The implication for politics of the “mobile factors” approach is just this: thescarce factor (labor, in the above example) will favor restricting trade, and theabundant factor (capital, in the above example) will have incentives to favorliberalizing trade, no matter where in the economy those factors are employed.

However, the interests of one of these groups offactor owners will in general be aligned with the interests of the owners of themobile factor(s). It seems probable that the stakes of the mobile factor ownerswill be smaller than those of the the specific factor owners.

But what if the capital used in the labor-intensiveindustry is different from the capital used in the capital- intensive industry? Tobring the example back down to earth, what if knitting machines cannot be usedto make microchips? Indeed, in many real world situations it seems intuitivelylikely that this will be the case: capital (or certain kinds of labor, for that matter)will not be able to flow easily from a declining sector to a rising sector. Adifferent set of assumptions is needed for this contingency. According to theassumptions of the Ricardo-Viner model (or “specific factors” model as it isoften called) factors of production are “specific” to a particular industry, andwhen that industry declines they cannot move to the rising industry

The assumption that factors are mobile between sectors of the economy is crucialto the derivation of the Stolper-Samuelson theorem. It is only because capitalcan flow from the import-competing (labor-intensive industry) to the capital-intensive industry that it is able to enjoy the effect of the increased productionof the capital-intensive good.

The logic may be somewhat involved, but the bottom lineis not: in this example of a capital-abundant country, labor loses and capitalwins from freer trade.

Combin-ing this prediction with the Stolper-Samuelson theorem yields the usual conclu-sion that, other things being equal, in a relatively capital-abundant country laborwill favor protection because it cannot be intensively used in exports, whilecapital will favor relatively free trade.

Stolper-Samuelson or “mobile factors” approach (centralin Rogowski’s work) and the Ricardo-Viner or “specific factors” model (centralto Frieden and others).

To bring it down to earth, a one milliondollar lobbying contribution from G M will likely have a large effect on tradepolicy. A ten dollar contribution from an individual auto consumer will havevirtually no effect. Thus even though the cost of the auto consumer’s contributionis negligible, the expected benefits are even more negligible.*

Two answers to the puzzle are possible. First, there maybe per person transaction costs in organizing groups. Second, if policy outcomesare probabilistic, members of large groups with small per person stakes andcontributions may suppose that their own contributions will be insignificant tothe political outcome and therefore not make them. On the other hand, membersof smaller groups, with their larger stakes and contributions per person, may seethat their contribution has a non-negligible impact on the likelihood that a policywill be enacted, and therefore they will make their contributions.

A protectionist measure provides large benefits to a small number of people, andcauses a very great number of consumers a slight loss. This circumstance makes iteasier to put a protection measure in practice

We conclude by showing the importance of distinguishingthe intensity of demand (how much you want something) from collective actionsproblems (the relative cost of organizing politically to get something, howevermuch you want it) and we speculate about the effects of factor mobility on thedifficulty of collective action and what impact this has on the politics of tradepolicy

may become weighing the good of the many, which isoften served by relatively free trade, against the good of the powerful few whichmay be served by restricting trade.

In domestic politics the conflict over these distributional consequences willreflect the trade policy coalitions that form around shared interests in liberali-zation as opposed to protection. Whether trade policies are taken to be chosendemocratically or imposed from above, whether those coalitions are engaged invote mobilization or protest, the balance between the opposed coalitions favoringfreer trade and those favoring protection creates the “demand” by society forliberalization or protection.

Anyone theorizing about “trading states”(states of trading societies) should consider the state’s problem of how to weighthe aggregate

Trade typically offers cheaper goods, with more choice for consumers andthe greatest economic output for society as a whole. But at the same time, it isalso very disruptive to individuals’ lives, tying their incomes to the vagaries ofinternational markets. In so doing, trade affects the distribution of wealth withinthe domestic economy, raising questions of who gets relatively more or less, andwhat they can do about it politically.

whichemphasizes the influence exerted by special-interest groups on policy makers by means ofpolitical contribution

Forthis reason, we were surprised to find that thestrict version of the G-H model is not "grossly"inconsistent with the data.

