Automated Market Timing

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Automated Market Timing

Post by Veiled »

I've been reading about tactical asset allocation because of the current CAPE ratio and because I haven't seen my first big dip yet (I lived through the tech bubble and the GFC, but had no real skin in the game). I'm keenly aware that one doesn't know his own risk tolerance until the first big bear.

As I read, I began to wonder whether, just as some use roboadvisors to automate tax loss/gain harvesting, eventually we might be able to automate tactical reallocation according to a pre-arranged decision by the investor. An automated process like this would allow an investor to remain hands-off and emotionless and continuously participate in the market, but at the same time avoid unnecessary volatility. Imagine, for instance, that I instruct this theoretical software to shift my AA according to the CAPE ratio given historical data. My AA looks crazy (very antithetical to the typical BH approach), as shown by bh7 in this thread:

Image

But my returns look more stable. The portfolio behaves much like a balanced portfolio with slightly more gentle dips in a backtest at portfoliovisualizer given a 10k investment over 32 years. See how it's a little easier to look at in 2008? You may object that it dips to a similar absolute value; I reply that a gentler dip might be helpful for a more nervous investor to stay the course. (Or I might reply, why would anyone want to see the meaningless extra inflation anyway, when we're all going to come down to the same place?)

Image
This is log scale and inflation-adjusted.

There are even more provoking graphs that I could create with market timing portfolios that exit the market or use other extreme strategies, but I think if you're on this forum can use your imagination.

In short, my question is: could automated tactical adjustment be useful? I'd get a stabler portfolio performance, I wouldn't screw up market timing by forgetfulness, fear, or greed, and I don't have to spend a lot of time thinking about it. I understand this is market timing, but would market timing be any more acceptable if it was planned, desired according to risk tolerance, and automated?

(Please be gentle to a newbie who is knowingly asking a market timing question here. It's an honest and theoretical question. My IPS is clearly laid out for strategic AA and this question is not for personal decision-making. It's academic discussion of a tool that doesn't exist, as far as I know. I predict the most common response will be "But you don't need to if you have enough time in the market with a balanced portfolio, look at your own graph!" But perhaps there are investors who'd prefer a stabler ride without giving up the excellent returns available in equities.)
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

CAPE has repeatedly been shown to be a very poor market timing indicator. Simple moving averages, like the 200 DMA, have proven themselves to be very effective for the last 50 years at least. See Feldman et al. (2015, Journal of Portfolio Management) for a detailed analysis of this.

For me, the appeal of market timing mechanisms such as moving average or momentum rotation is not in their ability to yield higher returns than buy-and-hold: it's in their ability to reduce downside risk. Even Larry Swedroe has admitted that he suspects that if rigidly applied, these methods may provide a better risk/return ratio than buy-and-hold. For the accumulating investor, the primary benefit of this could be that they could not have to endure watching their stock position drop 50% in an event like 2008. But I suspect that the distribution phase is where market timing methods could be the most beneficial. It's now well known that safe withdrawal rates are primarily as low as they are due to two factors: (1) long periods of lackluster returns and (2) significant drops in portfolio value early in the withdrawal phase. Market timing can very likely help to minimize the impact of #2, and some of the methods could potentially with #1 as well.

Analyses I've done using Portfolio Visualizer indicates that these methods could have helped retirees significantly boost their withdrawal rates, accumulate a larger portfolio over time, or both. I won't say how much it could have helped because people would really think I was off my rocker. Whether this will work in the future in a similar way is anyone's guess, like all aspects of investing.

Enter the 'that's only backtesting' crowd, to whom I'll remind that the "4% rule" is only the result of fairly extensive backtesting across time and geography.
The Sensible Steward
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

willthrill81 wrote: Sun Dec 10, 2017 12:46 pm CAPE has repeatedly been shown to be a very poor market timing indicator. Simple moving averages, like the 200 DMA, have proven themselves to be very effective for the last 50 years at least. See Feldman et al. (2015, Journal of Portfolio Management) for a detailed analysis of this.

For me, the appeal of market timing mechanisms such as moving average or momentum rotation is not in their ability to yield higher returns than buy-and-hold: it's in their ability to reduce downside risk. Even Larry Swedroe has admitted that he suspects that if rigidly applied, these methods may provide a better risk/return ratio than buy-and-hold. For the accumulating investor, the primary benefit of this could be that they could not have to endure watching their stock position drop 50% in an event like 2008. But I suspect that the distribution phase is where market timing methods could be the most beneficial. It's now well known that safe withdrawal rates are primarily as low as they are due to two factors: (1) long periods of lackluster returns and (2) significant drops in portfolio value early in the withdrawal phase. Market timing can very likely help to minimize the impact of #2, and some of the methods could potentially with #1 as well.

Analyses I've done using Portfolio Visualizer indicates that these methods could have helped retirees significantly boost their withdrawal rates, accumulate a larger portfolio over time, or both. I won't say how much it could have helped because people would really think I was off my rocker. Whether this will work in the future in a similar way is anyone's guess, like all aspects of investing.

Enter the 'that's only backtesting' crowd, to whom I'll remind that the "4% rule" is only the result of fairly extensive backtesting across time and geography.
Thanks for the reference on CAPE. I'm not necessarily wanting to discuss CAPE...there are lots of ways to guess at where the market's going, and I'm not married to CAPE.

I especially like that you said "the appeal of market timing mechanisms such as moving average or momentum rotation is not in their ability to yield higher returns than buy-and-hold: it's in their ability to reduce downside risk." That's what I think is beneficial, too.

And I know the backtesting crowd is coming. I was so scared to post this! It's honestly intimidating, because I'm inviting all the backtesting nerds AND market timing criticism at the same time. But I have to remember that no one knows what's happening next, ever! I just backtested because that's all I can do to explain why I think that this can "reduce downside risk" as you said.
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Automated Market Timing

Post by livesoft »

The only way you will know is if you try it out and keep extremely good records comparing your results to a benchmark. If you do this, then you have to decide how much of your portfolio to commit to this. I suggest that you do not commit 100% at the start.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

Veiled wrote: Sun Dec 10, 2017 12:52 pm
I especially like that you said "the appeal of market timing mechanisms such as moving average or momentum rotation is not in their ability to yield higher returns than buy-and-hold: it's in their ability to reduce downside risk." That's what I think is beneficial, too.

And I know the backtesting crowd is coming. I was so scared to post this! It's honestly intimidating, because I'm inviting all the backtesting nerds AND market timing criticism at the same time. But I have to remember that no one knows what's happening next, ever! I just backtested because that's all I can do to explain why I think that this can "reduce downside risk" as you said.
Using Portfolio Visualizer, a 7 month timing period (approximately equal to 200 DMA), investing in VTSMX and using VBMFX as the "out of market asset," since 1993, the timing portfolio would have had a worst year of -5.89%, compared to a worst year of -37.04% for the buy-and-hold portfolio. The Sharpe ratio of the timing portfolio was .91 compared to .54 for the buy-and-hold portfolio.

The historical ability of such timing strategies to reduce downside risk over the long-term (e.g. across the entire cycle) is very compelling. It's literally 'baked in' to the strategy.

That being said, one of the biggest risks of these approaches is that those rigidly implementing them are very like to underperform a buy-and-hold strategy when the market is doing well, perhaps substantially. From 2009 until today, buy-and-hold has trounced all of the major timing methods in terms of returns. But when you include bear markets such as 2000-2002 and 2008 in the analysis (e.g. 2000-2017), most of the major timing approaches actually outperform buy-and-hold in terms of annualized returns.

Would you be able to handle a decade of underperforming buy-and-hold? This is a personal question that is not entirely knowable. But I think that for many investors, the risk of underperforming the market for a lengthy period might be easier than watching their portfolio drop 30% or 40% in a bear market. Further, the more accurate comparison in terms of risk is not 100% stock, buy-and-hold, but perhaps something like 50/50 in stocks and bonds that are bought and held.

