Talk:Equity premium puzzle

From Wikipedia, the free encyclopedia

Contents 1 New Article 2 . 3 Please provide Reference 4 Wiki Article on Shiller 5 Reverted complete replacement of article 6 With all due respect... 7 Possible explanations 8 Implications

[edit] New Article

I was the one that changed the current article in the Wikipedia for the “Equity Premium Puzzle”. The reason I did was because I thought the current version was very imprecise, incomplete, and in many statements plain wrong. I thought most people would benefit from a more complete and detailed version.

For example, in the first line of the current article it says:

“The equity premium puzzle refers to the phenomenon that observed average annual returns on stocks over the past century are higher, by approximately 6 percentage points, than returns on government bonds.”

This statement make very little sense. The fact that the historical equity premium is about 6% does not constitute a puzzle per se. The puzzle comes from the fact that standard neoclassical growth theory can not replicate such a large historical equity premium. This point is layout very clearly in the version I proposed and placed in wikipedia.

Furthermore, contrary to what is stated in the wikipedia article, the puzzle says nothing about the existence of arbitrage opportunities. In fact this is a very common misinterpretation of the puzzle. Arbitrages are investment opportunities with zero initial cost that lead to positive profit with no risk in the near future. The fact that the equity premium is so large could well be a reflection of risks that we economist do not know how to model and quantify yet.

Another incomplete and poor statement was:

“A large number of explanations for the puzzle have been proposed. These include a contention that the puzzle is a statistical illusion, modifications to the assumed preferences of investors and imperfections.”

Which studies state that the puzzle is a statistical illusion? What does the writer mean by imperfections?

Also emphasizing Benartzi and Thaler (1995) as the main solution to the puzzle is very bias and incorrect. There are several studies with insightful and interesting solutions for the puzzle, written in the last 20 years! I don’t think the wikipedia article should bias towards one solution over all others. In the version I proposed, a large number of solutions are pointed out.

I have read the original article of Mehra and Prescott and even replicated most of their results. Plus have also read Shiller’s articles. In fact, Shiller is known for a different puzzle called the volatility puzzle. It would be incorrect to attribute Shiller’s work to the equity premium puzzle.

I suggest we incorporate into the new article I proposed, some of the things you found interesting from the initial version. I think many students of economics and finance would benefit from a more complete, detailed, and correct version of the equity premium puzzle.

Hope you find this helpful. I will think about a new version, but please let me know your thoughts first.

There was no doubt things you could incorporate. A better way to do that would probably be to build off the existing article, though, deleting where necessary and expanding where appropriate. The trouble with what you had in was that it looked and read like an essay and it deleted quite a bit of the useful information which is here now. But edit away, this isn't perfect by any means. Much of the info here was based off the thaler article because that's what I had studied at the time, not because I had any particular barrel to push. I'm sorry I can't be of much more help, it's been too long since I did this stuff and I no longer have my material. The thaler article might have references for things like the claims it was a statistical illusion, though, if you want to scan it (that might have been another article, though, I can't remember).

Also, I believe the article does mention arbitrage... the sentence after the one you quoted is "Economists expect arbitrage opportunities would reduce the difference between returns on these two investment opportunities to reflect the risk premium investors demand when investing in relatively more risky stocks." and I think the nature of the puzzle (in line with what you said) is explained reasonably well if you read the full first para (eg. "That is, economists predict the difference in returns between these two investments should be much smaller than 6 percentage points" and the statistical example). Psychobabble 10:05, 11 September 2006 (UTC)

I agree that the current article does not even do justice to the subject of the equity premium. There have been hundreds of papers on the topic (The original article has over 1000 citations), and citing just 2 papers as possible explanations sheds poor light on the puzzle. I would rather that no explanations be offered - and the reader instead be referred to the review articles. I believe that the article needs to be reworked quite a bit. —The preceding unsigned comment was added by 131.215.220.112 (talk) 18:29, 24 January 2007 (UTC).

[edit] .

I totally couldn't understand this! There's a premium for equities because they're more risky, okay, got that, so what's the puzzle? Here's hoping that one day someone who knows economics and can write will find this page!

