MOM is real, too bad it doesn't work

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
garlandwhizzer
Posts: 2521
Joined: Fri Aug 06, 2010 3:42 pm

MOM is real, too bad it doesn't work

Post by garlandwhizzer » Tue Nov 12, 2019 2:40 pm

The following is an article based on a study by Dimension Fund Advisors. It concluded that academic backtesting models of MOM demonstrate it is real and consistent, but that funds that try to harvest it fail due to increased trading costs and frictions. The funds that trade most frequently turn out to harvest the academic MOM premium premium well but to be the biggest losers in real MOM funds. I think we have to take our hats off to DFA, a firm that strongly advocates factor approaches, for publishing the unvarnished truth about MOM. Worth a read IMO whether or not you're a factor investor.

https://www.institutionalinvestor.com/a ... esn-t-Work

Garland Whizzer

User avatar
Forester
Posts: 564
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

Re: MOM is real, too bad it doesn't work

Post by Forester » Tue Nov 12, 2019 3:20 pm

Methodology may be why iShares MTUM beats PDP & QMOM, naive academic approach (high turnover, less liquid stocks, stale momentum) vs pragmatic approach (liquid megacap bluechips with some amount of momentum).

Uncorrelated
Posts: 171
Joined: Sun Oct 13, 2019 3:16 pm

Re: MOM is real, too bad it doesn't work

Post by Uncorrelated » Tue Nov 12, 2019 8:45 pm

Forester wrote:
Tue Nov 12, 2019 3:20 pm
Methodology may be why iShares MTUM beats PDP & QMOM, naive academic approach (high turnover, less liquid stocks, stale momentum) vs pragmatic approach (liquid megacap bluechips with some amount of momentum).
According to portfoliovisualizer and both FF, AA and AQR factor sets, MTUM has negative size exposure and PDP and QMUM have positive size exposure. All three funds have negative value exposure. It makes sense that MTUM outperformend since the size premium was negative in recent history.

Since fund inception the alpha was statistically insignificant for all funds with all factor sets, and you would expect trading costs to turn up in the alpha. I don't think you can conclusively say any of these funds is the best.

DFA screens for stocks with bad volatility before adding them to the index. Vanguards new mutlifactor fund VFMF uses the same approach. This avoids the large negative momentum exposure some of the value funds have (value stocks tend to have negative momentum) without significantly adding to trading costs. This is one of the areas where multifactor funds beat combinations of single factor funds in theory.

Does anyone have a link or DOI of the DFA paper?

User avatar
nisiprius
Advisory Board
Posts: 39483
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: MOM is real, too bad it doesn't work

Post by nisiprius » Tue Nov 12, 2019 8:54 pm

Uncorrelated wrote:
Tue Nov 12, 2019 8:45 pm
...Does anyone have a link or DOI of the DFA paper?...
The article says "...according to new research available to clients of Dimensional Fund Advisors..." so, probably not.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

ARoseByAnyOtherName
Posts: 364
Joined: Wed Apr 26, 2017 12:03 am

Re: MOM is real, too bad it doesn't work

Post by ARoseByAnyOtherName » Tue Nov 12, 2019 9:51 pm

The article says the reason this doesn’t work in practice is trading/transaction costs.

What’s the exact definition of trading costs here? Is it trade commissions, bid-ask spreads, or something else, or all of the above?

What I’m getting at: does this imply that an individual investor, trading individual stocks commission free, has a theoretical chance of capturing this premium?

I figured that trading costs at large funds would be lower than us little guys. But I don’t know how trading works when tens or hundreds of millions of dollars are involved.

cheezit
Posts: 261
Joined: Sat Jul 14, 2018 7:28 pm

Re: MOM is real, too bad it doesn't work

Post by cheezit » Tue Nov 12, 2019 10:48 pm

Uncorrelated wrote:
Tue Nov 12, 2019 8:45 pm
Forester wrote:
Tue Nov 12, 2019 3:20 pm
Methodology may be why iShares MTUM beats PDP & QMOM, naive academic approach (high turnover, less liquid stocks, stale momentum) vs pragmatic approach (liquid megacap bluechips with some amount of momentum).
Since fund inception the alpha was statistically insignificant for all funds with all factor sets,
That's not what I'm seeing; MTUM has sizeable and positive alpha and QMOM has large negative alpha. Also, the R^2 for the curve fit is bad for all of these funds no matter how many factors you throw at them.

