Thad Kousser: Ask anybody here what a “good” state legislature should look like. Can anybody actually answer that? Seth Masket: Campaigns can matter. In districts that Colorado’s wealthy Democrats targeted via 527s, Democratic candidates for state legislature did 4% better than in previous elections. Apparently, a team of four extremely wealthy Democratic donors singlehandedly swung the legislature to the Democrats. Adam Brown: Self-financed spending is not strategic. Candidates spend if they have it, regardless of their likelihood of victory. (Yes, that was a shameless self-promotion.)

David Konisky and Neal Woods: Smart state governments should encourage their biggest polluters to locate along state boundaries. That way, the state can reap the benefits of industry, but let all the pollution drift into neighboring states. Great theoretical story. Awesome maps showing locations of all polluters in each state. Trouble is, the presentation ended with Konisky saying that all the empirical tests produced null results. There’s no evidence that states are actually doing this. As far as the “gotcha” goes, what a letdown. But I suppose we should be glad about these null findings. Emily Huston: HAVA set minimal standards for voter identification, but allowed states to impose stricter standards. Why did some states impose strict standards but others did not? Emily threw a lot of spaghetti at the wall, but none of it stuck. The question remains unanswered. That’s two “null results” papers in one panel.

Chris Mooney: The coolest guy in state politics. Received several well-deserved honors, including a giant red pen to commemorate his work as founding editor of State Politics and Policy Quarterly.

Dan Smith and Michael McDonald apparently make a LOT of money as expert witnesses in lawsuits. And Bob Erikson looks surprisingly like the late Senator Ted Kennedy.

Boris Shor will release his common-space scores of legislators’ ideal points later this summer after a publication in LSQ comes out. Woot! (See an example of what you can do with his data.) Jim Battista and Megan Gall are assembling demographic data for all 7,380 legislative districts by matching census tracts to districts. Sounds painstaking. No word yet on whether they’ll release the data publicly so that we can all freeride. Battista/Gall’s data combined with Shor’s could be awesome.

Read tweets sent during the conference by searching Twitter for hashtag #sppc.

Broad bills ensure that general state policies remain current and fair, but legislators might avoid them for two reasons. First, they are technically complicated; if you wish to revise the state highway code, for example, you will need expert advice and probably a few studies. Second, they are politically difficult to pass; since they influence the entire state, you’ve got to work to bring a coalition of legislators on board with your proposal. By contrast, “district” bills are technically less complicated; the process of campaigning generally gives legislators all the information they need about some pressing local problem. District bills are also politically easier to pass; since they don’t have any impact outside of a small geographical area, other legislators have no reason to oppose most district bills.

By examining every bill introduced in 13 states in 1881, 1901, 1921, 1941, 1961, 1981, and 1997–that’s over 165,000 bills–Gamm and Kousser try to explain why some states produce so many more district bills than others. In Alabama, only 53% of bills had statewide impact; in Nebraska, 77% did. In general, Gamm and Kousser find that states pass more district bills when legislators have incentives to build up their reelection constituency or to make themselves stand out as an individual, but they pass more statewide bills when legislators have incentives to develop their influence and power within the state legislature. There’s more to it than that, though.

Hypotheses and Findings

Gamm and Kousser identify specific conditions under which legislators might prefer to introduce district bills or statewide bills.

Hypothesis 1 (one-party dominance): When the major parties are evenly balanced in a state, you’ll see more statewide bills; when one party dominates, you’ll see more district bills. This prediction is based on Key’s seminal 1949 book, Southern Politics in State and Nation. Key wrote that two-party competition produces “conditions favorable to government according to rule or general principle” (e.g. statewide bills); by contrast,  “in a loose, catch-as-catch-can [one-party] politics highly unstable coalitions must be held together by whatever means is available…. A loose factional system lacks the power to carry out sustained programs of action [e.g. statewide bills]” (pages 305, 308). In other words, two-party competition creates strong incentives for each party to form a statewide brand name by competing over statewide policies; in a one-party system, unstable factional coalitions within the supermajority party need to be held together with pork and other district bills. This hypothesis is confirmed in the analysis.

Hypothesis 2 (legislative salary): Legislators will introduce more district bills if they are paid more. California’s legislators earn six-figure salaries each year; Montana’s earn less than $10,000. Higher salaries create higher incentives to hold onto your job. Thus, higher salaries create incentives to deliver pork and other particularistic bills to voters in your district, even at the expense of pursuing the state’s general interest. This hypothesis is also confirmed in the analysis.

