I am collecting data on investment incentives around the world as part of a book project.  One big deal was just announced this week.  Big River Steel is investing over $1 billion in Arkansas.

According to this article, the incentive deal includes:

…$125 million for start-up costs: a $50 million loan to the company, $50 million for site prep, $20 million for piling and $5 million for bond insurance.

Other state incentives include:

• sales tax refunds on building materials, taxable machinery and equipment used in the project;
• a 4 percent income tax credit based on new payroll jobs for five years;
• $10 million from the Governor’s Quick Action Closing Fund;
• $5 million from the Department of Workforce Services Trust Fund to be used for training;
• an income tax credit for recycling equipment equal to 30 percent of eligible recycling costs that will include legislation that could extend the credit from three to 14 years;
• and a sales tax exemption on utilities that will include legislation to fully exempt sales tax associated with the sale of natural gas and electricity.

This is a huge deal, but the dollars often obscure the actual costs to the state.  There are big differences in a loan (even at a subsidized interest rate), tax exemptions (which reduce the tax burden in the future) and cash up front.

To me, the most troubling type of incentives are the “deal closting funds”.  Arkansas has one of the larger deal closing funds (although dwarfed by the Texas Enterprise Fund).  The last report I can find is that in 2011 it had $23 million in assets.  Essentially, the Governor’s office is providing a special, $10 million cash grant to a specific firm.

What could go wrong?

A year or so ago I put a freedom of information act request to Texas and Arkansas on their incentive programs.  Texas gave me the info on the accepted and rejected applications (they accepted most applications).  Arkansas couldn’t provide any info on rejected applications.  Dodgy.

Brazil’s automobile regime in the 1970s and 1980s was the poster child for failed economic development policies.  In short, Brazil combined high tariffs on imported cars with lucrative investment incentives to attract foreign producers.  These foreign producers were required to use “local content” in the form of local suppliers and in some cases local managers.

Sound good?  As you might predict, auto producers came to Brazil because of the protected market, but they build small boutique production facilities that made crappy cars that were sold at high prices.  This was expensive for consumers and tax payers and did little to generate economic development.  Policies of “import substitution” largely fell out of favor.

BMW recently announced the opening of a new production facility in Brazil.  Why?

Brazil’s new auto regime sounds a lot like the failed economic development policies of the past.  Brazil introduced a new tax on imported cars (from 25% to 55%) although they are exempting producers with Brazilian factories from this tax.  Although it seems like this tax will be gradually reduced.  Right now, cars like the Toyota Corolla cost almost twice as much in Brazil than in the US.

Brazil also has complicated local content rules and they are enticing firms with large investment incentives.

Sounds a lot like the failed policies of the past.

Dan Nexton kicked off a debate on the value of academic job talks.  Dan provides a nice follow-up that links to blog posts by Tom Pepinsky, Jeremy Wallace.  Tom Oatley has a provocative post on a potential historical explanation for the  emergence of the job talk.

Tom Oatley has a nice point that we should look for some data to address this debate. He focuses on a historical explanation.  But I also have a contemporary one on the use of talks as part of an evalaution process.

What are (job) talks for?

In my original post I mentioned that department speaker series often have the same format as job talks.  The point was that job talks may not be the most efficient ways to make a decision, but they: a) can complement other pieces of information for evaluation and b) they have other positive externalities.  We have speaker series because they are valuable to a dept.  A job talk can also provide some value outside of just evaluation.

I have another example.

At WashU we’re on the second week of listening to our Ph.D. students present their 3^rd year papers in the department.  These paper presentations take the form of a 20-30 min talk.  We actually have more of these presentations in a year than we have job talks.

I was on the committee that enacted this new requirement two years ago.  Our process for the 3rd year papers is that we have a faculty committee read the papers and we also included a public presentation of the work.  What was the logic of having a public presentation?

