played around with the data I seem to remember that there was *virtually no evidence at all* that this was true. But I might not be remembering it right, or maybe I missed something, and so on.
So, can anyone show me any reasonable empirical evidence that returns following a dip are greater? Take the opportunity here to school myself and other Bogleheads. If there isn't any good evidence, can we, I don't know, provide some wiki page or sticky thread to link to whenever someone suggests it?
(Please do not confuse this with the question "are returns greater if you buy at a market low"?)
https://awealthofcommonsense.com/2020/0 ... own-month/
"Save like a pessimist, invest like an optimist." - Morgan Housel |
"Pick a bushel, save a peck!" - Grandpa
-
- Posts: 2006
- Joined: Mon Mar 02, 2020 4:33 pm
Re: Any evidence that returns are greater after sudden drop?
Correlation does not imply causation.
It could simply be that both big gains and big drops tend to cluster together in the periods of high market volatility, not that drops make subsequent gains more likely.
It could simply be that both big gains and big drops tend to cluster together in the periods of high market volatility, not that drops make subsequent gains more likely.
- Uncorrelated
- Posts: 1237
- Joined: Sun Oct 13, 2019 3:16 pm
Re: Any evidence that returns are greater after sudden drop?
There is considerable evidence that volatility clusters, i.e. if yesterday was a large drop, there is increased volatility today which increases the probability of large returns (and also increases the probability of more large losses).
For individual stocks on monthly timescales, the effects of momentum are generally considered to be much stronger than mean reversion.
For individual stocks on monthly timescales, the effects of momentum are generally considered to be much stronger than mean reversion.
Re: Any evidence that returns are greater after sudden drop?
I don't see any evidence in favor, if that's what you mean. In any case, I did comparisons against the total mean and against the "not after a drop" case. E.g., take all months. Split them into those that follow a 5% drop, and those that don't. Are *future* returns (i.e. after the drop) better in the drop group than the non-drop group? Better than the overall average? Historical SP 500 data suggests not.firebirdparts wrote: ↑Mon Mar 09, 2020 7:19 pm I've been looking at index charts for 30 years. It's pretty easy to see the evidence, isn't it?
If there's any confusion, I guess it would be in the area of "compared to what?". Returns after the drop are greater than the average including the drop. If I'm wrong, just use a shorter average.
Last edited by rbaldini on Tue Mar 10, 2020 9:26 am, edited 4 times in total.
Re: Any evidence that returns are greater after sudden drop?
That's a fair question. Of course the sample size of months that have dropped more than 30% is extremely small (0?), so I'd have to broaden the analysis. Perhaps look at yearly results.siamond wrote: ↑Mon Mar 09, 2020 9:41 pm Now if you were to test the actual returns over 10 years after a drawdown of 30% or more (hence a bit more than a 'dip'), then I suspect you'll find more evidence of some form of mean reversion... Even though drawdowns are terribly misleading, by virtue of being relative.
Last edited by rbaldini on Tue Mar 10, 2020 9:33 am, edited 1 time in total.
Re: Any evidence that returns are greater after sudden drop?
Quite so. He says average long-term returns after dips are positive (it would be surprising if they weren't!), but he doesn't demonstrate that they are *greater that they would be otherwise*. But this is precisely what people are claiming when they say that there is some positive buying opportunity after a stock has dropped in price. I have yet to see any evidence.CyclingDuo wrote: ↑Mon Mar 09, 2020 10:10 pm Not sure Ben Carson's post last week in any way qualifies that returns following a dip are greater, but it does highlight what returns were after a big down month 1 year, 3 years, and 5 years hence since 1926...
https://awealthofcommonsense.com/2020/0 ... own-month/
Last edited by rbaldini on Tue Mar 10, 2020 9:32 am, edited 1 time in total.
Re: Any evidence that returns are greater after sudden drop?
Having looked at SP500 data I only see evidence to the contrary.
Re: Any evidence that returns are greater after sudden drop?
The comparison I did above was much simpler: are long-term returns greater immediately following a dip, or not? No need to look at rebalancing, etc. I haven't checked for statistical significance but the observed empirical trend does not support the hypothesis that they are greater.
Re: Any evidence that returns are greater after sudden drop?
Perhaps you've stumbled into something along the lines of the momentum factor.
Re: Any evidence that returns are greater after sudden drop?