Hypothesis 3 (session length): Legislators will introduce more statewide bills if they are in session longer. California’s legislators are in session almost year round; in other states, legislators meet only one or two months per year. Because statewide bills are more complicated to write and pass than district bills, you will see fewer statewide bills in states with shorter sessions. As it happens, however, the data do not bear out this expectation; session length appears to have no relationship with the types of bills that pass.

Hypothesis 4 (turnover): Legislators can increase their statute within the legislature by building coalitions and passing statewide bills. As turnover decreases and legislators serve longer, they gain incentives to develop their stature within the legislature. As such, decreased turnover should lead to increased statewide bills. That’s the expectation, anyway. In the analysis, Gamm and Kousser find the opposite: Higher turnover leads to more focus on statewide bills. Perhaps their logic was off. Perhaps what’s really going on is this: Legislators who know they won’t stay long have less need to build up their reelection constituency (with district bills), so they can instead focus on statewide needs. I’m sure Kousser would rather not say this, but…score one for term limits?

Hypothesis 6 (district size): If districts are larger, legislators will focus more on statewide bills. Larger districts include more cities and towns, meaning that an individual district bill will reach a smaller portion of the district. In this situation, a more effective way to show the entire district that their legislator is attentive to their needs might be to pass statewide bills rather than district bills. This hypothesis is confirmed in the analysis.

Yes, I skipped hypothesis 5. I found it unpersuasive, and it ended up being statistically insignificant in the final analysis.

Punchline: If you want your legislators to pass general policies that benefit the state as a whole, pay them less, make districts bigger, and strive for partisan balance. If you want your legislators to pass pork and other district-focused bills, pay them more, make districts smaller, and promote one-party government.

Comment and Criticism

Gamm and Kousser are right to look to state legislatures to answer this question. Too often, folks think that they have to look at Congress if they want to study legislative process. The trouble with that is that there are no cross-sectional variations in the variables that we might expect to matter. In any given year, the majority party margin is the same for every member of the U.S. House; so is the salary, session length, average turnover, and so on. Sure, these things vary over time–but ALL of them vary from year to year, making it hard to tell which over-time variation is influencing the results.

Sometimes, folks who give up on Congress try instead to compare countries to one another. This approach is useful, but it introduces a host of other complications (variations in culture, constitutions, voting rules, etc.). The state legislatures are the perfect venue for this research question. The authors are able to study all the variables they care about, but all their observations are similar by virtue of their status as members of a common political system.

Masket takes this a step further, arguing that local informal party organizations control nominations and through those nominations exert control over the legislative behavior of politicians.  He argues that “parties control the public behavior of their office holders by acting as gatekeepers to political office.”  While agreeing with Aldrich (1995), that parties help organizing the legislature, Masket argues that parties, and in conjunction party nominations, are primarily a mechanism by which concerned citizens hold legislators accountable for their actions.

Testing and Results

To test these arguments, Masket uses roll-call analysis of the California Assembly prior and subsequent to the 1953 decision to mandate party labels on primary ballots, effectively eliminating the ability of legislators to cross-file in both political primaries.  His analysis of roll call votes as well as votes for the Speaker of the Assembly shows that the imposition of these new regulations on primary elections had a marked effect of increasing a the partisan nature of the legislature, as politics became less bipartisan after the change, which Masket argues was not the desire of those within the legislature.

Inside the Smoke-Filled Rooms and Thoughts on the How

In addition to his quantitative analysis of legislator behavior, Masket also details the structure of 5 different local party organizations: Orange County Republicans, South Los Angeles Democrats, East Side LA County Democrats, West LA Democrats, and the local party organizations of Fresno County.  Through interviews, Masket details the ways in which party organizations exercise their influence using things such as donor networks, sample ballot mailings, and the mobilization of activist networks.

Unfortunately, none of these items that come out through interviews are easily testable, and while Masket presents ample evidence of local influence on the nomination process, it is difficult to determine how much influence these organizations have on the outcome of nominations.

Small Qualms

While Masket’s work shows clearly the change in the influence of parties after the change in primary election law, he doesn’t give any solid quantitative evidence as to how exactly those mechanisms work.  Is it the fundraising network?  Is it the power of mobilization?   While the interviewees claimed to have influence in all of these aspects, the heads of a campaign or a campaign organization has an incentive to make their role as significant as possible in order to increase their status as the gatekeeper.  Masket clearly demonstrates to the reader that local party organizations influence nominations, but falls a little short on convincing the reader as to the mechanism through which these organizations control nominations.

