I am co-authoring a paper with three of my graduate students (Noel Johnston, Chai-yi Lee, and Abudulhadi Sahin) on government decisions to break contracts with foreign investors.  Our main finding in the paper is that governments are actually less likely to break contracts during periods of economic crisis, and that governments that are dependent on other countries for foreign aid for support from the multilaterals (specifically IMF programs) are less likely to expropriate investors assets or income streams.  Our data analysis include a model that predicts expropriation events in a country-year, and a survival analysis of political risk contracts issued by the U.S. government political risk insurance agency (OPIC).  We’re revising our paper for submission over the next week or two.  Drop me an email if you want a copy.

A final part of the project is a series of case studies of “pre-claims” from the World Bank’s political risk insurance arm, the Multilateral Investment Guarantee Agency (MIGA).  The MIGA folks have bee really helpful in providing feedback on our project and providing information on the “pre-claims” of expropriation, where they document government actions taken that threatened investors and how this dispute was eventually settled.

These cases provide a window into both why governments “try” to break contracts and why they eventually back down.  The table at the end of this post provides basic info on the disputes.  I had the opportunity to talk to MIGA about a bunch of these cases.

Table 1 does provide some examples of government incentives to renege on contracts during periods of crisis, although there are two important points.  First, these are cases of pre-claims, where the government either ultimately backed down from the initial policy to a resolution with the investors or in a few cases the negotiation is still under way.  Second, these crisis-triggered expropriations are quite uncommon.  Only seven of the 34 cases are related to economic crisis, and three of these are related to the financial crisis in Argentina.

Other types of disputes are much more common. In some cases political change leads to an investment dispute, for example a new minister of mines in Guatemala denying tariff adjustments or a regime change in Guinea leading to the canceling of a telecommunications contract.  Also common are reviews of privatization programs or the revising of contracts written by previous regimes previous contracts.  Examples include privatized natural resources investment in the Democratic Republic of the Congo and Moldova. Political change in Ecuador let to a review of all water concession contracts.

Other patterns of investment disputes are related to what seem to be legitimate environment regulations.  In Costa Rica the government expropriation an investment with compensation to preserve a rain forest, was a dispute was over the level of compensation, not the legitimacy of the government action.  The government of Guyana’s canceling of a contract due to a mercury contamination from a mining operation is another clear example. A dispute in Benin over potential environmental and safety issues in a hotel investment that was being build over buried oil pipelines is an example of environmental issue leading to a dispute over who pays for unexpected costs to a project.

The most common pattern of these pre-claims is governments attempting to renegotiate terms of contracts, often on the tariffs that power and water providers can charge consumers or payments owed to firms from the government.  Why do governments want to rewrite these contracts?

Interviews with MIGA staff point to unbalanced contracts as one potential trigger for expropriation threats.  In a number of power contracts, investors pushed much of the risk onto the host government that eventually led to major financial losses by the government.  The contract on hydroelectric generation by AES in Uganda is a clear example of this pattern.  AES negotiated favorable terms for a power generation contract, which became obvious during a period of low rainfall.  The government attempted to renegotiate the contract, claiming that they were incurring major financial losses by making minimum payments to AES. Similar examples include the 2003 geothermal dispute in Nicaragua, 2003 and 2004 power disputes in Kenya and Guatemala, and 2007 dispute over the investment in a cotton gin in Afghanistan.

This wide variety of types of pre-claims provides evidence of exogenous shocks (crisis, environmental disasters), political change, and simply disputes between firms and governments.  There are also a number of cases that could be classified as “corruption”, often where government officials either attempted to extract from firm, or that the government was attempting to force our the firm in order to help a competitor. Given this wide range of triggers for the disputes, is there a common pattern to how these were successfully renegotiated?  To answer this question we drawn on a number of interviews with MIGA officials.

One of the major tools that can be used is to articulate how these claims, made public through MIGA, would have negative consequences for the country’s reputation.  Some of the clearest cases were the disputes in China, where in a couple of pre-claims local or provincial government officials took actions against a firm and MIGA contacted the central government to intervene.  The conclusion of the 1998 dispute in China was literally a public ceremony signifying a conclusion of the dispute that included the company and government officials.

While different in nature, the role of reputation in the 1998 dispute in Guatemala was important in resolving the issues at stake.  In essence, the energy minister was pushing for changes in a power contract.  MIGA consulted with the Ministry of Finance, articulating the potential financial consequences of expropriation behavior.  The political fight between these ministries is complicated, but the Minister of Finance eventually prevailed.

