Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
Interesting that no ETFs appear on that comprehensive list @msf listed. That’s what I was searching the web for. I did stumble on 4 or 5 market neutral ETFs that had already closed. And 2 or 3 that were still around.
Does it seem like a high proportion of ETFs fail to last for more than 5 or 6 years? Much greater rate of attrition than mutual funds ISTM. Might be wrong.
GAMNX -16% in 2022 // +39% in 2023. Yup. Sounds market neutral.
I was trying to illustrate what M*'s basic fund screener couldn't find because it wasn't being maintained and so didn't recognize new (often replacement) categories. As a fund screener, it wouldn't have found any ETFs even if M* had continued to maintain it.
For the record, M* calls BTAL equity market neutral.
[snip] Does it seem like a high proportion of ETFs fail to last for more than 5 or 6 years? Much greater rate of attrition than mutual funds ISTM. Might be wrong. [snip]
You are not wrong. The attrition rate for ETFs has been high.
"As of Dec. 13, this year’s nearly 500 exchange-traded fund launches have already broken the record set in 2021 (461). The ETF universe is more expansive than ever: Investors can choose from 3,487 ETFs. There have been 5,067 ETFs brought to market since SPDR S&P 500 ETF SPY launched in 1993. This means 31% of them have since closed." https://www.morningstar.com/etfs/best-worst-new-etfs-2023
MFO Premium Homepage has an interesting layout. The 1st row of the horizontal Menu Bar has all the FREE stuff. The 2nd row has all the SUBSCRIBER stuff.
MFO's Basic Screener (aka QuickSearch) is still free!
Yes it is, and it is a fine engine with several post-analysis criteria available (Great Owl, MFO risk,etc.). But just as with M*'s "new and degraded" premium investor screener,only post-analysis criteria are available.
Neither tool provides screens for funds based on annualized returns, though those figures are displayed in the result sets and one can sort them. Nor are other raw (pre-analysis) screening criteria like ER or AUM available.
My preference is to slice and dice raw data (annualized returns, ERs, etc.). My ideal would be a screener that let the user write their own queries - to have access to every data column, to be able to use logical connectors. For example: > $1B in AUM or (> $500M in AUM and < 3 years old).
After that tool vanished, M*'s basic fund screener was still available for awhile. It was a very weak tool. But it did have a limited ability to screen on a few raw data attributes. Now what?
My preference is to slice and dice raw data (annualized returns, ERs, etc.). My ideal would be a screener that let the user write their own queries - to have access to every data column, to be able to use logical connectors. For example: > $1B in AUM or (> $500M in AUM and < 3 years old).
With MFO premium you could download the dataset that includes something like AUM, and a whole lot of other fields. And then you could apply those criteria in your spreadsheet. If you're already thinking Boolean, you could probably learn how to apply those criteria in a spreadsheet or data query. Am I missing something?
One can download the entire MFO dataset, or as you seem to suggest, download a view that includes only a subset of columns (AUM and a whole lot of other, but not all, fields). And one can program a spreadsheet to sort and search based on various criteria. Better yet, import into a database and use its query engine.
Either way, this is circumventing the MFO screeners, not using them. In the picture on the right below, what this is doing is snarfing the box "Data" and attaching your own query engine to it.
FWIW, I believe that one can download the M* data using the Investor screener. Though because downloads are limited to 120 rows, one will wind up doing around 400-500 downloads depending on how clever one is in constructing small subsets to download. An absolutely insane exercise, albeit theoretically possible. I have better things to do for a week.
One can download the entire MFO dataset, or as you seem to suggest, download a view that includes only a subset of columns (AUM and a whole lot of other, but not all, fields). And one can program a spreadsheet to sort and search based on various criteria. Better yet, import into a database and use its query engine.
Either way, this is circumventing the MFO screeners, not using them. In the picture on the right below, what this is doing is snarfing the box "Data" and attaching your own query engine to it.
One of the things I used to do for 25 years was sell data. Being able to download all of the data for the price of an MFO subscription seems like more of a feature than a circumvention to me. But there were always people that wanted us to do the slicing and dicing for them. So we charged them more.
MFO's MultiSearch enables searches of nearly 40,000 ETFs, mutual funds, CEFs, ETNs, Interval Funds, Fund of Funds, and Insurance Funds employing literally hundreds of screening criteria across almost 100 evaluation periods.
Dating back to 1960. Plus, indexes back to 1926. While offering dozens of analytics.