A wise Latina

We like to think that in our form of government, political officials represent the citizens at large. Trouble is, it’s hard to know what “represent” means. Often, we talk about representation through two major lenses. “Descriptive” representation refers to whether people in government look like Americans generally (in terms of race, gender, maybe even age, occupation, class). “Substantive” representation refers to whether people in government make the sorts of decisions that Americans generally would make.

In a recent article, Bonneau and Rice take those two concepts into the world of judicial politics. Their basic question: Do black judges make different decisions than white judges? Bonneau and Rice provide a nice empirical answer to the question, but their interpretation of what their findings mean is confusing and less than persuasive.

Do black judges make different decisions than white ones?

Although there’s a large stack of existing articles asking the same question, previous work has produced mixed results. Bonneau and Rice argue persuasively that institutional context matters enough to resolve those mixed results in the literature.  Specifically, they argue that discretion matters. Some judges have discretion over which cases they hear; others do not. This contrast is most apparent when looking at state supreme courts. Some states have a three-tiered court similar to the federal system, with trial courts, intermediate courts of appeal, and then a state supreme court. Other states omit the intermediate courts of appeal. In either system, criminal defendants always have a right to appeal. In the absence of an intermediate court of appeals, then, the state supreme court has no discretion over which cases it will hear; it must hear all appeals.

Bonneau and Rice “contend that descriptive representation is translated to substantive representation when the court has discretion over their docket and thus an intermediate court of appeal is present” (p 3887).  Elsewhere, they state that “the link between descriptive and substantive representation relies on the expectation that minority and non-minority actors behave differently when faced with the same set of facts”–a curious statement that I discuss below, but one that I’ll accept for the time being. Putting it together, these two statements lead us to expect that race will have an effect (“minority and non-minority actors behave differently”) only when an intermediate court of appeal is present.

They test this argument by looking at the voting records of state supreme court judges, using a variety of appropriate controls (judge’s race, ideology, type of crime committed, etc). And what do they find? In states with an intermediate court of appeal, the judge’s race has no effect on voting; in states without an appeals court, race has a strong effect.

Wait–they find exactly the opposite pattern from what their theory implies. Didn’t they say that race should have an effect when there IS discretion (an intermediate court), not when there isn’t?

Perhaps I read too quickly, but nowhere did they acknowledge this curiosity or seek to explain it. We’re left, then, with an unexplained, unforeseen empirical finding. I know a fellow who likes to say, “The world is correlated at 0.3.” In other words, there are all sorts of correlations out there, but without a strong theoretical story explaining a particular correlation, we have no reason to suppose any particular finding is not random.

Bonneau and Rice are smart people, so I’m sure I missed something. But near as I can tell, the data did not support the theory.

Does substantive representation require that white and black judges vote differently?

Perhaps most curious about Bonneau and Rice’s article is this assertion, stated several times in different forms (including in the quotation given above):

“If there are no differences between white and African-American judges, then it also means that there is no substantive representation on the bench. After all, if African-American judges are deciding cases the same way as white judges, then neither group is representing the interests of minorities” (p 382).

Somebody help me understand why that needs to be true. Bonneau and Rice say nothing to back it up. Perhaps if black and white judges are deciding cases the same way, then BOTH groups are representing the interests of minorities. Must it be the case that only a “wise Latina” can make decisions about Latino defendants? Indeed, in their conclusion, Bonneau and Rice give evidence that would seem to indicate that both groups are reasonably fair to everybody:

“The vast majority of judges … are distinguished jurists with years of experience in the judicial system. These judges are also socialized the same way (both in law school and in the legal profession). This homogeneity in socialization and experience–largely unique to the judiciary–may serve to mitigate any racial differences that exist.”

I’m not trying to make an argument either way about whether the legal system is biased against minority defendants. I’m only saying that Bonneau and Rice need to back up their claim rather than merely assert it.

Parting thoughts

So where are we left?

• The authors find exactly the opposite of what their theory predicts–but they should have tried to explain this.
• The authors need to give further justification for one of their major theoretical assumptions: That substantive representation requires black and white judges to behave differently.

