Empirical Legal Studies

Indiana Law Reception in Montreal May 30

Friday, May 30, 2008
6:30 pm - 8:30 pm

Restaurant Julien
1191, Avenue Union
Montreal (Quebec) H3B 3C3

Readers of the ELS Blog are invited to join Dean Lauren Robel and members of the Indiana Law faculty for a reception in conjunction with the 2008 Law and Society Association Annual Meeting. The gathering is a great opportunity to network, socialize, and enjoy hors d'oeuvres and drinks. We hope you can join us.

We would appreciate an RSVP by May 26 to lawalum@indiana.edu or 812-855-9700; feel free to attend even if you are not able to RSVP now.

BJS Report on Felony Defendants

Another recent BJS report, Felony Defendants in Large Urban Counties, 2004, analyzes data collected from a representative sample of felony cases filed in the nation's 75 largest counties during May 2004. Murder cases were tracked for up to 2 years and all other cases for 1 year to provide an overview of the processing of felony defendants from case filing to disposition and sentencing. Data highlight the demographic characteristics of felony defendants and types of arrest charges. The report also includes in-depth information on the criminal record of felony defendants, including criminal justice status at the time of arrest and the number and type of prior arrests and convictions. It describes conditions of pretrial release (bail amounts, type of release bonds, and pretrial misconduct), adjudication outcomes (dismissal, diversion, guilty plea, trial conviction rates), and sentencing data for convicted felony defendants. Notable findings include:

• Two-thirds of felony defendants were charged with a drug or property offense.
• More than three-fourths of felony defendants had a prior arrest history, with 53% having at least five prior arrest charges.
• Fifty-seven percent of felony defendants received a pretrial release prior to adjudication.

Warren Buffett and Efficient Market Theory

I like Warren Buffett--or at least his public persona and mythology--for a lot of reasons.  He makes a lot of money without much use for computers or technology, he claims to invest only in what he understands (e.g., insurance, candy, etc.), he treats people with respect (including his shareholders), and he says that his taxes are too low.  Buffett also agreed to give away 85% of this wealth to the Bill & Melinda Gates Foundation upon his death.  He could have built a monument to himself, but instead he concluded that his money would go further by giving it to Gates. 

But my favor reason might be in the below abstract from a top-ranked paper on SSRN: that  Warren Buffett et al.'s 31 year track record is not consistent with Efficient Market Theory.    See Gerald S. Martin (American/Texas A&M) & John Puthenpurackal (UNLV), Imitation is the Sincerest Form of Flattery: Warren Buffett and Berkshire Hathaway:

We analyze Berkshire Hathaway's equity portfolio over the 1976 to 2006 period and explore potential explanations for its superior performance. Contrary to popular belief, we find Berkshire Hathaway invests primarily in large-cap growth rather than "value" stocks. Over the period the portfolio beat the benchmarks in 27 out of 31 years, on average exceeding the S&P 500 Index by 11.14%, the value-weighted index of all stocks by 10.92%, and a Fama and French characteristic-based portfolio by 8.56% per year. Although beating the market in all but four years can statistically happen due to chance, incorporating the magnitude by which the portfolio beats the market makes a luck explanation extremely unlikely even after taking into account ex-post selection bias. We find that Berkshire Hathaway's portfolio is concentrated in relatively few stocks with the top five holdings averaging 73% of the portfolio value. While increased volatility is normally associated with higher concentration we show the volatility of the portfolio is driven by large positive returns and not downside risk. The market appears to under-react to the news of a Berkshire Hathaway stock investment since a hypothetical portfolio that mimics the investments at the beginning of the following month after they are publicly disclosed also earns significantly positive abnormal returns of 10.75% over the S&P 500 Index. Our evidence suggests the Berkshire Hathaway triumvirates of Warren Buffett, Charles Munger, and Lou Simpson possess investment skill unlikely to be explained by Efficient Market Theory.

So here is my question:  If you know how to make money like Buffett et al., what would you do:  apply the strategy in the market and make supra normal returns, or publish it all in a finance journal to settle an academic debate (and ruining the strategy prospectively)?

