Dumb Quant - EBITDA/EV

[cobafdek]

For minority interest (in a perfect world) I would at least try to find out the industry P/B ratio and apply it to the book value of whatever minority interest was.

 

This adjustment makes good sense.  Do you have a reference for it (Damodaran?).

 

On the "bad data" stocks that show up in screens in other places: These honestly don't seem to be showing up in Compustat.  For real.

 

I figured this wouldn't grossly skew the data since most companies in the US have just one class of stock.

 

Even if the "bad data" stocks were included in the cohort, maybe there aren't enough of these bad apples to spoil the whole bag, and the final results might not be too different?  The key might be wide diversification, and the bag has to be big enough. 

 

So if you have any spare time (!), you could run the screen on Compustat unadjusted data to see how sensitive the results would be if one "erroneously" included the false positives.

 

 

 

[west]

For minority interest (in a perfect world) I would at least try to find out the industry P/B ratio and apply it to the book value of whatever minority interest was.

 

This adjustment makes good sense.  Do you have a reference for it (Damodaran?).

 

On the "bad data" stocks that show up in screens in other places: These honestly don't seem to be showing up in Compustat.  For real.

 

I figured this wouldn't grossly skew the data since most companies in the US have just one class of stock.

 

Even if the "bad data" stocks were included in the cohort, maybe there aren't enough of these bad apples to spoil the whole bag, and the final results might not be too different?  The key might be wide diversification, and the bag has to be big enough. 

 

So if you have any spare time (!), you could run the screen on Compustat unadjusted data to see how sensitive the results would be if one "erroneously" included the false positives.

 

On the adjustment, yes, that's Damodaran's technique.

 

On the bad data screen, you are correct on my spare time :).  I'm about two weeks backlogged with stuff to do, and then I'll be visiting relatives for Thanksgiving.  So maybe at the beginning of December?  PM about this around then and I'll see what I can do ;)

[west]

Thanks for your answer. I used my free time on screener.co already and was pleased with the data, but i am just to cheap to pay for data. Perhaps i should just do it.

I read on alpha architect that a shiller PE screen had very pleasing results, too. Do you think when you take the average EBITDA of the past 5-8 years you get better results? Or did you do that already?

 

I missed this post.

 

From everything I've seen, using a normalized EBITDA value underperforms using a TTM value.  However, I'm still in denial about this fact since it just doesn't make any sense to me.  So, until I collect more data, I'm just going to assume nothing about EV/TTM EBITDA versus EV/Normalized EBITDA :)

[NormR]

Thanks for your answer. I used my free time on screener.co already and was pleased with the data, but i am just to cheap to pay for data. Perhaps i should just do it.

I read on alpha architect that a shiller PE screen had very pleasing results, too. Do you think when you take the average EBITDA of the past 5-8 years you get better results? Or did you do that already?

 

I missed this post.

 

From everything I've seen, using a normalized EBITDA value underperforms using a TTM value.  However, I'm still in denial about this fact since it just doesn't make any sense to me.  So, until I collect more data, I'm just going to assume nothing about EV/TTM EBITDA versus EV/Normalized EBITDA :)

 

Seems that price to normalized book value works fairly well.  (I'm still trying to think that one through.)

[opihiman2]

By the way, when you did your back tests, how did you handle the sell and buy transactions?  Did you take transaction costs into account?  What about short term/long term taxes?  And, finally, did you rebalance the portfolio?  I think those are all important to consider when doing back testing vs indexing benchmarks.

 

Transaction costs were not considered.  Luckily, at least with IB and non-penny stock stocks, this is close to being true for me :)

 

I didn't consider taxes as well.  However, I did the study over one year periods exactly so that in theory people could take the Greenblatt approach to taxes, i.e., sell the losers one day before one year to get short-term capital losses on them, and sell the winners one day after one year to get long-term capital gains on them.

 

On rebalancing: I just assume that at the end of each year the positions were all liquidated to cash and put into whatever the top picks were at the current time.

 

That is not good.  The rebalancing means there is a huge turn over ratio.  Based on how this back test was constructed, the actual performance of all of those stock screening strategies is probably about 0.75 actual returns--I'm going off a some simple math, but I just found a research article that also said taxes in an active mutual fund will take off as much as 25% off CAGR.  But, you're right.  Using IBKR, transaction costs are negligible. 