At least one faculty member advocating for presentations did say argue that students would eventually have to give a job talk and this would be good practice.  This is a similar logic to Tom Oatley’s path dependence argument.  But the more important reasons (at least in my mind) were:

• Public presentations actually incentivize more effort.  Not only do you have to produce a paper; you also have to anticipate questions, know your data, and be able to explain your methods.  Papers can be narrow, but presenters have to almost always be more broad than their written work.
• This was a low cost way to communicate research not only to faculty but to other graduate students in the program.  First year graduate students probably don’t read these papers (or job market candidate files), but they can learn quite a bit during a research presentation.
• Attending these talks and really engaging the speaker is a way to signal to the students in our program that we take this process seriously and we our invested in the graduate program.

We could play with the format and have the presentation 5 min or 45 min.  The point is that these presentations serve multiple purposes.  If we just wanted to evaluate the work as efficiently as possible, we would circulate the papers to the faculty.  Hell, why not just circulate it to one or two faculty.  Who said an efficient process needs to impose uniformly high costs on all faculty members when delegation is possible?  But these presentations are more than just about evaluation.

Some of these reasons may not apply to a job talk.  But I think the quick point is that attending a job talk is a pretty low cost activity that teaches you about the speaker, the project, and informs the speaker about your department.

Making Better Decisions

My main point is that the job talks serve a evaluation function along with an intellectual and professional function.  But I think I might be alone in thinking about job talks in this way.  So let me go back to the big question on how to evaluate candidates.

I think there is a tremendous amount of uncertainty in junior faculty hiring for a number of reasons.

I’m all for a better search process.

The one common criticism is that job talks allow faculty to be lazy and make decisions based on the talk.  I’m not sure what the process looks like at other departments, but here is ours:

• Job candidates are selected by a search committee that has already read the work.
• Most of us have some form of a one-on-one or group meeting with the candidates.
• Many of us read at least some of the written work.
• The search committee has to circulate a report 72 hours before the faculty meeting when we discuss the candidates.
• We have a faculty meeting where the search committee sets the agenda and then we discuss the candidate(s).

Most meetings have a few faculty that had to miss the talk for teaching or travel (although most faculty attend).  Faculty also have a range of exposure to the written work.  Committee members read a ton and a bunch of faculty outside of the committee almost always engage the written work.

Is it plausible that removing the job talk would incentivize more people to read more and engage the written work?  Sure, it seems plausible to me.  But would this lead to a better decision?

I personally agree that some people overweigh the value of the job talk.  I think letters are more informative, but many people place way too much stake in them as well.  I’m mostly a “read the dissertation” type person, although others have cautioned me that dissertations can easily be largely ghost written by a dissertation advisor.  Who’s right?

My colleague Matt Gabel has some work on group decision-making in medicine.  The simple point is that the best decisions are made when individuals that make different types of errors are in the same room together.  It doesn’t matter who is right (or who is wrong).  What matters is how me make a decision as a group.

I think it is equally plausible that having people that over weigh the job talk or the letters or the dissertation in the same room leads to better decisions.  A decision solely based on a job talk would be a terrible decision.  But having a faculty member or two advocate for or against candidate based on the talk seems like it could be a part of a healthy discussion about hiring.

This last part about decision making is really just a mix of conjecture and my application of group decision making (or portfolio theory) to political science hiring.  I’m not 100% sold on the job talk, but like I said, I’m mostly a “read the dissertation” type person.  I also like to attend the job talks.

Dan Nexon blogged on problems with the academic job talk and possible alternatives.  Tom Pepinsky and Jeremy Wallace both have posts on the subject.  I posted a defense of the job talk.

I think Dan’s original questions was more broad.  Is the academic search process broken? What are the alternatives?

One of my favorite New Yorker articles is by Malcom Gladwell on the similarities in the difficulty of NFL teams predicting the future success of college quarterbacks and that of schools predicting who will be a good teacher.

Many universities pass over candidates that would have been great additions to their faculty.  More often, departments hire people who don’t pan out.  Again, sounds a lot like the NFL draft.

But evidence that mistakes are made in hiring doesn’t necessarily mean that the search process is broken.  It could mean that the search process is really hard, and there is a lot of randomness in this profession.