User avatar
nedsaid
Posts: 12802
Joined: Fri Nov 23, 2012 12:33 pm

Re: MOM is real, too bad it doesn't work

Post by nedsaid » Tue Nov 12, 2019 11:37 pm

Forester wrote:
Tue Nov 12, 2019 3:20 pm
Methodology may be why iShares MTUM beats PDP & QMOM, naive academic approach (high turnover, less liquid stocks, stale momentum) vs pragmatic approach (liquid megacap bluechips with some amount of momentum).
Another example of a "bad" factor product beating a "good" factor product.
A fool and his money are good for business.

andromeda2k12
Posts: 9
Joined: Sun Oct 13, 2019 10:22 pm

Re: MOM is real, too bad it doesn't work

Post by andromeda2k12 » Wed Nov 13, 2019 12:48 am

ARoseByAnyOtherName wrote:
Tue Nov 12, 2019 9:51 pm
The article says the reason this doesn’t work in practice is trading/transaction costs.

What’s the exact definition of trading costs here? Is it trade commissions, bid-ask spreads, or something else, or all of the above?

What I’m getting at: does this imply that an individual investor, trading individual stocks commission free, has a theoretical chance of capturing this premium?

I figured that trading costs at large funds would be lower than us little guys. But I don’t know how trading works when tens or hundreds of millions of dollars are involved.
To add on to this, where exactly do we see the trading costs manifest? The share price? Taxes? ER? I know this is somewhat of a basic question but I genuinely do not know how added trading costs specifically hurt a fund.

Uncorrelated
Posts: 171
Joined: Sun Oct 13, 2019 3:16 pm

Re: MOM is real, too bad it doesn't work

Post by Uncorrelated » Wed Nov 13, 2019 4:13 am

cheezit wrote:
Tue Nov 12, 2019 10:48 pm
Uncorrelated wrote:
Tue Nov 12, 2019 8:45 pm
Forester wrote:
Tue Nov 12, 2019 3:20 pm
Methodology may be why iShares MTUM beats PDP & QMOM, naive academic approach (high turnover, less liquid stocks, stale momentum) vs pragmatic approach (liquid megacap bluechips with some amount of momentum).
Since fund inception the alpha was statistically insignificant for all funds with all factor sets,
That's not what I'm seeing; MTUM has sizeable and positive alpha and QMOM has large negative alpha. Also, the R^2 for the curve fit is bad for all of these funds no matter how many factors you throw at them.
What I'm seeing at your link is that all alpha's are statistically insignificant. With statistically insignificant I mean that the t-stat is less than 2. The fact that 2 out of three funds have positive alpha should be concerning because it implies the fund captures alpha from things that are not explained by fund literature, such as idiosyncratic risk.

Anyway, the point here is that I seriously doubt that you can say that one of these funds is better than any other because they don't target the same universe of stocks. The difference in size exposure between MTUM and QMOM alone explains 1.68% of the return difference per year.

Uncorrelated
Posts: 171
Joined: Sun Oct 13, 2019 3:16 pm

Re: MOM is real, too bad it doesn't work

Post by Uncorrelated » Wed Nov 13, 2019 4:28 am

andromeda2k12 wrote:
Wed Nov 13, 2019 12:48 am
ARoseByAnyOtherName wrote:
Tue Nov 12, 2019 9:51 pm
The article says the reason this doesn’t work in practice is trading/transaction costs.

What’s the exact definition of trading costs here? Is it trade commissions, bid-ask spreads, or something else, or all of the above?

What I’m getting at: does this imply that an individual investor, trading individual stocks commission free, has a theoretical chance of capturing this premium?