 

So, the best CAGR is now down to 12.75% CAGR.  Better than indexing, but not by much.  It makes one wonder if it's worth the effort.

[frommi]

That is not good.  The rebalancing means there is a huge turn over ratio.  Based on how this back test was constructed, the actual performance of all of those stock screening strategies is probably about 0.75 actual returns--I'm going off a some simple math, but I just found a research article that also said taxes in an active mutual fund will take off as much as 25% off CAGR.  But, you're right.  Using IBKR, transaction costs are negligible. 

 

So, the best CAGR is now down to 12.75% CAGR.  Better than indexing, but not by much.  It makes one wonder if it's worth the effort.

 

But you only get your tax advantaged index fund returns when you never sell and don`t get dividends. I am under the impression that these studies are made for people who want to sell index funds. From the back of my head compounding of tax advantages really kick in after 5 year holding periods or so, so any value investing that sells near fair value has these problems.

[opihiman2]

Also, Compustat and CRSP databases are the best equities database out there.  They try their best to get rid of many known biases, for example, survivor-ship bias (which is a BIG but problem that most people don't realize when they run back tests on stupid finance sites, i.e. Yahoo, that allow you to export their data).  Although, even with best efforts, there are still some errors and bias in those databases.  Although, I think they are pretty negligible at this point.  If someone said they ran a back test using those two databases, I would be pretty confident in at least the data portion.  The actual test construction, well, sorry west, that was a poorly constructed back test. 

 

By the way, I've been following NormR's stock screening portfolio's over the years.  He has two, I believe, that both follow a value investing strategy.  Anyways, I've been looking at the tax implications of the strategy: hold for one year, sell at the end, run the screen again, invest again.  CAGR performance of the Graham screener run in real time over the past 12 years was a little over 20.1%.  It is pretty fantastic.  However, when one takes taxes into consideration (I'm using 25% capital gains tax), the CAGR comes down to 15.3%.  Much better than S&P, but still not as good as it seems.  You gotta take those taxes into account.

[opihiman2]

That is not good.  The rebalancing means there is a huge turn over ratio.  Based on how this back test was constructed, the actual performance of all of those stock screening strategies is probably about 0.75 actual returns--I'm going off a some simple math, but I just found a research article that also said taxes in an active mutual fund will take off as much as 25% off CAGR.  But, you're right.  Using IBKR, transaction costs are negligible. 

 

So, the best CAGR is now down to 12.75% CAGR.  Better than indexing, but not by much.  It makes one wonder if it's worth the effort.

 

But you only get your tax advantaged index fund returns when you never sell and don`t get dividends. I am under the impression that these studies are made for people who want to sell index funds. From the back of my head compounding of tax advantages really kick in after 5 year holding periods or so, so any value investing that sells near fair value has these problems.

 

Well, I think that's the point of indexing.  Invest for the long term, cost average over that period, and take a one time long term capital gains tax.  Whereas, an actively traded system has a compounded long term capital gains tax.  So, for example, in NormR's Graham screener.  At the end of 12 years, if someone followed the screen, someone would end up with 15% CAGR vs 20% CAGR as shown in the hypothetical tax free account.  But, if one were trading in an IRA and took out total compounded return, one only takes the long term tax hit once.  The tax costs are better.  The CAGR following this strategy is a little over 17.5%. 

 

So, taxes do matter quite a lot.  As for the dividends, that would happen even with the screener strategy, so that point is moot.

 

[west]

Also, Compustat and CRSP databases are the best equities database out there.  They try their best to get rid of many known biases, for example, survivor-ship bias (which is a BIG but problem that most people don't realize when they run back tests on stupid finance sites, i.e. Yahoo, that allow you to export their data).  Although, even with best efforts, there are still some errors and bias in those databases.  Although, I think they are pretty negligible at this point.  If someone said they ran a back test using those two databases, I would be pretty confident in at least the data portion.  The actual test construction, well, sorry west, that was a poorly constructed back test. 