That would be consistent with the finding that returns appear to be worse, yes. The problem is that the forum is absolutely brimming right now with the exact opposite idea: that the price drops of the last few days are a wonderful thing for those who are still accumulating, that "stocks are on sale", and so on. So I'm pleading with the forum to provide any convincing evidence supporting this viewpoint. I don't know of any.
Re: Any evidence that returns are greater after sudden drop?
Yes, I didn't mean a monthly drop. I meant a drawdown in the traditional definition (from the last high point, which may have been a while ago).rbaldini wrote: ↑Tue Mar 10, 2020 9:13 amThat's a far question. Of course the sample size of months that have dropped more than 30% is extremely small (0?), so I'd have to broaden the analysis. Perhaps look at yearly results.siamond wrote: ↑Mon Mar 09, 2020 9:41 pm Now if you were to test the actual returns over 10 years after a drawdown of 30% or more (hence a bit more than a 'dip'), then I suspect you'll find more evidence of some form of mean reversion... Even though drawdowns are terribly misleading, by virtue of being relative.
I agree it's a rather different question. Perceptions and facts driven by sudden moves are one thing (essentially what we are going through right now), perceptions and facts driven by a deeper/longer crisis are another.
I suspect there is a good deal of confusion between the two questions though, notably in the other thread which apparently triggered you to open this new thread. And this was my entire point...
Re: Any evidence that returns are greater after sudden drop?
(I meant to say "fair point", not "far point", for what it's worth).siamond wrote: ↑Tue Mar 10, 2020 9:31 am Yes, I didn't mean a monthly drop. I meant a drawdown in the traditional definition (from the last high point, which may have been a while ago).
I agree it's a rather different question. Perceptions and facts driven by sudden moves are one thing (essentially what we are going through right now), perceptions and facts driven by a deeper/longer crisis are another.
I suspect there is a good deal of confusion between the two questions though, notably in the other thread which apparently triggered you to open this new thread. And this was my entire point...
I might experiment with broadening the definition of a "sudden drop". I focused on a sudden drop on a short timescale (one month) because that appears to be what we are experiencing now - we're not in an extended draw down, yet. And *every time it happens*, people talk about "buying opportunity".
Re: Any evidence that returns are greater after sudden drop?
Nobody knows.... if it's a dip or the start of a huge drawdown. As nobody knows what kind of profit impact this virus is going to end up happening. Most bear markets play out over months/ years..... buying the "dip' in 2001 wouldn't have worked too well....as the bear market went on another 2 years.rbaldini wrote: ↑Tue Mar 10, 2020 9:35 am(I meant to say "fair point", not "far point", for what it's worth).siamond wrote: ↑Tue Mar 10, 2020 9:31 am Yes, I didn't mean a monthly drop. I meant a drawdown in the traditional definition (from the last high point, which may have been a while ago).
I agree it's a rather different question. Perceptions and facts driven by sudden moves are one thing (essentially what we are going through right now), perceptions and facts driven by a deeper/longer crisis are another.
I suspect there is a good deal of confusion between the two questions though, notably in the other thread which apparently triggered you to open this new thread. And this was my entire point...
I might experiment with broadening the definition of a "sudden drop". I focused on a sudden drop on a short timescale (one month) because that appears to be what we are experiencing now - we're not in an extended draw down, yet. And *every time it happens*, people talk about "buying opportunity".
Re: Any evidence that returns are greater after sudden drop?
And I thank you for doing that. We definitely need more of those "calm down, don't overreact, check hard facts" posts nowadays...
Disclaimer: I talk the talk, but I don't necessarily walk the walk. I rushed to do a Roth conversion last week. And I hesitated to do another round last night. Then sobered up. Those emotions are just so hard to control.
-
- Posts: 2241
- Joined: Fri Oct 19, 2012 11:24 am
Re: Any evidence that returns are greater after sudden drop?
Two thoughts:
1) are you using price data or total return?
2) to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
1) are you using price data or total return?
2) to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
“Doing nothing is better than being busy doing nothing.” – Lao Tzu
- Ben Mathew
- Posts: 2720
- Joined: Tue Mar 13, 2018 11:41 am
- Location: Seattle
Re: Any evidence that returns are greater after sudden drop?