BJS Report on Domestic Violence

The Bureau of Justice Statistics (BJS) recently released State Court Processing of Domestic Violence Cases, a report (along with links to underlying data) that examines the processing of domestic violence (DV) and non-domestic violence (non-DV) cases filed in May 2002 in 15 large urban counties. In particular, the study compares the domestic and non-domestic offenses of sexual and aggravated assault on 11 prosecution, conviction, and sentencing outcome measures. Data are also presented regarding court issued protection orders, guilty plea versus trial convictions, and the demographic characteristics of domestic violence defendants. Major findings include:

• A third of violent felony defendants were charged with domestic violence.
• Prosecuted domestic sexual assault defendants had a higher overall conviction rate (98%) than prosecuted non-domestic sexual assault defendants (87%).
• Domestic aggravated assault defendants (54%) were less likely to be granted pretrial release than non-domestic aggravated assault defendants (62%).

Best Practices of Structuring a Faculty Workshop (that includes students!)

Over at TaxProf, Paul Caron has posted an immensely thoughtful commentary by Neil Buchanan (George Washington Law) on how to structure a faculty tax policy workshop that includes students.  Having participated in empirical studies workshop at Illinois that followed a similar format,  I want to strenuously endorse Neil's comments and assert that his insights need not be limited to tax policy.  Neil writes:

[O]ne of the great values of holding tax policy workshops in law schools is that we embrace an interdisciplinary approach to scholarship. Bringing together scholars, convenors, and commentators with different scholarly backgrounds (within law as well as in related fields) is our greatest strength. We do ourselves and our students [emphasis added] a great favor by being mindful not to allow our discussions to be dominated by the methodologies or assumptions of any single field of study. It is very heartening to see the interdisciplinary approach to tax policy [my edit] scholarship flowering across the country. I hope that the trend of encouraging this through tax policy workshops [involving students] at more and more law schools continues.

My efforts to organize a mini-subfield conference within a conference at Law & Society each year are based on a similar model perfected by Neil Buchanan.   In my book, he is a real thought leader.

More accurate than the polls? A Model for Predicting IN and NC Primary Results

Wondering, like me, whether the polls in the Democratic primary can be trusted?  Over at the Social Science Statistics Blog, statistics whiz Kevin Bartz has posted two very interesting models for predicting results of the Indiana and North Carolina primary.   Bartz mines a huge amount of county-level data, including likely voter turnout, and his models purport to explain between 84% and 99% of prior results from Ohio, Pennsylvania, and Virginia.  (Because his S.E. is 5%, Bartz qualifies that the "results could be shaky at the county level.")

Here are Kevin's projections:

• Indiana: Clinton 53.5%, Obama 46.5%; turnout 950,000
• North Carolina: Obama 58%, Clinton 42%; turnout 1,200,000

16% seems like a huge margin for Obama--indeed, if the above numbers are even close to right, Obama (who I had for Con Law in the fall of 2000) will likely add to his lead in pledged delegates.  If so, then perhaps we can wind up a protracted primary season that is no longer serving the interests of the Democratic Party (granted, that is just my opinion).  Kevin explains the differences between his models and the polls:

How do my results stack up against the current polls? In Indiana, the RealClearPolitics average has Clinton +6%, only a point from my prediction. In North Carolina, the RCP average has Obama +8%, significantly below my predicted 16% victory. Two factors shed light on this discrepancy:

• In neighboring South Carolina, the polling average had Obama +11.6% and he won by 28.9%.
• In neighboring Virginia, the polling average had Obama +17.7% and he won by 28.2%.
• So perhaps my analysis isn't so crazy putting Obama above what the polls say in NC.

Here is what Kevin's predictions look like using GIS mapping:

Conducting Empirical Legal Scholarship Workshop, June 23-25

On Monday, June 23, 2008 through Wednesday, June 25, 2008 Lee Epstein and Andrew Martin will be teaching their annual Conducting Empirical Legal Scholarship workshop. This workshop will be held in Chicago, and is co-sponsored by Northwestern University and Washington University.

