• Selected
• Entire List

Abstract:
In this Article, Professors Hausman and Sidak propose a consumer-welfare model for the mandatory unbundling of telecommunications networks. Their approach, responsive to both the Supreme Court's 1999 decision in AT&T Corp. v. Iowa Utilities Board and the Federal Communications Commission's Second Further Notice of Proposed Rulemaking later the same year, reconciles the necessary and impair standards of Statute 251(d)(2) of the Telecommunications Act with the economic analysis of antitrust law. The essential facilities doctrine in antitrust law provides four necessary, but not sufficient, conditions for finding impairment. The authors add a fifth condition, responsive to the explicit text of Statute 251(d)(2), which addresses whether an incumbent local exchange carrier could exercise market power over end-users by restricting competitors' access to a requested telecommunications network element in a particular geographical market. The authors also recommend that necessary be interpreted to mean that competition in end-user services would be impossible unless the requested element were unbundled at a cost-based regulated price. This heightened standard, they argue, will protect the economic incentives to create the intellectual property embodied in elements that are proprietary in nature. The authors' proposed interpretation of Statute 251(d)(2) focuses on the effectiveness of competition in the end-user services market, rather than on the ability of a particular competitor to earn profits. Thus, the test adopts consumer welfare, rather than competitor welfare, as its touchstone.

Abstract:
Whenever feasible, market power determinations should rest on competitive benchmark prices rather than the typical market concentration approach. Government regulators in many countries have issued guidelines on the evaluation of market power in the merger context and other areas that define relevant markets and calculate market shares - along with a summary measure of market concentration, usually the Hirschman-Herfindahl index (HHI). However, competition authorities recognize that high concentration measures are generally not a sufficient condition to infer market power. Use of other structural factors in a market often does not lead to any clearer conclusion. We show that prices that consumers pay for the product in question often offer a superior quantitative measurement that leads to a clearer conclusion than the HHI approach. Further, because prices form the basis for the evaluation of consumer welfare (consumers surplus), they also provide important information for competition authorities, whose goal is typically the protection of consumer welfare. To demonstrate our argument, we examine a decision by the Irish telecommunications regulator, ComReg, which used the EU competition guidelines and the HHI approach to determine that Ireland's two largest mobile providers, Vodafone and O2, had joint dominance and were exercising significant market power. We demonstrate how our benchmark prices approach is superior to the HHI approach. We thank the ABA for granting permission to post the article.

Abstract:
In this article, we examine the rationales offered by telecommunications regulators worldwide for pursuing mandatory unbundling. We begin by defining mandatory unbundling, with brief descriptions of different wholesale forms and different retail products. Next, we examine four major rationales for regulatory intervention of this kind: (1) competition in the form of lower prices and greater innovation in retail markets is desirable, (2) competition in retail markets cannot be achieved with mandatory unbundling, (3) mandatory unbundling enables future facilities-based investment (stepping-stone or ladder of investment hypothesis), and (4) competition in wholesale access markets is desirable. We proceed by testing empirically the major rationales in the United States, the United Kingdom, New Zealand, Canada, and Germany. For each case study, we review the mandatory unbundling experience with respect to retail pricing, investment, entry barriers, and wholesale competition. We review the lessons learned from the unbundling experience. We also identify which rationales were incorrect in theory and which rationales were correct in theory yet were not satisfied in practice. For the second category of rationales, we attempt to provide alternative explanations for the failure of mandatory unbundling to achieve its goals.

antitrust, regulation, telecommunications, competition, unbundling, access pricing, investment, infrastructure

Abstract:
To date, most residential customers to the Internet have used dial-up modems with a top speed of about 56.6 kbps [kilobits per second]. In the past two years broadband access has become available via cable modems offered by the local unregulated cable provider and via digital subscriber lines (DSL) offered by the local regulated telephone company (the incumbent local exchange carrier [ILEC]) and competitors who resell DSL using the ILEC facilities. Cable modems and DSL offer access speeds about 10-30 times higher than dial-up access and are termed broadband Internet access. Although Federal Communication Commission (FCC) regulation required ILEC's to sell the use of their facilities to competitors at below-cost prices, no regulation of cable companies has occurred. This outcome is curious given that cable companies have a significantly greater incentive to distort competition as a result of their unregulated monopoly profits from their cable operations. This asymmetric regulation by the FCC has led to the open-access debate. The open-access debate involves the question about whether the cable providers should be required to provide access to competing broadband Internet service providers (ISP's) or whether cable providers can use exclusive contract with their affiliated ISP's.

Here, we consider the economic incentives and actions of the providers of broadband access with respect to limiting the usage of broadband access, including the potential competitive effects for cable television, a sector of the economy where, to date, system operators have been able to exercise significant market power. We answer the question of whether the price of narrowband Internet access constrains the price of broadband Internet access. We reject the hypothesis that the price of narrowband access does not affect the price of broadband access (transport) and ISP service is not rejected. Our finding is that lower narrowband access prices do not constrain the prices charged for broadband access.

Abstract:
The linkLine price squeeze case from the Ninth Circuit is the most important antitrust case that the Supreme Court could take during the Fall 2007 Term. Amici are professors and scholars in law and economics who have taught, or have conducted research on, antitrust law and the economics of industrial organization. They include William J. Baumol, Robert H. Bork, Robert W. Crandall, George Daly, Harold Demsetz, Jeffrey A. Eisenach, Kenneth G. Elzinga, Gerald Faulhaber, Franklin M. Fisher, Charles J. Goetz, Robert Hahn, Jerry A. Hausman, Thomas M. Jorde, Robert E. Litan, Paul W. MacAvoy, J. Gregory Sidak, Pablo T. Spiller, and Daniel F. Spulber.

We agree with the petitioners that the Ninth Circuit has generated an inescapable conflict among circuits, and that the Ninth Circuit's opinion below is incompatible with this Court's reasoning in Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398 (2004), Weyerhaeuser Co. v. Ross-Simmons Hardwood Lumber Co., 127 S. Ct. 1069 (2007), and Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209 (1993). We agree with Judge Gould's dissent in linkLine that Trinko "takes the issues of wholesale pricing out of the case," such that the plaintiffs' only possible remaining theory of harm would be predatory pricing at the retail level - which the plaintiffs did not allege. linkLine Commc'ns Inc. v. Pac. Bell Tel. Co. d/b/a/ AT&T Cal., Inc., No. 05-56023, 2007 U.S. App. LEXIS 21719, at *28-29 (9th Cir. Sept. 11, 2007) (Gould, J., dissenting). We also agree with Judge Ginsburg's opinion for the D.C. Circuit in Covad Communications Co. v. Bell Atlantic Corp., 398 F.3d 666 (D.C. Cir. 2005), which in turn embraces the conclusion of the Areeda-Hovenkamp treatise that "'it makes no sense to prohibit a predatory price squeeze in circumstances where the integrated monopolist is free to refuse to deal.'" Id. at 673-74 (quoting 3A Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 767c3, at 129-30 (2d ed. 2002)). The existence of a rule like linkLine has a pervasive impact on business behavior that, at the margin, affects competition and consumers. This deleterious effect extends beyond the telecommunications industry to affect all firms that do business in the Ninth Circuit. These reasons justify granting certiorari in linkLine and reversing the Ninth Circuit's decision.