Not necessarily. If returns are independent across time and have a positive expected return, markets will generally rise, but will be no likelier to rise after a large drop.RubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
Total Portfolio Allocation and Withdrawal (TPAW)
- CyclingDuo
- Posts: 6006
- Joined: Fri Jan 06, 2017 8:07 am
Re: Any evidence that returns are greater after sudden drop?
Are you looking for data confirming that an individual stock (I bolded the word stock since you used it in your statement), or the total market (index) does or does not provide opportunity via price discovery?rbaldini wrote: ↑Tue Mar 10, 2020 9:15 amQuite so. He says average long-term returns after dips are positive (it would be surprising if they weren't!), but he doesn't demonstrate that they are *greater that they would be otherwise*. But this is precisely what people are claiming when they say that there is some positive buying opportunity after a stock has dropped in price. I have yet to see any evidence.CyclingDuo wrote: ↑Mon Mar 09, 2020 10:10 pm Not sure Ben Carson's post last week in any way qualifies that returns following a dip are greater, but it does highlight what returns were after a big down month 1 year, 3 years, and 5 years hence since 1926...
https://awealthofcommonsense.com/2020/0 ... own-month/
In terms of the S&P 500, narrow the time reference down to the most recent 40 year period (1979-2019), and this storyboard put together by a particular person named Jeremy Schneider (https://www.instagram.com/personalfinanceclub/) who must be a BH member based on posts on his blog - or visit his blog Personal Finance Club link: https://www.personalfinanceclub.com/ - at least attempts to show the difference between an investor buying at the highs right before each drop, and one buying at the lows right after a big drop compared to another investor path (one that I much prefer after doing it for 30+ years) - buying all the time through the thick and the thin. Even with all of our dividend producing investments, I don't bother collecting the dividends and waiting for dips or drops using price discovery in an attempt to purchase at lower prices. I just reinvest them all immediately on automatic pilot (well, the financial institutions do that for us since we selected that option with all of our accounts).
Regardless, here is the storyboard with Tiffany (buys right before major drops), Brittany (buys at the lows after major drops), and Sarah (buys all the time no matter what the price).
https://imgur.com/gallery/BlK4jzM
Link to the spreadsheet that was used to create that illustration:
https://docs.google.com/spreadsheets/d/ ... edit#gid=0
It's always an interesting discussion whether it is framed as "buy low, sell high", or "buy on the dips", or "buy undervalued, sell overvalued", or "time in the market beats timing the market" and on and on. I'm not arguing against you or for you - just trying to clarify if you are speaking of an individual stock, or the market. We're in the camp of just continuing to dump more in on automatic pilot every week, every month, and every year while we are still working.
If you have data that confirms your thoughts - feel free to post it up. JL Collins and his stock series is also a good reference point to read if you have time and interest. I think, and I paraphrase, his strategy centers around the theme of "Buy as much VTSAX and as often as you can!".
CyclngDuo
"Save like a pessimist, invest like an optimist." - Morgan Housel |
"Pick a bushel, save a peck!" - Grandpa
Re: Any evidence that returns are greater after sudden drop?
ReturnRubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am Two thoughts:
1) are you using price data or total return?
Yeah, it's wrong. A run of coin flips over time tends to converge to 50% heads. This does not mean you get more heads after a run of tails. This is not to say that the market must behave like a coin flip. The point is that it does not mathematically "have to be greater" after a drop: future results could be independent. Or hell, they could have "momentum", in which case the expected future return would be lower!RubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am 2) to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
EDIT: Basically what Ben Mathew said above.
Last edited by rbaldini on Tue Mar 10, 2020 1:33 pm, edited 3 times in total.
Re: Any evidence that returns are greater after sudden drop?
Sorry to be unclear. Was looking at the whole market. Well, SP 500 to be specific.CyclingDuo wrote: ↑Tue Mar 10, 2020 12:10 pm Are you looking for data confirming that an individual stock (I bolded the word stock since you used it in your statement), or the total market (index) does or does not provide opportunity via price discovery?