Intestinal fortitude is required to consistently apply any strategy.
The Sensible Steward
Soon2BXProgrammer
Posts: 3289
Joined: Mon Nov 24, 2014 10:30 pm

Re: Automated Market Timing

Post by Soon2BXProgrammer »

it would be interesting to see if back-testing over the last 100 years.. you come out ahead on

EMA 13 vs 34 Month cross overs.. or MACD(12,26,9)....
http://stockcharts.com/public/1499522/c ... /325419138;

Note, not that i time the market doing this, but people do.
Earned 43 (and counting) credit hours of financial planning related education from a regionally accredited university, but I am not your advisor.
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

Veiled wrote: Sun Dec 10, 2017 12:34 pm In short, my question is: could automated tactical adjustment be useful? I'd get a stabler portfolio performance, I wouldn't screw up market timing by forgetfulness, fear, or greed, and I don't have to spend a lot of time thinking about it. I understand this is market timing, but would market timing be any more acceptable if it was planned, desired according to risk tolerance, and automated?
I think the question is whether you truly believe that CAPE or any other indicator can tell you whether the market is "overvalued" or dangerous.

If your program wrongly indicates that currently is a very good time to invest (e.g. CAPE is very low), chooses to go 100% equities and then the market crashes 50% then you won't have a very stable ride.

Or otherwise perhaps the market will remain "overvalued" for the next 30 years but no major crash ever happens and the program decides to keep you in bonds all of the time. You will have a very stable ride but no growth.

If the market is like a random walk with minimal mean-reverting then a fixed percentage allocation like 60/40 exposes you to less risk than a formula that varies e.g. your allocation between 40/60 and 80/20. Of course you can make the formula more conservative but you can do the same for the fixed percentage allocation.

Do you really have what it takes to determine which of the hundreds of market valuation metrics to pick? Remember they may be based on data-mining. How will you determine it is producing incorrect results and you need to adjust its parameters or pick another formula?

Assuming that the market is a random, unpredictable walk is basically the safest assumption you can make. In investing terms you only have a very limited amount of time until your retirement. Do you really want to gamble it on picking the right formula?
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

Soon2BXProgrammer wrote: Sun Dec 10, 2017 1:24 pm it would be interesting to see if back-testing over the last 100 years.. you come out ahead on

EMA 13 vs 34 Month cross overs.. or MACD(12,26,9)....
http://stockcharts.com/public/1499522/c ... /325419138;

Note, not that i time the market doing this, but people do.
Wow, exploring your link led me down quite the rabbit hole. Had no idea that there was so much data and so much ink spent in public on market timing. Are there private investors like you and I doing this? I always thought it was professionals messing with this stuff and that it would likely be proprietary. Anyway, I forayed into the brave new world enough to realize I don't know how to backtest EMA13 and EMA34 crossovers.

In a way, it doesn't matter that much. I'm not as interested in which market timing tool is the most accurate. I'm interested in whether BH would ever consider automatically adjusting their AA.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

smesman wrote: Sun Dec 10, 2017 1:38 pmDo you really have what it takes to determine which of the hundreds of market valuation metrics to pick? Remember they may be based on data-mining. How will you determine it is producing incorrect results and you need to adjust its parameters or pick another formula?
Myself? Definitely, definitely not. Never in my wildest daydreams would I consider myself able to create such a program. (Post right above proves I can't even figure out how to backtest without a googleable calculator.) But if a smart team at a respected brokerage added this feature to its robo-advisor, perhaps considering multiple ways of assessing valuation, I would think it worth considering.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

smesman wrote: Sun Dec 10, 2017 1:38 pm
Veiled wrote: Sun Dec 10, 2017 12:34 pm In short, my question is: could automated tactical adjustment be useful? I'd get a stabler portfolio performance, I wouldn't screw up market timing by forgetfulness, fear, or greed, and I don't have to spend a lot of time thinking about it. I understand this is market timing, but would market timing be any more acceptable if it was planned, desired according to risk tolerance, and automated?
I think the question is whether you truly believe that CAPE or any other indicator can tell you whether the market is "overvalued" or dangerous.

If your program wrongly indicates that currently is a very good time to invest (e.g. CAPE is very low), chooses to go 100% equities and then the market crashes 50% then you won't have a very stable ride.
CAPE has already been shown to be a very poor market timing indicator. See Feldman et al. (2015, Journal of Portfolio Management) for a detailed analysis of this.
smesman wrote: Sun Dec 10, 2017 1:38 pmIf the market is like a random walk with minimal mean-reverting then a fixed percentage allocation like 60/40 exposes you to less risk than a formula that varies e.g. your allocation between 40/60 and 80/20. Of course you can make the formula more conservative but you can do the same for the fixed percentage allocation.
smesman wrote: Sun Dec 10, 2017 1:38 pmAssuming that the market is a random, unpredictable walk is basically the safest assumption you can make. In investing terms you only have a very limited amount of time until your retirement. Do you really want to gamble it on picking the right formula?
The 'random walk' theory of stock returns has been repeatedly shown to be false for at least the last 20 years.
smesman wrote: Sun Dec 10, 2017 1:38 pmAssuming that the market is a random, unpredictable walk is basically the safest assumption you can make. In investing terms you only have a very limited amount of time until your retirement. Do you really want to gamble it on picking the right formula? Do you really have what it takes to determine which of the hundreds of market valuation metrics to pick? Remember they may be based on data-mining. How will you determine it is producing incorrect results and you need to adjust its parameters or pick another formula?
No one knows what market timing indicator will be the 'best' in terms of anything (e.g. returns, Sharpe ratio, best 'worst' year, max. drawdown), but simple indicators like the 200 DMA have been successfully used by investors for at least the last 50 years. This is not just 'backtesting' and 'mining the historic data until you find something that works'; asset classes have a strong tendency to perform in the short-term future as they have in the short-term past (i.e. the momentum effect).

Market timing has historically achieved the OP's stated goal: minimizing downside risk. For all it has going for it, buy-and-hold has repeatedly shown itself to expose investors to high levels of downside risk (e.g. 2008).

I'm not saying that market timing is 'magic' or that it will even work in the future as it has in the past. I'm also not a naysayer of buy-and-hold; it's clearly superior to what many investors do in practice. All I'm saying is that there are other, plausible investment strategies than buy-and-hold that might better meet certain investors' goals.
Last edited by willthrill81 on Sun Dec 10, 2017 2:01 pm, edited 1 time in total.
The Sensible Steward
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

Veiled wrote: Sun Dec 10, 2017 1:52 pm Myself? Definitely, definitely not. Never in my wildest daydreams would I consider myself able to create such a program. (Post right above proves I can't even figure out how to backtest without a googleable calculator.) But if a smart team at a respected brokerage added this feature to its robo-advisor, perhaps considering multiple ways of assessing valuation, I would think it worth considering.
Ok, so, how do you pick that smart team/brokerage? On past performance? If they underperform for years/decades will you move your money away?
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

Veiled wrote: Sun Dec 10, 2017 1:52 pm
smesman wrote: Sun Dec 10, 2017 1:38 pmDo you really have what it takes to determine which of the hundreds of market valuation metrics to pick? Remember they may be based on data-mining. How will you determine it is producing incorrect results and you need to adjust its parameters or pick another formula?
Myself? Definitely, definitely not. Never in my wildest daydreams would I consider myself able to create such a program. (Post right above proves I can't even figure out how to backtest without a googleable calculator.) But if a smart team at a respected brokerage added this feature to its robo-advisor, perhaps considering multiple ways of assessing valuation, I would think it worth considering.
Historically, there are a number of market timing strategies that are very easy to implement. One investment 'guru' back in the 1960s (not sure of the exact period) was advocating the 200 DMA approach for individual investors buying individuals stocks. He advocated that they use the closing price of the last day in the month and to approximately calculate the 200 DMA and buy/sell accordingly. They did this by hand, and it only took a few minutes per month per stock. Today, we have free tracking software that will do that for us. Portfolio Visualizer is a free one that will actually do this for you (except the actual trading of course).
The Sensible Steward
Soon2BXProgrammer
Posts: 3289
Joined: Mon Nov 24, 2014 10:30 pm

