Close

Complete Captcha

• 12 years - 53 posts - Latest - RSS
• Thread: 26 Goods vs 11 No Goods

1. Economist
   7c18
   

   The economic question is basic. Does leverage behave like a random walk? Does leverage mean revert?

   It is well known that if something is measured with error, then on average the error term washes out subsequently, creating one possible source of reversion. But it is far from the only possible source of mean reverting behavior. The underlying corporate cash flows could mean revert, or the managers could take offsetting actions. In fact there is direct evidence of off-setting issuing activity. How big it is, and how fast it is are debated. But it is there in the data.

   Go to FINRA (http://cxa.marketwatch.com/finra/MarketData/Default.aspx) and extract the debt issues. Then go to SDC and extract the issuing data. That will give you actual security issues by real firms.

   Then go to Compustat (or to Edgar and get the annual reports balance sheets if you really think that Compustat is too error prone). Calculate the leverage ratios for say the S&P 500 firms. Then examine the patterns in the subsequent issuing decisions and study whether, and how those issues are related to the leverage ratios -- and other factors. Then come back and tell us whether you still believe that it is all just a statistical regression to the mean.

   Cheap and easy simulation can illustrate that something is logically possible. It does not show that other things are not.

   12 years ago # QUOTE 2 Volod 1 Vlad !

2. Economist
   ce43
   

   I don't get what c182 is trying to say. Of course patterns may be genuine. But the LRZ shoddy job does not tell us whether their result is genuine or not. The burden of doing what c182 wants fell on LRZ not the posters, especially since they did not get published in a shitty place. I am not Ivo Welch, by the way. This thread is devolving into interesting punches and counterpunches but the basic problem remains, LRZ is crap (and didn't the Aussie paper show this anyway?).

   12 years ago # QUOTE 1 Volod 0 Vlad !

3. Economist
   788b
   

   c182: You can remove the effect of partitioning and get an estimate which is not expected to revert. Then, you need to correct for seasonal effects which yield reversion too. After all that is done, maybe there is something there. Maybe. LRZ failed because they did none of this and were, instead, amazed to see reversion which might very well be mechanical. It's like a caveman who is amazed by a flashlight: I wouldn't trust him to tell me if the bulb is using some new technology.

   People here provided code (simple code!) to replicate the effect. That didn't show it is possible, it showed that mechanical reversion is the null hypothesis you have to disprove. If you think the effect is real, how about you provide code to create it and show it is not purely mechanical?

   12 years ago # QUOTE 1 Volod 0 Vlad !

4. Economist
   3a98
   

   I observe actual issuing of financial securities that suggests that it is not just mechanical. This is not simulation based. It is data. Furthermore I do not see that measurement error is the right null.

   12 years ago # QUOTE 0 Volod 1 Vlad !

5. Economist
   6905
   

   ^ this pretty much summarizes the disgraceful bias in our profession
   If you are insider, you don't face the burden of proof. If you are not, well, you are against a MUCH higher bar.

   12 years ago # QUOTE 2 Volod 0 Vlad !

6. Economist
   6905
   

   Let's see now. With a986, we appear to not only be dealing with an insider, but also a dumb insider. The process is supposed to consider and rule out reasonable mechanical explanations before publication. It didn't. End of story 1. Let's consider this wikipedia example of the regression fallacy: "The student did exceptionally poorly last semester, so I punished him. He did much better this semester. Clearly, punishment is effective in improving students' grades." The convergence result in LRZ is similar to this assertion. To not rule out the fallacy explanation is bad research. End of story 2.

   12 years ago # QUOTE 2 Volod 0 Vlad !

7. Economist
   f23f
   

   "With a986, we appear to not only be dealing with an insider, but also a dumb insider." Isn't that a redundant statement in empirical corporate (at least at the median). **** me.

   12 years ago # QUOTE 1 Volod 0 Vlad !

8. Economist
   5f3d
   

   "I observe actual issuing of financial securities that suggests that it is not just mechanical. This is not simulation based. It is data. Furthermore I do not see that measurement error is the right null."
   ---
   If I can replicate what your data "say" with white noise, then what you think your data is saying is GARBAGE AND MEANINGLESS.

   This paper is exactly the type people hold up as an example of how stupid people are in finance. You are a pimple on the ass of out profession. GET OUT OF MY PROFESSION, YOU ****ING RETARD.

   12 years ago # QUOTE 2 Volod 0 Vlad !

9. Economist
   3a98
   

   http://en.wikipedia.org/wiki/Regression_fallacy

   "On the other hand, dismissing valid explanations can lead to a worse situation. ...
   In essence, misapplication of regression to the mean can reduce all events to a "just so" story, without cause or effect. (Such misapplication takes as a premise that all events are random, as they must be for the concept of regression to the mean to be validly applied.)"

   12 years ago # QUOTE 0 Volod 0 Vlad !

10. Economist
    5f3d
    

    Yes, I (88bf at work) admitted as much. I even told you how to correct for that effect (and others) in order to see if the effect is real. However, we don't just assume an effect is real; we must prove it.

    You might have been good enough to say thank you for the pointers and move on. If you are one of LRZ, you might have even piped down, admitted (later+elsewhere) that you screwed up a la Oster, and moved on. If you were *really* ethical, you would have even donated money to EJMR in thanks for such anonymous (and free) criticism. Instead, you are so blind you now think we should take reversion as proven until refuted?

    Let me repeat myself: GET OUT OF MY PROFESSION, YOU ****ING RETARD.

    12 years ago # QUOTE 1 Volod 0 Vlad !

11. Economist
    e934
    

    I'm convinced a986 is a dimwit. Nobody is "dismissing" any explanation. We are merely saying that it is bad research to not rule out a fallacy first. I agree that we would be better off without a986 types...Jeez we could hit one of these jokers as a referee or AE and Cam as an editor...and then we are truly ****ed...Also, a986's last post omits the example Wikipedia makes to ullustrate his point (the ellipse part):
    "After the Western Allies invaded Normandy, creating a second major front, German control of Europe waned. Clearly, the combination of the Western Allies and the USSR drove the Germans back.

    Fallacious evaluation: "Given that the counterattacks against Germany occurred only after they had conquered the greatest amount of territory under their control, regression to the mean can explain the retreat of German forces from occupied territories as a purely random fluctuation that would have happened without any intervention on the part of the USSR or the Western Allies." This is clearly not the case. The reason is that political power and occupation of territories is not primarily determined by random events, making the concept of regression to the mean inapplicable (on the large scale)."

    Drawing an analogy...it is completely obvious that the territory was acquired by military might and not at random. Unfortunately, there's nothing that obvious that says capital structure doesn't have randomness in it...so the quote is not directly applicable...Regardless of this, again, the onus is on LRZ to show otherwise...

    12 years ago # QUOTE 1 Volod 0 Vlad !

12. Economist
    dbd7
    

    Guys, this issue has been settled already: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1689713

    11 years ago # QUOTE 1 Volod 1 Vlad !

13. Economist
    2b19
    

    Yes, we know. Some doubted it was true. Q36 posted R code showing it was mechanical. This paper was known about then -- although it just hit JF.

    11 years ago # QUOTE 0 Volod 0 Vlad !

Reply

Post

Send Post »

Markup: a blockquote code em strong ul ol li.