Tests of the strict versions of trade models tra-ditionally yield disastrous results for the theo-ries under investigation-the poor empiricalperformance of the strict version of theHeckscher-Ohlin model is a good example

Of some interest, we believe, is also ourestimate of a key structural parameter of themodel, namely the weight attached by the gov-ernment to social welfare. We estimate thisweight to be many times higher than the weightattached to contribution

e measurement prob-lems associated with trade elasticities should not biasour results given that elasticities do not appear on theright-hand side of the esimating equa

Presumably,industries that are experiencing high unemploy-ment rates tend to be more vocal in the politicalarena, and manage to obtain more trade protec-tio

ur interpretation of this finding is that theUnited States is relatively open to trade, evenwhen nontariff barriers are accounted for

Thus, the positive correlation between importpenetration and protection documented in pre-vious papers does not constitute evidenceagainst the mod

heseworks take a reduced-form approach, in thesense of not being guided by a theoreticalmode

The model's predictionswith respect to the relevant coefficient signsare as follows: (i) trade protection should behigher in industries represented by a lobby,and in industries with a lower import elastic-ity; (ii) within the subset of organized indus-tries, protection should be higher in industrieswith lower import penetration, whereas in thegroup of nonorganized sectors, protectionshould increase with import pen

yields clear predictions for the cross-sectional struc-ture of import barriers

We concur with Jenkins, who points out that "the existence ofa crisis is no guarantee that a government will respond, and more importantly,that it will be successful in convincing interest groups that 'something must bedone.' "

Democracies choose lowerlevels of trade barriers, even when holding many other factors constant

his variable(IMF) was never significant. Increases in external pressure via the IMF have littlemeasurable effect on trade policy, and when they do (in the regressions on Sachs-Warner openness), their effect is to increase the closure of the economy, not to openit

These variables do not seemto matter significantly, either independently or jointly. Countries may respond dif-ferently to crises, sometimes raising trade barriers and other times lowering them.These results are not unexpected. Many claims about the role of crises come fromcase studies that often select on the dependent variable; that is, they explore casesof economic reform and often find that crises (of varying types) existed as well.But they rarely look at the cases where countries experienced crises and did notreform.106 As noted earlier, the debt crisis of the early 1980s did not lead to tradeliberalization, but to greater closure. The effects of crises may be highly contin-gent on the environment

We also lagged the main alternatives to our argument: the crisis variables(EC CRISIS and BP CRISIs) and an external pressure variable (IMF) for two and threeperiods for both dependent variable

Our central hypothesis is that more democratic countries should have fewer tradebarriers, ceteris paribus; an increase in democracy should prompt a reduction intrade barriers.

n sum, in developing countries where autocratic governments depend on sup-port from a small selectorate and thus are not responsive to the overall population,the governments can employ extensive protectionism.

tal. Those with "voting" rights in autocracies thus ownabove-average levels of capital, and benefit the most from high levels of protec-tion

ralization? Democratization ma movement toward majority rule with universal suffrage in contested eleWe call the group of actors who participate in the selection of political"the selectorate," following the terminology of Bueno d

The democratization ofmany East Central European countries, such as Poland and Hungary, was fol-lowed by economic reform, including massive trade liberalization.4

In no country were trade barriers reduced to zero, and thus weare not arguing that protectionist interest groups no longer mattered. But indemocratizing countries they mattered less than before, and thus leaders could lib-eralize more than previously

. Many scholars agree that in large parts ofLatin America, Africa, and Asia, the groups that gained from ISI were urban own-ers of industry (that is, capitalists) and urban, higher-skilled and often unionizedworkers (more capital-endowed workers); the losers tended to be less-skilled, poorerworkers, generally rural ones.24Even the debt crisis of the early 1980s was unable to shake this coalition fromits economically inefficient policies. Many scholars report that protection increasedin the wake of the debt crisis, especially in Latin America.