In our minds, an even larger reason than those described above makes it imperative that the Court take this case. The Ninth Circuit's decision in linkLine implicates the normative foundation of modern Sherman Act jurisprudence: that antitrust law exists to advance consumer welfare. We have three points to make.

First, any rule of price-squeeze liability that threatens liability based on the claim that the difference between a firm's upstream and downstream prices leaves downstream rivals insufficient margin substitutes a rule of competitor welfare for consumer welfare.

Second, properly understood, a price squeeze is a regulatory issue, which makes sense only as a rule of price regulation in an industry already subject to duties to deal and to control by institutionally competent regulators. Attempting to implement regulatory policy through section 2 of the Sherman Act is ill-advised, both because it makes no sense for courts to re-regulate deregulated or lightly regulated industries, and because courts lack the institutional competence to implement regulation.

Third, the Ninth Circuit's rule is of pressing concern precisely because it will deter efficiency-enhancing conduct and competitive pricing. Vertical integration and partial integration are ubiquitous, and firms need to be able to make decisions about such integration without the threat of liability. Vertically integrated firms likewise need to be free to cut retail prices (as long as the prices are not predatory) without concern for rivals - the point of Brooke Group. Moreover, the Ninth Circuit's standard is so vague and open-ended that it creates uncertainty and invites litigation; it also permits imposition of liability based on apparently subjective evaluation of disputed and hard-to-prove facts, which will lead to a substantial risk of false positives.

Abstract:
As part of the Modification of Final Judgment (MFJ) that implemented the divestiture of the Bell operating companies (BOCs) from AT&T on January 1, 1984, the BOCs were forbidden to carry telephone calls from one local access and transport area LATA) to another. Although the Telecommunications Act of 1996 superseded the MFJ, it retained the BOCs' interLATA prohibition and established, in section 271, a process - involving each state public utilities commission, the Federal Communications Commission (FCC), and the Department of Justice (DOJ), acting on a state-by-state basis - by which the BOCs could earn regulatory approval to enter the interLATA market within the regions in which they provide local exchange service. As of September 1, 2002, the BOCs had received section 271 authorizations to provide in-region interLATA service in fifteen states.

In this Article, we review the origin of section 271. We explain that the FCC and DOJ did not expect that BOC entry would lower prices for interLATA service. Next, we report an empirical analysis designed to estimate the effect that BOC entry has had in New York and Texas, two states where section 271 authorizations has been given. We have three major findings. First, we find that the average consumer received a savings of 8 to 11 percent on the monthly interLATA bill in the states where BOC entry occurred as compared to "control" states where BOC entry had not occurred. Second, we find that competitive local exchange carriers (CLECs) gained a substantial increase in cumulative share of the local exchange market in states where BOC entry occurred as compared to the control states. Third, we find that the average consumer experienced no significant change in her local bill in states where BOC entry into interLATA service occurred as compared to the control states. These empirical results suggest that BOC entry in New York and Texas has led to consumer benefits in terms of lower interLATA bills and greater effective choice for local exchange services in those states.

We explain how these empirical results are consistent with the economic theory of "double marginalization." Because this economic analysis is not part of the approach that the FCC and the DOJ take with respect to implementing section 271, it is not surprising that these two agencies did not expect price to fall after BOC entry into the interLATA market.

Reprinted from Antitrust Law Journal, vol. 70, no. 2, pp. 463-484 (2002), a publication of the American Bar Association Section of Antitrust Law.

Abstract:
In this article, we examine the open access debate in the context of cable services and broadband Internet services from an antitrust framework. Our analysis is prompted by the recent AT&T-MediaOne and AOL-Time Warner mergers, which raise issues concerning the impact of integrated cable content and Internet access to residential telecommunications. Economic analysis, demographic surveys and federal antitrust guidelines each indicate that the broadband Internet access market is distinct from the narrowband Internet access market. Emerging or competing technologies, such as satellite Internet services or digital subscriber lines, cannot discipline the broadband Internet access market over the relevant time horizons. Vertical integration increases the incentives and power of cable providers to discriminate against unaffiliated broadband content, thereby substantially decreasing consumer welfare. We conclude that the recent mergers of cable content and Internet access is the most current manifestation of the classic strategy of cable providers to control alternate channels of content distribution.

Abstract:
The legal framework under which patent damages are calculated changed substantially after the Federal Circuit decided Grain Processing Corp. v. American Maize-Products Co. in 1999. Perhaps the most important question in the typical lost profits analysis is determining the fraction of the infringing sales that constitute lost sales to the patent holder. The answer to this question usually depends on the set of non-infringing substitute products to which the customers of the infringing product could have turned in the but-for world where the infringing product was not available to them. Before Grain Processing, the case law as a legal matter generally restricted the set of non-infringing substitute products to include only products that were actually sold in the marketplace. For example, an infringer could claim that it would have continued to sell a non-infringing product that it had actually been selling and that this product would have captured some of the infringing sales, which would tend to limit the patent holder's lost sales. However, the infringer could not claim that it would have developed and introduced some new non-infringing product in the but-for world and that this product would have captured some of the infringing sales. Grain Processing eased this restriction on the set of non-infringing substitutes available in the but-for world by allowing an infringer to claim that it would have offered a non-infringing product that, although not actually sold in the marketplace, was technically feasible at the time and could have been made commercially available relatively quickly. The Grain Processing decision then went further and concluded that, in the particular case at issue, the plaintiff was not entitled to lost profits because the infringer's non-infringing product would have been identical from the point of view of customers (though more costly to the infringer). Damages were therefore calculated on a reasonable royalty basis only. Although Grain Processing has generated much scholarly commentary, we are unaware of any article considering the factor that we see as the decision's most important economic ramification: the grant of a free option to the infringer. Although it is widely appreciated how Grain Processing has made it more difficult for patent holders to claim lost profits damages, it is less well understood how Grain Processing has affected the incentives of companies to risk litigation by using patented technology (without a license) rather than to avoid infringement by using an economically inferior non-infringing technology. Whether the patent is valid and infringed is not known until the litigation takes place. A patent only provides the patent holder with the right to sue for infringement. A court decides whether the patent is valid and infringed. We find that the grant of a free option is contrary to the basic framework of the patent system in the United States. If a firm chooses to risk litigation and use the patented technology, it retains the option to switch to the non-infringing technology if the patent is later found to be valid and infringed. Of course, it will be liable for damages for the period of infringement. If, on the other hand, the firm chooses to use the non-infringing technology, it will not have the opportunity to learn whether the patent is valid and infringed. Thus, by choosing the patented technology, the firm keeps its options open, although at the risk of having to pay damages once the uncertainty regarding validity and infringement is resolved. Grain Processing has the effect of substantially decreasing this risk by decreasing the size of the damages award. If the patent is found to be valid and infringed, the firm can argue under Grain Processing that it would have switched to the non-infringing technology in the but-for world, thereby effectively making the switch retroactively. Grain Processing thereby makes the option essentially free. By providing potential infringers with increased option value if they use the patented technology, Grain Processing reduces the deterrent effect of litigation and therefore encourages infringement. Consequently, it reduces the returns to research and development, and so also the incentives to innovate.