Here's what you (and I think others) are missing: "buying at the highs right before each drop, and one buying at the lows right after a big drop" is not the proper comparison. It's not actionable because you never know if you are at a high or low "before" a big drop or rise. That requires you to know the future! You only know that past!CyclingDuo wrote: ↑Tue Mar 10, 2020 12:10 pm Personal Finance Club link: https://www.personalfinanceclub.com/ - at least attempts to show the difference between an investor buying at the highs right before each drop, and one buying at the lows right after a big drop compared to another investor path (one that I much prefer after doing it for 30+ years) - buying all the time through the thick and the thin
All you know is this: a notable drop just happened. You don't know if you are at a low yet, because you don't know the future. It could keep going down, or it could rebound. Based on this information, do you expect returns to be higher than average? The evidence suggests that you have no reason to believe this.
I posted the data I used above. It's just the SP500 history. I can post the code I used to analyze it, if you'd like. Or feel free to do the math yourself.CyclingDuo wrote: ↑Tue Mar 10, 2020 12:10 pm If you have data that confirms your thoughts - feel free to post it up.
Re: Any evidence that returns are greater after sudden drop?
Lets look at the long runRubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am
2) to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
a) If you invest 10k and get 10%/year for 99 years and then get -50% in year 100, how much money do you end up with? (62.6 million)
b) if you invest 10k, have a 50% drop, and then have 99 years of 10% growth how much money do you end up with? (62.6million)
Makes no difference if you buy at the top or the bottom over long periods of time. The only way you get better or worse results is from sampling issues (i.e. if you sample say 15 years, you will have some cases with 2 crashes and other with 1 crash).
For reference go back and look at how an April 2001 investor did. They bought in after the market dropped 20%. They make a 6% CAGR. What about the Sept 2002 crew (markets off 40%)? 9.37. You have people who bought at the bottom of one of the worst crashes ever where 15+ years later they still have had below (granted not by much) average returns. Obviously this is looking at just 1 time sequence. Right now the March 2009 is look looking at great 10 year returns. But who know what the 15, 20 or 30 year ones will look like.
Now the gotchas
a) if you are doing periodic investment or selling, sequence of returns matter
b) if you don't make it to year 100, you would much prefer for the drop to happen after you die:)
-
- Posts: 2241
- Joined: Fri Oct 19, 2012 11:24 am
Re: Any evidence that returns are greater after sudden drop?
Are they independent across time? I would imagine returns are NOT independent across time. I guess I have some reading to do to adjust my intuitionBen Mathew wrote: ↑Tue Mar 10, 2020 10:06 amNot necessarily. If returns are independent across time and have a positive expected return, markets will generally rise, but will be no likelier to rise after a large drop.RubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
“Doing nothing is better than being busy doing nothing.” – Lao Tzu
-
- Posts: 2241
- Joined: Fri Oct 19, 2012 11:24 am
Re: Any evidence that returns are greater after sudden drop?
Coin flip analogy not compelling to me as I would imagine expected returns are not independent flips over time. And also, I agree that in short term, momentum would lower expected (short term) return. Suspect that momentum is purely short term effect and not relevant to my thoughts about long term expected return being higher after asset sell off.rbaldini wrote: ↑Tue Mar 10, 2020 1:23 pmReturnRubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am Two thoughts:
1) are you using price data or total return?
Yeah, it's wrong. A run of coin flips over time tends to converge to 50% heads. This does not mean you get more heads after a run of tails. This is not to say that the market must behave like a coin flip. The point is that it does not mathematically "have to be greater" after a drop: future results could be independent. Or hell, they could have "momentum", in which case the expected future return would be lower!RubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am 2) to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
EDIT: Basically what Ben Mathew said above.
I will do some further reading, and follow this thread, and see whether my intuition gets adjusted.
“Doing nothing is better than being busy doing nothing.” – Lao Tzu
Re: Any evidence that returns are greater after sudden drop?
The independent random-walk model is pretty decent, though I don't think anyone believes it is exactly right. For intuition, here are some simple cases:RubyTuesday wrote: ↑Tue Mar 10, 2020 2:45 pm Are they independent across time? I would imagine returns are NOT independent across time. I guess I have some reading to do to adjust my intuition
If returns are independent, then the returns after a big drop are no different than after any other market behavior.
If returns are serially positively correlated - AKA they have momentum - then returns will be expected lower than average after a dip, so the long term will be ever-so-slightly worse after a dip.
If returns are serially negatively correlated, then returns will be expected higher than average after a dip, so the long term will be ever-so-slightly better after a dip.
Of course there are all sorts of other models that could apply too...