In many cases, powerful external actors also intervened.  The clearest example was the heavy involvement of President of the World Bank and the Prime Minister in Spain in the 2003 power dispute in Moldova.  According to MIGA sources, the government was harassing a Spanish power provider to entice the company to sell to a Russian company.  The President of the World Bank and Prime Minister of Spain directly sent letters, including a direct threat of cutting off World Bank, International Finance Corporation, and European Bank for Reconstruction and Development financial support.

The World Bank was also active in adjudicating the 2009 power dispute in Uganda.  But, as noted above, this was a relatively unbalanced contract in favor of the investor.  While the World Bank pushed for the Uganda government to moderate their claims, the World Bank wasn’t completely unsympathetic to the government concerns about the contract.  The contract was eventually rewritten with the firm taking more the risk in the electricity generation part of the contract, although the government took on a number of major risks at the distribution end of the contract.

Similar pressure was put on Benin by the Bank for their discriminatory treatment of a foreign cell phone provider.  This foreign firm was threatened with a major up front fee for future taxes to continue their operations, despite domestic providers not being included in this new fee plan.  The World Bank threatened to cut off future grants to Benin and the pressure on the foreign firm subsided.

The Democratic Republic of the Congo is one of the more complicated cases of foreign involvement, where the International Finance Corporation (IFC) and MIGA had involvement in a mining operation.  DR Congo was in the process of transforming their notoriously secretive mining contracts into a paradigm of transparency, signing onto the high profile Extractive Industries Transparency Initiative (EITI).  But the problem was on how to deal with the previous contracts.  Rather than providing a formal rule on how old mines would be treated, each mining operation engaged in one-on-one negotiations with the government.  This was a process rife with potential corruption, but the World Bank (IFC and MIGA) backed mine opted for the highest EITI standard.  This hardline stance by the World Bank lead to a major disagreement with the government.  The Bank negotiated hard, although the number of important post-conflict World Bank programs in DR Congo actually made threats of cutting them off from funding less credible than in the case of Benin and Uganda.

In some cases, international financial institutions not only provided the sticks, they provided carrots to help negotiate a settlement.  The ill-faded cotton gin dispute in Afghanistan was solved with money from multilaterals, while the Inter-American Development Bank provided funds to help cover power contracts that were costing the Guatemala and Nicaraguan governments scarce foreign currency.

The role of multilaterals isn’t a guarantee of stable relations between investors and governments.  Ecuador’s expelling of the World Bank from the country and Argentina’s tense relations with the IMF and Bank documented above provide evidence that multilateral involvement isn’t a panacea.  But the evidence does suggest that these institutions wield carrots and sticks that can be used to avoid expropriation events, even in cases of contracts that were very unfavorable to host governments.

Our cases study of the 34 “pre-claims” from MIGA compliments our statistical analysis in the paper.  We show that there is no absolute guarantee against the expropriation of investment.  Pressure to expropriate or breach contracts can come from a number of different factors.  Governments that are concerned about their reputation and/or governments dependent on multilaterals for financial support are much less likely to engage in expropriations.  At the other extreme, when a government such as Argentina as already alienated investors and multilaterals, there are few remaining constraints from reneging on contracts.

We’re still writing up these details and interviewing more people at MIGA on these disputes.  But it seems like these “pre-claims” are a window into government incentives to initially break contracts, and how diplomatic or economic pressure leads the governments to eventually back down.

Table 1: MIGA Pre-Claims (1998-2010)

Blogging has been light due to vacation, end of semester, and preparing for our second child to arrive (late May…we hope).

I’m trying to wrap up a book project with Eddy Maleksy on governments providing financial incentives to firms.  I took a quick look at the Canadian incentive data.

While some Canadian provinces ban certain types of incentives, our data reveals over 400 incentives given by provinces and cities from January 2010 to April 2013.  The vast majority of these incentives are provided by the two largest provinces, Ontario (170 incentives) and Quebec (147 incentives).  Three of deals top $100 million (Canadian), including $304 million (Canadian) for Shipbuilding in Nova Scotia, $142 million (Canadian) to Toyota’s expansion in Ontario, and $132 million (Canadian) for the upgrading of a paper mill in Quebec.  Unlike many of these U.S. incentives, the majority of Canadian incentives are in the form of subsidized loans.