Re: Automated Market Timing

Post by Soon2BXProgrammer »

Veiled wrote: Sun Dec 10, 2017 1:48 pm I always thought it was professionals messing with this stuff
Just because they get paid, doesn't mean they know what they are doing... also.. if they are strategic about it.. they would gamble differently with different sets of clients.. expect to lose some, and win big on some.. either way.. they take their cut.. and get paid for gambling with their clients money...
Earned 43 (and counting) credit hours of financial planning related education from a regionally accredited university, but I am not your advisor.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

smesman wrote: Sun Dec 10, 2017 1:57 pm
Veiled wrote: Sun Dec 10, 2017 1:52 pm Myself? Definitely, definitely not. Never in my wildest daydreams would I consider myself able to create such a program. (Post right above proves I can't even figure out how to backtest without a googleable calculator.) But if a smart team at a respected brokerage added this feature to its robo-advisor, perhaps considering multiple ways of assessing valuation, I would think it worth considering.
Ok, so, how do you pick that smart team/brokerage? On past performance? If they underperform for years/decades will you move your money away?
Oh, I see your point. I'm kicking the can down the road, but I'm still choosing something/someone and introducing extra risk on top of the market risk, just like picking single stocks or an active manager. (Did I get it?)
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

Soon2BXProgrammer wrote: Sun Dec 10, 2017 2:02 pm
Veiled wrote: Sun Dec 10, 2017 1:48 pmI always thought it was professionals messing with this stuff
Just because they get paid, doesn't mean they know what they are doing
Agreed, that's not a new idea to me.
Soon2BXProgrammer wrote: Sun Dec 10, 2017 2:02 pm... also.. if they are strategic about it.. they would gamble differently with different sets of clients.. expect to lose some, and win big on some.. either way.. they take their cut.. and get paid for gambling with their clients money...
:shock: That, on the other hand, is a totally new idea to me. Dang, I forget that not everyone has fiduciary duties.
Soon2BXProgrammer
Posts: 3289
Joined: Mon Nov 24, 2014 10:30 pm

Re: Automated Market Timing

Post by Soon2BXProgrammer »

Veiled wrote: Sun Dec 10, 2017 2:07 pm
Soon2BXProgrammer wrote: Sun Dec 10, 2017 2:02 pm... also.. if they are strategic about it.. they would gamble differently with different sets of clients.. expect to lose some, and win big on some.. either way.. they take their cut.. and get paid for gambling with their clients money...
:shock: That, on the other hand, is a totally new idea to me. Dang, I forget that not everyone has fiduciary duties.
Example. a FA talking to one client and the client expresses a concern about being near the peak.. might be put in:
http://www.leutholdfunds.com/individual ... tail/grzzx

and the next that thinks this bull has legs.. might be put in:
http://www.leutholdfunds.com/individual ... tail/lginx

both are big bets.. based on the what their client has asked for... if their client was right, client wins and they win... if the client was wrong... then the FA defends the position with that's what they asked for... (not that i like either of these funds.. but this could be something that is "picked" that a rational investor might not find/select themselves.. that is "secret sauce")
Earned 43 (and counting) credit hours of financial planning related education from a regionally accredited university, but I am not your advisor.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

Soon2BXProgrammer wrote: Sun Dec 10, 2017 2:15 pmExample. a FA talking to one client and the client expresses a concern about being near the peak.. might be put in:
http://www.leutholdfunds.com/individual ... tail/grzzx

and the next that thinks this bull has legs.. might be put in:
http://www.leutholdfunds.com/individual ... tail/lginx

both are big bets.. based on the what their client has asked for... if their client was right, client wins and they win... if the client was wrong... then the FA defends the position with that's what they asked for... (not that i like either of these funds.. but this could be something that is "picked" that a rational investor might not find/select themselves.. that is "secret sauce")
The impunity is what gets me. And OBs get sued if there's a shoulder dystocia without permanent deficits. :annoyed </off-topic>
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

Veiled wrote: Sun Dec 10, 2017 2:04 pm Oh, I see your point. I'm kicking the can down the road, but I'm still choosing something/someone and introducing extra risk on top of the market risk, just like picking single stocks or an active manager. (Did I get it?)
Yes, you've got it exactly!

The "boglehead" way will never be the optimal strategy in hindsight, but (assuming the market trends upwards) is guaranteed to always do okay (beat 50% of the market before, 80% of the market after costs). The average person who chooses a different strategy (e.g. timing or sector/stock picking) ends up losing (often considerably after costs). Mathematically speaking it has to be so. The only reason to go active is if you have specific knowledge or skills superior to the average investor.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

Thanks to everyone for the replies. Definitely does seem like a balanced portfolio like the 3-fund avoids excessive risk from any one asset class and avoids introducing unique risk from human error.
victw
Posts: 190
Joined: Sat Feb 20, 2016 3:07 pm

Re: Automated Market Timing

Post by victw »

Veiled wrote: Sun Dec 10, 2017 6:53 pm Thanks to everyone for the replies. Definitely does seem like a balanced portfolio like the 3-fund avoids excessive risk from any one asset class and avoids introducing unique risk from human error.
Veiled,

I think you have capitulated too soon.
Are you reading any of the behavioral economics books that are recommended on the forum? I try keep myself sober by reading them.

I do think there are additional strategies that can be employed.
Value average vs dollar cost averaging. I wish I had known about this much earlier - but even knowing about it later will help in the de-accumulation phase.

And possibly a dynamic asset allocation. McClung's backtesting on rebalancing has planted some kind of seed in my mind - but I haven't figured out how I would execute it.

livesoft employs some kind of system that I have yet to see a top/down write up on - he leaves cryptic clues.

I'm over the individual stock picking of my 30's - but I do think there are strategies that can be implemented beyond buy and hold. I do think your focus on the human error problem is key.

Vic
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

victw wrote: Mon Dec 11, 2017 12:10 pm
Veiled wrote: Sun Dec 10, 2017 6:53 pm Thanks to everyone for the replies. Definitely does seem like a balanced portfolio like the 3-fund avoids excessive risk from any one asset class and avoids introducing unique risk from human error.
Veiled,

I think you have capitulated too soon.
Are you reading any of the behavioral economics books that are recommended on the forum? I try keep myself sober by reading them.

I do think there are additional strategies that can be employed.
Value average vs dollar cost averaging. I wish I had known about this much earlier - but even knowing about it later will help in the de-accumulation phase.

And possibly a dynamic asset allocation. McClung's backtesting on rebalancing has planted some kind of seed in my mind - but I haven't figured out how I would execute it.

livesoft employs some kind of system that I have yet to see a top/down write up on - he leaves cryptic clues.

I'm over the individual stock picking of my 30's - but I do think there are strategies that can be implemented beyond buy and hold. I do think your focus on the human error problem is key.

Vic
I agree. Bogleheads seem to be anti-market timing because (1) it is certainly true that no one can predict the future and (2) most investors who use market timing do so subjectively, and that has a poor track record compared to buy-and-hold. While we cannot predict the future precisely, there are a great many areas of reality where we can make predictions that are far better than blind chance. The example I like to provide is of a coin that is weighted to come up heads 55% of the time. If you could bet $1 that it would come up heads and win $2 if it did, then you should play this game as much as you possibly can. You should know going into it that there will be many periods where tails comes up repeatedly, causing you to lose money. But that doesn't matter because you know that if you play enough times, you'll come out ahead. We cannot predict precisely whether heads or tails will come up, but the expected value is weighted in our favor. The same can be true of objective, 'automated' market timing. It's not voodoo or just data mining; it's based on the well known and documented momentum effect.
The Sensible Steward
User avatar
whodidntante
Posts: 13090
Joined: Thu Jan 21, 2016 10:11 pm
Location: outside the echo chamber

Re: Automated Market Timing

Post by whodidntante »

There is a genuine risk to being different than the market. When your annoying neighbor is bragging about how great the market is doing, and you are sitting on a loss, there aren't a lot of people who can stomach that for very long. And that's where a lot expected return models fail to be useful in practice.