Patent, damages, Grain Processing, infringement, but-for, non-infringing, option, free option, real option, innovation, royalty, lost profits

Abstract:
The benefits of competition among the long-distance interexchange carriers (IXCs) are not realized equally by all their customers. Despite the declines in rates under the discount plans, we document that basic message toll service (MTS) rates have been rising for several years. We show that poorer and less educated customers pay more than better educated and more affluent customers. We suspect that the reason for this correlation is that they are more apt to pay the MTS rates or other high rates, and we present some preliminary evidence that this tendency explains the correlation that we find. We also present evidence that the payment differences exist even after controlling for usage. These findings are significant because it seems likely to us that these two patterns (rising MTS rates and higher payments by the poor and the less educated) will each be ameliorated by the entry of the regional Bell operating companies (RBOCs) into long-distance markets - a state-by-state regulatory process that was nearly complete as of the beginning of 2004.

competition, regulation

Abstract:
While 2SLS is the most widely used estimator for simultaneous equation models, OLS may do better in finite samples. Here we demonstrate analytically that for the widely used simultaneous equation model with one jointly endogenous variable and valid instruments, 2SLS has smaller MSE error, up to second order, than OLS unless the R2 , or the F statistic of the reduced form equation is extremely low. We then consider the relative estimators when the instruments are invalid, i.e. the instruments are correlated with the stochastic disturbance. Here, both 2SLS and OLS are biased in finite samples and inconsistent. We investigate conditions under which the approximate finite sample bias or the MSE of 2SLS is smaller than the corresponding statistics for the OLS estimator. We again find that 2SLS does better than OLS under a wide range of conditions. We then present a method of sensitivity analysis, which calculates the maximal asymptotic bias of 2SLS under small violations of the exclusion restrictions. For a given correlation between invalid instruments and the error term, we derive the maximal asymptotic bias. We apply our results to IV estimation of the returns to education. We derive the bias in the estimated standard errors of 2SLS for the first time. This derivation also has implications for the test of
over-identifying restrictions.

Instrumental Variables, 2SLS, Weak Instruments, Returns to Education

Abstract:
This paper analyzes the second order bias of instrumental variables estimators for a dynamic panel model with fixed effects. Three different methods of second order bias correction are considered. Simulation experiments show that these methods perform well if the model does not have a root near unity but break down near the unit circle. To remedy the problem near the unit root a weak instrument approximation is used. We show that an estimator based on long differencing the model is approximately achieving the minimal bias in a certain class of instrumental variables (IV) estimators. Simulation experiments document the performance of the proposed procedure in finite samples.

dynamic panel, bias correction, second order, unit root, weak instrument

Abstract:
Around the world, regulators since 1996 have mandated that incumbent local exchange carriers (ILECs) offer competitors access to their network at regulated prices that reflect forward-looking cost. Regulated prices for unbundled network elements are based on total element long-run incremental cost (TELRIC), which in turn is calculated using engineering models that estimate the costs of a hypothetical carrier employing the most efficient telecommunications technology currently available and the lowest cost network configuration, given the existing location of the ILEC's actual wire centers. These cost models require detailed estimates of the equipment and installation prices of the numerous components that are used in a telecommunications network. When there is uncertainty about how these prices will change over the period for which costs and prices are required, the resulting cost estimates used for setting the regulated prices of unbundled network elements can be very inaccurate.

Similarly, when regulators in other jurisdictions are considering such rates as benchmarks, it is necessary to make adjustments to account for such large differences in critical input prices, so that the benchmark rates will be representative of the costs that actually will be incurred by efficient carriers offering unbundled elements in those jurisdictions. The precipitous rise in the price of copper since 2003 exemplifies this need to reevaluate the inputs used by regulators in their cost model, as well as the inferences drawn from those models. These increases differ from the type of constant annual expected input price growth (or decline) situation that some cost models used outside the United States have accommodated with tilted annuity methods. Rather than a gradual anticipated price increase, copper prices escalated rapidly and are likely to remain well above the levels that regulators used to set existing loop rates.

Accounting for such evidence would change the forward-looking costs of a hypothetically efficient ILEC network that one of the most prominent U.S. state regulatory commissions - the California Public Utilities Commission (CPUC) - established in 2006. Meanwhile, in 2007, the Commerce Commission in New Zealand has similarly employed a benchmarking methodology for the pricing of unbundled loops that fails to account for the increased price of copper. A global trend may be emerging among telecommunications regulators to ignore the input requirements of their own forward-looking cost models. Such a trend would be consistent with a version of regulatory opportunism in which regulators are forward-looking only when doing so produces lower regulated prices over time.

Abstract:
Fifteen scholars on auctions and telecommunications regulation urge the FCC to cancel bids made in, or permit winning bidders to opt out of, the reauction of the NextWave licenses in Auction 35.

For auctions to function efficiently, buyers and sellers must follow basic rules, including the rule that a seller deliver in a timely manner what the winning bidder has purchased. This rule has not been applied in Auction 35. The FCC auctioned something that it did not have - immediate access to the spectrum for the winning bidders. Thus, if the FCC forces the winning bidders to pay, they will sue the agency for forcing them to pay for something that they did not receive. Alternatively, their shareholders will sue the companies. Meanwhile, wireless carriers have invested in less efficient technologies to meet capacity needs.