Which is true? I don't know, but the historical market returns don't strongly support the negative case - at least at the monthly timescale I looked at
Last edited by rbaldini on Tue Mar 10, 2020 3:05 pm, edited 1 time in total.
-
- Posts: 2241
- Joined: Fri Oct 19, 2012 11:24 am
Re: Any evidence that returns are greater after sudden drop?
First, I’m not suggesting there’s a benefit to trying to time the market. This isn’t about that IMO. This is about whether expected return is higher after a “significant” decline. My intuition is that it is. The sampling issue is exactly what we are discussing.randomguy wrote: ↑Tue Mar 10, 2020 2:06 pmLets look at the long runRubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am
2) to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
a) If you invest 10k and get 10%/year for 99 years and then get -50% in year 100, how much money do you end up with? (62.6 million)
b) if you invest 10k, have a 50% drop, and then have 99 years of 10% growth how much money do you end up with? (62.6million)
Makes no difference if you buy at the top or the bottom over long periods of time. The only way you get better or worse results is from sampling issues (i.e. if you sample say 15 years, you will have some cases with 2 crashes and other with 1 crash).
For reference go back and look at how an April 2001 investor did. They bought in after the market dropped 20%. They make a 6% CAGR. What about the Sept 2002 crew (markets off 40%)? 9.37. You have people who bought at the bottom of one of the worst crashes ever where 15+ years later they still have had below (granted not by much) average returns. Obviously this is looking at just 1 time sequence. Right now the March 2009 is look looking at great 10 year returns. But who know what the 15, 20 or 30 year ones will look like.
Now the gotchas
a) if you are doing periodic investment or selling, sequence of returns matter
b) if you don't make it to year 100, you would much prefer for the drop to happen after you die:)
In other words, I’m simply saying that my intuition is that the expected long term return after a significant decline is higher than before the decline.
“Doing nothing is better than being busy doing nothing.” – Lao Tzu
Re: Any evidence that returns are greater after sudden drop?
Just noting that if expected long-term return where higher after a dip, it would potentially make sense to hold money out of the market and invest after drops. Which is a form of market timing, of course. It wouldn't necessarily be the best course of action (holding money out of the market has opportunity cost; transactions costs), but it's possible.RubyTuesday wrote: ↑Tue Mar 10, 2020 2:59 pm First, I’m not suggesting there’s a benefit to trying to time the market. This isn’t about that IMO. This is about whether expected return is higher after a “significant” decline. My intuition is that it is. The sampling issue is exactly what we are discussing.
In other words, I’m simply saying that my intuition is that the expected long term return after a significant decline is higher than before the decline.
In any case, I encourage you to check out the data yourself.
Re: Any evidence that returns are greater after sudden drop?
I would argue that only buying on days after a sudden drop is just a version of market timing. I don't hold any money out of the market for that type of timing. It's better to just invest the money if you have it, when it becomes available.
The arguments you hear about it being a great time to buy because stocks are on sale are often to counteract the natural fear that many people have to pull all their money from the market after a big decline. Creating the perspective that stock prices are low is a good thing fights the fear. It's okay to let your money ride and buy more as money becomes available. You're not throwing good money after bad by buying into a bear market.
But I wouldn't time the market by only buying after sudden drops as a strategy.
The arguments you hear about it being a great time to buy because stocks are on sale are often to counteract the natural fear that many people have to pull all their money from the market after a big decline. Creating the perspective that stock prices are low is a good thing fights the fear. It's okay to let your money ride and buy more as money becomes available. You're not throwing good money after bad by buying into a bear market.
But I wouldn't time the market by only buying after sudden drops as a strategy.
Re: Any evidence that returns are greater after sudden drop?
If the market is at 100 one day and then falls to 80, you are better off buying at 80. This is true whether the market climbs to 500 or falls to 20 by your retirement date.
Re: Any evidence that returns are greater after sudden drop?
This is another useless comparison. Yes, after you have observed a drop, you can say it was better to buy after than before. That would be useful information if you had a time machine. The actionable question is "are returns going to be greater starting now than if the drop hadn't happened?" If that's so, then it implies that market timing might make sense: returns are better after a dip, so wait for the dips! It does not appear to be true, however.
Look, it's simple: Go back and look at every month in the last 100 years (minus the most recent 10 years). Look at every month following a 5% drop, and then compute the return exactly 10 years later. Then do the same for all the other months. Is the mean return higher in the first group? No, it is not.
