July 15, 2021
For students taking my Law and Econ course and who want to delve deeper into the scholarly literature, the papers below are a great place to start. What follows are my summaries—which go beyond the abstracts. My inclusion here does not indicate full agreement with everything in every paper; but it signifies the importance of the paper to Law and Econ scholarship. I don’t include “The Nature of the Firm” or “The Problem of Social Cost,” as these, the foundational Law and Econ papers, are already very well-known. Likewise for classics “Toward a Theory of Property Rights,” “Property Rules, Liability Rules, and Inalienability: One View of the Cathedral,” and “The Role of Market Forces in Assuring Contractual Performance.” I don’t include many papers here that examine contracts extensively—but I do in this class.
The list reflects my bias for comparative institutional analysis. (For instance, there are many empirical L&E papers that ask (important) questions like: Does prison deter? or Do “three strikes laws” deter? With the exceptions of the Levitt and Wolfers/Donohue paper, these sorts of questions aren’t my focus). The average paper summarized here has over 825 citations as of this writing (”The Problem of Social Cost” has over 40,000 citations).
1.
Ellickson, Robert C. "Of Coase and cattle: Dispute resolution among neighbors in Shasta County." Stanford Law Review 38 (1985): 623-687.
Ellickson goes to Shasta County, California for an ethnographic investigation. There he talks with cattle ranchers and farmers to determine the applicability of the Coasean story. In this field work, he finds that the legal rules for liability change often, but the property rights allocations do not. This is not, however, because of side payments a la Coase. The patterns of activity simply do not change with the law. The transactions costs are not low; they are high. The high cost of tracking and enforcing formal legal rules make side payments non-economic. Instead, the individuals are constrained spontaneously-emergent, informal norms that lead to efficient outcomes. Enduring social relationships lead to mutual consideration. “Open” and “Closed” ranges—differential property rights allocations—did not change the behavior of Shasta County inhabitants. They instead relied on informal norms and punishments (gossip, harming trespassing cattle, paying for destroyed crops, etc.) that were self-enforcing. This was bargaining beyond the shadow of the law; not in it. People can get along and create efficient outcomes using informal norms without reference to formal laws. The law (often) matters less than you think.
2.
Rizzo, Mario J. "Law amid flux: The economics of negligence and strict liability in tort." The Journal of Legal Studies 9, no. 2 (1980): 291-318.
The classic “Austrian” paper on the then-emerging Law and Economics paradigm. Rizzo first critiques the prevailing “efficiency” standard in law and econ. This approach is most clearly seen in tort law where “negligence” is assessed on the basis of some Kaldor-Hicks dollar maximization standard. The clearest example is the Hand Rule. This approach also shows up in certain readings of Coase (1960). The so-called “normative Coase Theorem” would have judges decide in property disputes as if they were the market—that is, to allocate to the party who would come to possess the property right if transaction costs were negligible. This is the context for Rizzo’s critique.
The basic claim is that courts cannot engage in the cost-benefit analysis that standard L&E would have them do due to the Austrian problems of uncertainty, subjective value, and the passage of time. As a result, Rizzo recommends substituting the bright-line rule of strict liability for negligence regimes. Rizzo examines a few classic cases—Bolton v. Stone, Re Polemis, Vincent v. Lake Erie Transportation, and Rylands v. Fletcher—to illustrate the difficulties in applying the conventional efficiency standard. The substitution of strict liability makes causality the primary question in any tort case. It makes first-possession the primary question in any property dispute.
Once the problems of uncertainty, subjective value, and time passage are accounted for, ironically, a universally-applied strict liability regime is more efficient than is the typical efficiency standard that attempts to weight negligence (tort) or divine the party with the highest willingness-to-pay (property). Such an approach allows for what Rizzo elsewhere calls “praxeological coherence” (in contrast to the “logical coherence” of the standard approach). As he puts it:
“In a dynamic world in which the uncertainties of technological change, the ambiguities of foreseeability, and the absence of a unique objective measure of social all conspire to make the efficiency paradigm a delusion, the importance certainty in the legal order is clear. Strict liability obviates or minimizes the need for courts to grapple, if only implicitly, with such impossibly elusive problems as foreseeability, cheaper-cost avoider, social cost, and second best. It provides a series of basically simple, strict presumptions. The prim facie case is based on straightforward commonsense causal paradigms whereas the defenses and later pleas minimize the number of issues which must be considered in a given case.”