The conventional wisdom, however, has been that democracy is not propitious foreconomic reform, and indeed that nondemocratic countries should be more likelyto liberalize, as Chile in the 1970s suggests. As Geddes summarizes, "until recently,it was widely accepted that democracies, especially fragile, uninstitutionalized newdemocracies have difficulty carrying out economic liberalization because its costsmake it unpopular and hence politically suicidal to elected officials.

xplanations for this change have emphasized three factors. Some scholars claimthat economic crises have forced countries to reform and liberalize. Others haveclaimed that external pressures from the United States, Western countries in gen-eral, or international institutions, such as the World Trade Organization (WTO),IMF, and World Bank, are responsible. Finally, the spread of neoliberal policyideas is often credited with bringing economic liberalization. We examine theseclaims with respect to trade policy and introduce another factor that we think isimportant and underappreciat

mocracy. We argue that these two trendsare related: democratization of the political system reduces the ability of governments to use trade barriers as a strategy for building political support. Political lead-ers in labor-rich countries may prefer lower trade barriers as democ

enators whomatch a voter's trade policy preference can expect to see only a minimal boost inthe probability of re-election; while those who do not pay a minimal cost in votes

Additionally, higher levels of education, a proxy for high-skilled workers andthus voters with theoretically predicted free-trading preferences, are linked with ahigher propensity to provide an answer. However, there is not a significant differ-ence between the likelihood of the answers being correct or incorrect. Regionally,those in the Midwest and the South are far more likely to both offer an answerand provide a correct answer.

Second and unexpectedly, given the usefulnessof union organizations for information distribution, union membership increasesthe likelihood of not providing an answer, even after controlling for factors suchas education

4 In each case, the difference between trade policy and all other issues wasstatistically significant

y. CAFTA attracted a politically wide spec-trum of opponents: liberal presidential candidate Ralph Nader,19 conservative pres-idential candidate Pat Buchanan,20 and the Nobel Prize-winning economist JosephStiglitz,21 among others, all spoke out against CAFTA. Unlike many other billsthat mix together multiple policy areas, the CAFTA legislation focused almostexclusively on trade issues, making interpretation of politicians' vote choice eas-ier for voters. If voters do not hold politicians accountable over a recent, high-profile piece of trade legislation such as CAFTA, they are unlikely to do so formore typical bil

hus, salienceembraces two concepts: knowledge of deviations of the candidate's position fromthe voter's preferred policy and the relative importance the voter places on this

A finding of low salience in terms of knowledge or importance combined withlow accountability of politicians on trade policy issues would suggest the need tore-evaluate voter-inclusive models of trade policy, especially those with a genericcall to the voter.

hese models agree that votersmatter, but diverge on which voters matter. In majority voting models, character-izations by economic interest of the voter may differ, but once preferences andinstitutional rules are defined, aggregation is as simple as counting. Alternatively,influential voter models focus on variations in saliency - generally characterizedby diffuse versus specific interests - which may lead some voters to care moreabout the issue, to be more knowledgeable, and thus to hold politicians account-able for their policy decision

the day. The article finds trade policy salience to berelatively low in terms of stated importance, in voters' knowledge of their representatives' policy positions, and in its effect on voters' propensity to vote for the incumbent. The low salience of trade policy, particularly among highly affected groupcalls into question voter-driven models of

ctorate. Using data from a survey of 36,501 potetial voters in the 2006 U.S. midterm congressional elections, this article explores tsalience of trade policy for voters as a whole and for populations predicted to bemost affected by changing trad

8 Likewise, when unemployment isemphasized in local and national media, individuals' perceptions of the state ofunemployment also become more negative.69 Any tendency for the press to placegreater emphasis on the harmful impact of trade and less emphasis on the benefi-cial impact might help to account for the negatively skewed sociotropic percep-tions shown in Figure 1.