Interestingly, the pattern of incentives seems related to the electoral calendar.  Quebec’s most recent general election was held on September 2012 after the dissolving of parliament in August 2012.  While there is not direct evidence that the government was providing more generous incentives in the run up to the election, the descriptive data fits this pattern.  While 2012 had a similar number of incentives offered (38 incentives) to 2011 (42 incentives), the size of these incentives increased dramatically, from an average of under $3 million in 2010, under $5 million in 2011, to over $9 million in 2012.  The incentives thus far in 2013 (as of this week) were back to the pre-election levels of just over $5 million.

Perhaps this is just a spurious correlation.  What does the pattern look like in Ontario?  Fortunately for us, Ontario’s general election was held in October 2011, roughly a year earlier than Quebec’s general election.  For Ontario, we find an even clearer pattern, with the government offering many more incentives during their election year (72 incentives) than the year before elections (47 incentives) or the year after the elections (34 incentives).  The size of these incentives is equally striking.  In the year before the general election the average incentive was $1.62 million and $3.16 million after the election.  During the year of the elections the average was over $6 million.  Put another way, Ontario’s total dollars in incentives increased by over 400% during the election year.

Our point general argument is that politicans offer incentives to “claim credit” for investment that was coming to their province anyway.  This is some simple descriptive evidence that is consistent with our theory, although it could be consistent with a couple of alternative theories.  But something fishy is going on during election years.

The New York Times ran a series on investment incentives, including the Kansas City Border war.  I’ve also bloggedon the topic.  Kansas City is an extreme example of a competition for investment between two cities that share the same name:  Kanas City, MO and Kansas City KS.

As part of a book project with Eddy Maleksy, I had two of my undergrad RAs collect some data on this border war.

The states of Missouri and Kansas, along with cities in the Kansas City metro area, have fiercely competed for 67 firm investments just from 2010-2012. These 67 incentives have cost $312 million to the two states, with an average cost of over $4 million per investment.  These incentives were exceptionally generous, averaging over 50% of the capital expenditures of the firms and $37,000 per job created.  What is driving this border war?

Perhaps to the surprise of at least some readers, rarely did these firms provide direct contributions to the either the Missouri or Kansas governor’s election campaign.  Only 4 of the 67 firms provided direct contributions to the Governor’s campaigns in 2010 or 2012. The biggest player by far was engineering firm Burns & McDonnell.  Our data work for a book project (with Eddy Maleksy) identified direct contributions by employees to the Jay Nixon (Missouri-D) campaign of $137,000.  Documentation by OpenSecrets, an online campaign contribution tracking website provides a more complete picture of their contributions, including PAC contributions.  This company has a PAC that largely funds incumbent Representatives and Senators across the country.  The key point is that while this company is politically active in the state and Missouri and beyond, it is by far the exception.  The other three contributors provided $3,500, $2000, and $1000.

We also examined the contributions of employees of these 67 companies beyond the governor elections and coded any campaign contribution by employees of the firm to state politicians.  Only 6 more companies (a total of 10) provided contributions.  Burns & McDonnell and insurance broker Lockton Companies both gave roughly $22,000.  The remaining companies gave an average of just over $1000 each.  In summary, the 67 incentive recipients gave an average of $3,000 across all state and local elections in Missouri and Kansas from 2010-2012.

The use of campaign contributions, at least the direct ones we can track, doesn’t seem to be the deciding factor in shaping these incentive decisions.  Similar to an analysis of Texas investment incentives that we blogged about in the past, there is a strong correlation between the number of jobs created and the size of the incentives (correlation of 0.30).

Ok, if campaign contributions aren’t responsible for this incentive war, perhaps the tough competition for capital is driving this.  In most cases of these incentives, we couldn’t find evidence of competing incentive bids, or even concrete evidence that the firm was considering an alternative location. For 31/67 of these incentives, the firms were simply expanding existing operations in their previous locations.  For others, they were jumping to the other side of the border since “new” investment can receive incentives, while existing plants (without expanding) often can not.  We could only find evidence of claims of competition from other locations in 20 of the cases, with many of the companies citing “other Midwestern locations” or other vague claims of competitors.  In 6/20 cases, the competition was from across the state line in either Kansas or Missouri.  Only in two of the cases could we find direct evidence of competing offers, where two firms locating in the Kansas City, Kansas suburb of Overland received competing incentive bids from Missouri.