Similarly, on the classic Coasean example, Rizzo notes:
“Consider the famous case of the railroad operating near a farmer's land. The farmer has decided to let crops grow right up to the edge of his property even though this means that the sparks emitted by the train will destroy some of the crops. Should the farmer be allowed to recover for the damage to his crops? In a system of strict liability, there is no need for a problematic cost-benefit analysis of the value of more food versus the value of faster train service. The railroad's use of its property oversteps the physical bounds of that property by the emission of particles in the form of sparks. The farmer's use does not similarly overstep the physical bounds of his property, and his conduct can be taken as constant.”
3.
Peters, H. Elizabeth. "Marriage and divorce: Informational constraints and private contracting." The American Economic Review 76, no. 3 (1986): 437-454.
Peters presents a test of the Coase Theorem in divorce cases. Some states had unilateral or no-fault divorce schemes, which meant that one party could initiate divorce without the other’s consent. The alternative is mutual divorce or a costly legal battle. In her model, under symmetric information, the ex post value of each spouse’s opportunities is known to both. In this case, Coasean bargaining would suggest that marriages would be dissolved when it is efficient to do so. If the net benefit across both parties of retaining the marriage is higher, then the divorce will not happen and vice versa. In this low transaction costs world, a no-fault law would have no affect on the marriage rate. The no-fault law simply switches the property right of dissolving the marriage from one party to another. Under mutual divorce, the party valuing the marriage more has the property right; the one who wants out will buy it if he values the divorce outcome more than she values staying in the marriage. Under no-fault, the party who values marriage least has the property right; the one who wants to retain marriage will compensate if it is efficient to do so. The symmetric case predicts no change in divorce rates, but a change in compensation since the property rights allocation changes. Alternatively, an asymmetric information condition would preclude efficient bargaining, and we would expect the law to increase divorce rates. Spouses would enter a fixed wage contract under these conditions. Peters finds that the divorce rates are unaffected by the no-fault law, but that women’s compensation is lower in the no-fault states. She interprets this as evidence “for” the Coase Theorem.
My own view is that the Coase Theorem is not “testable.” It’s logically true when certain conditions are met. (Agreeing on just what those conditions are is a difficult task, as illustrated by the Block-Demsetz debate). What’s really being “tested” is whether transaction costs were sufficiently high to prevent bargaining and thus whether alternative property rights regimes matter for “outcomes.” Also, see Doug Allen’s insightful comment. He finds that Peters’ results are sensitive to geographic controls. That is, divorce rates do seem to differ between “fault” and “no-fault” divorce regimes. Furthermore, Allen points out that Peters assumed a zero-transaction cost world—which isn’t our world. In short, we should expect the divorce regime to matter for the number of divorces.
4.
Ellickson, Robert C. "A hypothesis of wealth-maximizing norms: Evidence from the whaling industry." Journal of Law, Economics, and Organization 5 (1989): 83-97.
The most fun paper on this list, and also criminally under-cited relative to its quality. It asks a critically important question: What is the source of property rights? The “legal centralist” tradition sees the state as “the exclusive creator of property rights.” Another tradition sees property rights emerging from “the workings of non-hierarchical forces.” Ellickson not only adopts this second view, but also makes the tougher case that norms which emerge are “wealth-maximizing.”
To argue these points, he examines the practices of Greenland and New England whalers from 1750 to 1870. As context, whaling communities tended to be tightknit (thus allowing informal enforcement to operate) and there was no maritime law which governed how one came to have rightful possession of a whale. There are many hypothetical rules that could govern who owns which whale. It turns out that the rules which emerged were well-suited to the particular challenges each group of whalers faced. Among British whalers, a “fast-fish, loose-fish” rule emerged. Under this rule, an unbroken line between harpoon and boat established ownership. This rule rewarded the ship which was the first to harpoon a whale, the most challenging part of the hunt. Furthermore, the right whales hunted in this region were sluggish and unlikely to break the line.
By contrast, American whalers converged on a different rule—”iron holds the whale.” Under this rule, the first harpooner of a whale was its owner, regardless of whether the line remained unbroken. Sperm whales were vicious fighters who often broke the line, but this rule incentivized getting one’s harpoon into a whale, which was no easy task. As mentioned, these rules emerged “spontaneously” and they worked remarkably well, as evidenced by the low degree of litigation over whale ownership.