Granted, it's not "free" but pretty close ... and, proceeds go to a good cause.
And, granted, it's ex-post, but show me an ex-ante screener and I will show you the moon. (On second thought, I will show you David Snowball.)
MFO's MultiSearch is a great tool with an extremely extensive database underlying it, including a lot of MFO-defined metrics. In no way was I trying to disparage it.
That said, I'm more interested in the underlying data than in the tool. You link to images of the UI to show the screening parameters available. I prefer to look at simple table of those parameters:
I was taught in programming language design that generally less is better; too many ways of accomplishing something confuses people. Not unlike the effect seen when 401(k)s have too many options. So using predefined sets of options via select boxes is a good screener design for many users.
Realistically, many people think of screeners this way - give me all the diversified foreign funds with large cap blend portfolios and a three year risk/reward rating of 5 (out of 5).
I prefer something more flexible and powerful like a SQL interpreter. The LexisNexis query language is another example of this sort of flexibility, though that language is designed more for document searches than database field searches.
I gave a simple example of a query above that most tools can't handle. Say I'm concerned about the stability of funds that are very small. I might search for funds over $1B in AUM. But recognizing that it takes time to build assets, I might make allowances. For young (under 3 year old) funds, I could accept funds with at least $500M in assets.
Of course I could do this in two separate searches (over $1B, young and over $500M), though I'd have to combine the result sets and weed out duplicates (young funds with over $1B in assets). I'd like to be able to do this in a single query. And I'd like to be able to set arbitrary thresholds, say $767M instead of $500M. That's not possible if I'm given only a choice of $500M, $1B, $10B, or $100B.
As to cost, let's see how the tax laws change in a couple of years. Right now, there are no above-the-line charitable deductions (as there were in 2021). And without a mortgage it's hard to itemize now.
One can still get a tax advantage by contributing via a donor advised fund or a QCD. Those must be pure contributions with no benefit received; they are not subscriptions. Proceeds do go to a good cause. I've contributed before and expect to do so again.
This is not a complaint. It is a fact of life. The subscription services are an inducement to contribute and so pragmatically they cannot be offered separately. If the services were offered separately, people would be less motivated to contribute beyond the cost of the subscription. (I also decline the tote bags offered from 501(c)(3) institutions X, Y, and Z.) https://www.mutualfundobserver.com/support-us https://www.mutualfundobserver.com/support-us/501c3/
Any search in our tools, and all the attendant data, can be downloaded. But currently we limit the size to 1000 funds, which seemed practical to me. We also have a link to an excel file with performance summary of all funds, oldest share class only ... about 15,000 funds. I know some subscribers just prefer that link! Spreadsheet style. I suppose we could generate a downloadable file of all the static data (no performance) for all funds in database. But then you still need performance data, which has infinite possibilities. I think we could come up with a customized screener or database based on subscriber request, but that is really a different product. Happy to chat more. c
Thanks Charles. I'd be glad to chat with you about some thoughts, though I just got down this rabbit hole responding to a comment that using raw data might not be circumventing a screener. Your post now addresses that in crystal clear terms:
We also have a link to an excel file with performance summary of all funds, oldest share class only ... about 15,000 funds. I know some subscribers just prefer that link! Spreadsheet style
No muss, no fuss, just circumventing all tools.
If I were still really deep into data analysis, I'd likely import that spreadsheet into a DBMS. These days I've got better things to do with my time: going on a river cruise out to the Black Sea, researching a trans-Canada railway trip, fighting with my insurance company on the network status of my healthcare providers (the insurer has processed over half of them incorrectly).
Comments
Does it seem like a high proportion of ETFs fail to last for more than 5 or 6 years? Much greater rate of attrition than mutual funds ISTM. Might be wrong.
GAMNX -16% in 2022 // +39% in 2023. Yup. Sounds market neutral.
For the record, M* calls BTAL equity market neutral.
"As of Dec. 13, this year’s nearly 500 exchange-traded fund launches have already broken the record set in 2021 (461). The ETF universe is more expansive than ever: Investors can choose from 3,487 ETFs.
There have been 5,067 ETFs brought to market since SPDR S&P 500 ETF SPY launched in 1993.