Thus in the majority of cases, there is little evidence that direct competitive forces, such as bids from alternative locations, were driving these incentives.

While we can not conclusively say what is driving this border war, our data collection efforts document that most of these firms do not provide any campaign contributions, nor where these incentives and obviously outcome of a bidding war.  In many cases, firms received only one offer, often to jump across the border to Kanas or Missouri.

Our book project extends the logic of this paper on the use of incentives for credit claiming by politicians.  Credit claiming, not campaign contributions are driving this incentive competition.

The Washington Post had a short article on the decline in corporate taxes paid by the Dow 30.

I have an ongoing research project with Adam Rosenzweig on the decline in U.S. corporate taxes.  We’re specifically interested in what is causing the decline and what government policies, if any, have been effective in slowing this decline.

But here is a quick picture on effective corporate tax rates of the Fortune 500 (from Compustat data).  We use the same methodology as this paper on lobbying and taxes. Below is a graph on the average effective corporate tax rates of the Fortune 500 with a trend line and the 95% confidence interval.

There clearly is a decline.  But what is the cause?  Is it learning by firms?  Changing rules on transfer pricing and use of tax havens?  Government policies that are increasingly generous in their deductions and allowances?

Today in my graduate International Political Economy class we discussed how firms influence the political process.

While the US media often focuses on campaign contributions in a quid pro quo exchange with firms, political science has rejected this simple approach.  Even in cases of government contracts (think Halliburton getting billions during the Iraq war) there is limited evidence that politicians trade votes for campaign contributions?

Really?  How come?

First, the empirical evidence simply doesn’t point to this type of exchange.  As documented in this classic paper(and first described by Tullock), there is too little money in politics.  In the last election cycle over $1 billion was spent by special interests.  But compared to $3.5 trillion in government spending or $615 billion in non-defense discretionary spending, political contributions are quite small.

If a company like Walmart could swing government policy, they should be investing millions and millions in campaigns.  They don’t.  They spend hundreds of millions in charitable contributions and advertising, and only thousands in political contributions.

Second, most of the studies cited in this study find very limited evidence of contributions affecting roll call votes.

That was a quick empirical point, but doesn’t explain the theory.  How come money isn’t more influential?  Thispiece by Gordon and Hafter makes three points (focusing specifically on regulation).

1. Reductions in regulations are a collective good for the industry.  Sure, chemical companies want looser environmental laws, but the lobbying by a single firm could provide the good for everyone.  The collective action problem means that this actually makes it less likely for this to be provided.
2. There is a classic problem of credible commitments.  A politician gets thousands of campaign contributions from a firm with the promise that the politician will shield the firm in the unlikely event of an environmental disaster.  The disaster strikes and then politicians run away from the firm.  Let’s call this example Congress and BP Deepwater Horizon.
3. More generally, politicians are getting campaign contributions to help win elections.  If the positions politicians have to take to get these contributions are unpopular, this can defeat the purpose of getting the contributions in the first place.

These are probably obvious points to most my reader(s), but worth noting.  Most firms invest more in crappy corporate art and definitely much, much more in advertising than in political contributions.  Why?  Because of collection action problems and that you can’t trust those lying politicians.

I don’t want to go too far.  The above cited Gordon and Hafter piece is an excellent study of how contributions can be used influence politics.  But the point is that contributions may matter, but they are not the main way that firms influence politics.

Years ago I had a MBA student in one of my Ph.D. classes.  His answer to these sort of questions was always something like, “why don’t you just ask the firms.”  One recent paper did something along these lines, by examining 250,000 Enron emails to see how important contributions were as a political strategy.  The quick point is that they aren’t very important in general (summary link).

Contributions have minimal influence, but firms can influence politics in other way.  Let me outline one quick mechanism. Work such as Eddy Malesky’s research on Vietnam show that foreign investment, by generating jobs and tax revenues, is supported by local elites.  Local are willing to bend or break central government rules for investors for these benefits.

Related work includes Pinto and Pinto’s study of how the industry of investment matters for politicians. Left politicians, favoring labor, will enact policies that will help encourage investment in labor intensive industries, while right politicians, representing capital, will represent capital intensive industries.

Finally, work by Facco and co-authors documents “politically connected firms”.  In most cases these are firms where executives or board members of a firm are also politicians.