If you think you're different and can benefit from tactical allocation, go ahead. The majority on this site will only criticize your efforts. That will be the least of your problems in successfully implementing tactical allocation. You can use their valid points to stress test your ideas.

I do tactical allocation myself, but in the form of overweighting assets with higher expected returns.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

whodidntante wrote: Mon Dec 11, 2017 12:54 pm There is a genuine risk to being different than the market. When your annoying neighbor is bragging about how great the market is doing, and you are sitting on a loss, there aren't a lot of people who can stomach that for very long. And that's where a lot expected return models fail to be useful in practice.
I agree, but that sword cuts both ways. In a bear market, it can be difficult for a buy-and-hold investor see their portfolio drop by 40% while their neighbor using some form of market timing only experienced a 10% drop or, even 'worse', is still in the black. I have no doubt that many investors who 'thought' they were buy-and-hold sold off in 2008.

Any investment strategy will underperform other strategies at points in time, and this underperformance can last for years. Sticking to any strategy requires intestinal fortitude.
Last edited by willthrill81 on Mon Dec 11, 2017 3:04 pm, edited 1 time in total.
The Sensible Steward
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

If we had a crystal ball and could check out which algorithm with what parameters would be doing best we could just pick that and be set!

However, in the real world we only have historical data to rely on. So we can use that to pick one of the infinite number of trading strategies possible (hoping it wasn't data-mined or already priced into the market) and make it go up against the all of the other strategies used by millions of investors around the world (that are using same data as you to decide to SELL when you BUY) and hope it does better than 80% of them (after costs). Or you can just index invest and invest in all possible strategies at once.

Most people end up underperforming the index. Why risk that? Your investment horizon is relatively short, you only have one try.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

smesman wrote: Mon Dec 11, 2017 1:53 pm If we had a crystal ball and could check out which algorithm with what parameters would be doing best we could just pick that and be set!
"The enemy of good is perfect."

No a priori strategy will always work perfectly. That doesn't mean that long used indicators like the 200 DMA haven't worked and worked well.
smesman wrote: Mon Dec 11, 2017 1:53 pmHowever, in the real world we only have historical data to rely on. So we can use that to pick one of the infinite number of trading strategies possible (hoping it wasn't data-mined or already priced into the market) and make it go up against the all of the other strategies used by millions of investors around the world (that are using same data as you to decide to SELL when you BUY) and hope it does better than 80% of them (after costs). Or you can just index invest and invest in all possible strategies at once.
You are equating index investing with buy-and-hold. They are completely distinct.
smesman wrote: Mon Dec 11, 2017 1:53 pmMost people end up underperforming the index. Why risk that? Your investment horizon is relatively short, you only have one try.
Your logic cuts both ways.

What if the last decade of your accumulation phase turned out to be a repeat of 2000-2009, when VTSMX had an annualized return of -.27%? Why risk that? By comparison, had you used the above 200 DMA strategy with TSM and TBM, you would have had an annualized return of 8.61%.
The Sensible Steward
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

willthrill81 wrote: Mon Dec 11, 2017 3:02 pm Your logic cuts both ways. What if the last decade of your accumulation phase turned out to be a repeat of 2000-2009, when VTSMX had an annualized return of -.27%? Why risk that? By comparison, had you used the above 200 DMA strategy with TSM and TBM, you would have had an annualized return of 8.61%.
Sure, but for every person who got a 8.61% return there was another person who got a -8.87% return (approximately, for the sake of argument). For every investment making more than the market return there is another investment making less. Index investing will always be square in the middle. Why gamble you pick the right method? Everyone thinks they are smarter and better than average.
willthrill81 wrote: Mon Dec 11, 2017 3:02 pm You are equating index investing with buy-and-hold. They are completely distinct.
There are 2 ways of momentum investing: time-series or cross-sectional. With cross-sectional you use an algorithm to decide which stocks to pick and with time-series you choose which time periods to pick. In the boglehead method, instead you pick all stocks for all time periods. Again this will always be in the center of all investment results (before costs) because for every time period or stock you decide to sell, another investor is deciding to buy and vice versa.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

smesman wrote: Mon Dec 11, 2017 3:24 pm
willthrill81 wrote: Mon Dec 11, 2017 3:02 pm Your logic cuts both ways. What if the last decade of your accumulation phase turned out to be a repeat of 2000-2009, when VTSMX had an annualized return of -.27%? Why risk that? By comparison, had you used the above 200 DMA strategy with TSM and TBM, you would have had an annualized return of 8.61%.
Sure, but for every person who got a 8.61% return there was another person who got a -8.87% return (approximately, for the sake of argument). For every investment making more than the market return there is another investment making less. Index investing will always be square in the middle. Why gamble you pick the right method? Everyone thinks they are smarter and better than average.
I'm not referring to every potential strategy that exists. There are only a few 'major' market timing strategies out there (e.g. moving averages, relative strength momentum, dual momentum, adaptive allocation).

I'm not concerned about the investors who, for instance, thought they were 'buy-and-hold' investors and yet sold in early 2009 and never reentered the market. Such investors likely experienced a much lower annualized return than what you specify.
smesman wrote: Mon Dec 11, 2017 3:24 pm
willthrill81 wrote: Mon Dec 11, 2017 3:02 pm You are equating index investing with buy-and-hold. They are completely distinct.
There are 2 ways of momentum investing: time-series or cross-sectional. With cross-sectional you use an algorithm to decide which stocks to pick and with time-series you choose which time periods to pick. In the boglehead method, instead you pick all stocks for all time periods. Again this will always be in the center of all investment results (before costs) because for every time period or stock you decide to sell, another investor is deciding to buy and vice versa.
You are assuming that all investment strategies are 'created equal' (i.e. they will all eventually revert to the mean return of the market). But it is quite possible that a number of strategies exist which can outperform the median returns of the market (e.g. buy-and-hold) over the long-run. Conversely, it is well known that many strategies exist which are sub-optimal to buy-and-hold, and Bogleheads are quick to acknowledge this.
The Sensible Steward
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

willthrill81 wrote: Mon Dec 11, 2017 4:00 pm I'm not concerned about the investors who, for instance, thought they were 'buy-and-hold' investors and yet sold in early 2009 and never reentered the market. Such investors likely experienced a much lower annualized return than what you specify.

...

Conversely, it is well known that many strategies exist which are sub-optimal to buy-and-hold, and Bogleheads are quick to acknowledge this.
Most bogleheads would say that timing the market is always bad. By getting in and out of the market you are losing your diversification across time (time in the market) and yes you may get lucky and miss a crash or get unlucky and miss a recovery/boom.
willthrill81 wrote: Mon Dec 11, 2017 4:00 pm You are assuming that all investment strategies are 'created equal' (i.e. they will all eventually revert to the mean return of the market). But it is quite possible that a number of strategies exist which can outperform the median returns of the market (e.g. buy-and-hold) over the long-run.
I'm just assuming that all people are in the stock market to make money and for every trade there is a buyer and a seller.

It might be true that there is a significant number of people trading using an easily exploitable strategy (e.g. all trading on an instinct based 300 day moving average which gets beaten by a 200 day moving average). But if that is consistently and demonstrably true why wouldn't it already be priced in by the market? In fact, there are a ton of momentum funds and hedge funds doing such strategies. How do you know you won't be beaten by 190 day moving average trading strategy the next 20 years? Will you adjust based on performance or what?