I figured that trading costs at large funds would be lower than us little guys. But I don’t know how trading works when tens or hundreds of millions of dollars are involved.
To add on to this, where exactly do we see the trading costs manifest? The share price? Taxes? ER? I know this is somewhat of a basic question but I genuinely do not know how added trading costs specifically hurt a fund.
They turn up in the share price. Every time the ETF does a trade the NAV is reduced by a little to account for the spread and transaction costs of the trade.

Under MIFID II all EU funds are required to disclose their best guess of actual transaction costs, but vanguard UK is the only one that I can find the documents for. See https://www.vanguard.nl/portal/site/por ... fid-priips and then any of the documents under "Vanguard funds' cost and charges information as required by MiFID II". The transaction costs for any normal fund is in the range of 0.00% to 0.05% per year. But the vanguard momentum fund has estimated transaction costs of 0.22% per year.

ARoseByAnyOtherName
Posts: 364
Joined: Wed Apr 26, 2017 12:03 am

Re: MOM is real, too bad it doesn't work

Post by ARoseByAnyOtherName » Wed Nov 13, 2019 6:17 am

Uncorrelated wrote:
Wed Nov 13, 2019 4:28 am
andromeda2k12 wrote:
Wed Nov 13, 2019 12:48 am
ARoseByAnyOtherName wrote:
Tue Nov 12, 2019 9:51 pm
The article says the reason this doesn’t work in practice is trading/transaction costs.

What’s the exact definition of trading costs here? Is it trade commissions, bid-ask spreads, or something else, or all of the above?

What I’m getting at: does this imply that an individual investor, trading individual stocks commission free, has a theoretical chance of capturing this premium?

I figured that trading costs at large funds would be lower than us little guys. But I don’t know how trading works when tens or hundreds of millions of dollars are involved.
To add on to this, where exactly do we see the trading costs manifest? The share price? Taxes? ER? I know this is somewhat of a basic question but I genuinely do not know how added trading costs specifically hurt a fund.
They turn up in the share price. Every time the ETF does a trade the NAV is reduced by a little to account for the spread and transaction costs of the trade.

Under MIFID II all EU funds are required to disclose their best guess of actual transaction costs, but vanguard UK is the only one that I can find the documents for. See https://www.vanguard.nl/portal/site/por ... fid-priips and then any of the documents under "Vanguard funds' cost and charges information as required by MiFID II". The transaction costs for any normal fund is in the range of 0.00% to 0.05% per year. But the vanguard momentum fund has estimated transaction costs of 0.22% per year.
From the link above:

“Transaction costs: All costs and charges incurred as a result of the acquisition and disposal of underlying investments within the fund. This includes broker commissions, entry and exit charges, spreads, stamp duty, transactions tax, and foreign exchange costs. (Please note that when rounded to 2 decimal places, Transaction costs less than 0.01% will be displayed as 0.00%).”

Note that is for a European fund.

So what about individual stocks, bought and sold my individual investors? Of the above I think commissions, spreads, and tax would apply to us individual investors in the US buying stocks of US companies. (That said I don’t know what entry and exit charges are.)

Broker commissions are effectively zero nowadays.

Spreads are a factor but seem like they could be mitigated somewhat by limit orders and such?

Tax seems like the biggest cost here.

What am I missing?

Uncorrelated
Posts: 171
Joined: Sun Oct 13, 2019 3:16 pm

Re: MOM is real, too bad it doesn't work

Post by Uncorrelated » Wed Nov 13, 2019 6:46 am

ARoseByAnyOtherName wrote:
Wed Nov 13, 2019 6:17 am
The spread is the biggest factor. Getting your orders to reliably execute at a spread of zero is not trivial. In fact, that is where high frequency traders and market makers make all their money.

User avatar
nisiprius
Advisory Board
Posts: 39483
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: MOM is real, too bad it doesn't work

Post by nisiprius » Wed Nov 13, 2019 6:54 am

Transaction costs would be reflected by index tracking error, specifically underperforming the index they track. I think one of the benefits of index funds is that the effect of transaction costs can't be hidden. And in my cynical opinion that might be one of the reasons why increasingly sophisticated factor funds are not index funds. Many modern factor funds basically don't have any realistic benchmarks, a fact which doesn't bother enthusiasts. (The riposte is that when a fund is outperforming a total market index fund is, why would you care whether it is matching some index, or exactly what it is doing?) To be sure, there are, perhaps, legitimate reasons for not trying to track an index. Dimensional Fund Advisors has said (for decades) that it does not track indexes because they can get superior results by "patient trading;" not so much buying low, but avoiding having an index force them to buy high, when there's no reason for a long-term buy-and-hold investor to care about short-term tracking accuracy.