Previous Work

Because most of the states that adopted the initiative in this period were agrarian, western, and mostly white, historians have traditionally answered this question by claiming that the initiative was most likely to be adopted in “younger states with a strong history of populism and antimonopoly sentiments, more homogeneous [white] populations…, more radical (but weaker) political parties, and stronger interest groups” (pg 334). Smith and Fridkin contend that these conclusions have been based on deeply flawed logic and methods.¹

In particular, scholars have erred in asking why some states adopted the initiative and others did not; such an approach focuses on voter demand for direct democracy. But since the initiative is usually adopted via legislative referendum, in which legislators invite voters to choose whether to adopt the initiative, scholars should instead be asking why legislators in some states chose to offer the initiative to voters, while legislators elsewhere did not.²

Theory and Findings

Smith and Fridkin argue that legislators were most likely to offer the initiative when states experienced close partisan competition. When a minority party gains tenuous control of the legislature, or when a majority feels that it might become the minority soon, then legislators gain an incentive to weaken the legislature by empowering the median voter. This is their main causal argument, which is supported by an event history analysis.

Other conditions also play a role. In particular, legislatures in newer states are more like to offer the initiative than legislatures in older states. The authors attribute this difference to the weaker party organizations in younger states–organizations that might have less ability to rein in independent-minded legislators. Likewise, legislators were more likely to offer the initiative if there were more third-party legislators or if there was a strong interest group pressure in a particular state for instituting the initiative.

Contrary to previous work, the authors do not find any evidence of a contagion/diffusion effect. Nor do they find that initiative adoption was more common in racially and ethnically homogeneous states. Significantly, they find that dummies for the western and southern region are not statistically significant in their models; controlling for other variables apparently eliminates the regional bias in initiative adoption.

Critique

This paper’s methodological rigor makes it a welcome addition. As the authors point out, most previous work on this topic has been somewhat impressionistic and ad hoc. Event history analysis is the right way to study adoption of the initiative. Their shift from studying adoption to studying legislative support for the initiative is very smart.

That being said, a few things seem off to me. First, why did initiative adoption largely cease after 1918? Imperfect as previous work may be, its emphasis on western culture and populism seems to have an answer to this question. Smith and Fridkin don’t so much as speculate; they simply restrict their analysis to 1898-1918. But if some omitted variable caused initiative adoption to cease in 1918, is it possible that the same omitted variable, if measured, would change the results in Smith and Fridkin’s analysis?

Also, it seems that the role of governors may be downplayed. The authors start well (p 339) by listing several Democratic and Republican governors who pressured their legislators to pass an initiative reform; they even note that Republican Hiram Johnson of California made this a major goal. Yet inexplicably, the authors “control” for these considerations by including a dummy for whether the governor belonged to a third party. They provide no justification of this measure which, unsurprisingly, is not statistically significant–leading them to conclude (without basis) that gubernatorial action plays little role in adoption of the initiative. But if we believe that executives can pressure legislators by “going public” or through other means, then we should be looking more at the role governors may have played.

Third, they take no account (empirically) of variations in initiative institutions. Initiatives vary widely in (1) how hard they are to use and (2) how easily the legislature can ignore them (Bowler and Donovan 2004). Why did some states (California, Oregon) pass powerful initiatives, but other states (Utah, Idaho) passed weak ones? The authors acknowledge this variation, but do not seek to explain it. My speculation: The “party competition” variable that the authors discuss might be two different things. If a minority (perhaps allied with third parties) seizes control for a short while, or if a majority genuinely fears becoming a minority, then we might see an effort to create a truly powerful initiative. But if a majority sees its power slipping and wants to do something popular to build up support, then we might see a weak initiative created purely for its symbolic value. Thoughts?

Every study has weaknesses, and the flaws in this one are by no means fatal, nor are they necessarily more numerous or severe than what is typical for an article in a top journal. But I think they point to some avenues for further research.

It is one of the happy incidents of the federal system that a single courageous state may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country.

These claims have inspired a literature on “diffusion,” mostly focused on the American states, that has asked whether there is evidence that the states do, in fact, learn from one another. For the most part, the existing literature has concluded that diffusion occurs regularly among the American states, although there is some disagreement as to the mechanism.¹

By contrast, Volden, Ting, and Carpenter argue that none of the existing work (including their own previous work on diffusion, apparently) has provided any evidence of policy diffusion. The methods and assumptions used in previous research cannot differentiate between innovation (isolated experimentation by myopic states) and diffusion (learning from experiments in other states).

Overview of the Model

To correct this problem, the authors present two formal models of experimentation. Both begin from the same basic setup: States have policy makers (legislators, bureaucrats, whatever) that can be placed along a unidimensional ideological line (i.e. we can classify the state as conservative or liberal). Within a particular issue area, there is a status quo policy and a proposed experimental policy. Each policy has two characteristics. First, it can be placed along the ideological line, and this “spatial” characteristic is common knowledge–that is, everybody agrees as to which policy’s goals are more liberal or conservative. Second, each policy has a “valence”–that is, each policy might be more or less effective at reaching its stated goals.

