I am an economist with broad research interests, mainly at the intersection of micro, development, law and economics, and behavioural economics. My interests in micro center on asymmetric information, institutions (including credit market institutions, inheritance norms, and intermediaries like auditors or the medieval "law merchant"), (applied) game theory and mechanism design. My interests in development include microfinance, affirmative action, female labor force participation, and structural transformation. My interests within law and economics mainly focus on the economics of crime, (including work on maritime piracy, gun control, and policing) and litigation (the legal process). Currently, I am an Associate Professor at the Centre for International Trade and Development at the School of International Studies, Jawaharlal Nehru University. Previously, I was an assistant professor at the School of Economics at the Singapore Management University. This is a picture of me.
I earned my PhD (in Economics) from Princeton University in 2005. My thesis title was "Malfeasance and the Market: Essays in Corporate Cheating." Before that, I went to St Stephen's College, Delhi for my bachelors degree (B.A honors in Economics, 1998), and to Jawaharlal Nehru University for a M.A in Economics with specialization in the world economy (2000).
A Brief Description of Some Selected Research Papers
I describe a subset of my research papers below. My entire publications list of 21 journal articles, a book and two book chapters is available here. My entire list of both published and unpublished research is here.
A. Selected Recent Publications
1. "Revisiting the volunteer's dilemma: group size and public good provision in the presence of some ambiguity aversion" Economics Bulletin (2020), 40: 1308-1318.
Conventional game theory dictates that in a volunteer’s dilemma, the probability of public good provision decrease in group size. However, experimental evidence does not support this: the probability of public good provision approaches one in large groups. I build a theoretical model addressing this puzzle, where a fraction of the group has maximin preferences, while the rest are expected utility maximizers. In small groups, the probability of public good provision may decrease with group size, but it reaches one in larger groups. While randomization is common in small groups, larger groups have some individuals who always volunteer, and others who never volunteer (another experimentally observed phenomenon).
2. "Should Jurors Deliberate?" Review of Law and Economics (2019), accepted.
Does the accuracy of verdicts improve or worsen if individual jurors on a panel are barred from deliberating prior to casting their votes? I study this question in a model where jurors can choose to exert costly effort to improve the accuracy of their individual decisions. I find that, provided the cost of effort is not too large, it is better to allow jurors to deliberate if jury size exceeds a threshold. For panels smaller than this threshold, it is more effective to instruct jurors to vote on the basis of their private information, without deliberations, and to use a simple majority rule to determine the collective decision (regardless of the voting rule used with deliberations). The smaller the cost of paying attention, the larger the threshold above which the switch to allowing deliberations becomes optimal. However, if the unanimity rule had to be maintained under the no-deliberations system, it would be better to allow deliberation. The results apply to binary decision making in any committee where the committee members incur some effort in reviewing the evidence. Examples are arbitration panels and tenure and promotion committees and some board of director meetings on issues such as whether to dismiss a CEO. As an extension I consider the case where jurors differ in their costs of effort.
3. "Pretrial Beliefs and Verdict Accuracy: Costly Juror Effort and Free Riding", B.E. Journal of Theoretical Economics, (2019), accepted.
If jurors care about reaching the correct verdict, but also experience costs to paying attention during the trial, even a small effort cost generates interesting interactions between pretrial beliefs and verdict accuracy. I demonstrate the existence of a strong free riding effect; jurors respond to a more informative prior by reducing their probabilities of paying attention, to the extent that over a non-empty range, a more informative prior will be associated with poorer verdicts. Pretrial beliefs can depend on several factors: I consider two – the extent of discovery during the pre-plea bargaining process, and the efficiency of the police. My results imply that more liberal discovery rules, which result in a less noisy plea bargaining process, will actually be complemented by greater juror effort over a range, resulting in more accurate verdicts. In contrast, greater police efficiency will, over a range, elicit a sufficient drop in juror effort such that verdicts are less accurate. Thus, improving discovery has added benefits over a range, while the benefits of exogenous improvements in policing may be dampened.
4. "Malice and Patience in Rubinstein bargaining", Research in Economics (2019), 73:264-270.