. First, many have found that mass media coverage of the economy playsan important role in forming these perceptions

Consequently, the effects of education reflect differences in attitudesabout the extent to which the United States should take an activist stance in inter-national affairs and a general tendency to think less of out-groups relative toin-groups

The estimated coefficients of the two variablestapping exposure to ideas about the benefits of trade are both positive, econo-mists' view of trade has a statistically significant effect, whereas the influenceof economics class is marginally significant (/? = .105). Nonetheless, economicknowledge has a small impact on trade attitudes: taking an economics class orunderstanding that economists argue that free trade is beneficial increases the pre-dicted value of the KN dependent variable by only 2 to 3 percent, holding con-stant the remaining variables in th

Our initial results - in the first column of Tables 1 , 2, and 3 - provide clear evi-dence of a gender gap, with women holding more protectionist attitudes than m

t. Individuals who attended graduate school are much more supportiveof open trade than both people with only a high school education and those withsome college educat

, "In your current (or most recent) job,what kind of work do you do?

Our key independent variables fall into three broad categories: (1) indicators ofthe individual characteristics suggested by the factor endowments and specific fac-tors models, (2) indicators of respondents' perceptions of how trade influencestheir family's financial conditions as well as the country as a whole, and (3) indi-cators tapping potential noneconomic influences on trade preferences, includingnationalism, ethnocentrism, and attitudes toward intervention in the a

Thus, to the extentthat trade preferences are similar to attitudes toward other aspects of economicpolicy, they will stem from people's perceptions of the collective impact that tradepolicy has on the nation. This pattern has been dubbed sociotropic influence becauseof the tendency it suggests for relying on collective-level information rather thanpersonal experien

Empirical research suggests that education affects political pref-erences in at least two ways: (1) through occupational prominence and position insocial networks, and (2) through cognitive proficiency and analytical ability

Equally, an individual employed in a nontradedsector of the economy should be more protrade than someone employed in anindustry that faces substantial competition from foreign import

Open trade increases the demand for skilledworkers, since the United States has a comparative advantage in the production ofgoods that use this input intensively. As the demand for such workers rises, so dotheir wages. Conversely, heightened trade decreases the demand for low-skilled laborin the United States and depresses its wage rate since goods produced using suchlabor intensively can be produced more efficiently abroad.5

psychologically based models that address how individuals process informationabout the economy, and how they form attitudes about people and places beyondtheir borders

Consistent withprevious research on the impact of self-interest on policy attitudes, our data showlittle support for either the factor endowments or the specific factors model. Bycontrast, perceptions of how trade affects the country as a whole - what are oftenreferred to as "sociotropic" perceptions - play a substantial role in shaping atti-tudes about foreign commerce, a role that is largely independent of self-interest.

First, the factor endowments approach emphasizes that, in a given country,trade benefits those individuals who own factors of production that are in abun-dant supply relative to the remainder of the world, and harms owners of factorsthat are in scarce suppl

works. We find little support for either of these mod-els using two representative national surveys of A

The specific factors or Ricardo-Viner model posits that trade preferences depend onthe industry in which a person

factor endowments or Heckscher-Ohlin model posits that these preferences are affected primarily by a

The likely opponents of globalization among firmsare, in contrast, large in number, relatively weak in their pref-erences, and lacking in political powe

We propose, then,that the rise of intra-industry trade—and of the superstar ex-porter divorced from the concerns of the weaker firms in hisindustry—has contributed to the steady erosion of trade bar-riers over the past several decades

Concern about import com-petition and export opportunities are positively correlated,suggesting the importance of intra-industry trade flows.

We find that attitudes are bestexplained by firm-level characteristics regardless of industryfeatures.

Finally, it is worth noting that firms generally rated changesin wage rates, whether among skilled or unskilled workers,as the least important determinants of their trade attitudes.Among responding firms, 73% and 78% of respondents re-ported that changes in skilled and unskilled workers wageswere “not important” or only “somewhat important” in eval-uating trade policy changes.

Firms that feel that their products are easy toreplace with varieties made by competitors are extremelyworried about the impact of import competition; firms thatfeel that their products are more difficult to find substitutesfor are much less concerned.