I personally think that most of us understate how much luck has to do with success in our profession.  Maybe most professions.

I hate to defend the status quo, since there are a lot of seemingly irrational aspects of the search process.  But I would also like to see some evidence that we would make better decisions with a different process.  To do this, we have to at least evaluate what types of errors are made in the search process and how alternative processes would limit this type of error.

This Duck of Minerva blog post takes on the issue of the academic job talk.  For the non-academic readers, a job talk is simply a 30-45 min presentation of a job market candidate’s work.  Thanks for reading mom.

As noted in the post, the conventional wisdom is that these talks make or break a candidate coming right out of grad school.  I’d also add that there is probably way too much weight placed on the job talks of tenured faculty as well.

But I think the post missed two really important elements.

How do we make decisions?

I think one of the most important aspects of a job talk is it is just one of a number of factors that can be weighed to make a hiring decision.  As Daniel Nexon note’s, there might be better ways to evaluate research, teaching, or the personal attributes of a candidate.  For example, simply reading the written work or holding a seminar on the research could be a better way to evaluate the merits of the scholarship.

I attended a meeting in the Fall at WashU for search committee chairs.  In this meeting we had a social psychologist present research on how women and minority candidates are disadvantaged in the search process.  One important lesson was that people tend to ask women and minority candidates different types of questions in the one-on-one meetings.  These questions also tend to be of the “harder” type.  The simple advice was to either ask the same opening questions of the candidate or let them set the agenda.

This is just one example of how bias can influence decision-making.  I can imagine similar types of bias occurring when reading an application packet.  Do you look at the CV?  The letters?  Which working papers do you read?  How to you weigh the different factors?  Is it the same for every candidate?

One of the few saving graces of a job talk is that the candidate gets to the set the agenda.  Sure, faculty can interrupt in the talk, but the candidate selects the topic, how it is framed, what evidence will be presented, and how this fits into a broader research agenda.  This is obviously an imperfect system, but the point is that it can help compliment the other pieces of information about an applicant.

What if efficiency?

If someone asked me how I could produce more research per day, I would say I would like more time to produce research.  But given I teach classes, sit on committees, and try to help select candidates for jobs at WashU there are many days when I don’t get any time to do research.

But many of the activities that we engage in as academics that compliment each other.  I taught Introduction to International Relations last semester.  Presenting basic material on international relations actually gave me a few new research ideas, helped me recruit a couple of top notch research assistants, and forced me to try to frame many of the classic debates in international relations as relevant for current public policy.  All of these factors have had spillovers on my research.

I could easily say the reverse about my research and my teaching.  My research has had me travel to a number of countries around the world to interview individuals and collect data for research.  These experiences are incredibly useful in the classroom.

We’ve had a number of American Politics job talks at WashU over the past few years.  This is the one field (other than political theory) that I haven’t sat on a hiring committee.  But I have learned a tremendous amount of information from these talks that I brought into the classroom and has influenced my own research.

Job talks can perform many of the same functions as a speaker series.  Candidates bring in work and get criticisms and suggestions to make the work better.  The audience not only gets to hear the work of the candidate, but also the response of other faculty on this work.

Brining in a candidate to teach a course probably wouldn’t serve the same function.  We’d have a better sense of how good they are at teaching, but it is doubtful that the audience would get the same research benefits form the talk, and definitely wouldn’t engage the speaker in the same way.

Dumping the job talk would probably have many of the same costs and benefits of axing speaker series.  Job talks may not be the most efficient to make recruitment decisions.  But that doesn’t mean that they’re not an efficient tools for a department.

The Bigger Picture

Both of these points aren’t a defense of the job talk as the sole or primary way of making hires.  But they do give the candidate the ability to set the agenda.  Plus it is efficient for me in the sense that I can both learn something about a candidate’s research and teaching ability while also learning something that helps my own teaching and research.  Seems pretty valuable to me.

Here is another story on invesmtent incentives gone bad.  A major development project in North St. Louis received $40 million in incentives years ago and has failed to deliever after promises of major investment and high quality job creation.