Although each proposal’s “spatial” characteristics are assumed to be common knowledge, “valence” is known only for the status quo–the experimental proposal’s valence is unknown. Thus, policy makers have a choice: They can stick with the status quo (with known valence) or they can switch to the experimental policy (with unknown valence). If they choose to experiment, then in “period two” (e.g. the next legislative session), when the experimental policy’s valence is known, they can choose to stay with the new policy or revert to the old one.

After setting up this basic model, the authors derive two models from it. The first is a decision-theoretic model that assumes each state exists in isolation; states may innovate similar policies, but there cannot (by definition) be diffusion. The second is a game-theoretic model that assumes each state can learn from policy experiments in other states; either innovation or diffusion can occur.

The decision-theoretic model

In this model, states do not have the option of learning from other states’ experiments.

In an extremely liberal state, the policy makers will choose the most liberal policy proposal, regardless of expectations about valence. (Likewise for extremely conservative states). But in moderate states, policy makers will balance valence against ideology. A moderate conservative would prefer an efficient but liberal policy over an inefficient but conservative policy; a moderate liberal would prefer an efficient but conservative policy over an inefficient but liberal policy. These are the conditions under which innovation occurs.

Thus, we would expect to see only moderate states experimenting. If they learn that the policy is inefficient, then they would revert to the previous policy–the one that is ideological preferable.

The game-theoretic model

In this model, states do have the option of learning from other states’ experiments.

If states can learn about a proposed policy’s valence by observing policy experiments in other states, then the incentive to experiment drops. Experimentation is risky; if you can learn from others’ mistakes rather than having to make those mistakes yourself, then why experiment at all? As such, those policy makers willing to experiment will fit into a narrower ideological range than those willing to experiment in the decision-theoretic model.

Implications for the Literature

Previous work on “diffusion” has not appreciated the differences between these two models. All of the evidence for policy diffusion presented in previous work can be explained in terms of the decision-theoretic model. In order to conclude that diffusion actually occurs, we must find evidence of behaviors that are predicted by the game-theoretic model and NOT by the decision-theoretic model.

The literature has presented five different causal mechanisms to explain “diffusion,” but all five mechanisms can be explained with the decision-theoretic model:

• Walker says some states are inherently more disposed to innovate than others.
• Gray says diffusion happens when states face similar policy problems.
• Others say that diffusion happens when neighboring states are ideologically similar.
• Still others say that diffusion happens when any states (neighboring or not) are ideologically similar.
• A final argument is that diffusion happens when policy advocates take their arguments to multiple states.

As the authors put it: “Much of the empirical work to date has not adequately distinguished [game-theoretic] learning-based policy diffusion from [decision-theoretic] myopic individual adoptions.”

The authors conclude by listing specific empirical implications of their model that future research should evaluate in order to determine whether or not diffusion actually occurs.

Critique

Much of the work on diffusion was published in the 1960s and 1970s. I suspect that modern political scientists have hesitated to take up the question again because they thought it was settled. This article demonstrates persuasively that the question is far from settled. This is the article’s most important contribution–to point out that we do not yet have any evidence that states do (or do not) act as “laboratories of democracy” that learn from one another. This question is normatively important; if states do not learn from one another, then we lose an argument for devolution.

However, I was less than satisfied with the authors’ empirical suggestions. They conclude by pointing out several ways that we can use their theory to empirically determine whether states follow the decision- or game-theoretic model. Unfortunately, these empirical implications are extremely nuanced and may be difficult (or nearly impossible) to apply in practice.

In the absence of direct democracy, political parties might not have much reason to promote moderate policies. Republican legislators would generally prefer policies to the right of the median, while Democratic legislators would prefer policies to the left. Whichever party has the legislative majority has a variety of tools at its disposal to help it push policy away from the median and towards its own ideal point. These tools include control of the legislative agenda, control of committee chairmanships, and so on. In addition, the need to appease major campaign donors and partisan activists increases the incentive to use these tools. As a result, it is not the legislative majority that governs; it is the majority of the majority party that governs. When legislative control shifts from Republicans to Democrats, you might see a large shift in policy outcomes–even if only a couple of legislative seats changed hands.