By betting on this strategy you're opening yourself up to so many risks. You're losing diversification (across stocks, exposure and/or time), you're getting higher costs, you're opening yourself up to front running and investor risk. Why risk it? Why not work harder, longer or save a few % more instead of trying to squeek out a few more % by obsessing over your stocks.
larryslocum1982
Posts: 45
Joined: Sat Dec 09, 2017 9:36 pm

Re: Automated Market Timing

Post by larryslocum1982 »

I have been investing for 40 years

CAPE or other valuation measurs are not good ways to time the market.
You will lose your shirt timing the market with moving averases.

I have spent a lot of time testing publicly available timing methods. They do not work. Perhaps that is why they are public.

I have repeatedly seen several talented individuals very good at market timing. None of them discloses enough details to make these fo it yourself. They sell their signals for a hefty price.
Some of them mean reverse after a few years. I know only a couple of individuals who have not mean reversed.

Market timing is not something an average investor should do on his own.If you know of someone who is an expert at market timing over a long time, then it is worth seeking their assistane. Otherwise stick with index funds like the three fund portfolio.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

smesman wrote: Mon Dec 11, 2017 4:50 pm
willthrill81 wrote: Mon Dec 11, 2017 4:00 pm I'm not concerned about the investors who, for instance, thought they were 'buy-and-hold' investors and yet sold in early 2009 and never reentered the market. Such investors likely experienced a much lower annualized return than what you specify.

...

Conversely, it is well known that many strategies exist which are sub-optimal to buy-and-hold, and Bogleheads are quick to acknowledge this.
Most bogleheads would say that timing the market is always bad. By getting in and out of the market you are losing your diversification across time (time in the market) and yes you may get lucky and miss a crash or get unlucky and miss a recovery/boom.
Losing "diversification across time" can be a very good thing when the market is in a downward trend.

There is certainly an element of randomness involved in the performance of any strategy, but the ability to profitably capitalize on human behavior is not entirely based on luck.
smesman wrote: Mon Dec 11, 2017 4:50 pm
willthrill81 wrote: Mon Dec 11, 2017 4:00 pm You are assuming that all investment strategies are 'created equal' (i.e. they will all eventually revert to the mean return of the market). But it is quite possible that a number of strategies exist which can outperform the median returns of the market (e.g. buy-and-hold) over the long-run.
I'm just assuming that all people are in the stock market to make money and for every trade there is a buyer and a seller.
The number of buyers and sellers, relative to each other, is constantly shifting due to changes in prices. In some situations, there are more sellers than buyers, causing prices to fall, sometimes dramatically.
smesman wrote: Mon Dec 11, 2017 4:50 pmIt might be true that there is a significant number of people trading using an easily exploitable strategy (e.g. all trading on an instinct based 300 day moving average which gets beaten by a 200 day moving average). But if that is consistently and demonstrably true why wouldn't it already be priced in by the market? In fact, there are a ton of momentum funds and hedge funds doing such strategies. How do you know you won't be beaten by 190 day moving average trading strategy the next 20 years? Will you adjust based on performance or what?
Remember, "the enemy of good is perfect."

I don't know a prior what the optimal timing period will be going forward. But extensive historical data, combined with behavioral finance, indicates that a 200 DMA or something akin to it is at least extremely likely to have a better return/risk (i.e. Sharpe) ratio than buy-and-hold going forward and also very likely to have higher returns. That is not my own 'private' conclusion; it's supported by academic research as well that I've already cited.
smesman wrote: Mon Dec 11, 2017 4:50 pmBy betting on this strategy you're opening yourself up to so many risks. You're losing diversification (across stocks, exposure and/or time), you're getting higher costs, you're opening yourself up to front running and investor risk. Why risk it? Why not work harder, longer or save a few % more instead of trying to squeek out a few more % by obsessing over your stocks.
It's true that there are new risks opened by market timing, but there has been a tremendous reduction in other risks in the historic record, as I've already posted. To what extent this will be the case going forward is anyone's guess, but I put more weight in the historic record than do many here.

A "few more percent" over a number of year can easily be huge and certainly could be for me. But as I've said repeatedly, I see the real value of market timing to be its ability to reduce downside risk rather than to increase overall returns. If I can prevent a 50% decline in my portfolio and achieve the same returns as buy-and-hold with minimal additional effort (literally a few minutes per month with free software like Portfolio Visualizer, no "obsessing" needed :)), I'm perfectly fine with that. Obviously, many are not okay with it, and I'm fine with that too. If everyone were implementing the same strategy, really bad things could happen.
Last edited by willthrill81 on Mon Dec 11, 2017 7:27 pm, edited 1 time in total.
The Sensible Steward
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

victw wrote: Mon Dec 11, 2017 12:10 pm
Veiled wrote: Sun Dec 10, 2017 6:53 pm Thanks to everyone for the replies. Definitely does seem like a balanced portfolio like the 3-fund avoids excessive risk from any one asset class and avoids introducing unique risk from human error.
I think you have capitulated too soon.
You're right, I'm still thinking about this; however, I realize that any attempt to modify behavior based on a guess at the future is risky. The index investor never guesses at the future.

But I still wonder.... Once upon a time buying the whole market was not something the average investor could do but an innovation (the index fund) changed that. Currently, active management doesn't usually outperform index funds. But could an innovation change that, too?
smesman wrote: Mon Dec 11, 2017 4:50 pmMost bogleheads would say that timing the market is always bad. By getting in and out of the market you are losing your diversification across time (time in the market) and yes you may get lucky and miss a crash or get unlucky and miss a recovery/boom. ... By betting on this strategy you're opening yourself up to so many risks. You're losing diversification (across stocks, exposure and/or time), you're getting higher costs, you're opening yourself up to front running and investor risk. Why risk it? Why not work harder, longer or save a few % more instead of trying to squeek out a few more % by obsessing over your stocks.
1. I proposed a way to stay in the market 100% of the time (not the stock market, but in stock/bonds)
2. My proposal doesn't lose diversification permanently and arguably maintains some measure of diversification since the investor would re-allocate into total bond (or REITs or some other instrument) where there is internal diversity
3. I ask about automation precisely because I don't want to obsess.

I also like willthrill's point about the logic of returns cutting both ways. If I'm (hypothetically) happy to watch my wallet grow less for one or two decades in exchange for gentler dips, couldn't this be useful? I argue that new investors are more likely to know whether they can be patient while being exceeded (from various experiences in life), than they are to know how strong their steel is in the face of a bear market.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

willthrill81 wrote: Mon Dec 11, 2017 7:14 pmIf I can prevent a 50% decline in my portfolio and achieve the same returns as buy-and-hold with minimal additional effort (literally a few minutes per month with free software like Portfolio Visualizer, no "obsessing" needed :)), I'm perfectly fine with that. Obviously, many are not okay with it, and I'm find with that too. If everyone were implementing the same strategy, really bad things could happen.
Beat me to explaining what I had in mind for no obsessing. What can I say? I'm a millenial, I want a hassle-free UI. :wink:
pkcrafter
Posts: 15461
Joined: Sun Mar 04, 2007 11:19 am
Location: CA
Contact:

Re: Automated Market Timing

Post by pkcrafter »

larryslocum1982 wrote: Mon Dec 11, 2017 7:00 pm I have been investing for 40 years

CAPE or other valuation measurs are not good ways to time the market.
You will lose your shirt timing the market with moving averases.

I have spent a lot of time testing publicly available timing methods. They do not work. Perhaps that is why they are public.

I have repeatedly seen several talented individuals very good at market timing. None of them discloses enough details to make these fo it yourself. They sell their signals for a hefty price.
Some of them mean reverse after a few years. I know only a couple of individuals who have not mean reversed.