It isn't momentum, but an illustration of the kinds of problems funds get into when they try to trade smaller and less liquid issues--which of course an individual investor would get into as well, only more so--was given by Rick Ferri in The Truth about Micro-Cap Index Funds.
Last year was a recovery year for U.S. stocks. The Wilshire 5000 Index, a broad benchmark of U.S. common stocks, returned 29.3% in 2009.

[But] the Wilshire US Micro-Cap index was up 47.6%, and the CRSP Decile 10 Index, which measures the smallest micro-cap stocks, registered a whopping 81.7% gain....

You may think that with micro-cap returns like these, micro-cap index fund investors would be jumping for joy. But they're not. Why? Because the returns of all micro-cap index funds were well below these returns, so low that they didn't even reach the Wilshire 5000 return.... [He goes on to tabulate returns for four micro-cap index funds and ETFs: BRSIX, 26.0%. IWC, 23.7%. FDM, 20.0%. PZI, 12.1%]
Whether or not mutual funds and ETFs can achieve, in real life, the results obtained in spreadsheet tabulations of cost-free paper indexes, has always been an issue.
Last edited by nisiprius on Wed Nov 13, 2019 7:39 am, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

lazyday
Posts: 3617
Joined: Wed Mar 14, 2007 10:27 pm

Re: MOM is real, too bad it doesn't work

Post by lazyday » Wed Nov 13, 2019 7:30 am

ARoseByAnyOtherName wrote:
Tue Nov 12, 2019 9:51 pm
Is it trade commissions, bid-ask spreads, or something else, or all of the above?
Market Impact costs might be the most significant. Buying pushes up the market price, selling pushes it down. Spreads can also matter, they can be quite wide for smallcaps.
andromeda2k12 wrote:
Wed Nov 13, 2019 12:48 am
where exactly do we see the trading costs manifest? The share price? Taxes? ER?
Absolutely not in the ER. Even the commissions don't show up in the ER. You can find commissions costs in documents required by the SEC, usually found on the fund website, and if not it should be on the SEC site. As I recall, I think it's in the SAI, statement of additional information. Commissions are probably a small part of costs though.

All trading costs will show up in the fund share price. You can compare fund returns to index returns to estimate total costs, but even that can miss some costs. Some costs will show both in the share price and in the index itself. Imagine a popular, simple smallcap index with a policy of rebalancing on Jan 1. At the end of the year, the largest company in the index is no longer a smallcap stock, so it will have to leave the index at the start of the new year. Since it's a popular index, many shares will be sold on the first trading day. Arbitrageurs will anticipate this, and sell shares before the end of the year, driving the price down before the stock leaves the index or the fund.

Fund companies and index providers try to reduce this problem in various ways, but I imagine it's still significant in some cases.
ARoseByAnyOtherName wrote:
Tue Nov 12, 2019 9:51 pm
What I’m getting at: does this imply that an individual investor, trading individual stocks commission free, has a theoretical chance of capturing this premium?
Possibly, since we usually won't cause market impact with our trades. We would need to avoid trading close to the times that big money is trading. Maybe don't follow an index exactly, or use the most obvious trading rule like exactly 12-1 months of momentum.

lazyday
Posts: 3617
Joined: Wed Mar 14, 2007 10:27 pm

Re: MOM is real, too bad it doesn't work

Post by lazyday » Wed Nov 13, 2019 7:48 am

nisiprius wrote:
Wed Nov 13, 2019 6:54 am
I think one of the benefits of index funds is that the effect of transaction costs can't be hidden. And in my cynical opinion that might be one of the reasons why increasingly sophisticated factor funds are not index funds.
Funds that don't follow an index do seem to have the disadvantage that it's difficult to evaluate them. I believe that some funds manage costs better than others, but without an index, how can we tell?