 

I'm confused, but I'd like to understand what you're trying to say.  Why was my back test poorly constructed?  If there's something wrong about it, I'd like understand why it's wrong so I can avoid making the same mistake in the future.

 

On taxes, and all other talk about the "conclusions" from this "study" (in quotes because calling what I did a study is a bit of stretch):  I am merely posting data here!  I've said it before, but I'll say it again: Any conclusions from the data are left to the reader to make! :)

[oddballstocks]

Do taxes matter?  I know more than 50% of our portfolio is in tax advantaged accounts.  Considering how much money an American can stuff in there a year I don't know why people wouldn't be investing in there.  I can buy and sell all day long with no tax consequences in an IRA.

 

In our taxable account I'm more aware of taxes, but I will always sell a company when it hits fair value or is over valued rather than sitting on it for tax reasons.

 

As Walter Schloss said, don't let the tax tail wag the dog..

[Hielko]

So, taxes do matter quite a lot.

For some people, yes. Not for everybody. As a Dutch citizen I don't pay any capital gains taxes* so it doesn't matter how often securities are traded. And I bet plenty of institutions can trade tax free, and as Nate points out: even US individuals can do that inside an IRA.

 

* That's technically not 100% true since we pay a 30% tax on a 4% hypothetical annual capital gain (so it's absolutely irrelevant what you exactly own and how much you trade). Think it's a pretty decent system since it's so much easier w.r.t. reporting and calculating taxes.

[cobafdek]

So, the best CAGR is now down to 12.75% CAGR.  Better than indexing, but not by much.  It makes one wonder if it's worth the effort.

 

Therein lies the irony and the larger point:  it was hardly any effort, and still you get pretty good results!

 

However, when one takes taxes into consideration (I'm using 25% capital gains tax), the CAGR comes down to 15.3%.  Much better than S&P, but still not as good as it seems. 

 

I admit I'm a slacker to be contented with 15% annually, but if I knew I could achieve such a result long-term, I could have retired more than 20 years ago.  I guess my expectations are different, and I'm not swinging for the fences trying to achieve Buffett-like 20+% returns.  So to my mind, your numbers support this lazy approach.

 

The actual test construction, well, sorry west, that was a poorly constructed back test. 

 

Again, you may be inadvertently supporting west's work!  That's the lovely paradox I see, that a "poorly constructed back test" can have such great results.  It's the Dumb Quant approach.  I was being only half-jocular earlier when I suggested leaving in the "bad data" in order to push the boundaries to see how well a Really Dumb Quant approach might perform.

 

This field is rife with paradox and counterintuitive findings.  That's why I'm glad guys like west are looking at the data independently.  One unexpected result from quant research is that within the basket of net-net stocks, the group that had a net loss in the prior year performed better than those that had a positive net income.  Who woulda thunk it?  Nothing is obvious or cleanly predictable in economic/social science.  (A lot of you guys on this board don't need to work so hard!)

 

 

 

[NormR]

I might also caution west on the micro-cap end of things.  The data might be good and yet result in rather over optimistic return estimates due to bid-ask spread / trading volume effects.  I'm not against investing in small fry, but many studies run into real trouble there.

 

I'd also say that the ranking of the "best ratio" to use has a habit of varying over time.

 

Nonetheless, I'm rather fond of several simple quant strategies.   

[opihiman2]

Also, Compustat and CRSP databases are the best equities database out there.  They try their best to get rid of many known biases, for example, survivor-ship bias (which is a BIG but problem that most people don't realize when they run back tests on stupid finance sites, i.e. Yahoo, that allow you to export their data).  Although, even with best efforts, there are still some errors and bias in those databases.  Although, I think they are pretty negligible at this point.  If someone said they ran a back test using those two databases, I would be pretty confident in at least the data portion.  The actual test construction, well, sorry west, that was a poorly constructed back test. 

 

West, a good back test will take things like slippage, taxes, rebalancing, and transaction costs into account.  One of the more important things is slippage.  Not everyone will get a trigger in real time that says, "Stock X just met the screen, buy NOW at this price!"  In reality, people will end up buying the equity at a different price, sometimes dramatically so, than what the screen in a back test would have bought it at.  So, you really need to run a back test with pricing parameters that are intervals of 5-10% around the price.  The overall CAGR will then also be a range of #'s.  When I don't see all of these things in a simple back test, it really makes me question the overall methodology, and I pretty much just ignore the results. 