Market timing is not something an average investor should do on his own.If you know of someone who is an expert at market timing over a long time, then it is worth seeking their assistane. Otherwise stick with index funds like the three fund portfolio.
Hey Larry, you just might have the makings to be a true Boglehead! :thumbsup

Paul
When times are good, investors tend to forget about risk and focus on opportunity. When times are bad, investors tend to forget about opportunity and focus on risk.
larryslocum1982
Posts: 45
Joined: Sat Dec 09, 2017 9:36 pm

Re: Automated Market Timing

Post by larryslocum1982 »

I like being a boglehead.
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

Veiled wrote: Mon Dec 11, 2017 7:19 pm But I still wonder.... Once upon a time buying the whole market was not something the average investor could do but an innovation (the index fund) changed that. Currently, active management doesn't usually outperform index funds. But could an innovation change that, too?
It is impossible for more than half of active managers to beat the index. I think the only exception is some funds that are managed in a very tax efficient way (e.g. you could get a tax benefit by only investing in "green" companies as a European investor) and are still low cost.
Veiled wrote: Mon Dec 11, 2017 7:19 pm 1. I proposed a way to stay in the market 100% of the time (not the stock market, but in stock/bonds)
2. My proposal doesn't lose diversification permanently and arguably maintains some measure of diversification since the investor would re-allocate into total bond (or REITs or some other instrument) where there is internal diversity
I don't see how keeping a small % in the stock market would make a difference. If you're invested 80% when the stock market crashes and 20% when it recovers you're still going to have lost a bunch of money.
Veiled wrote: Mon Dec 11, 2017 7:19 pm I also like willthrill's point about the logic of returns cutting both ways. If I'm (hypothetically) happy to watch my wallet grow less for one or two decades in exchange for gentler dips, couldn't this be useful? I argue that new investors are more likely to know whether they can be patient while being exceeded (from various experiences in life), than they are to know how strong their steel is in the face of a bear market.
I encourage you to go to portfoliovisualizer.com and play with a couple of timing methods (make sure to change their indexes, parameters and time periods to get a good sense of how they work).

You will see that sometimes they work, sometimes they don't. E.g. the moving average strategy works for slow trends, but if the market crashes very quickly it will only be in time to lock in your losses, then if it recovers quickly you will miss the recovery. If you change the numbers of days it is averaging over it will always work at some point for historical data.

I think for almost any strategy you can make an argument why the counter strategy should work. Like above during/after a crash some people will be selling (because of negative momentum) and others buying (stocks are underpriced! buying opportunity). At some point prices will stabilize when there are enough people in both camps.
H-Town
Posts: 5876
Joined: Sun Feb 26, 2017 1:08 pm

Re: Automated Market Timing

Post by H-Town »

Veiled wrote: Sun Dec 10, 2017 12:34 pm As I read, I began to wonder whether, just as some use roboadvisors to automate tax loss/gain harvesting, eventually we might be able to automate tactical reallocation according to a pre-arranged decision by the investor. An automated process like this would allow an investor to remain hands-off and emotionless and continuously participate in the market, but at the same time avoid unnecessary volatility. Imagine, for instance, that I instruct this theoretical software to shift my AA according to the CAPE ratio given historical data. My AA looks crazy (very antithetical to the typical BH approach), as shown by bh7 in this thread:
In theory, I like the idea of having a trading software that can optimize AA based on certain set parameters. This is ignoring transaction cost and the feasibility of actual apply this in real-time trading. This system will need to remove human completely from the decision making process, not even click to confirm a trade. Even if you have a software, you still have submit trade order on Vanguard/Fidelity website. And it is too easy for human to override the trade order from the software.

For an average Joe, it's better for them to buy and rebalance every quarter.
Time is the ultimate currency.
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

willthrill81 wrote: Mon Dec 11, 2017 7:14 pm Losing "diversification across time" can be a very good thing when the market is in a downward trend.
Unfortunately you don't know whether the downward trend will continue when you already lost a lot of money and are deciding to get out of the market.

There is certainly an element of randomness involved in the performance of any strategy, but the ability to profitably capitalize on human behavior is not entirely based on luck.
willthrill81 wrote: Mon Dec 11, 2017 7:14 pm The number of buyers and sellers, relative to each other, is constantly shifting due to changes in prices. In some situations, there are more sellers than buyers, causing prices to fall, sometimes dramatically.
Like during a correction, exactly when you're proposing to sell because of the negative momentum?
willthrill81 wrote: Mon Dec 11, 2017 7:14 pm I don't know a prior what the optimal timing period will be going forward. But extensive historical data, combined with behavioral finance, indicates that a 200 DMA or something akin to it is at least extremely likely to have a better return/risk (i.e. Sharpe) ratio than buy-and-hold going forward and also very likely to have higher returns. That is not my own 'private' conclusion; it's supported by academic research as well that I've already cited.
The problem is that the market is not like a weather event which you can study and be sure it will remain the same. It consists of people like you who are constantly incorperating all kinds of data and new research to get the optimal trading strategy. How will you know your 200 DMA strategy is not the optimal strategy (anymore)?
willthrill81 wrote: Mon Dec 11, 2017 7:14 pm If I can prevent a 50% decline in my portfolio and achieve the same returns as buy-and-hold with minimal additional effort (literally a few minutes per month with free software like Portfolio Visualizer, no "obsessing" needed :)), I'm perfectly fine with that. Obviously, many are not okay with it, and I'm fine with that too.
There is no way you can prevent a large decline in your portfolio in all stock market conditions except by holding a fixed amount of safe assets.

Of course you can put something like stop loss order at a 10% loss, but after it triggers you have to at some point get into the market again (perhaps after it recovered) and then put in another stop loss order which may again be triggered. Ultimately making you miss most of the recoveries.
smesman
Posts: 97
Joined: Thu Apr 27, 2017 7:11 am

Re: Automated Market Timing

Post by smesman »

Now I'd like to leave you with this thought: 80% of active managers, people who make it their entire job to learn everything there is to know and use powerful trading computers, underperform the market.

Now you guys are intelligent individuals, and you think it is an interesting puzzle to get the return without the risk. But everyone wants that! The only way to get a lot less risk, is to accept a lot less expected return. You have to take the risk if you want growth, up to the risk that you're willing and able to take.

Maybe there are some strategies or managers that you are able to find that improve your returns a little bit without extra risk.

But there are many more things you can invest in. If you use your intelligence and drive to invest in your work and get a better job, then your increased salary will probably give you way more returns for no risk. Or perhaps you can invest it in your personal life and become more frugal or happier without the need of money. If everyone did that I think the world would be a lot more successful than people wasting huge amounts of time and resources trying to take money from each other.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

smesman wrote: Tue Dec 12, 2017 5:07 pm
willthrill81 wrote: Mon Dec 11, 2017 7:14 pm Losing "diversification across time" can be a very good thing when the market is in a downward trend.
Unfortunately you don't know whether the downward trend will continue when you already lost a lot of money and are deciding to get out of the market.
That's true, but time has shown that when the market is below its 200 DMA, its performance has had a strong tendency to be poor, much poorer than average. Rather than phrase this as buy-and-hold versus 200 DMA, I think an equally compelling argument could be why I should hold an easily disposed of asset during periods where the returns have had a strong tendency to be poor.