What doe these three studies have in common?  In all three of them, politicians have strong incentives for certain firms to do well.  Politicians are willing to champion these firms, either because they affect an important constituency, or that they directly impact a politicians own asset performance.

These are just a couple of quick points from a single week of my grad IPE class.  Thought I would share.

My many posts on investment incentives have mostly focused on the US.  I’ve been collecting some info on incentives in Canada, the UK, Brail and now South Africa.

South Africa is an interesting case where the majority of the incentives are provided through a government agency based on this tax incentive act.

This act directly specifies what types of investments are eligible for incentives and at what level.

• R900 million in the case of any Greenfield project with a preferred status;
• R550 million in the case of any other Greenfield project;
• R550 million in the case of any Brownfield project with a preferred status;
• R350 million in the case of any other Brownfield project;
• An additional training allowance of R36 000 per employee may be deducted from taxable income; and
• A maximum total additional training allowance per project, amounting to R20 million, in the case of a qualifying project, and R30 million in the case of a preferred project.

Basically, new investment (greenfield) is privileged over mergers and acquisitions or expansions of existing investment (brownfield).

Further details include:

• Upgrade an industry within South Africa (via an innovative process, cleaner production technology or improved energy efficiency);
• Provide general business linkages within South Africa;
• Acquire goods and services from small, medium and micro-sized enterprises (SMMEs);
• Create direct employment within South Africa;
• Provide skills development in South Africa; and
• In the case of a Greenfield project, be located within an Industrial Development Zone (IDZ).

Sounds fair.  How does it work in practice?  This is the latest press release.

Not a lot of detail on these investments, but I wanted you to focus on one stat.  The claim is that these investments will create 1,618 direct jobs and 25,448 indirect jobs.  Indirects jobs are the jobs that will be created by suppliers or other businesses benefiting from the investment.

This is a crazy number of indirect jobs.  My conversations with people in US economic development offices have indicated that they have hard rules (and some specific software) on calculating the indirect benefits of investment.  Automobile investment is the best type to get, where the multiplier is 7-8 indirect jobs for every direct job at the most.  The South African average multiplier is an average of over 15.

What is going on?  This week I saw five announcements of incentives in South Africa.  Below are the companies and tax incentives in U.S. dollars.

1. Mamba Cement: $20.59 million
2. Omnia Group: $8.71 million
3. Sepkahu Fluoride: $20.59 million
4. Sappi Southern Africa: $20.59 million
5. Lomotek Polymers:  $2.51 million

What do these investments have in common?  First, all five are South African companies, so don’t blame foreign MNCs for this.  Second, all of these investments are classified as being investments in the “Basic Materials”.  Mamba Cement is pretty self-explanatory.  Omnia, Sepkahu and Sappi are chemical plants and Lometek makes “second generation” pallets.

Do these investments justify job multipliers of 15?  I don’t have any details about South Africa, but the United States Bureau of Economic Analysis has standard multipliers by industry (and by region).  Here are multipliers for investments in California.

Don’t bother clicking on the document.  The quick story (presented in Column 6 of the many pages of tables) is that for every direct job created, most industries created about 2 total jobs.  A few industries hit multipliers of 6 or 7 (automobile production is 6.45).  Petroleum refineries are a 9 and Electronic Computer manufacturing tops 13, and on type of finance (funds, trust, etc.) hits 14.  That is the maximum.

The industries most closely related to the three chemical investments are chemicals are 3-7, and cement at 4.45.  I don’t have a multiplier for “second generation” pallets, but most plastic manufacturing has multipliers of 2-3.  My brother worked at a pallet company in Wisconsin.  2 is generous.

What is the story here?  Looks like the South African government is providing very large incentives (even by US standards) to local companies using inflated indirect job numbers.  Why?  That is a question for future research.

I received a media inquiry today about US investment abroad.  One of the questions was why does the Netherlands receive so much US foreign direct investment (FDI)?  Even more surprising might be the top five countries (in 2011) were: the Netherlands, UK, Luxembourg, Bermuda, and Canada.

The answer to the original question is the Dutch Sandwich.  Or if you want to get more complicated, a Double Irish with a Dutch Sandwich.

What?

The Dutch Sandwich is basically a way for firms to establish a Dutch holding company, route their foreign income through the holding company, and then send the money to a tax haven.  The outcome of this is that companies can, legally, move foreign profits from countries with high (or non-zero) tax rates to tax havens with zero tax rates.  You can get more complicated and move money from Ireland to the Netherlands and then to a tax haven.