The paper is a great example of pursuing research in the “Mengerian” tradition which asks, “How can it be that institutions which serve the common welfare and are extremely significant for its development come into being without a common will directed toward establishing them?”
For more on rules outside the formal legal system, probably the most-cited classic is Lisa Bernstein’s “Opting out of the legal system: Extralegal contractual relations in the diamond industry.”
5.
Glaeser, Edward L., and Andrei Shleifer. "Legal origins." The Quarterly Journal of Economics 117, no. 4 (2002): 1193-1229.
Part of a group of papers on comparative legal systems that have been among the most cited in economics of the last two decades. See “The Economic Consequences of Legal Origins” for an overview of these papers.
Many of the world’s countries inherited either the English common law or the French civil law tradition. Even in the present day, English common law heritage countries appear more favorable to free markets, property rights security, and contractual enforcement. French civil law heritage countries seemingly evince more regulation, worse property rights protection, government corruption, and the like. This paper attempts to answer two questions: Why did legal systems evolve differently in England vs. France? Secondly, why do these differences seem to generate variation in social and economic outcomes?
LaPorta et al. (2008) summarize the answer to the first question:
“They argue that England was a relatively peaceful country during this period, in which decentralized dispute resolution on the testimony of independent knights (juries) was efficient. France was a less peaceful country, in which high nobles had the power to subvert decentralized justice, and hence a much more centralized system, organized, maintained, and protected by the sovereign, was required to administer the law. Roman law provided the backbone of such a system. This view sees the developments of seventeenth and eighteenth centuries as reinforcing the structures that evolved over the previous centuries.”
Glaeser and Shleifer categorize the consequences of alternative legal origins into buckets: civil procedure and social outcomes. The differences in civil procedure are numerous, but here’s one example that highlights what Rizzo might call “praxeological coherence”:
“Absent bright line rules and other guides for adjudicators, precedents may serve to remind judges and juries where the law has drawn lines previously. Despite precedents, it is common for advocates in common law systems to draw subtle distinctions between cases, unlike in the civil law systems, where similarities are sought by a judge [Damasïka 1986]. Nonetheless, precedents may serve to eliminate excessive unpredictability, which may be a natural consequence of the importance of individual trials and of particular sentiments of the juries. “Certainty is achieved in the common law by giving the force of law to judicial decisions, something theoretically forbidden in civil law” [Merryman 1969, p. 51].”
On the question of social outcomes, Glaeser and Shleifer wish to explain why civil law heritage countries tend to exhibit more interventionism and offer less security of property rights protection. In short, they argue that civil law systems, when transplanted, do not comport as effectively with “liberalism,” as do common law systems.
Put simply, regulations and controls [civil law] are much more vulnerable to misuse by
the sovereign than is community justice [common law].
Before this empirically-oriented debate, economists debated the merits of common vs. civil law. See Todd Zywicki’s excellent “Gordon Tullock’s Critique of the Common Law” for a helpful analysis of this debate.
6.
Farnsworth, Ward. "Do parties to nuisance cases bargain after judgment? A glimpse inside the cathedral." The University of Chicago Law Review (1999): 373-436.
Farnsworth colorfully details twenty nuisance cases and asks if there is post-judgment bargaining in the shadow of the law. Presumably, many types of transaction costs are low. There aren’t technical feasibility costs or other physical restraints on exchange. However, in none of the twenty cases did either of the parties bargain or expect that bargaining would have happened if the case had gone the other way. This was largely due to ‘animosity.’ Farnsworth does not think that animosity or personal disputes should be counted among the costs in transactions costs. In my view, the paper would benefit from a healthy dose of subjective value. At the end of the day, parties did not find it worthwhile to negotiate swaps to property titles—we can’t necessarily know why. The value of the paper lies showing that post-judgment bargaining is likely rare.
7.
Weinstein, Mark I. "Share price changes and the arrival of limited liability in California." The Journal of Legal Studies 32, no. 1 (2003): 1-25.