 

I'm confused, but I'd like to understand what you're trying to say.  Why was my back test poorly constructed?  If there's something wrong about it, I'd like understand why it's wrong so I can avoid making the same mistake in the future.

 

On taxes, and all other talk about the "conclusions" from this "study" (in quotes because calling what I did a study is a bit of stretch):  I am merely posting data here!  I've said it before, but I'll say it again: Any conclusions from the data are left to the reader to make! :)

[opihiman2]

Do taxes matter?  I know more than 50% of our portfolio is in tax advantaged accounts.  Considering how much money an American can stuff in there a year I don't know why people wouldn't be investing in there.  I can buy and sell all day long with no tax consequences in an IRA.

 

In our taxable account I'm more aware of taxes, but I will always sell a company when it hits fair value or is over valued rather than sitting on it for tax reasons.

 

As Walter Schloss said, don't let the tax tail wag the dog..

 

Hell yeah taxes matter.  I don't think most people have a sizeable amount of their liquidity in a tax advantaged account like an IRA.  Besides, I think it's very important to disclose these things in the methodology and conclusion, otherwise, it would be highly misleading.  I mean, 17% CAGR sounds fantastic, and to the average Joe with no deeper understanding of these things, they may just plow right into it without knowing that the taxes will take a significant chunk of the performance.  They will end up heavily disappointed. 

 

I'm just saying, if you're going to provide some deeper statistics like draw downs, you should be complete and more accurate by including the tax implications of the strategy.  Churning the entire portfolio over every year IS a big disclosure one should make.

[opihiman2]

I might also caution west on the micro-cap end of things.  The data might be good and yet result in rather over optimistic return estimates due to bid-ask spread / trading volume effects.  I'm not against investing in small fry, but many studies run into real trouble there.

 

I'd also say that the ranking of the "best ratio" to use has a habit of varying over time.

 

Nonetheless, I'm rather fond of several simple quant strategies. 

 

Yep, and that's the real paradox of these back tests, especially in small micro-caps.  Actually utilizing the strategy may cause the strategy to perform differently. 

 

Anyways, by the way, Norm, it was your site that gave me the push and courage towards automated investing.  I haven't touched a 10 q in ages.  I'm much better off for it.

[NormR]

Anyways, by the way, Norm, it was your site that gave me the push and courage towards automated investing.  I haven't touched a 10 q in ages.  I'm much better off for it.

 

Thanks opihiman2, I hope it has worked well for you.  We have a bit of a minority opinion around these parts.  But there is always something to learn too :)

[oddballstocks]

Do taxes matter?  I know more than 50% of our portfolio is in tax advantaged accounts.  Considering how much money an American can stuff in there a year I don't know why people wouldn't be investing in there.  I can buy and sell all day long with no tax consequences in an IRA.

 

In our taxable account I'm more aware of taxes, but I will always sell a company when it hits fair value or is over valued rather than sitting on it for tax reasons.

 

As Walter Schloss said, don't let the tax tail wag the dog..

 

Hell yeah taxes matter.  I don't think most people have a sizeable amount of their liquidity in a tax advantaged account like an IRA.  Besides, I think it's very important to disclose these things in the methodology and conclusion, otherwise, it would be highly misleading.  I mean, 17% CAGR sounds fantastic, and to the average Joe with no deeper understanding of these things, they may just plow right into it without knowing that the taxes will take a significant chunk of the performance.  They will end up heavily disappointed. 

 

I'm just saying, if you're going to provide some deeper statistics like draw downs, you should be complete and more accurate by including the tax implications of the strategy.  Churning the entire portfolio over every year IS a big disclosure one should make.

 

No interest in getting into a nit-pick fight but I have one thought. Everyone I know who isn't related to finance or investing in any manner ONLY has money in tax advantaged accounts. Most Americans can't even put away the $27k you get to shelter each year. So in a sense arguing over taxes limits the discussion to people who have wealth outside of retirement savings.