Last edited by rbaldini on Tue Mar 10, 2020 3:35 pm, edited 1 time in total.
-
- Posts: 2241
- Joined: Fri Oct 19, 2012 11:24 am
Re: Any evidence that returns are greater after sudden drop?
The useless comparison is basically what my intuition was clinging to... that fact that after observing a crash, it was better to buy After than It would have been before. As you state, this isn’t actionable. And I don’t invest this way, or do anything like market timing. Waiting for a crash before investing is a losing game IMO.rbaldini wrote: ↑Tue Mar 10, 2020 3:23 pmThis is another useless comparison. Yes, after you have observed a drop, you can say it was better to buy after than before. That would be useful information if you had a time machine. The actionable question is "are returns going to be greater now than if the drop had never happened?" If that's so, then it implies that market timing might make sense: returns are better after a dip, so wait for the dips! It does not appear to be true, however.
Look, it's simple: Go back and look at every month in the last 100 years (minus the most recent 10 years). Look at every month following a 5% drop, and then compute the return exactly 10 years later. Then do the same for all the other months. Is the mean return higher in the first group? No, it is not.
“Doing nothing is better than being busy doing nothing.” – Lao Tzu
- firebirdparts
- Posts: 4411
- Joined: Thu Jun 13, 2019 4:21 pm
- Location: Southern Appalachia
Re: Any evidence that returns are greater after sudden drop?
Maybe we should go 100% equities after not a drop. I never thought about looking at it month by month. I suppose as another posted suggested you are looking at "momentum."rbaldini wrote: ↑Tue Mar 10, 2020 9:12 am I don't see any evidence in favor, if that's what you mean. In any case, I did comparisons against the total mean and against the "not after a drop" case. E.g., take all months. Split them into those that follow a 5% drop, and those that don't. Are *future* returns (i.e. after the drop) better in the drop group than the non-drop group? Better than the overall average? Historical SP 500 data suggests not.
This time is the same
- Ben Mathew
- Posts: 2720
- Joined: Tue Mar 13, 2018 11:41 am
- Location: Seattle
Re: Any evidence that returns are greater after sudden drop?
Whether or not returns are independent across time is an empirical question (the evidence suggests that they are pretty close to independent across time, though there are anomalies like momentum over certain time horizons). But I understood your claim in the OP to be a mathematical one--that if markets are rising over the long term, then long term returns will be higher after a drop. So that is what I was responding to.RubyTuesday wrote: ↑Tue Mar 10, 2020 2:45 pmAre they independent across time? I would imagine returns are NOT independent across time. I guess I have some reading to do to adjust my intuitionBen Mathew wrote: ↑Tue Mar 10, 2020 10:06 amNot necessarily. If returns are independent across time and have a positive expected return, markets will generally rise, but will be no likelier to rise after a large drop.RubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
Last edited by Ben Mathew on Wed Mar 11, 2020 11:06 am, edited 1 time in total.
Total Portfolio Allocation and Withdrawal (TPAW)
-
- Posts: 2241
- Joined: Fri Oct 19, 2012 11:24 am
Re: Any evidence that returns are greater after sudden drop?
In case I haven’t said it, I concede the point. My intuition is not trust worthy!Ben Mathew wrote: ↑Wed Mar 11, 2020 11:03 amWhether or not returns are independent across time is an empirical question (and the empirical answer is that they are pretty close to independent, but there seem to be anomalies like momentum over certain time horizons). But I understood your claim in the OP to be a mathematical one--that if markets are rising over the long term, then long term returns will be higher after a drop. So that is what I was responding to.RubyTuesday wrote: ↑Tue Mar 10, 2020 2:45 pmAre they independent across time? I would imagine returns are NOT independent across time. I guess I have some reading to do to adjust my intuitionBen Mathew wrote: ↑Tue Mar 10, 2020 10:06 amNot necessarily. If returns are independent across time and have a positive expected return, markets will generally rise, but will be no likelier to rise after a large drop.RubyTuesday wrote: ↑Tue Mar 10, 2020 9:47 am to my intuition, it seems to me if the market is generally rising, albeit in a noisy and volatile manner, the average long term return after a significant drop has to be greater than the average long term return. Seems like math... but my intuition isn’t what it used to be.