Here is a story on how Google uses this strategy to minimize their tax burden.

There are clearly examples of firms doing this, but how much do these practices affect global FDI patterns?  Attached is the 2011 US Bureau of Economics Analysis Report on US outward and inward FDI. See my highlights on page 32.

Here is the key quote from the report:

For the third consecutive year, the position in the Netherlands was the largest—at $595.1 billion, or 14 percent of the total. Most of the position increase and 77 percent of the position in the Netherlands was accounted for by holding companies, which likely invested the funds in other countries; see the box “Indirect Ownership in the Statistics on U.S. Direct Investment Abroad.”

The biggest FDI recipient largely attracts foreign investment to avoid taxes.  Bermuda and Luxembourg aren’t attracting manufacturing investments.  These are mostly holding companies as well.

One final note.  The establishment of Dutch subsidiaries isn’t only for tax purposes.  The Dutch have an extensive network of investment treaties around the world.  Establishing an operation in the Netherlands can also give you access to use these treaties and “forum shop” different legal venues if you have an investment dispute.

UPDATE:  I had a tax lawyer friend who had a couple of comments on my blog post.  The original version of this post used the terms “evade” rather than “avoid” taxes.  This is importance since the firms are engaging in legal means of reducing their tax burden.  I also changed the language that firms are reducing their “foreign income” subject to taxation, not necessarily their US income.  I’m not 100% if firms are using these strategies to also reduce their US income, or if they are just reducing their foreign income tax burden.  This is above my pay grade/my time commitment to blogging.

As part of a book project with Eddy Malesky I did a little poking into financial incentives offered to firms investing in the EU.  The EU governs incentives through “state aid” rules, constraining the ability of wealthier regions from giving aid to firms.  The conventional wisdom is that these rules limit the fiscal bidding wars for firms.

I was surprised to see so many large incentives given over the past 10 years.  Here is a quick list just from the European auto producers over the past 10 years.

• In 2004 Fiat received grants of €8 million, €33 million, and €47 million for investments in Italy.
• In Spain, Renault received grants of two grants of €30 and one of €33 for in 2004, €45 million in 2005, €13million and €32 million in 2006, and €17 and €18 million in 2010.
• Puegeot received a €58 million grant for their investment in Galicia, Spain in 2006.
• Volvo and Volkswagen got €4 and €9 million incentives for a Spanish investments in 2005 and 2006.
• BMW received €14 for their investment in the Ober-Österreich region of Austria and €4.9 million for an investment in Warwichshire, UK.
• In 2008, Renault obtained a comparatively modest €3 million in France and a massive €25.5 million direct grant from Romania.
• The Czech Republic has gave a €22 million grant and €29.5 million tax exemption to Skoda in 2008 and 2011.
• Volkswagen and Fiat received €16 million and €26.99 million in grants from Poland
• Opel grabbed €20 and 22.5 million in grants and tax benefits from Denmark and Hungary in 2010.
• The Rolls-Royce Group (ok, not all auto production) have racked up a total of €49 million in grants and other incentives in the UK between 2010-2012.

These examples are just from the operations of European auto producers, ignoring the massive incentives given to Ford, GM, Nissan and Toyota, among others.  It also doesn’t address the massive incentives offer to other iconic manufacturing companies such as: Abbott, Alcoa, Amazon, BP, Bridgestone, Caterpillar, Coca-Cola, Cadbury, Caesars Hotel, H&M, Dell, Dupont, Eli Lily, GlaxcoSmithKline, Hewlett-Packard, IBM, Intel, Johnson and Johnson, Merck, Michelin, Pfizer, Pirelli Tyres, Samsung, and Sony.

My favorite ones is cash strapped Greece providing a €28.9 million subsidy for a Formula One race track in 2012.

As the Director of Graduate Studies at WashU in Fall 2012, I completed the US News survey of political science grad programs.  For those of you unfamiliar with the methodology, DGSes and department chairs fill out a survey ranking all schools on a 1-5 scale and write in the top ten departments for each subfield.

This is really tough to do.  You try it.  Take all of the Ph.D. programs in your state and rank them on a 1-5 scale.

The rankings for 2013 literally didn’t change from 2009.  Ok, there are a few more ties in 2013, but the order of schools is the same.