What is the effect of limited liability on share prices? It’s a hard question to answer because, in many places, stock markets and limited liability developed concurrently. This paper examines the effect of California moving from unlimited to limited liability between 1929 and 1931. The “man on the street view” is that limited liability would be unambiguously positive both for corporations and their shareholders, leading to a higher willingness-to-pay for that corporations shares. The Coase Theorem predicts, however, that creditors would simply charge higher prices to the firm in a regime of limited liability. Little benefit would accrue to the shareholders. Thus, the null hypothesis is that the move to limited liability would have no effect on share prices. Weinstein examines publicly traded companies in CA. He looks at two groups: a small sample of those in CA traded on the NYSE compared to similar non CA firms and then compared all publicly traded firms with the average percentage change of all NYSE-listed firms. For the first group, he does not find significance on the “California” dummy. However, this sample size is small and thus inconclusive.
In the larger groups (including companies unlisted on the stock exchanges), he finds little evidence of a positive/pervasive effect of the change. In a last attempt to find an effect, he divides his CA firms into banks and nonbanks. He finds no significant difference in the rate of return to banks and nonbanks, a finding consistent with his hypothesis. He considers objections to his findings: selection effects with the type of firms that had chosen to locate in CA and sample size issues. Additionally, there might have been “de facto” limited liability before the change. Another possibility is that the change didn’t matter for firms in CA, but this seems unlikely given the wide breadth of industries in CA. While he finds no effect, he says it’d be a mistake to say limited liability is of no consequence, particularly due to the concerns of de facto limited liability existing.
The least-cited paper on this list, but it should be better known.
8.
Hansmann, Henry, and Reinier Kraakman. "Toward unlimited shareholder liability for corporate torts." Yale LJ 100 (1990): 1879-1934.
The authors argue normatively in this piece. Limited liability is believed to create incentives to excessive risk-taking since a corporation does not bear the full costs of its actions. The existing literature provides no comparison of limited and unlimited liability for torts. They conclude that the disadvantages of limited liability outweigh any advantages. They argue this is especially relevant in a world where firms are increasingly imposing larger and larger risks (like carcinogens and environmental disasters). The Article systematically explores the case for retaining a rule of limited shareholder liability by comparing it with pro rata shareholder liability for torts. They argue that there’s no reason unlimited liability would discourage investment except for in those firms that impose a net loss in society—which is where we’d want to discourage investment. They believe that unlimited liability would, in fact, depress share prices, but that this price would better reflect the full social cost of corporate activities. This is contrary to the prediction (and findings) of Weinstein (2003).
They do acknowledge, however, several arguments for limited liability. First, shareholders might attempt to hide assets in order to evade liability. Second, it would it not be easy to implement unlimited liability in one jurisdiction when limited liability governed a neighboring jurisdiction. Third, limited liability might serve as a necessary check on the expansionist tendencies of the courts in deploying liability rules. They conclude by saying that the weaknesses of limited liability should call for greater examination.
9.
Ringleb, Al H., and Steven N. Wiggins. "Liability and large-scale, long-term hazards." Journal of Political Economy 98, no. 3 (1990): 574-595.
This paper analyzes the application of liability to large-scale, long-term hazards (like asbestos). These hazards have two key features: first, there is a long time between exposure and harm; second, there are large damages when harm finally emerges. The latter creates a strong incentive to avoid liability, while the former creates possible avenues to avoid paying damages. Specifically, the strategy of vertically divesting risky production tasks is examined in this paper. By examining startup firms between 1967 and 1980—the period where liability laws were changing in the US—the authors find that, ceteris paribus, liability laws appear to have led to large increases of small firms in hazardous sectors. This is an attempt to avoid liability by either a. the small firm going out of business before suits are filed or by b. becoming judgment-proof before the awarding of damages. Tort was becoming increasingly popular (as opposed to regulation) because it does not require regulators to know extent of the damage beforehand.
One way for firms to shield themselves against tort is to spin off risky activities into small firms. This is particularly valuable when the injured is limited in compensation to the total value of assets of the injuring firm. Because the authors find evidence for their hypothesis, they argue that liability may not lead to large damage awards in long-run equilibrium. Since the damages can “pile up” over a long period of time, this creates an incentive to avoid liability. The temporal separation between injury and damages also creates the opportunity to avoid liability. One way to avoid liability is to shut down the hazardous part of the company and then buy the required input from a small, specialized producer. This leads to an equilibrium where small firms handle hazards, leading to improper incentives to take care. The null hypothesis is that worker exposure to occupational hazards (measured by exposure to carcinogens here) will be unrelated to the entry of small firms. The results strongly reject this hypothesis. Furthermore, exposure is a major determinant of small firm entry. A one percentage increase in exposure is associated with an increase of 25,000 small firms. One suggested response is to assess damages earlier. Another is to apportion damages according to market share.