A very strong pattern emerges: only the largest and mostsuccessful exporting firms consider trade politics to beworthy of interest and expending their political capital. Ex-port status alone does not generate such interest, but a highintensity of export participation does

Firm-level factors, especially export ac-tivity and intensity, are strongly predictive of support fortrade liberalization, while various proxies for industry-widecomparative advantage are not.

. Lobbying is also common-place: 35.9% of our respondents reported contacting the Of-fice of the President, the Legislative Assembly, or the Ministryof Foreign Trade (COMEX) to discuss public policy.

Prediction 2. The largest exporters will consider tradepolicy to be a much more significant issue than eithersmaller exporters or nonexporters. This preference inten-sity will also manifest itself in patterns of information-seeking and lobbying on trade policy, where we expectthe largest firms to be most active.

The reach of our theory is limited to industrieswith some measure of product differentiation, and we wouldexpect that where products are entirely homogeneous anindustry’s preferences over trade would look like the theRicardo-Viner predictions, that is, driven by industry-levelfactors rather than firms’ ability to export.

However, among all firms that exported, thelargest 1% (about 1,700 firms total) accounted for 80.9% of allUS exports (Bernard et al. 2009).

Less productive firmsgenerally must charge higher prices in order to recoup theirhigher costs, and they therefore are less likely to find exportmarkets for their relatively expensive products once tradecosts are included (

We then turn to testing our predictions. We expect to seefirm-level characteristics—export status, and especially export-intensity—to be positively correlated with support for tradeliberalization.

The remainder of firms—who are notwinners from liberalization—are, in contrast, small in size andmany in number, and they face costs diffused over a broadrange of firms.

These firms are the core constituents for trade liberalization,and their enormous share of export sales give them the in-centive, intensity, and resources to supercharge their pro-trade political activity.

we find that firmfeatures are generally more important predictors of attitudes toward trade liberalization than industry-wide compar-ative advantage.

s shown in column (3) of Table 7, pridein a country’s democracy is positively correlated withpro-trade attitudes.

Whatever thereason, the bottom line that emerges from this regression is that the main cleavage inpreference formation over trade lies not between the two tradable sectors but betweentradables and non-tradables.

he poorest countries in the sample – Bangladesh, Nigeria, Armenia,Georgia – are those where more skill is associated with less pro-trade views. None ofthe richer countries exhibits this reversal.

The result is striking: there is a very strong and tightrelationship between a country’s per-capita GDP and the magnitude of the correspond-ing estimated marginal eDect of education (the coePcient of per capita GDP is 1.53percentage points per US$10,000, robust t-statistic = 4:97, signi/cant at 1% level). Thericher a country, the more positive is the impact of a marginal increase of educationon the probability of pro-trade attitudes.

ccording to the Stolper–Samuelsontheorem, skilled workers in all sectors of the economy will gain and unskilled workerswill lose.

We shall keep age, gender, citizenship andeducation as controls in all speci/cations

Older individuals appear tobe more protectionist, but the estimates are not always signi/cant. 17 Citizenship in thecountry is negatively associated withpro-trade sentiments, while political aPliationwith the right has a positive and signi/cant impact on pro-trade attitudes

n other words, communitarian-patrioticvalues tend to foster protectionist attitudes, but this tendency is moderated when thebroader institutions of society are perceived to be working well

For example, individuals who consider themselves part of the upperclasses tend to be more favorable to trade than those who consider themselves to befrom the lower classes.

We /nd support for the speci/c-factors model as well. In particular, individuals em-ployed in import-competing industries are more likely to be in favor of trade restrictions(compared to individuals in non-traded sectors).

High degrees of neighborhood attachment and nationalism/patriotismare associated withprotectionist tendencies.

relative economic status

e /nd that pro-trade prefer-ences are signi/cantly and robustly correlated with an individual’s level of human capital, in themanner predicted by the factor endowments model. Preferences over trade are also correlated withthe trade exposure of the sector in which an individual is employed: individuals in non-tradedsectors tend to be the most pro-trade, while individuals in sectors with a revealed comparativedisadvantage are the most protectionist

The other message of our empirical analysis is that political-economy modelersmight consider asset ownership.