If two players playing a Rubinstein alternating offers game are highly malicious (getting a high utility from “malice” in every period when the other player does not obtain a share in a fixed pie), and highly patient, no equilibrium with an agreement exists and players choose perpetual disagreement. This does not change if the players are subjected to a known deadline after which the pie will be appropriated by outside agencies or disappear: perpetual disagreement is still the only outcome. If in addition players are required to pay endogenously determined fines if they fail to reach agreement, players with discount factors in a certain range do reach agreement, but only at the deadline. However, infinitely patient players would never reach agreement for any feasible level of one-time fines. The result contrasts with spiteful or envious preferences. Our results highlight a novel reason for failure to resolve property disputes.
5. "Malice in Pretrial Negotiations", International Review of Law and Economics (2019), 58:25-33.
This is the first paper to incorporate two-sided malice into a complete information model of pretrial bargaining. Each party obtains malice utility from financial losses incurred by the other: psychological dissatisfaction caused to the other party by depriving them of a resolution also contributes to a utility from malice in every period in which there is no resolution. Introducing this feature in a complete information model a la Spier (1992) where a trial date is exogenously set and a plaintiff makes offers to a defendant in every period before the trial, and where there are both bargaining costs before the trial and litigation costs during the actual trial, I find that (i) some cases are never settled, but proceed to trial, (ii) settlement occurs either immediately or on the courthouse steps. This is in line with empirics, and happens without any form of private information or divergent expectations, contrary to traditional explanations. Since there is ample experimental evidence of malice as a motive, and since it is even more relevant between disputing parties, the new explanation that the paper suggests for both settlement failure and the timing of settlements is likely significant. The results are robust to a change in bargaining protocol to alternate offers, to the trial date being endogenous in an infinite horizon model, and to the introduction of costly discovery and uncertainty of plaintiff victory.
6."Arbitration, Indivisible assets and the Possibility of Malice", Bulletin of Economic Research (2018), 71:404-415.
Many real-life contexts involve allocating an indivisible prize between two claimants. One (or both) claimants may derive "malice utility" from depriving the other claimant. I study an arbitrator who does not know if malice is involved in a particular dispute, but wishes the prize to be allocated to the party with the higher intrinsic valuation for it and discuss two simple mechanisms that achieve his purpose regardless of whether malice is actually present. I examine several extensions.
7. "Malice in auctions and commitments to cancel", Economics Bulletin (2018), 38:1623-1631.
A bidder is “malicious” if his valuation includes, apart from his intrinsic valuation for the object, a malice component from depriving other bidders. A bidder has malice even if the auction is cancelled as this prevents the other bidder from getting the object. I consider a second price sealed bid auction of a single object with two potentially malicious bidders. Valuations, including malice components, are private knowledge. Suppose the seller commits to cancel the auction if either bidder drops out. Then, while bidders with standard preferences never drop out, malicious bidders do so over a non-empty parameter range, and the auction is cancelled despite the absence of any entry fees or reserve prices. This could also be true of spiteful bidders. However, if valuations are common knowledge, the bidder with the higher intrinsic valuation never drops out in an auction with two spiteful bidders (given the seller’s commitment to cancel if he does): however, in an auction with two malicious bidders, both might drop out. Experimenters might thus be able to distinguish between (i) standard bidders versus malicious or spiteful ones, and (ii) spiteful versus malicious bidders. For specific distribution functions, the tendency to drop out is increasing in the bidder’s malice.
8. "Secret ballots and costly information gathering: the jury size problem revisited", International Review of Law and Economics (2018), 54:58-67.
Suppose paying attention during jury trials is costly, but that jurors do not pool information (as in contemporary Brazil, or ancient Athens). If inattentive jurors are as likely to be wrong as right, I find that small jury panels work better as long as identical jurors behave symmetrically. If not paying attention makes error more likely than not, jurors may coordinate on two different symmetric outcomes: a “high attention” one or a “low attention” one. If social norms stigmatize shirking, jurors coordinate on the high attention equilibrium, and a smaller jury yields better outcomes. However, increasing the jury up to a finite bound works better if norms are tolerant of shirking, in which case coordination on the low attention outcome results. If jurors always act as if they are pivotal, a larger jury may work better. Allowing deliberations is efficient if the jury panel is relatively large, and if the police and prosecution are effective. However, barring deliberations is better at smaller jury sizes, specially if the police and prosecution are not too efficient.
9. "Malice in the Rubinstein bargaining game", Mathematical Social Sciences (2018), 94:82-86.