They just got their first tenant.  Dollar General.

Blogging has been very light lately do to a bunch of administrative responsibilities.  Two things.

1.  I am kicking around the idea of putting together an edited vol on null results are failed experiments.  Drop me an email if you have thoughts or are interested in contributing.

2.  Jamie Monogan circualted the following email on the Political Methodology listserv on registering experiments:

The most recent issue of Political Analysis (vol. 21, issue 1) features a Symposium on Research Registration, or the idea that scholars in many cases can prespecify a research design prior to observing the outcome variable. The merits of preregistration are still debated, and the discipline currently does not have a comprehensive central registry. However, we write at this time to draw people’s attention to a handful of proto-registries that are available to researchers. From our experience, the best way to figure out what you think about registration is trying it in one of your own projects!  In particular, any researcher who is interested in self-registering a study is welcome to take advantage of the the Political Science Registered Studies Dataverse (http://dvn.iq.harvard.edu/dvn/dv/registration).

This dataverse is a fully-automated resource that allows researchers to upload design information, pre-outcome data, and any preliminary code. Uploaded designs will be publicized via a variety of free media. List members are welcome to subscribe to any of these announcement services, which are linked in the header of the dataverse page.  Besides this automated system, there are also a few other proto-registries of note:

EGAP: Experiments in Governance and Politics: <http://e-gap.org/design-registration/> The EGAP website has a registration tool that now accepts and posts detailed preanalysis plans. In instances when designs are sensitive, EGAP offers the service of accepting and archiving sensitive plans with an agreed trigger for posting them publicly.

J-PAL: The Abdul Latif Jameel Poverty Action Lab: <http://www.povertyactionlab.org/Hypothesis-Registry> J-PAL has been hosting a hypothesis registry since 2009. This registry is for pre-analysis plans of researchers working on randomized controlled trials, which may be submitted before data analysis begins.

The APSA Experimental Research Section: <http://ps-experiments.ucr.edu/> The experimental research section of APSA hosts a registry for experiments at its website. (Please note, however, that the website is down for maintainence until February.)  Additional information about the method and the resources are printed in this month’s issue of Political Analysis. Additionally, anyone who has questions about these resources is welcome to contact Jamie Monogan off-list (monogan@uga.edu).

Amazon has announced the location of a major distribution center in Robbinsville, N.J. with incentives that are “valued” at $13.755 million.  I put valued in quotes because I don’t fully understand what is going on.

The Robbinsville Township Council (and any city in N.J.) has the ability to offer firms a payment in lieu of taxes deals (PILOT) where the company pays gradually increasing amounts of money to the city over time instead of paying statutory taxes.  This amounts to a subsidy, but the value of the subsidy is really based on a counterfactual of what the company would have paid without the PILOT program.

It gets more complicated.  The land and building is actually going to be owned by KTR capital partners who bought it from Matrix Development.  Martix sold their land because dealing with Amazon is a very complicated affair.  But only “urban renewal” projects are eligible for PILOT programs.  So KTR formed a new subsidiary.  Check out the uncreative name: KTR NJ Urban Renewal LLC.

The tone of my post is one of skepticism.  But this is a great investment for New Jersey that will generate a lot of new jobs and generate some new sales tax revenue.  But this opaque policy does worry me a bit.

I’m been avoiding blogging as I finish up some research projects and spend lots of time playing with my son.  This is a quick entry about a failed survey experiment.

In academia, like many professions, we are judged on outputs, not inputs.  You can work long hours and have very little to show for it.  This is most obvious in the publication process, where you start a project and fingers crossed you take it to the stage of a completed manuscript.  Then it is rejected at a journal.  I’d guess that most of my published papers get published after rejections at the first two journals.  Some never make it.  The advice I got in grad school was to grow a thick skin and keep trying.

In some cases, research projects don’t even make it to that stage.  I have a research project on agriculture projection that I blogged about here and here.  Basically, I find that framing US agriculture subsidies relative to other countries has a massive impact on individual preferences.