The FCC has said that its current policy toward Auction 35 seeks to "protect the integrity" of the spectrum auction process. The opposite is already occurring. The FCC increases uncertainty in the wireless market if it holds carriers accountable for winning bids for licenses that the agency cannot deliver. Bidders will discount their future bids accordingly, and auction revenues will fall. That outcome does not benefit consumers, taxpayers, workers, or shareholders.

Abstract:
Economists have long recognized that certainty of contract is essential to a healthy economy. Long-term forward contracts, in particular, help reduce financial risk. Those contracts can only accomplish that goal, however, if parties know the contracts will be enforced.

From an economic and policy standpoint, long-term energy contracts should be abrogated only in truly exceptional circumstances. The mere fact that a price seems too high in retrospect does not justify abrogating contracts voluntarily agreed to by sophisticated buyers and sellers. Nor do generalized claims of - market dysfunction - at the time the contract was formed.

Abstract:
We develop a new specification test for IV estimators adopting a particular second order approximation of Bekker. The new specification test compares the difference of the forward (conventional) 2SLS estimator of the coefficient of the right-hand side endogenous variable with the reverse 2SLS estimator of the same unknown parameter when the normalization is changed. Under the null hypothesis that conventional first order asymptotics provide a reliable guide to inference, the two estimates should be very similar. Our test sees whether the resulting difference in the two estimates satisfies the results of second order asymptotic theory. Essentially the same idea is applied to develop another new specification test using second-order unbiased estimators of the type first proposed by Nagar. If the forward and reverse Nagar-type estimators are not significantly different we recommend estimation by LIML, which we demonstrate is the optimal linear combination of the Nagar-type estimators (to second order). We also demonstrate the high degree of similarity for k-class estimators between the approach of Bekker and the Edgeworth expansion approach of Rothenberg. An empirical example and Monte Carlo evidence demonstrate the operation of the new specification test.

Abstract:
We develop a new specification test for the IV estimators adopting a particular second order approximation of Bekker (1994). The new specification test compares the difference of the forward (conventional) 2SLS estimator of the coefficient of the right hand side endogenous variable with the reserve 2SLS estimator of the same unknown parameter when the normalization is changed. Under the null hypothesis that conventional first order asymptotics provides a reliable guide, the two estimates should be very similar. Our test sees whether the resulting difference in the two estimates satisfies the results of second order asymptotic theory. Essentially the same idea is applied to develop another new specification test using second-order unbiased estimators of the type first proposed by Nagar (1959). If the forward and reverse Nagar-type estimators are not significantly different we recommend estimation by LIML, which we demonstrate is the optimal linear combination of the Nagar-type estimators (to second order). We also demonstrate the high degree of similarity for k-class estimators between the approach of Bekker (1994) and the Edgeworth expansion approach of Rothenberg (1983). Empirical example and Monte Carlo evidence are provided.

Abstract:
The linkLine price squeeze case pending in the Supreme Court for the Fall 2008 Term is one of the most significant antitrust cases on monopolization law that the Court has taken in years. Amici are professors and scholars in law and economics who have taught, or have conducted research on, antitrust law and the economics of industrial organization. They are William J. Baumol, Robert H. Bork, Robert W. Crandall, George Daly, Harold Demsetz, Jeffrey A. Eisenach, Kenneth G. Elzinga, Richard A. Epstein, Gerald Faulhaber, Franklin M. Fisher, Charles J. Goetz, Robert Hahn, Jerry A. Hausman, Keith N. Hylton, Thomas M. Jorde, Robert E. Litan, Paul W. MacAvoy, Sam Peltzman, J. Gregory Sidak, Pablo T. Spiller, and Daniel F. Spulber.

We agree with the petitioners that the Ninth Circuit has generated an inescapable conflict among circuits, and that its opinion is incompatible with the Supreme Court's decisions in Trinko, Weyerhaeuser, and Brooke Group. We agree with Judge Gould's dissent from the Ninth Circuit's decision in linkLine that Trinko "takes the issues of wholesale pricing out of the case," such that the plaintiffs' only possible remaining theory of harm would be predatory pricing at the retail level - which the plaintiffs did not allege. We also agree with Judge Ginsburg's opinion for the D.C. Circuit in Covad Communications Co. v. Bell Atlantic Corp., which in turn embraces the conclusion of the Areeda-Hovenkamp treatise that "it makes no sense to prohibit a predatory price squeeze in circumstances where the integrated monopolist is free to refuse to deal." The existence of a rule like linkLine has a pervasive impact on business behavior that, at the margin, affects competition and consumers. This deleterious effect extends beyond the telecommunications industry to affect all firms that do business in the Ninth Circuit. These reasons justify reversing the Ninth Circuit's decision.

In our minds, an even larger reason than those described above makes it imperative that the Court reverse this decision. The Ninth Circuit's decision in linkLine implicates the normative foundation of modern Sherman Act jurisprudence: that antitrust law exists to advance consumer welfare. We have three points to make.

First, any rule of price-squeeze liability that threatens liability based on the claim that the difference between a firm's upstream and downstream prices leaves downstream rivals insufficient margin substitutes a rule of competitor welfare for consumer welfare.

Second, properly understood, a price squeeze is a regulatory issue, which makes sense only as a rule of price regulation in an industry already subject to duties to deal and to control by institutionally competent regulators. Attempting to implement regulatory policy through section 2 of the Sherman Act is ill-advised, both because it makes no sense for courts to re-regulate deregulated or lightly regulated industries, and because courts lack the institutional competence to implement regulation.

Third, the Ninth Circuit's rule is of pressing concern precisely because it will deter efficiency-enhancing conduct and competitive pricing. Vertical integration and partial integration are ubiquitous, and firms need to be able to make decisions about such integration without the threat of liability. Vertically integrated firms likewise need to be free to cut retail prices (as long as the prices are not predatory) without concern for rivals - the point of Brooke Group. Moreover, the Ninth Circuit's standard is so vague and open-ended that it creates uncertainty and invites litigation; it also permits imposition of liability based on apparently subjective evaluation of disputed and hard-to-prove facts, which will lead to a substantial risk of false positives.

Abstract:
We examine the consumer-welfare implications of Google's project to scan a large proportion of the world's books into digital form and to make these works accessible to consumers through Google Book Search (GBS). In response to a class action alleging copyright infringement, Google has agreed to a settlement with the plaintiffs, which include the Authors Guild and the Association of American Publishers. A federal district court must approve the settlement for it to take effect. Various individuals and organizations have advocated modification or rejection of the settlement, based in part on concerns regarding Google's claimed ability to exercise market power. The Antitrust Division has confirmed that it is investigating the settlement. We address concerns of Professor Randal Picker and others, especially concerns over the increased access to “orphan books,” which are books that retain their copyright but for which the copyright holders are unknown or cannot be found. The increased accessibility of orphan books under GBS involves the creation of a new product, which entails large gains in consumer welfare. We consider it unlikely that Google could exercise market power over orphan books. We consider it remote that the static efficiency losses claimed by critics of the settlement could outweigh the consumer welfare gains from the creation of a valuable new service for expanding access to orphan books. We therefore conclude that neither antitrust intervention nor price regulation of access to orphan books under GBS would be justified on economic grounds.