In all eight models reported in Table 6, our estimates of the effect of countytrade exposure on individual homeowners’ trade preferences are consistent withour theoretical expectations. We estimate that, for homeowners, an increase incounty trade exposure from its sample mean to one standard deviation above itsmean increases the probability of supporting trade restrictions by between 0.029and 0.039.

For example, Model 1 indicatesthat increasing Occupation Wage from its mean to one standard deviation above itsmean reduces the probability of supporting trade restrictions by 0.074 on average

The results of our logistic regressions for Models 1 through 8 strongly supportthe hypothesis that individuals’ skill levels determine trade-policy preferences.Little evidence is found consistent with the hypothesis that industry of employ-ment influences policy preferences.

To summarize: in regions with a high concentration of activity in comparative-disadvantage sectors, homeowners should oppose freer trade because its contrac-tionary effects in the region tend to reduce homeowners’ welfare by lowering6housing demand and thus housing values

Trade policy is one of the forces affecting the level of regional economicactivity. Freer trade tends to shrink comparative-disadvantage industries andexpand comparative-advantage ones

To summarize, in HO models factors evaluate trade policy based on their factortype while in RV models factors evaluate trade policy based on their industry ofemployment. What do these two models predict about trade-policy preferences inthe United States? Many studies (e.g., Leamer, 1984) have documented that theUnited States is well endowed with more-skilled labor relative to the rest of theworld. According to the HO model, then, in the United States more-skilledworkers should support freer trade while less-skilled workers should oppose it. Incontrast, the RV model predicts that U.S. workers employed in comparative-advantage sectors should support freer trade while those in comparative-dis-advantage sectors should oppose it.

Second, our units of observation are individuals, not more-aggregatedgeographic units such as counties or Congressional districts.

Analyzing 1991 Congressional votes on granting presidential‘‘fast track’’ trade-negotiating authority, Kaempfer and Marks (1993) find Housevotes significantly correlated with the average wage in House districts (but not sofor Senate votes and average state wages).

Using an individual-level survey of Canadianvoters, Beaulieu (1996) finds that factor type rather than industry of employmentbest explains votes in the 1988 Canadian federal election, an election widelyregarded as a national referendum on the Canadian–U.S. Free-Trade Agreement(CAFTA).

Assets like housing do not fit into standard trade theory because theyare neither currently employed factors nor currently produced goods. We find thatindependent of factor type or industry of employment, home ownership in countieswith a manufacturing mix concentrated in comparative-disadvantage

First, we find that factor type dominatesindustry of employment in explaining support for trade barriers. Lower skill,measured by education or average occupation earnings, is strongly correlated withsupport for new trade barriers.

Ricardo–Viner (RV)framework where some or all factors cannot move to other sectors

In contrast, in a Heckscher–Ohlin (HO)framework where factors move costlessly across sectors, factor incomes tend tovary by factor type.

Second, we find that home ownership also matters forindividuals’ trade-policy preferences.

First, we find that factor type dominates industry of employment inexplaining support for trade barriers.

nternational institutions can performinformation-provision functions that allow states to overcome collective-action problemsand therefore have an impact on patterns of behavior, even if these institutions do notregulate, enforce, or otherwise take on the characteristics associated with “strong”institutions on the domestic level.

s Keohane’s title suggests, much of the early workon institutions developed in reaction to the predictions of early hegemonic stability theorythat international economic cooperation should plummet as U.S. power declined in the1970s. However, in many areas of the international economy cooperation appeared to bestable or to increase. At the same time, the international economy had become morehighly institutionalized, with organizations such as the GATT and IMF gainingprominence.