“Doing nothing is better than being busy doing nothing.” – Lao Tzu
Re: Any evidence that returns are greater after sudden drop?
This I can answer. Yes, but only if you're omniscient.rbaldini wrote: ↑Tue Mar 10, 2020 3:04 pmJust noting that if expected long-term return where higher after a dip, it would potentially make sense to hold money out of the market and invest after drops. Which is a form of market timing, of course. It wouldn't necessarily be the best course of action (holding money out of the market has opportunity cost; transactions costs), but it's possible.RubyTuesday wrote: ↑Tue Mar 10, 2020 2:59 pm First, I’m not suggesting there’s a benefit to trying to time the market. This isn’t about that IMO. This is about whether expected return is higher after a “significant” decline. My intuition is that it is. The sampling issue is exactly what we are discussing.
In other words, I’m simply saying that my intuition is that the expected long term return after a significant decline is higher than before the decline.
In any case, I encourage you to check out the data yourself.
https://ofdollarsanddata.com/even-god-c ... aging/amp/
So really, no. You cannot beat dumping it all in as you receive it without some advantage the rest of us don't have.
Re: Any evidence that returns are greater after sudden drop?
Since 1974, the five year rolling return is actually smaller during the ten year period after the drawdown ends. Defined as the last year the rolling return is negative. You can see this yourself via portfolio visualizer.rbaldini wrote: ↑Mon Mar 09, 2020 3:34 pm So, can anyone show me any reasonable empirical evidence that returns following a dip are greater? Take the opportunity here to school myself and other Bogleheads. If there isn't any good evidence, can we, I don't know, provide some wiki page or sticky thread to link to whenever someone suggests it?
So, actually, the best time to buy is when you’ve got the money.
-
- Posts: 18
- Joined: Mon Mar 09, 2020 2:02 pm
Re: Any evidence that returns are greater after sudden drop?
Interesting!
I think the heart of the general intuition is...
S&P will achieve price X in 2040 -- that's fixed. To get there, we will need to generate x% annual returns.
Now that there has been a 20% drop, the S&P will need to generate y% > x% returns -- so I'm better off investing more now.
Or
S&P will average x% returns over the next 20 years. A 20% drop over the last month means each annual return needs to be higher to make up and still achieve x%.
All of that is assuming the corona virus event has not in reality adjusted future growth/prices 20 years out.
Based on your analysis, it seems that maybe this reasoning is flawed -- but I'm not seeing it yet.
I think the heart of the general intuition is...
S&P will achieve price X in 2040 -- that's fixed. To get there, we will need to generate x% annual returns.
Now that there has been a 20% drop, the S&P will need to generate y% > x% returns -- so I'm better off investing more now.
Or
S&P will average x% returns over the next 20 years. A 20% drop over the last month means each annual return needs to be higher to make up and still achieve x%.
All of that is assuming the corona virus event has not in reality adjusted future growth/prices 20 years out.
Based on your analysis, it seems that maybe this reasoning is flawed -- but I'm not seeing it yet.
Re: Any evidence that returns are greater after sudden drop?
So, actually, the best time to buy is when you’ve got the money and you can see something worth buying.Lee_WSP wrote: ↑Wed Mar 11, 2020 11:37 amSince 1974, the five year rolling return is actually smaller during the ten year period after the drawdown ends. Defined as the last year the rolling return is negative. You can see this yourself via portfolio visualizer.rbaldini wrote: ↑Mon Mar 09, 2020 3:34 pm So, can anyone show me any reasonable empirical evidence that returns following a dip are greater? Take the opportunity here to school myself and other Bogleheads. If there isn't any good evidence, can we, I don't know, provide some wiki page or sticky thread to link to whenever someone suggests it?
So, actually, the best time to buy is when you’ve got the money.
Re: Any evidence that returns are greater after sudden drop?
I redid the test using Simba's sheet to see what the 10 & 15 year rolling returns were.Lee_WSP wrote: ↑Wed Mar 11, 2020 11:37 amSince 1974, the five year rolling return is actually smaller during the ten year period after the drawdown ends. Defined as the last year the rolling return is negative. You can see this yourself via portfolio visualizer.rbaldini wrote: ↑Mon Mar 09, 2020 3:34 pm So, can anyone show me any reasonable empirical evidence that returns following a dip are greater? Take the opportunity here to school myself and other Bogleheads. If there isn't any good evidence, can we, I don't know, provide some wiki page or sticky thread to link to whenever someone suggests it?