K20, K21, L40, L41, L50, O34

Abstract:
We examine the consumer-welfare implications of Google’s project to scan a large proportion of the world’s books into digital form and to make these works accessible to consumers through Google Book Search (GBS). In response to a class action alleging copyright infringement, Google has agreed to a settlement with the plaintiffs, which include the Authors Guild and the Association of American Publishers. A federal district court must approve the settlement for it to take effect. Various individuals and organizations have advocated modification or rejection of the settlement, based in part on concerns regarding Google’s claimed ability to exercise market power. The Antitrust Division has confirmed that it is investigating the settlement. We address concerns of Professor Randal Picker and others, especially concerns over the increased access to 'orphan books,' which are books that retain their copyright but for which the copyright holders are unknown or cannot be found. The increased accessibility of orphan books under GBS involves the creation of a new product, which entails large gains in consumer welfare. We consider it unlikely that Google could exercise market power over orphan books. We consider it remote that the static efficiency losses claimed by critics of the settlement could outweigh the consumer welfare gains from the creation of a valuable new service for expanding access to orphan books. We therefore conclude that neither antitrust intervention nor price regulation of access to orphan books under GBS would be justified on economic grounds.

Abstract:
Panel data based on various longitudinal surveys have become ubiquitous in economics in recent years. Estimation using the analysis of covariance approach allows for control of various "individual effects" by estimation of the relevant relationships from the "within" dimension of the data. Quite often, however, the "within" results are unsatisfactory, "too low" and insignificant. Errors of measurement in the independent variables whose relative importance gets magnified in the within dimension are often blamed for this outcome. However, the standard errors-in-variables model has not been applied widely, partly because in the usual micro data context it requires extraneous information to identify the parameters of interest. In the panel data context a variety of errors-in-variables models may be identifiable and estimable without the use of external instruments. We develop this idea and illustrate its application in a relatively simple but not uninteresting case: the estimation of "labor demand" relationships, also known as the "short run increasing returns to scale" puzzle.

Abstract:
Consumers often benefit from increased competition in differentiated product settings. In this paper we consider consumer benefits from increased competition in a differentiated product setting: the spread of non-traditional retail outlets. In this paper we estimate consumer benefits from supercenter entry and expansion into markets for food. We estimate a discrete choice model for household shopping choice of supercenters and traditional outlets for food. We have panel data for households so we can follow their shopping patterns over time and allow for a fixed effect in their shopping behavior. We find the benefits to be substantial, both in terms of food expenditure and in terms of overall consumer expenditure. Low income households benefit the most.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
Hausman (2003) discusses four sources of bias in the present calculation of the CPI. A 'pure price' index based approach of surveying prices as done by the BLS cannot succeed in solving the problems of bias. We discuss economic and econometric approaches to measuring the first order bias effects from outlet substitution bias. We demonstrate the use of scanner data that permits implementation of techniques that allow the problem to be solved. In contrast, the current BLS procedure does not treat correctly outlet substitution bias and acts as if Wal-Mart does not exist. Yet, Wal-Mart offers identical food items at an average price about 15%-25% lower than traditional supermarkets. The BLS 'links out' Wal-Mart's lower prices. We find that a more appropriate approach to the analysis is to let the choice to shop at Wal-Mart be considered as a 'new good' to consumers when Wal-Mart enters a geographic market. This approach leads to a continuously updated expenditure weighted average price calculation. We find a significant difference between our approach and the BLS approach. Our estimates are that the BLS CPI-U food at home inflation is too high by about 0.32 to 0.42 percentage points, which leads to an upward bias in the estimated inflation rate of about 15% per year.

Abstract:
The Consumer Price Index (CPI) attempts to answer the question of how much more (or less) income does a consumer require to be as well off in period 1 as in period 0 given changes in prices, changes in the quality of goods, and the introduction of new goods (or the disappearance of existing goods). In this paper I explain the theory of cost-of-living indices and demonstrate how new goods should be included using the classical theory of Hicks and Rothbarth. The correct price to use for the good in the pre-intro- duction period is a `virtual' price which sets demand to zero. Estimation of this virtual price requires estimation of a demand function which in turn provides the expenditure function which allows exact calucation of the cost of living index. The data requirements and need to specify and estimate a demand function for a new brand among many existing brands requires extensive data and some new econometric methods which may have proven obstacles to the inclusion of new goods in the CPI up to this point. As an example I use the introduction of a new cereal brand by General Mills in 1989-Apple Cinnamon Cheerios. I find the virtual price is about 2 times the actual price of Apple Cinnamon Cheerios and that increase in consumer surplus is substantial. Based on some simplifying approximations, I find that CPI may be overstated for cereal by about 25% because of its neglect of the effect of new brands. When I take imperfect competition into account I find that the increase in consumer welfare is only 85% as high with perfect competition so CPI for cereal would still be 20% too high

Abstract:
The signatories to this document are economists who have studied telecommunications, auctions, and competition policy. While we may disagree about the stimulus package, we believe that it is important to implement mechanisms that make stimulus spending as efficient as possible. To that end, we have come together to encourage the National Telecommunications Information Agency (NTIA) and Rural Utilities Service (RUS) to adopt auction mechanisms to allocate broadband stimulus grants.

The broadband stimulus NOI asks which mechanisms NTIA and RUS should use to distribute grants and how those mechanisms address shortcomings in traditional grant and loan programs. In this note we explain why procurement auctions are more efficient and more consistent with the stimulus goals of allocating funds quickly than a traditional grant review process. We recommend that NTIA/RUS use procurement auctions to distribute at least part of the stimulus funds.

The American Recovery and Reinvestment Act (ARRA) requires NTIA/RUS to distribute $7.2 billion in broadband subsidies. The broadband component of the Act has dual, and not entirely consistent, objectives of providing immediate economic stimulus and improving broadband service. NTIA/RUS faces a formidable challenge in determining how to spend the money quickly and efficiently in ways that meet these goals. The traditional grant application process is long, complicated, and involves subjective and arbitrary decisions regarding which projects to fund. In other words, requesting and reviewing grant applications is not an effective way to implement the plan.