I agree only a wild risk seeker would prefer 0/100,000 to 51,209. I imagine a lot of people would prefer $48,791 to 50% chance of $0/$100,000. That is what you would expect. I think the question is actually 50,000/100,000 vs 51,209 and I have changed it back to that. However, I suggest there is missing information: the risk premium of 6% is an annual 6% so the information required is the time period. crandles 14:26, 10 September 2005 (UTC)

The article which this example is taken from (Benatzi and Thaler) explain how this is derived. And yes, it is (as initially written) an implication of indifference between a bet with a 50% chance of $50,000 or $100,000 and a certain payoff of $51,209 which is a crazy high level of risk aversion.Psychobabble 22:07, 26 November 2005 (UTC)

Answer: I'd propose you have a look at the original paper. There is no doubt that most people are risk averse. If they are risk averse, it means they want to be compensated for taking risks. Hence, market prices of risky assets (assumed to be shares) must be such that expected payoffs are higher than with risk-free assets (assumed to be interest bearing papers).

The question is, how big is risk aversion. Mehra and Prescott looked at standard risk aversion as implied from other research (sorry, don't have the paper in front of me, so no details can be provided here). With this level of risk aversion and the risk of shares, it is implied that shares would only pay of 0.35% more than treasury bills. The puzzle is that premium was found to be 6.18%, so people seem to be way more risk averse with respect to shares than in other circumstances.

Hope this helps.

[edit] Please provide Reference

Can someone supply a reference for the quantitative risk adversion implication (the 100% of 51k vs 50% of 50k/50% of 100k) example. I think it is important to demonstrate how this figures was arrived at and assumptions behind it. o/w it just seems like a number out of thin air. novacatz 04:07, 25 November 2005 (UTC)

I wrote that part of the article and the reference is the Benartzi and Thaler artic/le in the references section. Is it necessary to footnote an article which is in the references section? Psychobabble 22:02, 26 November 2005 (UTC)

I think in this case it would be handy, for instance -- I wanted to dig up the working behind that calc. and sure the ref is there (one among four) -- it would be nice to know directly which reference to look at. novacatz 13:36, 5 December 2005 (UTC)

[edit] Wiki Article on Shiller

It seems that Shiller might be the first in this arena, not the authors cited? Here is a fragment: Long before he became famous for his book Irrational Exuberance, Shiller developed his thinking on this subject in a series of studies published in academic journals, most notably with an article entitled Do stock prices move too much to be justified by subsequent changes in dividends?, published in the American Economic Review in 1981.

There is a link to excess volatility, which I'll try to spell out in a revision, but the equity premium is a distinct phenomenon.JQ 01:52, 29 August 2006 (UTC)

[edit] Reverted complete replacement of article

A change a while back replaced the entire article as it stood with an essay that was in fairly poor shape, and had attracted a cleanup notice. I took a quick look and decided that it would be better to go back to the previous version, then incorporate anything worthwhile from the new one. I'll try to tackle this before too long.JQ 01:50, 29 August 2006 (UTC)

Gotcha. It did look like someone had simply cut-pasted their uni economics essay onto the page. Psychobabble 01:49, 29 August 2006 (UTC)

[edit] With all due respect...

Although the replacement article was in poor style for an encyclopedia entry, it does have certain advantages over the current version. First, I think it is remarkable that the current article does not so much as mention Mehra and Prescott in the body of the article. This was the spur for all subsequent research, it absolutely should be prominently placed. Second, the list of articles at the end of the replacement article makes a much better account of the attempts to explain the epp than the current article, which mentions only Benartzi and Thaler and the Ben-Haim book. If someone came to this article without knowing much about the epp, they'd leave believing these to be the only two attempted explanations. That's a plain awful summary of the research in this area over the last 20 years or so.

Also, the contribution about the Ben-Haim explanation (by Mr Ben-Haim himself?) is in terrible style. "We model this behavior..."? This is not a working paper.

Jackie hayes 05:39, 31 October 2006 (UTC)

I'm pretty sure some of this was in earlier versions. I've put (back) a reference to Mehra and Prescott, and will try and do some more when I get time. I agree that the Ben-Haim stuff needs work, and to be put in the context of the many proposed explanations.JQ 09:01, 31 October 2006 (UTC)

There are way too many explanations that try to resolve the equity premium, neither of them being successful. The original Mehra-Prescott paper has about 1500 citations now. There are several approaches - habit formation, alternative preferences, survivor bias etc. Moreover, the puzzle is not so much about the equity premium per se, but the fact that the implied risk-free rate by the model is so high that the equity premium predicted by the model is too small. Recent work focuses on non-representative agent models such as overlapping generation models, with constraints on borrowing, bequest motivations etc.