This means 31% of them have since closed."
https://www.morningstar.com/etfs/best-worst-new-etfs-2023
HOME BLOG QUICKSEARCH OWLS ALARM PROFILES DEFINITIONS MFO
LOGIN ANALYZE MULTISEARCH PORTFOLIOS LAUNCHES AVERAGES FAMILIES
Neither tool provides screens for funds based on annualized returns, though those figures are displayed in the result sets and one can sort them. Nor are other raw (pre-analysis) screening criteria like ER or AUM available.
My preference is to slice and dice raw data (annualized returns, ERs, etc.). My ideal would be a screener that let the user write their own queries - to have access to every data column, to be able to use logical connectors. For example:
> $1B in AUM or (> $500M in AUM and < 3 years old).
M*'s premium fund screener was great at this. It provided access to a plethora of underlying data categories and let you build queries using ands and ors
https://screen.morningstar.com/v2/AdvFunds/data_definition.html?field=Sector+Weightings
After that tool vanished, M*'s basic fund screener was still available for awhile. It was a very weak tool. But it did have a limited ability to screen on a few raw data attributes. Now what?
Either way, this is circumventing the MFO screeners, not using them. In the picture on the right below, what this is doing is snarfing the box "Data" and attaching your own query engine to it.
FWIW, I believe that one can download the M* data using the Investor screener. Though because downloads are limited to 120 rows, one will wind up doing around 400-500 downloads depending on how clever one is in constructing small subsets to download. An absolutely insane exercise, albeit theoretically possible. I have better things to do for a week.
Dating back to 1960. Plus, indexes back to 1926. While offering dozens of analytics.
Granted, it's not "free" but pretty close ... and, proceeds go to a good cause.
And, granted, it's ex-post, but show me an ex-ante screener and I will show you the moon. (On second thought, I will show you David Snowball.)
Here is link to png screenshot: https://www.screencast.com/t/8bTIcDOTE0b
Here is link to same screenshot but as a pdf download: https://1drv.ms/b/s!AmF1g-5ef-jI1ttKkPDZ6jeu2U9gDg?e=LjVZqv
That said, I'm more interested in the underlying data than in the tool. You link to images of the UI to show the screening parameters available. I prefer to look at simple table of those parameters:
https://www.member.mfopremium.com/wp-content/themes/twentynineteen-child_202106v3/olib/media/multisearch_screening_parameters.pdf?v20210728.1
I was taught in programming language design that generally less is better; too many ways of accomplishing something confuses people. Not unlike the effect seen when 401(k)s have too many options. So using predefined sets of options via select boxes is a good screener design for many users.
Realistically, many people think of screeners this way - give me all the diversified foreign funds with large cap blend portfolios and a three year risk/reward rating of 5 (out of 5).
I prefer something more flexible and powerful like a SQL interpreter. The LexisNexis query language is another example of this sort of flexibility, though that language is designed more for document searches than database field searches.
I gave a simple example of a query above that most tools can't handle. Say I'm concerned about the stability of funds that are very small. I might search for funds over $1B in AUM. But recognizing that it takes time to build assets, I might make allowances. For young (under 3 year old) funds, I could accept funds with at least $500M in assets.
Of course I could do this in two separate searches (over $1B, young and over $500M), though I'd have to combine the result sets and weed out duplicates (young funds with over $1B in assets). I'd like to be able to do this in a single query. And I'd like to be able to set arbitrary thresholds, say $767M instead of $500M. That's not possible if I'm given only a choice of $500M, $1B, $10B, or $100B.
As to cost, let's see how the tax laws change in a couple of years. Right now, there are no above-the-line charitable deductions (as there were in 2021). And without a mortgage it's hard to itemize now.
One can still get a tax advantage by contributing via a donor advised fund or a QCD. Those must be pure contributions with no benefit received; they are not subscriptions. Proceeds do go to a good cause. I've contributed before and expect to do so again.
This is not a complaint. It is a fact of life. The subscription services are an inducement to contribute and so pragmatically they cannot be offered separately. If the services were offered separately, people would be less motivated to contribute beyond the cost of the subscription. (I also decline the tote bags offered from 501(c)(3) institutions X, Y, and Z.)
https://www.mutualfundobserver.com/support-us
https://www.mutualfundobserver.com/support-us/501c3/
If I were still really deep into data analysis, I'd likely import that spreadsheet into a DBMS. These days I've got better things to do with my time: going on a river cruise out to the Black Sea, researching a trans-Canada railway trip, fighting with my insurance company on the network status of my healthcare providers (the insurer has processed over half of them incorrectly).