This paper incorporates malice into the Rubinstein alternating offers bargaining game. Malicious players obtain a positive payoff in every period in which the other player does not obtain any piece of the pie. This “malice payoff” is independent of whether the malicious player himself obtains the pie or not. I identify a unique SPNE of the bargaining game. With two equally malicious players, the equilibrium shares of the proposer and respondent are more equal than under traditional Rubinstein bargaining. Intuitively, this is because the respondent has the right of first rejection. However, this solution requires an upper bound on the players’ patience; malicious players who are also infinitely patient would not participate in the bargain in the first place. This is in contrast both to the case of “spiteful” preferences (where a player’s spite payoff is decreasing in the share that the other player gets) and “envious” preferences (one-sided inequality aversion), in both of which an interior bargaining solution can occur even if the discount factor approaches one. With one-sided malice, malice confers bargaining advantage. With two-sided malice, and unequal malice parameters, the proposer may obtain a higher or lower share than in the traditional Rubinstein game, and may end up with a lower share than the respondent (even if both have equal discount factors).
10. "Costly Leader Games with a Probabilistically Non-Strategic Leader", International Game Theory Review (2017), 19: 175008.
I consider a two-person costly leader game – in which the follower endogenously chooses whether to buy information about the leader, and a follow-up action – with a twist. With a known probability, the leader is a non-strategic type. A strategic type leader may choose to masquerade as a non-strategic type, at a cost. I show that, if the follower’s cost of information is not too large, and the probability of the leader being non-strategic is neither too large nor too small, this game has no pure strategy equilibrium. Moreover, the equilibrium of the simultaneous-move complete information game is inaccessible as follower information cost converges to zero. There is no equilibrium outcome in which leader advantage is destroyed: however, a mixed strategy equilibrium exists which does preserve leader advantage in the sense that payoffs and strategies converge to those of the sequential-move complete information equilibrium as information cost tends to zero. My results differ from some traditional results in costly leader games, and are due to the interaction of two forces, the “type uncertainty” and the “money down the drain” effects. To my knowledge this is the first paper to integrate behavioural types into costly leader games (other papers considering heterogeneity in type do not consider non-strategic players).
11. "Testing for Malice", Economics Bulletin (2017), 37: 327-335.
Consider two parties disputing claims over an indivisible prize. A malicious claimant may or may not intrinsically value the prize for its own sake, but always derives pleasure “malice utility” from depriving the rival claimant. I devise a method for detecting malice in experimental settings. I derive a simple mechanism which allows third parties (such as experimenters) to distinguish whether (i) both claimants bear each other malice (two-sided malice) (ii) whether only one claimant bears the other malice (one-sided malice) and if so, the identity of this malicious claimant, and (iii) whether neither of the claimants are maliciously motivated. I show that, with slight modifications, this mechanism is applicable both to the case where the claimants know each other and to the case where they are strangers. I also discuss a method whereby the experimenter may infer an upper bound on the malice of the less malicious party in the case of two-sided malice.
12. "Moral Hazard, Bertrand Competition and Natural Monopoly", Journal of Economics (2017), 121: 153-171..
In the traditional model of Bertrand price competition among symmetric firms, there is no restriction on the number of firms that are active in equilibrium. A symmetric equilibrium exists with the different firms sharing the market. I show that this does not hold if we preserve the symmetry between firms but introduce moral hazard with a customer-sensitive probability of exposure; competition necessarily results in a natural monopoly with only one active firm. Sequential price announcements and early adoption are some equilibrium selection mechanisms that help to pin down the identity of the natural monopolist. The insights of the basic model are robust to many extensions.
13."Malicious Litigation", International Review of Law and Economics (2016), 47:24-32.
It has long been recognized that some plaintiffs sue defendants out of malice, but malicious litigation has not been previously modeled in the law and economics literature. I construct a simple model of malicious litigation, wherein malice is defined by the plaintiff’s obtaining some utility whenever the defendant incurs costs. When plaintiffs are malicious, they are more likely to file even non-meritorious suits; both probability of filing and the plaintiff’s settlement payoff increase in the plaintiff’s malice. However, if the defendant is also malicious, obtaining utility when the plaintiff incurs litigation expenses, settlements may fail even with complete information. Two-sided malice deters filing over a certain parameter range; outside it, it raises the ratio of cases that go to trial instead of being resolved through settlement. Giving the defendant the right to call for a bar on settlement is less effective at deterring malicious lawsuits relative to non-malicious “negative-expected-value” (NEV) or “nuisance” suits. However, combining the optional settlement bar with a “commitment requirement” stipulating that the plaintiff commit to going to trial (rather than withdraw) whenever the defendant opts to defend discourages malicious litigation for a wider range of parameters.