10.
Kessler, Daniel, and Mark McClellan. "Do doctors practice defensive medicine?" The Quarterly Journal of Economics 111, no. 2 (1996): 353-390.
Kessler and McClellan start by noting the two roles of the medical malpractice system: compensating those who suffer negligent injuries and giving incentives for physicians to take efficient care. They cite the classic 1990 Harvard study that found that many more individuals are harmed by negligent care than receive compensation. The authors note that the incentives created by the liability system can lead doctors to be “too careful”—they might try to avoid risky services and perform unnecessary extra tests to protect themselves from liability. They use data on all elderly patients in the United States who had heart issues from 1984, 1987, and 1990. They look at the different in treatment and other outcomes in areas that did and did not adopt liability changes over the period. They find that limiting the liability by limiting award amounts lowered the cost of healthcare to patients but did not adversely affect health outcomes (including death within a year and typical complications). By reducing the potential litigation risk to physicians, the price of health care goes down while the quality remains, suggesting that doctors do indeed practice defensive medicine in response to the unlimited upside risk of liability.
11.
Klerman, Daniel. "Jurisdictional competition and the evolution of the common law." University of Chicago Law Review 74 (2007): 1179-1226.
Klerman looks at English legal history in the 17-1800’s. In this common law regime, judges received fees for cases, and there was overlapping jurisdiction. As such, judges and their courts competed to hear more cases, because the incentive structure was to hear more cases. Thus, competition led to swifter, smoother proceedings. However, it also led to difficulty for defendants to prevail. He shows that after 1799, when fees were abolished, the pro-plaintiff bias of the courts dissipates. The pro-defendant aspects of common law are actually 19th century phenomena that came after fees were abolished.
For more on competing court systems, see here. For the workings of private legal systems, here’s the classic place to start. For the best paper on private governance published in a top journal, see here.
12.
Fisman, Raymond. "Estimating the value of political connections." American Economic Review 91, no. 4 (2001): 1095-1102.
During the late 1990’s, as Indonesia’s economy spiraled downward, some were concerned that political connections, rather than fundamental issues of productivity were the true determinants of firm profitability. Testing this hypothesis is difficult. Defining political connectedness is likewise difficult. There are also endogeneity issues. Fisman uses an event study of India to investigate whether political connectedness is an important driver of profitability. He uses variations in Indonesian president Suharto’s health to compare the returns of firms with differing degrees of political exposure. Fisman finds that well-connected firms suffer more than less-connected firms in response to a severe rumor of Suharto’s health. His measure of political connectedness is the “Suharto Dependency Index” put together by a geopolitical consulting firm. One robustness check examined the relationship between other types of “bad” news and the returns of connected firms. In this case, there was connection between the bad news and the firm’s returns, suggesting that it was bad news about Suharto’s health specifically that is driving the results. These results suggest that political connections play a large role in many of the world’s largest economies, as most indices rank many other countries as being more corrupt than Indonesia.
One caution with papers like this. The mainstream of the economics profession muddies the distinction between “theory and history.” What this paper discovers isn’t some universal law about the relationship between political connections and firm outcomes. Rather, it’s an important piece of economic history, which demonstrates the extent to which political connections mattered in this historical episode. And it inspires us to examine similar questions in other episodes, though the constraints in those other episodes may very well differ.
For more Fisman on corruption, see here.
13.
Miron, Jeffrey A., and Jeffrey Zwiebel. "The economic case against drug prohibition." Journal of Economic Perspectives 9, no. 4 (1995): 175-192.
Since drug use is clearly destructive both to personal and communal health, many have advocated for government prohibition of production, sale, and consumption of drugs. A large number of economists have suggested, however, that many of the ills associated with drugs are actually a function of prohibition, rather than the drugs themselves. This paper argues for that position. The most direct effect of prohibition is a leftward shift of the supply curve for drugs. It likely does the same thing for the demand curve due to penalties for consuming drugs, but the shift is likely less than the supply curve because (among other reasons) penalties are almost always more severe for suppliers than demanders. Prices rise and equilibrium consumption falls. Prohibition lowers the marginal cost and raises the marginal benefit of violence because participants cannot rely on legal system and also because, as criminals, the opportunity cost of violent acts is lower. (One might reasonably question Miron and Zwiebel’s position that inability to rely on the formal legal system increases violence, since there are many cases of parties voluntarily opting out of the legal system to rely on other means of dispute resolution). The view is that prohibition, not consumption, increases violence. Under alcohol prohibition, the murder rate spiked.