2013                2009

Harvard                      1                      1

Princeton                    2                      2

Stanford                     2                      3

Michigan                    4                      4

Yale                            4                      5

Cal                             6                      6

Columbia                   7                      7

UCSD                        7                      8

MIT                            8                      9

Duke                          10                    9

UCLA                         10                    10

Chicago                     12                    11

UNC                           13                    13

WashU                        13                    13

Rochester                   15                    13

Wisconsin                   15                    15

NYU                            15                    17

Ohio State                   15                    17

What gives?  Is it that reputations of top programs really don’t change?  Or that this survey is so poorly designed most of us are going to simply recall the past rankings as a heuristic to rank programs?

Actually, there is a methodology reason for this.  Turns out that for the first time ever (with no explanation) they averaged the survey from 2008 (the 2009 rankings) with the survey in 2012 to make the 2013 rankings.

Why?  It could be that with only 50 responses, they got some odd rankings that didn’t make sense.  Or that a few top programs got panned, and for whatever reason they averaged the response.

UPDATE: Someone pointed out that there is already a discussion of this issue at Political Science Rumors.  This is what I get for not following the rumor blogs

Many years ago I started a project examining firm-level data on foreign investment.  This data is from the U.S. Bureau of Economic Analysis (BEA) on the operations of all of the 20,000+ foreign affiliates of U.S. multinationals.  This paper, on the taxation of multinationals, has been finally published at International Studies Quarterly.

I wanted to briefly document my experience with this project since it related to a number of discussions on “big data” in the social sciences (here is one good post on big data).  I know, 20,000 obsevations isn’t a lot, but this can be used as time-series data and there are other aspects of this data that are similar to “big data”.  Hold on.

Here is a couple of very quick bullet points on my experience with this paper.

• I found out about this incredible firm-level data set by reading a few econ papers that used it.  To my dismay, the data is confidential, housed in Washington, DC.  So I had to petition to the BEA to see the data.  After a few months, I got the ok from the BEA and then went through the process of getting a security clearance to use this data.  A few more long months.
• Once I was given access to the data (and a special sworn employee of the BEA) I had to comply with the rules of using the data.  The data is housed in DC and has to be used on site.  I had to fly to DC every time I wanted to run a regression.  Good thing my college roommate lives in DC.
• All of the data are housed in different MS Access files and that only old versions of Stata were available on the computers that I could work with.  No downloading files from the internet (R, do files, etc).  Putting together the data set and even running a few simple regressions was a lot more difficult than I expected.
• This data set was a gold mine, but like most mining operations, extracting anything from it is really, really messy.  How do you “clean” data that isn’t comparable to other data sets?  For example, I had way too many zero observations for one variable in the data.  Were these true zeros or just missing values that were coded as zero?  I went to the BEA papers archives and pulled a sample of paper forms to double check the coding.
• I rarely get a paper get accepted on the first go.  This means for every time the article this article was reviewed, I had to plan a trip to DC to run another set of regressions.  I went to APSR, IO, AJPS (R&R that was rejected), JOP and then finally ISQ.
• Given the barriers to replication, article reviewers and one NSF panel were probably harsher on this project than my others.  I can’t say that I blame them, but I got at least one negative comment from every journal and grant review process on the inaccessibility of the data.

The paper I wrote with this data won an award for best political economy paper at APSA in 2008.  It is now forthcoming in International Studies Quarterly in 2013.

What is my experience with “big data”?

1. The barriers to entry are really high. You probably already knew this.
2. Data quality is a serious issue.  When using a cross-national dataset, I look at the individual observations to make sure nothing looks odd.  It took much more legwork to verify the quality of this data.
3. The potential for “data mining” is much, much lower that you would think.  This relates to point 4.
4. There is no way to let the data “speak” to you.  It is a confusing mess that you really need to have a plan on how to analyze it.
5. Control variables or other important variables aren’t often available at the level of analysis that you’re examining.
6. Because this is “new” data, many of the standard methods of data analysis might not apply.

My only concrete suggestion is that theory is even more important when using “big data”.  You can only really harness the richness of complicated micro data if you have clear micro theories.

Barriers to entry can create rents for a researcher, but they also make it much more difficult to replicate your results.  This means that journal reviewers and grant reviewers can hold this against you, and the ultimate impact of your work might be lower.  This isn’t a suggestion.  It is a warning.

In the end, I’m really like this paper and I am really grateful for the folks at the BEA for giving me access to this project.  But this was a tough slog.