So, actually, the best time to buy is when you’ve got the money.
The only dips in the rolling returns are during deep drawdowns where the rolling returns for the 5 years from start of the draw-down. So, the worst time to invest is right before the draw-down, but any other time is perfectly fine as far as 10 or 15 year rolling returns go.
Data tested from 1871 to 2019.
edit:
So, basically the rolling returns dip right around the draw down period and then return to an "average" level. So, avoiding the dip is the only way to maximize the rolling returns.
Last edited by Lee_WSP on Wed Mar 11, 2020 11:56 am, edited 1 time in total.
Re: Any evidence that returns are greater after sudden drop?
Not greater, and possibly worse. They seem to want to return to an "average" level achieved before and after the dip, but the dip really wrecks havoc with the rolling returns.
Re: Any evidence that returns are greater after sudden drop?
The problem with your first statement is this faith that it will end at some fixed number, regardless of what happens in the mean time. Yes, it will end *somewhere*, but we don't know where. In reality, what happens in the meantime affects our estimate of where it will eventually end up. Imagine it on a shorter timescale: you might say "I think the SP500 will end up at $2800 on 5pm Friday". It's a fine estimate. Suppose it then drops down $100 in the next hour. Do you think "Wow! Given that it's going to go back up to $2800, returns between now and Friday are going to be even better!" No: you probably revise your estimate to some extent. If this doesn't make sense to you, imagine that it's still $2500 on Friday noon. Or Friday 4pm. Do you still expect a sudden jump?gone_forever wrote: ↑Wed Mar 11, 2020 11:42 am Interesting!
I think the heart of the general intuition is...
S&P will achieve price X in 2040 -- that's fixed. To get there, we will need to generate x% annual returns.
Now that there has been a 20% drop, the S&P will need to generate y% > x% returns -- so I'm better off investing more now.
Or
S&P will average x% returns over the next 20 years. A 20% drop over the last month means each annual return needs to be higher to make up and still achieve x%.
All of that is assuming the corona virus event has not in reality adjusted future growth/prices 20 years out.
Based on your analysis, it seems that maybe this reasoning is flawed -- but I'm not seeing it yet.
The problem with the second statement is basically the gambler's fallacy: we MUST see an x% annual return by date y, so if we just saw a low return, it must imply higher-than-average returns in the future to catch up! It is *possible* for this to to be true (compensatory market behavior, etc), but it is not a mathematical necessity. And it does not appear to be supported by data. Coin flips tend to approach 50% over a large sample size. That does not mean that the coin remembers past heads so that it can do more tails in the future. (Not saying that the market necessarily behaves like a coin flip, just using it as an example of the fallacy.)
-
- Posts: 18
- Joined: Mon Mar 09, 2020 2:02 pm
Re: Any evidence that returns are greater after sudden drop?
Hmm -- I'm having a hard time seeing is that we should all expect a 20% smaller balance 20 years from now than we would have a couple months ago. Imagine being in 2040 and thinking "I'm 25% away from my target -- it's too bad that corona virus permanently shaved off 20% of the value of all companies back in 2020.
That being said, I expect you're right -- just need some time to think about it.
Perhaps I'm treating the corona virus as 0% risk of having a lasting impact, when in reality it's >0% and if it were 0% and everyone knew it, the drop would *only* reflect lost revenue and potential bankruptcies over the next few months. By thinking stocks are on sale, I'm assuming the market is over-reacting to the risk and failing to take sufficient advantage of other peoples over-reactions.
This resolves another thing I've been thinking lately is -- why after a 10% jump do we not think stocks are over priced; but after a 10% drop we think they're on sale?
Thanks for your analysis.
That being said, I expect you're right -- just need some time to think about it.
Perhaps I'm treating the corona virus as 0% risk of having a lasting impact, when in reality it's >0% and if it were 0% and everyone knew it, the drop would *only* reflect lost revenue and potential bankruptcies over the next few months. By thinking stocks are on sale, I'm assuming the market is over-reacting to the risk and failing to take sufficient advantage of other peoples over-reactions.
This resolves another thing I've been thinking lately is -- why after a 10% jump do we not think stocks are over priced; but after a 10% drop we think they're on sale?
Thanks for your analysis.