Procurement auctions, in contrast, provide a mechanism that can allocate grant money quickly, efficiently, and according to well-defined rules. As a result, procurement auctions offer NTIA/RUS the most promising method of maximizing broadband improvement while also creating some level of “temporary, timely, and targeted” stimulus. We therefore strongly recommend that NTIA/RUS adopt procurement auctions as its preferred method of distributing grants.

This memo has three parts. First, it explains why the traditional grant application process is unsuitable for this task and why procurement auctions are better suited. Second, it sketches out a procurement auction plan. This plan is intended to be a starting point from which auction design experts would proceed to build and implement a fully functional auction. Finally, we explain that even if policymakers are skeptical of procurement auctions, one could be implemented quickly as part of an initial tranche of stimulus funding in order to test its efficacy relative to traditional approaches. This approach would allow NTIA/RUS to quickly expand upon or modify the procurement auction program in subsequent funding rounds.

Abstract:
Over 75% of Federal tax revenue is raised through the income tax
and FICA taxes. The potential effects on labor supply and
economic welfare are important because of the large and
increasing reliance on direct taxation. Over the past few years
significant legislative changes have occurred with respect to
taxation of labor: the 25% tax cuts, indexation, the tax credit
for working spouses, and likely increases in FICA taxation. I
review recent econometric work which measures the effect of taxes
on labor supply and which analyzes the likely effects of tax law
changes on labor supply and economic welfare.

Sections 1 and 2 develop the theory and econometric techniques
for models of labor supply with taxes. Section 3 discusses the
various tax systems in the U.S. In Section 4, I present
empirical estimates for husbands' and wives' labor supply
functions. The economic cost of the tax system is also
estimated. In Section 5 the individual questionnaire data for
high income individuals is reviewed. Lastly, in Section 6
evidence from the negative income tax experiments and for social
security beneficiaries is considered. These latter groups face
extremely high marginal tax rates so that evidence beyond that
contained in other surveys of labor supply is provided.

Abstract:
Cellular telephone is an example of a new product that has significantly affected how Americans live. Since their introduction in 1983, cellular telephone adoption has grown at 25-35% per year such that at year end 1996 about 42 million cellular telephones are in use in the U.S. However, cellular telephone has not been included in the construction of the CPI, and the CPI will not include cellular telephone until 1998 or 1999. This neglect of new goods leads to an upward bias in the CPI. The analysis of the paper demonstrates that the gains in consumer welfare from a new product such as cellular telephone can be substantial. The paper also gives an approximation result which the BLS could use to calculate gains in consumer welfare from new products for use in the CPI. The BLS telecommunications CPI estimates that since 1988, telecommunications prices have increased by 8.5% or an increase of 1.02% per year. This estimate ignores cellular service. A corrected telecommunication services COLI that includes cellular service decreased from 1.0 in 1988 to 0.903 in 1996 for a decrease of 1.28% per year. Thus, the bias in the BLS telecommunications services CPI equals approximately 2.3 percentage points per year. The neglect of new products in the CPI can lead to significant biases.

Abstract:
No abstract is available for this paper.

Abstract:
Telecommunications regulation in the U.S. is replete with a system of subsidies and taxes. Because of budgetary spending limits, Congress is unable to increase general taxes to pay for social programs and thus funds these programs from taxes on specific sectors of the economy. In this paper I consider the Congressional legislation which established a program so that all public schools and libraries in the U.S. will receive subsidized service to the Internet. The cost of the program is estimated to be $2.25 billion per year. Congress passed legislation that directed all users of interstate telephone service to pay for the program. Using analytical methods from public finance, I calculate the efficiency cost to the economy of the higher taxation of interstate telephone services to fund the Internet access discounts. I estimate the cost to the economy of raising the $2.25 billion per year to be at least $2.36 billion (in addition to the $2.25 billion of tax revenue), or the efficiency loss to the economy for every $1 raised to pay for the Internet access discounts is an additional $1.05 to $1.25 beyond the money raised for the Internet discounts. This cost to the economy is extremely high compared to other taxes used by the Federal government to raise revenues. I discuss an alternative method by which the FCC could have raised the revenue for the Internet discounts which would have a near zero cost to the economy.

Abstract:
This paper measures for the first time the economic efficiency effects of the taxation of wireless services, which are taxed by federal, state, and local governments at relatively high rates in the range of 14%-25%. The paper concludes such taxes are a much greater drain on the economy than their direct costs. The taxes identified in this paper cost the economy $2.56 billion more than the $4.79 billion they raise in tax revenues. These taxes are raised from wireless consumers and thereby suppress demand for service, imposing an efficiency loss on the economy of $0.53 for every $1 currently raised in taxes. Prospective taxes will impose an efficiency loss of $0.72-$1.14 per additional dollar of tax revenue raised.

Abstract:
This paper extends earlier work on the RID to patents relationship (Pakes-Griliches 1980, and Hausman, Hall, and Griliches,1984) to a larger but shorter panel of firms. The focus of the paper is on solving a number of econometric problems associated with the discreteness of the dependent variable and the shortness of the panel in the time dimension. We compare weighted nonlinear least squares as wellas Poisson-type models as solutions to the former problem. In attempting to estimate a lag structure on R&D in the absence of a sufficient history of the variable, we take two approaches: first, we use the conditional version of the negative binomial model, and second, we estimate the R&D variable itself as a low order stochastic process and use this information to control for unobserved R&D. R&D itself turns out to befairly well approximated by a random walk. Neither approach yields strong evidence of a long lag. The available sample, though numerically large, turns out not to be particularily informative on this question. It does reconfirm, however, a significant effect of R&D on patenting (with most of it occuring in the first year) and the presence of rather wide and semi-permanent differences among firms in their patenting policies.

Abstract:
In this paper, we discuss the econometric models put forward by both sides in the FTC v. Staples case and explore why those models canceled each other out, leaving the documents to play a central role in the court's decision.

Antitrust, Econometrics, Merger Review

Abstract:
This paper analyzes the competitive effect of a new product introduction. We break the overall competitive effect into two parts: the effect on the prices of existing products due to increased competition, and the effect of having additional product variety. Using data from both before and after the introduction, we directly estimate the price effects and the additional variety effect. Then, using only the estimated post-introduction demand structure, along with an assumed model of competition, we estimate the price effects indirectly. By comparing the "indirect" and "direct" estimates, we assess the validity of alternative models of competition for the industry.

Abstract:
Patent litigation has become an increasingly important consideration in business strategy. Damage awards in patent litigation are supposed to compensate the patent owner for economic harm created by infringement and are therefore important for protecting returns to innovation. We analyze the effects that a recent court decision in the United States, called Grain Processing, has had on the incentives of potential infringers to infringe and innovators to innovate. We find that Grain Processing has decreased the expected value of damages awards in patent cases by conferring a free option on infringers. Grain Processing also concluded that the patent owner in the case did not suffer lost profits due to the infringement because the infringer would have adopted an (inferior) non-infringing technology had it not infringed. We demonstrate that this conclusion is inconsistent with standard economic models.