As part of another survey project, I decided to field an internet survey in India using Amazon’s Mechanical Turk.  A few papers have documented the use of mTurk for US surveys , but very little has been done on the use of mTurk outside of the US.  A great resource for these papers is this blog.  But the one unpublished paper I found showed that about 40% of the “workers” on mTurk are in India.  Unlike the US sample, mTurk workers in India are overwhelmingly male and college educated.  Not a big surprise, but this would be a serious challenge to publication with this same.

A few days ago I fielded a 2-3 minute survey through mTurk at the cost of $0.25 per survey response using Qualtrics to program the survey.  Paying 1,000 respondents plus Amazon’s fee came to a total bill of $275 for the survey.  With the exception of explaining this all to WashU’s human subjects folks over the course of the month, this is as easy as it gets.

My experimental question framed India’s agriculture policies as either more or less generous than their neighbors.  No need to get into the details.  For both groups, almost exactly 90% of respondents supported increasing agriculture protecton.

Previous research has suggested that less educated individuals and women tend to be more in favor of trade protection.  My survey was over 65% men and 80% of respondents had completed college and the majority of respondents came from relatively rich regions.  My expectation that a more representative sample would generate at least as high of a level of support for agriculture protection.  There is almost no variance to explain and my experimental treatments had literally zero effect.

I think this is intellectually interesting, but a professional roadblock.  What do I do with these results?  Do I write a paper based on this survey?  I’m 100% sure this paper wouldn’t get published.  I’m thinking about adding this to my paper on the US survey experiment, but I’m not sure how exactly to fit this into the paper.

But it is an experiment I designed and these are clearly null results.  A few people in political science have pitched the idea of registering experiments and I’ve even talked with a few colleagues about getting a journal special issue to commit to publishing work based on the experimental design, not on the results (and thus precommitting to publish, or at least not hold against an author, null results).

This is a pretty minor project, but since I blogged about the US results I thought I would be intellectually honest and present the India results.  I also found similar null results in the UK, but probably for different reasons.

I’ve blogged too much about countries, states, and cities providing incentives for firms.   Here is a quick Missouri edition.  Two firms received incentives from the city and state over the past week or so.

Hudson Bay (Lord and Taylor) received a big incentive package to expand their operations just outside of downtown St. Louis.

Here is the quick description of their incentive package:

The incentive package includes $3.8m in Missouri Quality Jobs program tax credits and $2.2m in Brownfield Redevelopment program tax credits to the building’s owner, BEB Management. The funds will be awarded provided the companies meet job-creation and investment requirements for each program. Hudson’s Bay Company must create 177 new, full-time jobs that pay a wage that is at or above the county average wage and offer insurance, within five years.

The Missouri Quality Jobs program is a state level program, but where did the other $2.2 million come from?  Is this is good deal for St. Louis?

The company is expanding their operations and the St. Louis board of alderman voted to extend extra incentives as part of program to encourage revitalization of “blighted areas”.  Forget about the fact that this building isin a nice location downtown and the company is renovating the additional floors of the building they already occupy (are these floors blighted), I haven’t been able to get any real details about this incentive package.  I eventually found thebill, introduced by Alderman Phyllis Young.  But nothing else.  My email to Phyllis has yet to be returned.

A second incentive was offered to Vistar, a food and beverage distributor located in Kansas City Kansas has jumped to border to move 16 miles to Kansas City, Missouri.  This comes with a $700,000 in block grant financing and $100,000 in other incentives.  Is this another example of the “border war” between the two states?

A news story I found from the Shawnee local paper (the original location of the firms) claims that incentives weren’t central to this location decision.

What do make of these two case?  One point is that it isn’t clear if incentives had any impact on the location decision of the two firms.  One simply expanded and the other moved to a better location.  It is hard to know what would have happened without these expensive incentive packages.

The other point is more about conducting research in this area.  Even with my local knowledge, I struggled to find basic information on these incentive packages.

Still waiting for an email from Phyllis.