There are several reasons why underground industries are easily cartelized. Among them, the MC of using violence to enforce the cartel is lower. This same reason means it’s harder for small entrants to compete (not true in legal markets). Cartelization prevents dissipation of profits, which reduces the MC of violence further, since there’s a larger prize to be had. It also provides real resources with which to bribe police. Prohibition leads to increased uncertainty about product quality, since buyers can’t complain about product quality without incriminating themselves. The result is more accidental overdoses/poisonings. Accidental over-doses are also more likely due to the Allen-Alchian effect. This was observed empirically during alcohol prohibition. There is an increase of property crime under prohibition, which the authors attribute to the inelastic demand for drugs (people are finding alternative ways to acquire income). Since prices are artificially high, this effect is exacerbated.
The paper also conducts a “normative analysis” of prohibition in which they argue against the common views that drug users are uninformed and the view that all drugs are highly addictive. They also argue that the negative health consequences of drugs have been overstated. Next, they tackle the objection that consumption generates negative externalities. First, they point out that prohibition might incent shifting to subs (alcohol, tobacco). Second, enforcement is targeted more at light users, but externalities are important for heavy users. Lastly, they do a historical tour of Prohibition. Prior to drug criminalization, they point out the US was “not a country of drug addicts.” In the conclusion, they discuss alternative policy options. Among them, medicalization allows true addicts (who have a high inelasticity) to more safely consume drugs, under the supervision of a medical professional.
Image from Mark Thornton’s The Economics of Prohibition:
14.
Wolfers, Justin and John Donohue. Uses and abuses of empirical evidence in the death penalty debate. Stanford Law Review (2005) 58: 791-846.
A tricky and tangled literature. In the 1972 “Furman” case, the Supreme Court ruled that death penalty statutes were unconstitutional. Ehlrich (1975) claimed, however, that each execution saved eight lives. Partially due to this work, the Supreme Court re-established the use of the death penalty in the following year. The current state of the debate is one of extreme disagreement. Katz, Levitt, Shustorovich (2003) find no evidence of deterrence. Sunstein and Vermeule (2005) claim to have identified a powerful deterrent effect. This paper finds the evidence of deterrence is surprisingly fragile. The authors are certain that the effects are not large, but are uncertain whether they are positive or negative. Whether one gets a positive or negative effect is highly sensitive to minor changes in econometric specifications. Ehlrich’s results, for instance, relied heavily on data between 1963 and 1969, but when the data was restricted to 1935-1962, the deterrent effect disappears. The fact that there’s so much variation in the correlation between homicides and execution over time suggests that the deterrent effect itself may be changing over time. One would expect it to be weakened over time due to administrative red tape, but the correlations would suggest that, if anything, it got stronger during the 1990s.
Comparing Canada and the U.S. shows that “Furman” was not responsible for the spike in US homicides, since there was a comparable spike in Canada that had not been practicing the death penalty for years. Similarly, there is close co-movement between states that had the death penalty and those that didn’t. The paper employs time and stated fixed effects on data from 1960 to 2000, running a WLS regression, clustering standard errors at the state level. They find the coefficient on the dummy to be insignificant. Even when they account for some death penalty states being essentially de facto non-death penalty, they still cannot get the coefficient to be statistically significant. Similarly, they find that the re-establishment of the death penalty was associated with an increase in homicides in some states and a decrease in others. It is difficult to exploit variation in the intensity with which death penalty laws are applied because the annual number of executions fluctuates very little, whereas the annual number of homicides does.