Often, the buy-and-hold (BAH) crowd espouses "In order to time the market correctly, you have to be right twice: when to sell and when to buy." This is only half true. Timing will come out ahead of BAH as long as you buy at a lower price than when you sold. While no indicator is perfect, strategies like the 200 DMA have shown themselves to outperform BAH over the entire business cycle. But remember, I find the primary value of timing to be a reduction of volatility, not an improvement in returns. This is precisely the view of Paul Merriman as well.
smesman wrote: Tue Dec 12, 2017 5:07 pm
willthrill81 wrote: Mon Dec 11, 2017 7:14 pm The number of buyers and sellers, relative to each other, is constantly shifting due to changes in prices. In some situations, there are more sellers than buyers, causing prices to fall, sometimes dramatically.
Like during a correction, exactly when you're proposing to sell because of the negative momentum?
Right. Poor returns tend to be followed by more poor returns.
smesman wrote: Tue Dec 12, 2017 5:07 pm
willthrill81 wrote: Mon Dec 11, 2017 7:14 pm I don't know a prior what the optimal timing period will be going forward. But extensive historical data, combined with behavioral finance, indicates that a 200 DMA or something akin to it is at least extremely likely to have a better return/risk (i.e. Sharpe) ratio than buy-and-hold going forward and also very likely to have higher returns. That is not my own 'private' conclusion; it's supported by academic research as well that I've already cited.
The problem is that the market is not like a weather event which you can study and be sure it will remain the same. It consists of people like you who are constantly incorperating all kinds of data and new research to get the optimal trading strategy. How will you know your 200 DMA strategy is not the optimal strategy (anymore)?
Again, I don't know what the optimal strategy will be going forward; I don't even know what was truly optimal in the past. But I do know that the 200 DMA strategy (it's not "mine" by the way; it was 'discovered' and utilized long before my birth) has had statistically better returns and lower volatility than BAH for at least the last 50 years. And unlike many on this forum, I place more value on long-term trends such as this.
smesman wrote: Tue Dec 12, 2017 5:07 pm
willthrill81 wrote: Mon Dec 11, 2017 7:14 pm If I can prevent a 50% decline in my portfolio and achieve the same returns as buy-and-hold with minimal additional effort (literally a few minutes per month with free software like Portfolio Visualizer, no "obsessing" needed :)), I'm perfectly fine with that. Obviously, many are not okay with it, and I'm fine with that too.
There is no way you can prevent a large decline in your portfolio in all stock market conditions except by holding a fixed amount of safe assets.
smesman wrote: Tue Dec 12, 2017 5:07 pmOf course you can put something like stop loss order at a 10% loss, but after it triggers you have to at some point get into the market again (perhaps after it recovered) and then put in another stop loss order which may again be triggered. Ultimately making you miss most of the recoveries.
Your last two statements are contradictory because they are mutually exclusive. Stop losses and other sell signals absolutely, by definition, prevent large declines in portfolios. The question then becomes whether one can objectively determine when to buy back into the market at such a time that one would come out at least even with BAH.

Let's examine a historic instance of this. Using a 7 month (roughly equivalent to 200 day) moving average method with VTSMX and VBMFX, an investor would have had an annual return in 2008 of -3.55%, while BAH had -37.04%. In 2009, the timing approach yielded a return of 31.27%, while BAH had a return of 28.70%. So in this instance, timing beat BAH hands down on two counts: its decline was tiny compared to BAH, and it caught most of the recovery.

Interestingly, the market 'bottomed out' in March, 2009, and the timing method signaled to buy back in in May, 2009, which turned out to be remarkably accurate. In fact, the timing method resulted in being out of stocks completely from Dec., 2007, until May, 2009, with the exception of one month, June, 2008.

To be fair, during bull markets, timing seems to underperform BAH in terms of returns, though its minimization of downside risk seems to hold. For instance, from 2009 until now, the above timing method has had an annualized return of 14.66%, while BAH has returned 15.40%. Also, timing resulted in a worse 'worst' year; it returned -3.45% in 2015, while BAH returned .29%. However, the maximum drawdown of timing for this period was -12.10%, while BAH was -17.84%. Still, considering its robust ability to historically reduce significant drawdowns, I'm personally very willing to sacrifice a little return to get that protection.

Again, I don't know if this method will yield returns equal to or greater than BAH going forward, but the past is very clear that it has minimized downside risk over many periods in the past while providing very respectable returns when compared to BAH. And for an investor in the withdrawal phase, I see this as potentially being of great value.

YMMV. :)
The Sensible Steward
H-Town
Posts: 5876
Joined: Sun Feb 26, 2017 1:08 pm

Re: Automated Market Timing

Post by H-Town »

smesman wrote: Tue Dec 12, 2017 5:34 pm Now I'd like to leave you with this thought: 80% of active managers, people who make it their entire job to learn everything there is to know and use powerful trading computers, underperform the market.

Now you guys are intelligent individuals, and you think it is an interesting puzzle to get the return without the risk. But everyone wants that! The only way to get a lot less risk, is to accept a lot less expected return. You have to take the risk if you want growth, up to the risk that you're willing and able to take.

Maybe there are some strategies or managers that you are able to find that improve your returns a little bit without extra risk.
Don't confuse stock picking with optimizing portfolio. This is two completely different things. The goal is to achieve and stay at the efficient frontier to get the highest expected return for a certain given level of risk. This takes away human factor after you set the right parameters with the computer. Go research Modern Portfolio theory if you want. It gives way to the bogleheads' principles and the 3-fund portfolio.
Time is the ultimate currency.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Automated Market Timing

Post by willthrill81 »

smesman wrote: Tue Dec 12, 2017 5:34 pm Now I'd like to leave you with this thought: 80% of active managers, people who make it their entire job to learn everything there is to know and use powerful trading computers, underperform the market.

Now you guys are intelligent individuals, and you think it is an interesting puzzle to get the return without the risk. But everyone wants that! The only way to get a lot less risk, is to accept a lot less expected return. You have to take the risk if you want growth, up to the risk that you're willing and able to take.

Maybe there are some strategies or managers that you are able to find that improve your returns a little bit without extra risk.
Fund managers face three big hurdles that the individual investor doesn't have to deal with. First, most active managers must remain in a particular asset class. A large cap growth fund that is actively managed, for instance, must stay largely invested in large cap growth stocks. They cannot move the majority of their funds into 'safe' assets during periods of negative momentum. Hedge funds are potentially not subject to this restriction, but they still must grapple with the second hurdle.

Second, most retail investors will not tolerate underperformance, in terms of returns, for very long. If a fund significantly underperforms its benchmark for multiple years in sequence, many, usually most, of the investment in that fund heads for the exits. As I've repeatedly said, timing methods have often gone through significant periods where they underperformed buy-and-hold. Most investors are not that patient and will not hold out for a bear market to demonstrate the value of an objective timing method.

Third, many timing strategies are, as one writer put it, "deceptively simple." Despite many decades of data on the topic, most fund managers want something more complicated and subjective. There are numerous reasons for this, one of which being that subjective approaches are manager dependent; if the method is successful, the manager has job security (and often gets paid far more as well). Conversely, if a manager consistently used something as simple as a 200 DMA method that turned out to be successful, they would justly fear that they could be easily replaced by a very simple computer algorithm. Another reason is that fund managers, like many individual investors, just believe that a simple timing approach just can't work due to its parsimony. "It can't be that simple!"
The Sensible Steward
User avatar
unclescrooge
Posts: 6264
Joined: Thu Jun 07, 2012 7:00 pm

Re: Automated Market Timing

Post by unclescrooge »

I don't think any discussion on automated market timing is complete without reference to Meb Faber and his paper "A Quantitative Approach to Tactical Asset Allocation",which you can read here:https://papers.ssrn.com/sol3/papers.cfm ... _id=962461

The popular 2007 paper shows how marketing timing "models have performed well in real-time, achieving equity like returns with bond like volatility and drawdowns."

I highly recommend it to anyone interested in Tactical Asset Allocation.

He even created an ETF to replicate the performance (ticker: GTAA). Unfortunately, this didn't really work in real life and was replaced by GMOM, which has also had poor performance.

Here's an interested blog post on the poor performance: http://www.moneysense.ca/invest/the-fai ... et-timing/
You also need to ask yourself how patient you will be when the active strategy under-performs, as all of them will eventually. According to Faber and Richardson’s own data, the GTAA strategy unperformed the buy-and-hold model in 12 of the 17 years from 1975 through 1991. And including 2013 (using the Cambria ETF as a proxy) it has now fallen short in five straight years, and in eight of the last 11. Active investors may be attracted to low volatility, but they also want comparable performance. How many would continue with any active strategy after multiple years of dismal results like that?

Successful investing is about more than the funds in your portfolio: it also requires a change in thinking. To stick to an indexing strategy over a lifetime, you need to let go of the idea that you can improve it with some kind of active overlay. Before you think you’re adding value with market timing or picking individual securities, remember that the probability of success is low and the potential payoff is small: even the best active strategies cannot hope to achieve more than an incremental outperformance over the long term.