I suggest just cleaning up the entry to explain just the puzzle, and leave out all the explanations. The possible explanations can be a separate wikipedia entry - perhaps several entries. Therealgandalf 22:28, 8 October 2007 (UTC)

I think this move is at best, premature, and have reversed it. At the moment it just produces two stubby and unsatisfactory articles. If the article becomes so long that the explanations section is unmanageable, it might be appropriate. But, as you say, we need to improve the description of the puzzle as well as the explanations.JQ 22:55, 8 October 2007 (UTC)

[edit] Possible explanations

The section on possible explanations

• was written by several different people over a period of several months;
• cites reliable sources;
• is mostly understandable, at least to me, a person with no training in business whatsoever.

It might not be perfect—few pages in WP are—but it's certainly not original research. If you have a problem with it, either improve it yourself or state your complaint here, rather than just deleting the whole section. (Why do I have to explain things like this to an admin?) --Zvika 07:06, 9 October 2007 (UTC)

I've had a first go at a rewrite, first cut only.JQ 12:23, 9 October 2007 (UTC)

JQ, I like your edits. I do think that self-citing is against Wikipedia guidelines, and in general frowned upon. There are any number of works that cover this topic, and I think it would be unfair to the literature to just cite one paper. This is precisely why I separated the entries. Could you please remove the self-citations ? Therealgandalf 22:25, 9 October 2007 (UTC)

As regards the guidelines and self-citation, WP:COS says this is OK. I've drawn from my paper because it's a survey to which I (unsurprisingly) know well and have easy access to. But it would be good to have more from Kocherlakota and Mehra, and I'll try to add some when i get a chance.JQ 22:44, 9 October 2007 (UTC)

My working understanding of the process is based on the divergence of cost and income, the average performance of a company may be found to run the National average GDP growth +/- and by extention globally, individual corporate growth may be found to be a microcosim of the theory.

The result obviously suggests that not much has been done to improve investor perception of risk in the equities market, one thing is clear, positive fundamental performance will continue to attract intrest, as to whether we can measure the emotions of investors, now that's another thing... —Preceding unsigned comment added by Lloydastewart (talk • contribs) 20:20, 21 February 2008 (UTC)

Positive and negative pressures at play as demand and supply are finite, should the theory be true I will name it". —Preceding unsigned comment added by 72.252.36.120 (talk) 17:30, 19 February 2008 (UTC)

[edit] Implications

Admittedly, I am not familiar with the paper cited in this section. Nevertheless, it seems that these are not direct implications of the equity premium and/or high risk aversion and that they must rest on additional assumptions. For instance, the benchmark model of the firm would involve managers maximizing shareholder value. Managers acting myopically would rest on what is assumed about managerial compensation. This doesn't seem to have much to do with the equity premium puzzle. I question whether this section warrants inclusion in the article. If an editor feels it does, perhaps he could better illustrate how the equity premium puzzle is connected to these implications. Wik-e-wik (talk) 23:47, 29 April 2008 (UTC)

I have now read the article. It speculates on the implications various explanations of the equity premium puzzle might have for policy issues. As I suspected, the implications are not implied by the existence of the equity premium puzzle, but rather by various explanations of the puzzle. In my opinion this is recent research that does not belong in an encyclopedic article. I'm removing this section and the reference the paper. Wik-e-wik (talk) 00:56, 9 May 2008 (UTC)

I've reverted. If you wish to pursue this line, quote the policy under which recent (published and peer-reviewed) research doesn't belong in Wikipedia. You are correct in saying that the implications of the equity premium differ, depending on the explanation. Maybe rather than blanking material, you could edit to include this point. JQ (talk) 04:20, 9 May 2008 (UTC)

I have no problem with relevant peer reviewed recent research being included in wikipedia and in fact find it one of wikipedia's strengths. My issue is with the relevancy of the information contained in this section. My main point is that this is a list of implications two authors deduced from certain explanations of the existence of an equity premium. Since the implications follow from explanations rather than the puzzle itself this section seems tangential to me. I have no interest in getting involved in an edit war and as such won't redelete this section. However, I would urge you to follow the guidelines in WP:COI and yield to the consensus view when citing your own work.Wik-e-wik (talk) 20:04, 9 May 2008 (UTC)