However, I think avoiding an index isn't just about hiding your costs. Rather than retype, here's what I posted elsewhere:
lazyday wrote:
Tue Nov 05, 2019 7:27 am
For international small cap and for EM, the biggest advantage of Avantis over other [multifactor] ETFs might be that it does not use an index. This might significantly reduce market impact costs.

Perhaps a fund like ISCF tries to use creation and redemption baskets instead of trading, but I doubt this can be done as often and as efficiently when bound to an index and a rebalancing policy. Couldn’t an Avantis fund avoid trading entirely by patiently using redemption baskets to get rid of stocks it no longer wants?

Even if a fund avoided all trading while following an index, there may still be hidden costs in the index itself because of front running.

I just wish Vanguard would offer ex-US funds like the US multifactor VFMF. It’s difficult to evaluate funds that don’t follow an index, but I trust Vanguard and would be quick to buy.

User avatar
nisiprius
Advisory Board
Posts: 39483
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: MOM is real, too bad it doesn't work

Post by nisiprius » Wed Nov 13, 2019 10:55 am

lazyday wrote:
Wed Nov 13, 2019 7:48 am
...Funds that don't follow an index do seem to have the disadvantage that it's difficult to evaluate them. I believe that some funds manage costs better than others, but without an index, how can we tell?

However, I think avoiding an index isn't just about hiding your costs....
Agree to both. But, the question arises, for anything that isn't an index fund, how do we evaluate them other than on past performance?

Whether or not it is a good thing to track the S&P 500, it is possible to judge, pretty objectively, that

Source

Image

that, given that the job is "tracking the S&P 500" (in yellow), that PREIX does the job better than MXVIX, that VFIAX does it better than PREIX, and that all three of them do it better than RYSOX.

But as the arguments upthread shows, it's not clear whether MTUM is beating other momentum funds because it is better at tracking momentum, or because it's worse. For example, is QMOM doing its job? Well, according to Alpha Architect, the job of QMOM is to track the "QMOM index" (I'm not kidding), and the prospectus explains "ADDITIONAL INFORMATION ABOUT THE INDEXES: Each Index was developed by Messrs. Wesley R. Gray and John Vogel of the Adviser," and that's it. I take it for granted, but am willing to be educated, that no other mutual fund or ETF tracks the QMOM index, and I take it for granted, but am willing to be educated, that 99.9% of all investors must rely on Alpha Architect as the only source of the index values. What is odd, but I admit that I haven't done a thorough search, is that I can't find those values, nor can I find a tabular comparison of the performance of QMOM vis-a-vis its index. Can anyone else?
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
nedsaid
Posts: 12802
Joined: Fri Nov 23, 2012 12:33 pm

Re: MOM is real, too bad it doesn't work

Post by nedsaid » Wed Nov 13, 2019 12:25 pm

nisiprius wrote:
Wed Nov 13, 2019 10:55 am
lazyday wrote:
Wed Nov 13, 2019 7:48 am
...Funds that don't follow an index do seem to have the disadvantage that it's difficult to evaluate them. I believe that some funds manage costs better than others, but without an index, how can we tell?

However, I think avoiding an index isn't just about hiding your costs....
Agree to both. But, the question arises, for anything that isn't an index fund, how do we evaluate them other than on past performance?

Whether or not it is a good thing to track the S&P 500, it is possible to judge, pretty objectively, that

Source

Image

that, given that the job is "tracking the S&P 500" (in yellow), that PREIX does the job better than MXVIX, that VFIAX does it better than PREIX, and that all three of them do it better than RYSOX.