There is more information available about the workshop here:

http://www.law.northwestern.edu/faculty/conferences/empiricalworkshop.html

The Conducting Empirical Legal Scholarship workshop is for law school and social science faculty interested in learning about empirical research. The instructors provide the formal training necessary to design, conduct, and assess empirical studies, and to use statistical software (Stata) to analyze and manage data. Participants need no background or knowledge of statistics to enroll in the workshop. Topics to be covered include research design, sampling, measurement, descriptive statistics, inferential statistics, and linear regression.

Please visit the website for all logistical details.

The Match(ing) Game

Matching methods of various sorts have become a hot topic in the social sciences lately, in part because they offer a means of untangling causal effects from observational data (see, e.g., here).

Now the latest issue of Statistics in Medicine has an interesting paper by Peter Austin surveying the application of propensity score matching in the medical literature between 1996 and 2003.  The paper is followed by a series of illuminating responses to the paper, and a rejoinder by Dr. Austin.

Beyond what the papers say about the (possible mis)use of such techniques by medical researchers, they collectively provide a nice conceptual overview of propensity score matching (and matching methods in general).  Particularly clear is the discussion of balance tests (of the sort outlined here), which Austin argues are central to the correct use of such methods.   

Storage Solutions for Large Datafiles

Over the years, I have collected a lot of datasets that I have attempted to archive on the university file server.  While computer space is not very expensive these days, it is not free.   These files have, collectively, far exceeded the standard faculty allotment.   

Fortunately, we finally found a simple solution:  The HP Personal Media Drive, which is an external hard-drive with 500GB of memory.  Years worth of files only take up a small fraction of the space.  And just by connecting an USB port, I can quickly backup my entire laptop hard-drive.  It is also very portable.  That is a lot of peace of mind for under $200. 

Interactions

Regina Branton (Rice--Poli Sci) maintains a very helpful and accessible tutorial on interactive variables in multiple regression models (here). Many of the discussions include informative examples.

"[E]mpiricism gone astray"

A very interesting and helpful post (and comments) by John Donohue (Yale) over at the Balkinization blog illustrates some all-too-common problems with empirical research. To illustrate some of these problems, John comments on a recent paper by Royce de Rohan Barondes (Missouri-Columbia), provocatively titled: Want Your Opinions Questioned or Reversed? Hire a Yale Clerk.

As John notes, "unfortunately, counterintuitive empirical results almost always turn out to be wrong if they are not based on an appropriate empirical methodology for the inquiry at hand. In my opinion, the methodology of the Barondes (paper) is flawed, and the conclusions drawn from this research are either incorrect or unfounded." (To be fair, Barondes responds to some of John's critiques in comments to John's post.)

Large Law Firm Hiring -- Introducing the "Funnel Cloud"

Why do law students follow U.S. News ranking?  In a story in last week's National Law Journal, Andrew Morriss and I claim that USNWR provides a rough guide to future employment prospects.   Yet, drawing upon data on hiring patterns from the NLJ 250 (the 250 largest firms based on number of attorneys) and employment outcomes from the ABA-LSAC Official Guide, Andy and I demonstrate the serious financial perils of overreliance on USNWR.

The story has a lot of data.  The chart below, however, ended up on the editing room floor.  I call it the "funnel cloud."  It shows the percentage of students at each law school that were hired by a NLJ 250 firm.  [Side note: Why is the NLJ 250 a good barometer?  Because the starting salaries make law school tuition look like a bargain.  Based on data made available to us through the Law Firms Working Group,  the weighted average (based on firm size) starting salary at an NLJ250 firm for 2007 was $152,424.]