Abstract:
Four sources of bias in the Consumer Prices Index (CPI) have been identified. The most discussed is substitution bias, which creates a second order bias in the CPI. Three other changes besides prices changes create first order effects on a correctly measured cost of living index (COLI). (1) Introduction of new goods creates a first order effect of 'new good bias' (2) Quality changes in existing goods will lead to 'quality' bias, which has first order effects (3) Shifts in shopping patterns to lower priced stores can create first order 'outlet bias'. I explain in this paper that a 'pure price' based approach of surveying prices to estimate a COLI cannot succeed in solving the 3 problems of first order bias. Neither the BLS nor the recent report C. Schultze and C. Mackie, eds., At What Price (AWP, 2002), recognizes that to solve these problems, which have been long known, both quantity and price data are necessary. I discuss economic and econometric approaches to measuring the first order bias effects as well as the availability of scanner data that would permit implementation of the techniques. Lastly, I review recent research that demonstrates that these sources of bias are large in relation to measured inflation in the CPI.

Abstract:
No abstract is available for this paper.

Abstract:
Current methods of estimating the random coefficients logit model employ simulations of the distribution of the taste parameters through pseudo-random sequences. These methods suffer from difficulties in estimating correlations between parameters and computational limitations such as the curse of dimensionality. This article provides a solution to these problems by approximating the integral expression of the expected choice probability using a multivariate extension of the Laplace approximation. Simulation results reveal that our method performs very well, in terms of both accuracy and computational time.

Abstract:
This paper extends earlier work on the R&D to patents relationship (Pakes-Griliches 1980, and Hausman, Hall, and Griliches, 1984) to a larger but shorter panel of firms. Using both non-linear least squares and Poisson type models to treat the problem of discreteness in the dependent variable the paper tries to discern the lag structure of this relationship in greater detail. Since the available time series are short, two different approaches are pursued in trying to solve the lag truncation problem: In the first the influence of the unseen past is assumed to decline geometrically; in the second,the unobserved past series are assumed to have followed a low order autoregression. Neither approach yields strong evidence of a long lag. The available sample, though numerically large,turns out not to be particularly informative on this question. It does reconfirm, however, a significant effect of R&D on patenting (with most of it occurring in the first year or two) and the presence of rather wide and semi-permanent differences among firms in their patenting policies.

Abstract:
Over the period 1960 - 1983 the proportion of federal tax revenue raised by taxation of labor supply has risen from 57-77 percent. In this paper, we specify and estimate a model of family labor supply which treats both federal and state taxation. Husbands and wives labor supply are treated jointly rather than in aseparate manner as in previous research. A method to calculate the virtual wage for nonworking spouses is used within a utility maximizing framework to treat correctly the joint family labor supply decision. Joint family efforts are found to be important. The efficiency cost (deadweight loss) of labor taxation is estimated to be 29.6% of tax revenue raised. The effect of the new 10% deduction to ease the marriage tax for working spouses leads to a prediction of 3.8% increase in wives labor supply and a .9% decrease in husbands labor supply.Overall taxes paid are predicted to decrease by 3.4%.

Abstract:
The goal of the paper is to set forth general guidelines that we believe would enhance the usefulness of future social experiments and to suggest ways of correcting for inherent limitations of them. Although the major motivation for an experiment is to overcome the inherent limitations of structural econometric models, in many instances the experimental designs have subverted this motivation. The primary advantages of randomized controlled experiments were often lost. The major complication for the analysis of the experiments was induced by an endogenous sample selection and treatment assignment procedure that selected the experimental participants and assigned them to controlversus treatment groups partly on the basis of the variable whose response the experiments were intended to measure. We propose that to overcome these difficulties, the goal of an experimental design should be as nearly as possible to allow analysis based on a simple analysis of variance model. Although complexities attendant to endogenous stratification can be avoided, there are inherent limitations of the experiments that cannot. Two major ones are self-determination of participation and self-selection out, through attrition.But these problems, we believe, can be corrected for with relative ease if endogenous stratification is eliminated. Finally, we propose that as a guiding principle, the experiments should have as a first priority the precise estimation of a single or a small number of treatment effects.

Abstract:
This paper evaluates the effects of the 1986 Tax Reform Act on household labor supply and savings. It describes the tax bill`s effects on incentives to work and to save, and uses recent econometric estimates of labor supply and savings elasticities to describe the reform`s impact on household behavior. Two factors lead us to conclude that the new law will have small aggregate effects. First, most households experience only small changes in their marginal tax rates. Forty-one percent of the taxpaying population will face marginal tax rates as high, or higher, under the new law as under the previous tax code. Only eleven percent of taxpayers receive marginal tax rate reductions of ten percentage points or more. Second, plausible estimates of both the labor supply and savings elasticities suggest that even for those households that receive rate reductions, behavioral changes will be small. Our analysis suggests that the tax reform will increase labor supply by about one percent, and slightly reduce private savings.

Abstract:
Not all people with health problems are disabled. Some individuals with severe physical or mental impairments, such as blindness or limb amputation, continue to hold jobs and generally function satisfactorily.They constitute, however, a group of potentially disabled individuals who might apply and qualify for Disability Insurance or other disability-related benefits if they were to lose their jobs or to decide that employment offered an inadequate financial or non-pecuniary reward. Thus, disability, or a health-related inability to work, is more than a medical problem but involves motivational and attitudinal factors. We specify a model of the application process, which we model as choice under uncertainty about approval of an application for Disability Insurance. We specify the possible outcomes to the choice process of an individual in which the probability of acceptance for Disability Insurance is a key consideration. We then estimate a joint model of labor supply and application to the Disability Insurance program based on the 1972 survey. We then compare our results to the observed time series applications process since 1976. Lastly, we estimate the sensitivity of the application process to the probability of acceptance and the level of benefits.

Abstract:
This paper considers estimation of a transformation model in which the transformed dependent variable is subject to classical measurement error. We consider cases in which the transformation function is known and unspecified. In special cases (e.g. log and square-root transformations), least-squares or non-linear least-squares estimators are applicable. A flexible approximation approach (based on Taylor expansion) is proposed for a parametrized transformation function (like the Box-Cox model), and a semi-parametric approach (combining a semi-parametric linear-index estimator and non-parametric regression) is proposed for the case of an unspecified transformation function. The methods are applied to the estimation of earnings equations, using wage data from the Current Population Survey (CPS).