But as the arguments upthread shows, it's not clear whether MTUM is beating other momentum funds because it is better at tracking momentum, or because it's worse. For example, is QMOM doing its job? Well, according to Alpha Architect, the job of QMOM is to track the "QMOM index" (I'm not kidding), and the prospectus explains "ADDITIONAL INFORMATION ABOUT THE INDEXES: Each Index was developed by Messrs. Wesley R. Gray and John Vogel of the Adviser," and that's it. I take it for granted, but am willing to be educated, that no other mutual fund or ETF tracks the QMOM index, and I take it for granted, but am willing to be educated, that 99.9% of all investors must rely on Alpha Architect as the only source of the index values. What is odd, but I admit that I haven't done a thorough search, is that I can't find those values, nor can I find a tabular comparison of the performance of QMOM vis-a-vis its index. Can anyone else?
If at first you try and don't succeed, try creating your own index. Or better yet, use the Groucho Marx solution and switch indexes! Something out there that you ought to be able to beat.
A fool and his money are good for business.

User avatar
Forester
Posts: 564
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

Re: MOM is real, too bad it doesn't work

Post by Forester » Wed Nov 13, 2019 12:39 pm

QMOM = Russell 1000. Top decile 12mo (minus last month), then split in half on volatility to arrive at 50 stocks.

The obvious problem is stale, waning momentum. The next issue may be equal-weighting. Weighted average market cap of QMOM is $18.5billion. MTUM is $168.5billion. So QMOM is much smaller and in turn buying smaller companies, maybe this is also a drag. For some reason Alpha Architect didn't include a recent momentum filter, it was well known in quant circles that 6-month backtests better than 12-month. It's an ETF you would create if you wanted something academically elegant.

MTUM is still buying from the top three momentum deciles, but it also considers recent 6-month momentum, plus you have the expertise and Goliath scale of iShares when it comes to trading megacap liquid US stocks.

andromeda2k12
Posts: 9
Joined: Sun Oct 13, 2019 10:22 pm

Re: MOM is real, too bad it doesn't work

Post by andromeda2k12 » Wed Nov 13, 2019 7:13 pm

Uncorrelated wrote:
Wed Nov 13, 2019 4:28 am
andromeda2k12 wrote:
Wed Nov 13, 2019 12:48 am
ARoseByAnyOtherName wrote:
Tue Nov 12, 2019 9:51 pm
The article says the reason this doesn’t work in practice is trading/transaction costs.

What’s the exact definition of trading costs here? Is it trade commissions, bid-ask spreads, or something else, or all of the above?

What I’m getting at: does this imply that an individual investor, trading individual stocks commission free, has a theoretical chance of capturing this premium?

I figured that trading costs at large funds would be lower than us little guys. But I don’t know how trading works when tens or hundreds of millions of dollars are involved.
To add on to this, where exactly do we see the trading costs manifest? The share price? Taxes? ER? I know this is somewhat of a basic question but I genuinely do not know how added trading costs specifically hurt a fund.
They turn up in the share price. Every time the ETF does a trade the NAV is reduced by a little to account for the spread and transaction costs of the trade.

Under MIFID II all EU funds are required to disclose their best guess of actual transaction costs, but vanguard UK is the only one that I can find the documents for. See https://www.vanguard.nl/portal/site/por ... fid-priips and then any of the documents under "Vanguard funds' cost and charges information as required by MiFID II". The transaction costs for any normal fund is in the range of 0.00% to 0.05% per year. But the vanguard momentum fund has estimated transaction costs of 0.22% per year.
Interesting, 0.22% doesn't seem absurdly high though considering the theoretical premium right?

User avatar
nisiprius
Advisory Board
Posts: 39483
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: MOM is real, too bad it doesn't work

Post by nisiprius » Wed Nov 13, 2019 7:44 pm

Forester wrote:
Wed Nov 13, 2019 12:39 pm
QMOM = Russell 1000.
QMOM isn't trying to track the Russell 1000 index. It is trying to track the "QMOM Index." That is what Alpha Architect says. So, can you point us to a place where the curious can find published values for the QMOM Index, so that we can see how closely QMOM is tracking it?
Last edited by nisiprius on Wed Nov 13, 2019 9:22 pm, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
Forester
Posts: 564
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

Re: MOM is real, too bad it doesn't work

Post by Forester » Wed Nov 13, 2019 9:15 pm

nisiprius wrote:
Wed Nov 13, 2019 7:44 pm
Forester wrote:
Wed Nov 13, 2019 12:39 pm
QMOM = Russell 1000.
MTUM isn't trying to track the Russell 1000 index. It is trying to track the "QMOM Index." That is what Alpha Architect says. So, can you point us to a place where the curious can find published values for the QMOM Index, so that we can see how closely QMOM is tracking it?
My interpretation, is that "index" in this case is the same thing as "universe". There's no question of mis-representation, Alpha Architect don't pretend to be in the same corner of the market as Vanguard.