But this is exactly the sort of thing that the Progressive movement sought to end. A purpose of direct democracy was to force legislators to produce policy outcomes closer to what the median voter would want–regardless of which party has the legislative majority. The purpose of Phillips’s article is to ask whether they succeeded.

There are two main mechanisms by which direct democracy can be a median-enhancing institution (from Gerber 1996). The first mechanism is direct; voters can impose specific legislation on the legislature, as happened with California’s famous Prop 13. The second is indirect; even if voters never use the initiative, its presence acts as a deterrent against extreme behaviors by the legislature.

Phillips finds evidence that direct democracy does matter. Party government is weaker in states with the citizen initiative than in states without. Here’s how he did it:

• Dependent variable: The state tax burden.¹
• Main independent variables: First, a few dummies to measure partisan control.² Second, a dummy indicating whether a state has direct democracy. Third, interactions between these two sets of variables.

As expected, he finds that the partisan dummies have strong relationships with the state tax burden. Democratic governments tax the most; Republicans tax the least; divided governments are in the middle. But here’s the rub: these relationships disappear in states with direct democracy.

The existing literature on the relationship between partisan control and size of government has had mixed results. Phillips contends that these mixed results can be explained, at least in part, by looking at direct democracy.

I have only one complaint with this article. I’m not convinced that a simple dummy variable can adequately measure direct democracy. From state to state, there are huge variations in how easy it is to use the initiative process, leading to huge differences in how frequently the process is used. This article could be more persuasive if it discussed this problem, which Phillips does not even mention. In fact, this problem is severe enough that, to me at least, it undermines the entire argument.

Nicholson-Crotty presents evidence that it does not. Instead, he finds that states (indirectly) refund a significant proportion of federal funds to state taxpayers. When states receive money from the federal government, they use it to reduce state tax rates.

More precisely, Nicholson-Crotty finds that an increase in grant monies (X) leads to a decrease in the state’s taxation effort (Y), a measure (from the Advisory Commission on Intergovernmental Relations) of the state’s effective tax rate relative to the amount of money the state government could (hypothetically) tax. As X goes up, Y goes down.

This doesn’t mean that states don’t spend the money the way Congress wanted it spent. But it does mean that the states are cutting their own spending in the particular area, and possibly in other areas, keeping the state’s overall spending somewhat constant despite the influx of federal funds.

Now, this relationship isn’t perfect. Nicholson-Crotty considers three factors that might affect the extent to which federal monies simply get returned to state taxpayers.

1. First, the type of grant. An unconditional grant gets refunded to taxpayers at the highest rate.
   A matching grant (wherein the state receives more federal money if it spends more of its own money on the program) gets refunded to taxpayers at the lowest rate.
2. Oversight. If Congress exercises lots of oversight about a particular grant program, we might expect a different pattern. (But it turns out, in the final analysis, that oversight matters little.)
3. Preferences. State governments that have a strong prior preference for a Medicaid-style program will be more inclined to spend the federal government’s money without cutting money elsewhere.

As always, there are several control variables derived from previous work, such as partisanship, fiscal centralization, wealth within each state, urbanization, tax/spending limits, and election year effects.

And as always, there are a few methodological shortcomings in this article. Nicholson-Crotty addresses two–missing data and endogeneity–and applies reasonable remedies. The endogeneity issue is interesting; federal money is most likely to go to those states that are most likely to spend it–that is, to liberal states.

A third, which he does not discuss, is that his first intermediate variable (grant type) does not vary sufficiently. He examines only matching grants. Although the level of matching varies (from $.50 on the dollar to $.77), I would like to see more variance here. This is a bigger oversight than you might think. Given NIcholson-Crotty’s findings, members of Congress might like to know what they can do to minimize the problem and ensure that federal grants have the intended effects. An easy solution implied by the article is to switch from block grants to matching grants with a highly elastic matching rate–but this article does not directly tell us how much we can gain from such a change.

The political implications of this study are broader than the author seems to realize. To see why, consider these two points from the article:

• When discussing endogeneity problems, he tells us that federal money is mostly likely to go to the “tax and spend” states.
• When discussing his results, he tells us that a major effect of federal grants is to reduce a state’s tax burden; the state gets to have the same amount of spending (or a little more) as before, but at the expense of taxpayers in other states.

Putting it all together, Nicholson-Crotty is telling us that federal grants-in-aid are little more than a redistribution of the income tax burden from liberal states to conservative ones. No wonder Republicans have been so interested in their “devolution revolution.”

For other views on the “devolution revolution,” see here and here.