More important, the whole pursuit is an enormous distraction from what’s really important, which is saving regularly, diversifying widely, keeping costs low and tuning out the chatter of those who promise what they can’t deliver.
So read the great book by Meb Faber "the Ivy Portfolio" and follow that instead of TAA.
2pedals
Posts: 1988
Joined: Wed Dec 31, 2014 11:31 am

Re: Automated Market Timing

Post by 2pedals »

Market timing works for stocks the are cyclical and predictable. Unfortunately stocks are not cyclical and are not predictable. When back testing with an analytical tool, you may find that some stocks act like they are cyclical and predictable and appear to "look good" for a money making scheme. This "looking good" behavior in my experience is a "false flag" and can not be capitalized on in a meaningful way for a long term investor.
Thesaints
Posts: 5087
Joined: Tue Jun 20, 2017 12:25 am

Re: Automated Market Timing

Post by Thesaints »

The only advantage for the individual retail investor is a superior agility consisting in the ability to move positions without affecting the price of the underlying securities.
That advantage pales compared to the lower transaction costs of institutional investors.

The theory that professional managers eschew strategies that are "too simple" borders on the bizarre.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

thangngo wrote: Tue Dec 12, 2017 4:45 pmIn theory, I like the idea of having a trading software that can optimize AA based on certain set parameters. This is ignoring transaction cost and the feasibility of actual apply this in real-time trading. This system will need to remove human completely from the decision making process, not even click to confirm a trade. Even if you have a software, you still have submit trade order on Vanguard/Fidelity website. And it is too easy for human to override the trade order from the software.
Third parties such as this hypothetical software manufacturer might be able to work directly with the brokerage website and trade without investor input. Aggregators like personal capital already have access to sensitive data...why not eliminate the user from this part of the decision, just as the user is eliminated from individual stock picks in index investing? (I think many hackles will rise about that due to privacy and security concerns, but this doesn't have to be about you. If someone preferred to let software handle the trading and gave away their passwords to allow that, would you think it disastrous?)

And just like employers can get lower health insurance rates than individuals, perhaps this hypothetical software manufacturer could negotiate lower costs than the individual investor, since it could be making thousands more trades per month.
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

unclescrooge wrote: Tue Dec 12, 2017 6:27 pm I don't think any discussion on automated market timing is complete without reference to Meb Faber and his paper "A Quantitative Approach to Tactical Asset Allocation",which you can read here:https://papers.ssrn.com/sol3/papers.cfm ... _id=962461 The popular 2007 paper shows how marketing timing "models have performed well in real-time, achieving equity like returns with bond like volatility and drawdowns." I highly recommend it to anyone interested in Tactical Asset Allocation. He even created an ETF to replicate the performance (ticker: GTAA). Unfortunately, this didn't really work in real life and was replaced by GMOM, which has also had poor performance. Here's an interested blog post on the poor performance: http://www.moneysense.ca/invest/the-fai ... et-timing/
You also need to ask yourself how patient you will be when the active strategy under-performs, as all of them will eventually. According to Faber and Richardson’s own data, the GTAA strategy unperformed the buy-and-hold model in 12 of the 17 years from 1975 through 1991. And including 2013 (using the Cambria ETF as a proxy) it has now fallen short in five straight years, and in eight of the last 11. Active investors may be attracted to low volatility, but they also want comparable performance. How many would continue with any active strategy after multiple years of dismal results like that?
So read the great book by Meb Faber "the Ivy Portfolio" and follow that instead of TAA.
Thank you for the excellent paper and the review of the ETF's performance. I think its performance is much like what I expected in my OP--lower returns and gentler dips--so it is not as "disappointing" to me as it may be to you. If an investor committed to systematic market timing is willing to accept that market timing will never perfectly time, isn't that expectation of disappointment fundamentally similar to the index investor's expectation of reversion to the mean?

Thanks to many of the references in the wiki and many of the wise people on this forum, I am not duped when something promises higher returns than the market. This thread is not about that...it's about creating gentler dips for the nervous investor. If there is a way to market time without emotion that helps my risk tolerance, isn't that useful?

P.S. for honesty's sake: I did not completely understand the paper at first read. I'm still working on it. I feel that someone should completely read a point before responding and I wanted to explain that the paper will take more time than a polite response dictates.
User avatar
David Jay
Posts: 14569
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Automated Market Timing

Post by David Jay »

Veiled:

One of our BH board members, Nisiprius, has a great riff on the general topic of beating the market. It impacted me so much that I use a portion of it in my signature line, I hope it gets you thinking as well. The remainder of this post is a direct quote:

In order to get the "risk premium" you really do have to take the risk. There's no magic formula for vastly reducing the risk of your stock market investments by simply avoiding short periods of obvious danger. If there were, everybody would do it and the stock market would not be risky and there would be no risk premium.

The stock market is risky. It is a form of greed to believe that there is some easy, effective way to get the risk premium without actually taking the risk. Many people go through a stage in which they engage in delusional thinking:
a) The stock market has had high returns. That's good, I'd like that.
b) The stock market has had high risk. I don't like that.
c) I can't accept "a" and "b" are a package deal.
d) The stock market is risky for everybody else, but not for me, because I can do something clever and get the return without the risk.
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
User avatar
Topic Author
Veiled
Posts: 226
Joined: Sun Oct 15, 2017 7:12 am
Location: Texas

Re: Automated Market Timing

Post by Veiled »

David Jay wrote: Tue Dec 12, 2017 7:45 pm Veiled:

One of our BH board members, Nisiprius, has a great riff on the general topic of beating the market. It impacted me so much that I use a portion of it in my signature line, I hope it gets you thinking as well. This is a direct quote:

In order to get the "risk premium" you really do have to take the risk. There's no magic formula for vastly reducing the risk of your stock market investments by simply avoiding short periods of obvious danger. If there were, everybody would do it and the stock market would not be risky and there would be no risk premium.

The stock market is risky. It is a form of greed to believe that there is some easy, effective way to get the risk premium without actually taking the risk. Many people go through a stage in which they engage in delusional thinking:
a) The stock market has had high returns. That's good, I'd like that.
b) The stock market has had high risk. I don't like that.
c) I can't accept "a" and "b" are a package deal.
d) The stock market is risky for everybody else, but not for me, because I can do something clever and get the return without the risk.
Maybe I shouldn't have used the words "market timing" in my thread title. It's too loaded with baggage for everyone. I'm not interested in higher returns. I can see that market timing can't guarantee me higher returns 100% or 50% or even 1% of the time. I'm interested in a smoother ride including the lower returns. Sorry to 'lash out' in response to your post. Thank you for replying and for passing on Nisipirius' wisdom. I agree with the idea that accepting a higher premium = accepting higher risk. That's built into every market, from groceries to equities. But I don't think I'm asking the same question.

But I'm new at this, so I could be really wrong. Am I asking the same question? Is "a smoother ride" just another form of a better return?
2pedals
Posts: 1988
Joined: Wed Dec 31, 2014 11:31 am

Re: Automated Market Timing

Post by 2pedals »

For me a "smoother ride" is an asset allocation that fits my risk tolerance so that I can stay the course during large markets drops.

"One thing that I strongly urge: Don't ever, ever, ever if you're an investor think of being out of the market or in the market," Bogle said
User avatar
David Jay
Posts: 14569
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Automated Market Timing

Post by David Jay »

Veiled wrote: Tue Dec 12, 2017 7:53 pm I'm interested in a smoother ride including the lower returns.
The problem is your suggested robo tactical allocation method is in fact market timing. The appropriate way to have lower risk and a smoother ride is to set your equity allocation appropriately. Not to attempt market timing. As Bogle has said (my highlight):

"The idea that a bell rings to signal when investors should get into or out of the stock market is simply not credible. After nearly 50 years in this business, I do not know of anybody who has done it successfully and consistently. I don't even know anybody who knows anybody who has done it successfully and consistently."

If this was easy enough to put it into a robo system, how come no Wall Street types can pull it off consistently?
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
Post Reply