14. Microfinance Competition: Motivated Microlenders, Double Dipping and Default", Journal of Development Economics (2013),105:86-102 (with P. Roy Chowdhury).
In this paper, we use a Salop-type circular model with a continuum of borrower types and two sources of lender-borrower asymmetric information (the first concerning borrower type, and the second stemming from lenders' inability to observe if their borrowers are borrowing from other lenders, or consuming their loans) to construct a tractable model of double dipping (multiple-loan taking) and default. Our lenders are microfinance institutions (MFIs) that care both about borrower welfare and about their own profits. First, we establish conditions for an equilibrium where some borrowers always take multiple loans and default; we also characterize another no-default equilibrium which exists for different parameter values. Next, we study the impact of increased MFI competition on interest rates, default, and borrower welfare. We find surprising results, for example, we see that (i) even though competition among relatively profit-oriented MFIs decreases interest rates, it nonetheless raises borrower default, (ii) this increased default occurs, paradoxically, because of a reduction in transaction costs experienced by borrowers, and not because of a dilution in the quality of the borrower pool, (iii) competition among MFIs which care more about borrower welfare actually pushes up interest rates. We also obtain comparative static results about a change in MFI orientation, showing that increasing profit orientation may create mission drift, reducing borrower welfare. We derive policy implications regarding a cap on MFI interest rates, and on subsidies to MFIs.
15. "Guns and Crime Revisited", Journal of Economic Behavior and Organization (2013), 94: 1-10 (lead article).
In this paper, I build a theoretical model that explores the links between different empirical findings on crime (a) the drop in U.S crime rates since the nineties, (b) the heavy increase in policing and in sentencing, (c) a drop in gun use by private householders, and (d)a rise in other types of private precautions, such as keeping less cash on hand, or greater reliance on home or car alarms. In my model, I focus on potential victims' tradeoff between different types of precautions, and study the interaction between this tradeoff and the intensity of policing. My results are consistent with observed trends.
16. "Who Will Monitor the Monitors?Informal Law Enforcement and Collusion at Champagne",Journal of Economic Behavior and Organization, (2012),83(2): 261-277.
In this paper, I revisit the question of how large-scale impersonal trade became sustainable in the absence of a legal mechanism that could discipline cheats. An important role in this process was played by authorities at medieval trade fairs like the Champagne Fairs, as shown by Milgrom, North and Weingast (1990). However, this only begs the question of why people should trust such authorities, given their vulnerability to collusion. I tackle this issue, first showing that Milgrom et al's model is not collusion-proof, even if it were extended to allow for the possibility of collusion. Secondly, I construct a model showing how the fairs could have survived, as they did for several centuries, in spite of their vulnerability to collusion. My model incorporates historically relevant features such as competition from rival fairs and the role of traders' guilds. Moreover, my model sheds light on other historical facts such as the gradual shift away from community-based punishments towards individual liability.
17. "Reinterpreting King Solomon's Problem: Malice and Mechanism Design", Journal of Economic Behavior and Organization, (2014),98: 125-132.
In this paper, I argue for a new interpretation of King Solomon's Problem (two mothers claiming the same child) in terms of one of the two claimants being malicious. A malicious claimant places no intrinsic value on the object but derives pleasure from depriving the rival claimant. The traditional literature on King Solomon's problem, in contrast, ignores malice and interprets the problem as one of allocating an indivisible object between a high-value and a low-value claimant. My new interpretation permits a considerably simpler solution. I derive a mechanism that induces truthful revelation following a single round of elimination of weakly dominated strategies, and no monetary transfers. I consider extensions where the malicious claimant also places some positive, but low, value on the object, and also allow for two-sided malice. I discuss several real-life examples, including malicious patents and asset disputes between separating spouses or extended family members.
18. "Crime and Moral Hazard: Does More Policing Necessarily Induce Private Negligence?" Economics Letters, (2012),115(3): 455-459 (with A. Guha).