Abstract:
We consider the retirement behavior of civilian employees of the United States government. Unlike previous studies, this investigation is based upon a data set containing fairly complete and accurate information about the Social Security and employer-provided pensions for which employees are (or ultimately will be) eligible. These data permit us to specify the financial aspects of individual retirement decisions with a reasonable degree of precision. A large fraction of civil service pensioners is eligible to receive Social Security benefits because a part of their working careers was spent in Social-Security-covered employment. The prevalence of double pension coverage among government employees has raised serious equity questions about the treatment of civil servants by Social Security, and these questions have led to various suggestions for pension reform. Partly, the reform proposals have been put forward due to the perceived unfairness of "double dipping" which arises from the double pension coverage of government employees. Our analysis finds: (1) Both the amount of a Federal pension entitlement and the expected wait until the pension commences affect the timing of retirement from the Federal service. (2) The rate of anticipated wage growth significantly affects individual decisions to remain in Federal employment. (3) Workers who are eligible to ultimately receive Social Security in some cases show a different pattern of retirement than do workers not vested in Social Security. However, our analysis does not reveal any massive shift of Federal workers into Social-Security-covered employment in order to benefit from the "tilt" in the Social Security formula.

Abstract:
Seasonal adjustment procedures attempt to estimate the sample realizations of an unobservable economic time series in the presence of both seasonal factors and irregular factors. In this paper we consider a factor which has not been considered explicitly in previous treatments of seasonal adjustment: measurement error. Because of the sample design used in the CPS, measurement error will not be a white noise process, but instead it will be characterized by serial correlation of a known form. We first consider what effect the serially correlated measurement error has on estimation of the non-seasonal component in seasonal adjustment models. We also consider the effect of measurement error on the widely used seasonal adjustment process X11. X11 which is the seasonal adjust procedure used by the BLS will implicitly reduce the effect of measurement error because of the averaging process used. However, this treatment will not be optimal in general. We therefore specify a seasonal adjustment model which takes explicit account of the measurement error. For examples on the unemployment rate, we find that X11 does almost as well as the optimal filter on some series but its efficiency is less than 10% for the teenage unemployment series. We also find that optimal treatment of the measurement error which accounts for the serial correlation can reduce the overall mean square error of the seasonally adjusted series below the variance of the measurement error which is often used as the benchmark for the sampling procedure.

Abstract:
This paper considers both theoretical questions and empirical measures of the effects of various policies of income and payroll taxation on labor supply. It emphasizes deadweight loss as the correct criterion of taxation evaluation, rather than merely output effects. Distributional issues are also discussed. Simulations are done for the Kemp-Roth tax reform proposals to calculate both revenue effects and changes in deadweight loss. Deadweight loss calculations are also done for an equal yield progressive linear income tax.

Abstract:
A Poisson specification of the relationship between atmospheric pollution and lost work days is estimated.An important feature of the procedure is control for city-specific effects. A major source of ambiguity in interpreting the results of observational data on pollution versus health status or death rates is that pollution in a city may be correlated with other characteristics ofthat city that affect these outcomes but are not controlled for in the analysis. Or, individual attributes of residents may be correlated with pollution levels but notaccounted for in the analysis. Our results suggest a statistically significantand quantitatively important effect of total suspended particulates on work days lost. A standard deviation increase in total suspended particulates is associated with approximately a ten percent increase in work days lost. As a concomitant of our analysis, we also find a substantial relationship between smoking by others in the individual`s household and work days lost by non-smokers.

Abstract:
Modern work in labor supply attempts to account for nonlinear budget sets created by government tax and transfer programs. Progressive taxation leads to nonlinear convex budget sets while the earned income credit, social security contributions, AFDC, and the proposed NIT plans all lead to nonlinear, nonconvex budget sets. Where nonlinear budget sets occur, the expected value of the random variable, labor supply, can no longer be calculated by simply `plugging in` the estimated coefficients. Properties of the stochastic terms which arise from the residual or from a stochastic preference structure need to be accounted for. This paper considers both analytical approaches and Monte Carlo approaches to the problem. We attempt to find accurate and low cost computational techniques which would permit extensive use of simulation methodology. Large samples are typically included in such simulations which makes computational techniques an important consideration. But these large samples may also lead to simplifications in computational techniques because of the averaging process used in calculation of simulation results. This paper investigates the tradeoffs available between computational accuracy and cost in simulation exercises over large samples.

Abstract:
In this paper, we consider parameter estimation in a linear simultaneous equations model. It is well known that two-stage least squares (2SLS) estimators may perform poorly when the instruments are weak. In this case 2SLS tends to suffer from the substantial small sample biases. It is also known that LIML and Nagar-type estimators are less biased than 2SLS but suffer from large small sample variability. We construct a bias-corrected version of 2SLS based on the Jackknife principle. Using higher-order expansions we show that the MSE of our Jackknife 2SLS estimator is approximately the same as the MSE of the Nagar-type estimator. We also compare the Jackknife 2SLS with an estimator suggested by Fuller (Econometrica 45, 933-54) that significantly decreases the small sample variability of LIML. Monte Carlo simulations show that even in relatively large samples the MSE of LIML and Nagar can be substantially larger than for Jackknife 2SLS. The Jackknife 2SLS estimator and Fuller's estimator give the best overall performance. Based on our Monte Carlo experiments we conduct informal statistical tests of the accuracy of approximate bias and MSE formulas. We find that higher-order expansions traditionally used to rank LIML, 2SLS and other IV estimators are unreliable when identification of the model is weak. Overall, our results show that only estimators with well-defined finite sample moments should be used when identification of the model is weak.

Weak instruments, higher-order expansions, bias reduction, Jackknife, 2SLS

Abstract:
We evaluate the competitive and governance effects of "duality." Duality refers to the joint membership (e.g., by banks) in competing associations or joint ventures (e.g., Visa and MasterCard). We first show that the not-for-profit nature of the associations along with the usage-based fees they charge yield productive efficiency. We then analyze the impact of (i) membership exclusivity, when the associations remain not-for-profit, and (ii) the conversion into for-profit systems. We illustrate the results in the case of a double-differentiation model that is of independent interest. Finally, we discuss extensions to (i) endogenous system differentiation, and (ii) agency considerations.

Abstract:
We model demand for four cephalosporins and compute own- and cross-price elasticities between branded and generic versions of the four drugs. We model demand as a multi-stage budgeting problem, and we argue that such a model is appropriate to the multi-stage nature of the purchase of pharmaceutical products, in particular the prescribing and dispensing stages. We find quite high elasticities between generic substitutes and also significant elasticities between some therapeutic substitutes.