A second type of uncertainty in IPE is about the relationship between policies andoutcomes, or causal relationships.17 Some refer to this type of uncertainty as lack ofprecise knowledge about “the state of the wor

n most settings, signals sent by states are most effective inchanging the beliefs of others if these signals are costly. Otherwise, intended recipientsare likely to dismiss signals as cheap talk. Analyses of some international monetaryregimes sees them in this light, as commitment to a fixed exchange rate or a currencyunion involves a political cost that sends a signal about the intentions of governments(Giavazzi and Pagano 1988). However, in some circumstances even cheap talk can be aneffective signa

To simplify somewhat, two types ofuncertainty are important for IPE: uncertainty about the preferences of others, and aboutcausal relationships

Governments of large countries aremore likely to be protectionist than governments of small countries, for example; thelatter are likely to value exchange-rate stability more highly than the latter

Treating states as units – although not identical units– scholars ask how the constraints and opportunities offered by the international system,and processes of interaction with other states, influence decisions and outcomes

strategic interaction at theinternational level.

The recent upsurge of interest in standards that relate to trade – environmental,labor, health and safety – has also led to the creation of delegated institutionalmechanisms to deal with these issues, such as the independent agencies established underthe auspices of NAFTA. In the European Union, trade policymaking is delegated to theEuropean Commission and, within this, to industry-specific groups that includerepresentatives from many member states. Trade policy is not unique.

A constitutional structure, for example,that gives trade policy authority to the more localist legislature will produce differentoutcomes from one that delegates all trade policymaking to the executive

Institutions perform two general functions: aggregation and delegation. Someinstitutions aggregate interests in ways that affect the ability of groups to organize and theweight they will have in the political process. Other institutions delegate decision-making authority to particular actors, similarly changing the weight of various interests in

Variation in the organization of economic interests has important implications forthe making of policy. An obvious example is the difference between labor ormanagement interests that are organized on sectoral as opposed to class lines. Wheresectoral forms of association prevail, there is likely to be much more pressure forparticularistic policies: trade protection, export subsidies, controls on or subsidies toforeign investment. Where class politics prevails, policies are less likely to focus onindustry-specific benefits and more likely to involve concerns of broader interest to laboror capital as a whole.

Public opinion may be hostile to devaluation, but even more hostile to the expenditurereductions and tax increases necessary to defend the exchange rate

Another possibility is to focus not onsociety as a whole, but on the median voter, and attempt to establish the expected policypreferences of such pivotal actors.4 This is standard in studies of income redistribution,where it is common to assume that the net expected benefits of redistributive policies onthe median voter will determine their electoral viability.5 The generally applicable pointis that if the median voter – more generally, the politically pivotal portion of the public –can, roughly speaking, be identified, it may also be possible to identify the policypreferences of this segment of the population

More recent developments look at scale economies, typicallyto argue that larger firms should prefer more open trade

In this domestic analysis, we identify three analytical steps. First, scholarsspecify the economic interests at stake. Second, they characterize the organization ofthese interests. Third, they investigate how these interests are mediated through politicalinstitutions.

that pairs of democracies have moresuccess in lowering barriers to trade than pairs that match a democracy with an autocracy.Statistical work bears out this proposition.

Here, two factors get the most attention: theinterests of the legislature, particularly how its ideal point differs from that of thepresident; and the degree of uncertainty associated with international negotiations.

Jack Snyder called “synergy”

ynergy represents new possibilities that open up as the result of interactionbetween the domestic and international levels, and might occur when negotiators are ableto use international negotiations to create previously unattainable outcomes, thusenlarging their domestic win-sets

For example, Rodrik (1997) hasargued that economic integration has reduced the ability of governments to tax capital,thus limiting the scope for government policies to deal with the social dislocations thatglobalization itself creates. Others believe that these effects are less limiting, stillallowing for different national economic-policy paths (Garrett 1998). All thesearguments share a common causal approach connecting the international and domesticlevels: globalization, understood as developments in international economic integration,alter the choices available to national governments, which in turn affects national policy(and, one could continue, international outcomes).

to sustain controls on cross-borderinvestment.