In this paper, we show that far from inducing potential victims to neglect their safety, an increase in policing can actually induce people to take a higher level of costly precautions against being attacked. We derive conditions under which this happens, in three separate situations.In the first, where individuals may directly spend on private security measures, and may also "diversify" by splitting their assets into spatially separate lots, reducing the prize that can be seized in a single attack. In the second, individuals may diversify but may not spend on direct security measures (for example, due to bans on firearm possession). In the third, individuals can spend on security measures, but cannot diversify, as is the case with an asset that is indivisible, such as an Old Master.
19. "Pirates and Fishermen: Is Less Patrolling Always Bad?" Journal of Economic Behavior and Organization (2012), 81(1): 29-38.
In this paper, I model Somali pirates' time allocation problem between piracy and a non-violent alternative, fishing, whose returns are also sensitive to patrolling intensity. I show that increasing patrolling beyond a threshold can actually lead to an equilibrium with more pirate attacks. Lowering patrolling below this threshold results in an equilibrium where pirate attacks and patrolling costs are both lower. The model helps explain the puzzling fact that pirate attacks had been on the rise in spite of a rise in patrolling intensity; moreover, this did not reflect an inefficiency in patrolling, as patrols were foiling a larger portion of total pirate attacks.
20. "Pirates and Traders:Some Economics of Pirate-Infested Seas",Economics Letters (2011),111(2): 147-150 (with A Guha).
In this paper, we show that even risk-neutral traders have the option of costly "self-insurance" on a "pirate-infested sea" (one where there is an elastic supply of pirates prepared to attack whenever it is profitable to do so). Moreover, if patrolling is high enough, the incentive for traders to self-insure strengthens, so that the combination of private and public precautions can eliminate piracy.
21. "The Persistence of Goodness", Journal of Institutional and Theoretical Economics, (2012),168(3):432-443 (with A. Guha).
In this paper, we derive game theoretic underpinnings of the level of goodness in a population. "Good" people act in a way which is beneficial to others' payoffs, and do so even without the rule of the law or reputational effects. We show that virtue does not die out, even in a Darwinian world of the survival of the fittest; for appropriate parameters, a certain level of good "behavior" emerges as an evolutionarily stable equilibrium. The long run equilibrium proportion of good behavior does not depend on the level of intrinsic goodness.
22. "Borrower Targeting Under Microfinance Competition with Motivated MFIs", Developing Economies, (2014), 52(3):211-240 (with P. Roy Chowdhury).
This paper is motivated by the observation that even socially motivated MFIs often do not target the poorest of the poor. We examine how increased competition for donor funds among MFIs that only care about borrower welfare affects the poorest borrowers' access to microfinance. We find that borrower targeting depends on intra- borrower inequality, the technology available to borrowers, and the feasibility of double-dipping (taking loans from multiple lenders).Without competition, even a motivated MFI may lend to the not-so-poor in preference to the poor borrowers. If double-dipping is feasible, competition may encourage lending to the poor. The presence of double-dipping is critical for MFI competition to have this positive effect. When double-dipping is feasible, MFI coordination may worsen borrower targeting whenever inequality is intermediate. We discuss policy implications dealing with double dipping, MFI co-ordination and competition.
B. Selected Recent Working Papers
1. "Vulnerability to Exposure as a Determinant of Market Structure", under review.
Focusing on repeated games with imperfect public information and moral hazard, I show that if the probability of the informed party being detected and exposed increases in the number of its transactions, a natural monopoly tends to emerge even in the absence of traditional barriers to entry. I apply this to settings where the informed party is a "seller" of (i) "credibility-enhancing services" (such as an external auditor, vulnerable to collusion with client firms whose credibility they are supposed to boost), (ii) experience goods, particularly those where failure has disastrous consequences (nuclear reactors, planes, engines), or (iii) services where negligence or dishonesty may be detected ex-post, eg in the hospital industry or the insurance industry. Subsidiary results demonstrate that while auditors may lack bargaining power at the hiring stage, they will obtain considerable bargaining power while demanding bribes. Moreover, if an auditor does collude, he does so with his entire clientele.
2. "Grandparents as Guards: Inheritance and Post-Marital Residence in a World of Uncertain Paternity".