The white paper on the ETF page describes the methodology, bottom 60% of NYSE stocks by market cap are eliminated https://etfsite.alphaarchitect.com/wp- ... _final.pdf

It's easy to work out the size of their starting universe since they buy from a cheapest decile split exactly in half. I guess the correct benchmark for QMOM would be midcap or midcap growth.

User avatar
nisiprius
Advisory Board
Posts: 39483
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: MOM is real, too bad it doesn't work

Post by nisiprius » Wed Nov 13, 2019 9:22 pm

Forester wrote:
Wed Nov 13, 2019 9:15 pm
nisiprius wrote:
Wed Nov 13, 2019 7:44 pm
Forester wrote:
Wed Nov 13, 2019 12:39 pm
QMOM = Russell 1000.
QMOM isn't trying to track the Russell 1000 index. It is trying to track the "QMOM Index." That is what Alpha Architect says. So, can you point us to a place where the curious can find published values for the QMOM Index, so that we can see how closely QMOM is tracking it?
My interpretation, is that "index" in this case is the same thing as "universe". There's no question of mis-representation, Alpha Architect don't pretend to be in the same corner of the market as Vanguard.

The white paper on the ETF page describes the methodology, bottom 60% of NYSE stocks by market cap are eliminated https://etfsite.alphaarchitect.com/wp- ... _final.pdf

It's easy to work out the size of their starting universe since they buy from a cheapest decile split exactly in half.
So, where can we find published values for that index, so that we can compare QMOM to it, regardless of whether or not we expect QMOM to track it perfectly? Is the white paper so complete and the calculation so easy that you can calculate it yourself and post it for us?
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
nisiprius
Advisory Board
Posts: 39483
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: MOM is real, too bad it doesn't work

Post by nisiprius » Wed Nov 13, 2019 9:32 pm

Index values for the QMOM index might be available at this site. Registration is free, but I can't register because I am not a financial professional and you must affirm that you are one in order to become a user. Could someone who is a financial profession register and report on whether values for the QMOM index are available there?
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

YRT70
Posts: 400
Joined: Sat Apr 27, 2019 8:51 am

Re: MOM is real, too bad it doesn't work

Post by YRT70 » Thu Nov 14, 2019 3:15 am

Factor Momentum Is Real, Researchers Argue
AQR’s lab lobs new findings into factor investing’s hottest debate.
In the new study, Kelly and Gupta tested the factor momentum strategy by timing individual factors based on past one-month returns. Momentum-based factor timing resulted in positive excess returns for 61 of the 65 factors, and statistically significant alpha for 47 of them. When all the factors were combined, the factor-timing strategy earned an annual Sharpe ratio of 0.84 — a result that only decreased slightly when the window of past returns was lengthened from one month to one year and five years.

Even after accounting for transaction costs, which Kelly and Gupta acknowledged would “eat into” the performance of any momentum strategy, they found that the factor timing strategy delivered outperformance. The finding is in contrast with recent research by Dimensional Fund Advisors, which suggested that fees and trading costs wiped out any benefits earned using momentum.

“My research shows an implementable benefit of using momentum,” Kelly said. “That’s true for cross-section momentum” — a version of the momentum factor commonly applied to stocks — “and factor time-series momentum.”
https://www.institutionalinvestor.com/a ... hers-Argue

Uncorrelated
Posts: 171
Joined: Sun Oct 13, 2019 3:16 pm

Re: MOM is real, too bad it doesn't work

Post by Uncorrelated » Thu Nov 14, 2019 4:38 am

andromeda2k12 wrote:
Wed Nov 13, 2019 7:13 pm
Uncorrelated wrote:
Wed Nov 13, 2019 4:28 am
Interesting, 0.22% doesn't seem absurdly high though considering the theoretical premium right?
That's true, I expected the costs to be higher. I don't know how accurate these figures are, there is a fair bit of guesswork involved in estimating the transaction costs for illiquid securities.