In my opinion, the above chart has at least two takeaways: (1) the funnel cloud formation shows large law firm employment payoffs are non-linear and that the vast majority of schools offer similarly modest, but not insignificant, entree to this sector; (2) based on the volume of green and red at the top of the chart, most large firms prefer to recruit deeper into the class at a Top 20 school (and will pay a price premium of $160,000 per year) rather than shifting their model to lower ranked schoolss.    53% of all new NLJ250 jobs between 2005 and 2007 went to Top 20 biggest feeder schools from 2005.  See also Leigh Jones, Hiring More Deeply into Top Schools, Nat'l L. J., Apr. 14, 2007.  You would think that by 2007, the Top 20 would have been tapped out.  Not so.

The empirical findings of this 2007 article by Dinovitzer and Garth, which is based on After the JD data, suggest that this elitism has unintended negative consequences in terms of high attrition rates. 

Here are links to other graphics in our recent NLJ story:

Employment trends for law school grads: The top 100 out of 194 ABA-approved schools ranked by percentage of 2005 graduates at NLJ 250 firms.

By Region (defined by states containing Top 10 legal markets):

• New York
• California
• Illinois
• Texas
• Washington, D.C.
• Massachusetts
• Georgia
• Pennsylvania
• Alabama to Kentucky
• Louisiana to Nevada
• Ohio to Wyoming
• Composite report of all 11 regions
• Methodology

CELS 2008 -- Submission Deadline Relief

Information from the CELS 2008 Organizing Committee:

The Organizing Committee has begun reviewing papers submitted for the September 12-13, 2008 Conference on Empirical Legal Studies to be held at Cornell Law School in Ithaca, NY.  We thank all of you who have submitted papers.  We anticipate another first-rate meeting of empirically minded scholars, so please plan to attend.

In light of a number of requests for extensions and mindful of a deadline that falls near the end of a semester, the CELS Organizing Committee will continue to accept papers for consideration for the conference for a limited period after the April 15 deadline.  Papers submitted after April 15 will be considered for inclusion in the conference subject to the availability of space.   We expect a hard cutoff date of May 16.

Authors who have already submitted papers may use the SSRN system to submit a revised version. Peer reviewers will consider whichever version is available at the time they undertake their peer reviews.

For more details about the CELS 2008 conference, visit the conference web site.

Judging Prosecutors & Selection Effects

Mark Ramseyer (Harvard), Eric Rasmusen (Indiana), and Manu Raghav (Wash. & Lee) recently circulated an interesting paper, Convictions versus Conviction Rates: The Prosecutor's Choice, that compares the two main ways many assess a prosecutor's effectiveness. Prosecutors' dual roles--deciding which cases to prosecute and convincing courts to convict criminal defendants--complicate assessment efforts. For many "it is natural to suppose that a prosecutor's conviction rate--the ratio of convictions to cases prosecuted--is a sign of his [or her] competence. Prosecutors, however, choose which cases to prosecute. If they prosecute only the strongest cases, they will have high conviction rates. Any system which pays attention to conviction rates, as opposed to the number of convictions, is liable to abuse." The paper explores the dilemma flowing from selection effect. From the abstract:

"We model the tradeoffs theoretically in two models, one of a benevolent social planner and one of a prosecutor rewarded directly for his conviction rate as well as caring about convictions and personal goals. We also look at anecdotal evidence from Japan and detailed U.S. data drawn from county-level crime statistics and a survey of all state prosecutors by district. We find that prosecution rates vary little with budget, but conviction rates do increase, and that the conviction rate declines in the number of cases prosecuted and with the crime rate of a district."

ELS, Baze v. Rees & Deterrence

Dave Hoffman has an interesting post over at Concurring Opinions on the dispute between Justices Stevens and Scalia (and Thomas) over whether the death penalty deters.

ELS, Senate Confirmation Hearings & the NYT Op-Ed Page

Today's NYT includes an op-ed commenting on an empirical piece, co-authored by Lori Ringhand (Kentucky--soon-to-be Georgia) along with our very own ELS blog editors Jason Czarnezki and William Ford, forthcoming in Constitutional Commentary. The authors report that:

"Our results indicate that confirmation hearings statements about a nominee's preferred interpretive methodologies provide very little information about future judicial behavior. Inquiries into specific issue areas - such as the rights of criminal defendants - may be slightly more informative."