I build a game theoretic model to explain why (a) a majority of agrarian societies developed a combination of patrilineal inheritance (through the male line), patrilocality (with a young couple moving in with the husband's parents), and high "monitoring" of young women's sexual behavior by in-laws, and (b) a significant minority nonetheless developed the opposing triad of matrilineal inheritance, matrilocality (where women lived on in their parental homes after marriage) and relative sexual freedom for women. I build on anthropological evidence and my model involves an intergenerational game where productive effort on the family farm is chosen by young adults while the level of sexual monitoring exerted on co-resident young women is chosen by the old. I demonstrate that my hypothesis is superior to alternative explanations that bypass the factor of paternity uncertainty.
3. "Endogenizing Giving, Reciprocity and Inequality Aversion: Cross-Cultural Differences in Ultimatum Games", under review.
Not only do players of ultimatum games not exhibit rational maximizing behavior, but there is also considerable cross- cultural difference in these players' behavior patterns. Motivated by this, I use a costly conflict game to study how different types of behavior may optimally emerge in different societies, depending on parameter values like the initial distribution of strengths, the cost of conflict and the extent of uncertainty.
C. Selected Older Publications
1."Trade,Growth and Increasing Returns to Infrastructure:The Role of the Sophisticated Monopolist", Review of International Economics (2009),17(5):1053-1065 (with A Guha).
In this paper, we model the role of infrastructure providers' beliefs in the ability of an economy to graduate from being a primary product exporter to an exporter of manufactures. Equilibrium in international trade with increasing returns in infrastructure depends on whether the infrastructure provider is “naïve” or sophisticated. A monopolist produces infrastructure under decreasing cost using fixed equipment. Unlike similar work, we derive a unique closed-economy equilibrium. In a small open economy, with “naïve” infrastructure provider(s), multiple equilibria obtain. The industrial export potential of the economy depends on unexhausted economies of scale, and equilibria are possible where manufactures are exported despite an autarky price higher than the world price. With a sophisticated infrastructure provider, even an open economy has a unique equilibrium, which, at least as long as economies of scale are unexhausted, also involves more industrialization than the “naïve” equilibria. Access to the unlimited world market is necessary for significant industrialization but is not sufficient: one may also require “Schumpeterian” entrepreneurs, monopolists with a panoramic vision of the economy and of their catalytic role in it.
2. "Target Saving in an Overlapping Generations Model",B.E Journal of Macroeconomics (2008),8(1):1-24(with A Guha).
In this paper, we derive microfoundations to explain the observed fact that many people save towards a fixed target level of retirement resources, neither more nor less. We then find, using an overlapping generations model, that such saving then leads to multiple and unstable equilibria. Under static expectations, it also leads to a well- defined dynamics, including possible historical traps, implosions involving ever-diminishing capital stock and ever-increasing interest rates, and the feasibility of optimal one-time interventions.
3."Maids and Mistresses:Migrating Maids and Female Labor Force Participation",Economics Bulletin (2007),10(11):1-9.
In this paper, I examine the impact of immigrant maids on formal workforce participation by local women in a model with a competitive factory sector and a household sector which uses both immigrant maids and family labor. I find some surprising results. Relaxing immigration restrictions on maids does not necessarily increase the labor force participation of these maids' mistresses. Reducing taxes on hiring immigrant maids may not increase female formal labor force participation; instead, imposing such taxes, where there are none, might serve to increase female work force participation, depending on immigrants' and locals' relative propensities to consume household sector output.
4."Female Labor Force Participation and Labor-Saving Gadgets",Journal of International Trade and Economic Development (2005),14(4):483-495.
I show under what conditions women would migrate out of the household sector into formal sector jobs, in response to increased ability to use labour saving household gadgets, which raise the productivity of female labour engaged in household tasks. I model a small open economy with three outputs: one labour-intensive manufactured export (cloth), one capital-intensive intermediate good (gadgets) and one non-traded ‘household-sector good’ (meals) which requires both female labour and household gadgets for production. A terms-of-trade improvement capturing greater world demand for labour-intensive manufactured exports enables greater adoption of labour-saving household gadgets in response to changing relative prices. If the elasticity of substitution between female labour and household gadgets exceeds a threshold, this will result in women migrating from the household to formal sector employment. What matters is not the actual date of invention of these labour-saving appliances (female labour force participation may not grow significantly until long after) but their increased adoption by the small economy in response to changing relative prices.