The authors also re-test the results of Dezhbakhsh and Shepherd (2004), showing that their finding of a deterrent effect is not robust to the exclusion of Texas. In a test of another pair’s paper, they find that a moratorium on the death penalty in Illinois did not cause substantial divergence in homicides between Illinois and the rest of the country. All of these studies might fall prey to endogeneity: re-instating the death penalty might be part of a broader “get tough on crime” movement. They additionally examine a paper that makes use of instrumental variables such as: number of prison admissions, police payrolls, judicial expenditures, etc…, but find these to be weak since their variation should affect the homicide rate by other channels than just through the death penalty. In the words of the conclusion: “We have surveyed data on the time series of executions and homicides in the United States, compared the United States with Canada, compared nondeath penalty states with executing states, analyzed the effects of the judicial experiments provided by the Furman and Gregg decisions comparing affected states with unaffected states, surveyed the state panel data since 1934, assessed a range of instrumental variables approaches, and analyzed two recent state specific execution moratoria. None of these approaches suggested that the death penalty has large effects on the murder rate.”
One wonders if there’s a way to exploit variation in shifts between sentencing and execution. Ceteris paribus, this lag would dampen the deterrence effect of the death penalty.
15.
Levitt, Steven D. "Understanding why crime fell in the 1990s: Four factors that explain the decline and six that do not." Journal of Economic Perspectives 18, no. 1 (2004): 163-190.
Levitt attempts to answer: “why did crime fall in the 1990’s?” He lists a number of relevant factors and a number of irrelevant factors. When reading this paper, it’s critically important to separate positive from normative analysis. Here are the things that do not explain the decline: the strong economy of the 1990’s, changing demographics, better policing strategies, gun control laws, concealed weapons laws, and increased use of the death penalty. However, the things that did affect the drop in crime: decreasing crack epidemic, increased abortion, increasing numbers of police, and increasing prison population. Crime decreased in magnitude steadily through the 1990’s, with the homicide rate dropping almost 50 percent over the decade from ~9 to ~5 per 100,000 residents. Almost every type of crime decreased between 20-60 percent from 1991 to 2000. A stronger economy seemingly has little impact on crime, despite the increased labor market opportunities raising the opportunity costs of crime in theory. (This is a puzzling result that, to my mind, suggests ceteris is not paribus).
Likewise, Levitt argues that the changing demographics do affect crime, but the changes in this period produce a net null effect. There was an aging population, which should reduce crime (baby boomers). However, there is also a surge in the adolescent population which was expected to increase crime. If anything, the demographic change reduced crime slightly. “Better” police strategies do not seem to be a factor, largely because there is insufficient evidence. New York City is the main locality that innovates in police strategy, but it is unclear that the effects here are different from other large cities after accounting for the changing size of the police force. “Reasonable people could disagree on the impact, but I think it’s small.” Gun control laws are ineffective—especially gun buyback programs. The marginal gun bought back is the gun that has least value in other uses (i.e., in crime). Cases of DC and Chicago present little evidence of gun control success. Concealed carry laws that would allow individuals to carry guns should theoretically increase the costs of crime. However, there is no evidence that there is a lasting effect, if there is even a temporary shock. The increased use of the death penalty doesn’t fly as an explanation, because the quantity of death sentences is too small even if there is a large (6-7 fewer murders per death sentence) deterrence effect.
The increase in police over the period is believed to have reduced crime. Levitt suggests somewhere between 5-6 percent decrease in the crime rate (which amounts to about 10-20% of the total decrease in crime, since it dropped between 20-50 percent). About 1/3 of the decline in crime can be explained by the rising prison population. He discusses the joint mechanisms of the incapacitation effect and the deterrence effect—both appear to be important. The declining crack epidemic accounts for about 15% of the drop in crime. Crack cocaine was introduced around 1985 and corresponded to gang activity. Finally, Levitt cites the legalization of abortion as a major explanation. This is by far his most controversial argument, and it has been subject to harsh critique. Abortions presumably occur in family situations that would be unhealthy, and thus remove a portion of the population who are more likely to commit crime. Relatedly, a reduction in unwanted births can lead to less crime in and of itself—people perhaps become more desperate in attempts to ‘make ends meet.’ Legalized in 1973, abortion timeline is consistent with the coming-of-age of a generation. As a larger proportion of the would-be criminal population is aborted, crime is reduced. Together, these four reasons account for 90 to over 100 percent of the reported decrease in crime.
Levitt suggests that the real puzzle is why crime did not falter sooner. Indeed, economists and criminologists had been predicting the 1990’s would be a “bloodbath,” but this failed to materialize.
But that only goes to show the poverty of social scientific point predictions.