User avatar
Forester
Posts: 564
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

Re: MOM is real, too bad it doesn't work

Post by Forester » Thu Nov 14, 2019 5:45 am

nisiprius wrote:
Wed Nov 13, 2019 9:22 pm
Forester wrote:
Wed Nov 13, 2019 9:15 pm
nisiprius wrote:
Wed Nov 13, 2019 7:44 pm
Forester wrote:
Wed Nov 13, 2019 12:39 pm
QMOM = Russell 1000.
QMOM isn't trying to track the Russell 1000 index. It is trying to track the "QMOM Index." That is what Alpha Architect says. So, can you point us to a place where the curious can find published values for the QMOM Index, so that we can see how closely QMOM is tracking it?
My interpretation, is that "index" in this case is the same thing as "universe". There's no question of mis-representation, Alpha Architect don't pretend to be in the same corner of the market as Vanguard.

The white paper on the ETF page describes the methodology, bottom 60% of NYSE stocks by market cap are eliminated https://etfsite.alphaarchitect.com/wp- ... _final.pdf

It's easy to work out the size of their starting universe since they buy from a cheapest decile split exactly in half.
So, where can we find published values for that index, so that we can compare QMOM to it, regardless of whether or not we expect QMOM to track it perfectly? Is the white paper so complete and the calculation so easy that you can calculate it yourself and post it for us?
OK that would be interesting to compare the "index" to the live fund but probably in the end, QMOM is a flawed product and might suggest cross-sectional momentum is only practical & scalable with megacaps.

QMOM had half the performance of MTUM in the late '16 early '18 run up and even under-performed the S&P. It should be overhauled or put out of its misery.

User avatar
CULater
Posts: 2325
Joined: Sun Nov 13, 2016 10:59 am

Re: MOM is real, too bad it doesn't work

Post by CULater » Thu Nov 14, 2019 11:25 am

I once saw a snipe, but couldn't catch it before it disappeared into the forest. :(
On the internet, nobody knows you're a dog.

User avatar
hdas
Posts: 1275
Joined: Thu Jun 11, 2015 8:24 am

Re: MOM is real, too bad it doesn't work

Post by hdas » Thu Nov 14, 2019 11:42 am

Forester wrote:
Thu Nov 14, 2019 5:45 am
nisiprius wrote:
Wed Nov 13, 2019 9:22 pm
Forester wrote:
Wed Nov 13, 2019 9:15 pm
nisiprius wrote:
Wed Nov 13, 2019 7:44 pm
Forester wrote:
Wed Nov 13, 2019 12:39 pm
QMOM = Russell 1000.
QMOM isn't trying to track the Russell 1000 index. It is trying to track the "QMOM Index." That is what Alpha Architect says. So, can you point us to a place where the curious can find published values for the QMOM Index, so that we can see how closely QMOM is tracking it?
My interpretation, is that "index" in this case is the same thing as "universe". There's no question of mis-representation, Alpha Architect don't pretend to be in the same corner of the market as Vanguard.

The white paper on the ETF page describes the methodology, bottom 60% of NYSE stocks by market cap are eliminated https://etfsite.alphaarchitect.com/wp- ... _final.pdf

It's easy to work out the size of their starting universe since they buy from a cheapest decile split exactly in half.
So, where can we find published values for that index, so that we can compare QMOM to it, regardless of whether or not we expect QMOM to track it perfectly? Is the white paper so complete and the calculation so easy that you can calculate it yourself and post it for us?
OK that would be interesting to compare the "index" to the live fund but probably in the end, QMOM is a flawed product and might suggest cross-sectional momentum is only practical & scalable with megacaps.

QMOM had half the performance of MTUM in the late '16 early '18 run up and even under-performed the S&P. It should be overhauled or put out of its misery.
There must be more to the story given that VFMO doesn't do too bad either. Cheers :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

Post Reply