In addition to affecting domestic interests, the international economy might alsoaffect domestic institutions, for example by making a previously feasible policy difficultto sustain.

nterests are aggregated through institutions, which also delegateresponsibility for decisionmaking to particular actors

Like all of IPE, the interaction of domestic and international conditions can beanalyzed in terms of three factors: interests, institutions, and information

The core of the domestic-international connection is the impact of domesticinstitutions and interests on international interaction, and vice versa. Our ultimate goal isa simultaneous understanding of this mutual causation, recognizing feedback effects atboth levels -- a general equilibrium model, rather than a partial equilibrium one in whichone level is held fixed while the other varies

foreign direct investmen

The research frontier for IPE is thus defined in termsof empirical puzzles as well as a theoretical agenda, although we do not cover more thana sample of the empirical work that has been done over past decades. One relativelywell-developed empirical subject area, from which we draw many of our illustrativeexamples, is international trade.

Scholars typically see alternative approaches ascomplementary or applying under different, specifiable, conditions. Other disagreementsmight be debates about the relative weights that we should attribute to alternativeexplanatory variables. These types of scholarly differences are the hallmark of a maturefield of research that has moved from sweeping attempts at self-definition to formulatingrefutable analytical claims and evaluating their fit with empirical regularities.

he biggest challengefacing the field is understanding the simultaneous interaction of domestic andinternational factors in determining foreign economic policies and international economicoutcomes

A study of the height ofMexican children whose mothers worked in ma-quiladoras (export factories) in Mexico dramaticallyillustrates the power of a good job. 11 Maquiladorasgenerally have the reputation of being exploitativeand paying poor wages.

This is the paradox of the poor and their busi-nesses: They are energetic and resourceful andmanage to make a lot out of very little.

During the two hours we spent with this house-hold, we saw two customers. One bought a single ci-garette, the other a few sticks of incense. Clearly,the marginal return of increasing the size of the in-ventory a little was potentially extremely high, espe-cially if the family could try to buy something thatthe other shops in the same village did not supply.396/591

The overall return on a business, on the otherhand, is the total revenue net of operating expenses(the costs of materials, any wages you pay to yourworkers, and so on).

Compared to other ex-tremely poor households that were not selected toparticipate, the beneficiaries have more animals andother business assets; they earn more from livestockand other animals, but they also work longer hoursand earn more from working for others

In particular, we went backto the story of Ben Sedan, who was convinced thathe had no use for more money. Fouad drew up aclearly feasible business plan for him. He could takea loan, build a stable, and buy four young cows.They would not need to graze in a field: They couldbe fed in the stable. Within eight months, he couldsell the cows for a hefty profit. Fouad was persuadedthat if someone explained this to him, Ben Sedanwould see the wisdom of this plan and take out aloan

Muhammad Yunus, founder of the world-famousGrameen Bank, often describes the poor as naturalentrepreneurs. Combined with the late businessguru C. K. Prahalad’s exhortation to businessmen tofocus more on what he called the “bottom of thepyramid,” 1 the idea of the entrepreneurial poor ishelping to secure a space within the overall anti-poverty policy discourse where big business andhigh finance feel comfortable getting involved.

dawn the women went to the beach, where they col-lected wet sea sand. They then laid it evenly on thestreet before the real traffic began. When the carsstarted driving over the sand, the heat from theirwheels dried it. All they had to do then was occa-sionally to scrape off the top layer of sand, now dry.By nine or ten, they had a quantity of dry sand,which they brought back to the slum to sell in smallpackets made from discarded newspapers: The localwomen used the dry sand to scrub their dishes.This, the uncle reckoned, was true entrepreneur-ship: If you have very little, use your ingenuity tocreate something out of nothing

The microfinance movement has demonstratedthat, despite the difficulties, it is possible to lend tothe poor. Although one may debate the extent towhich MFI loans transform the lives of the poor, thesimple fact that MFI lending has reached its currentscale is a remarkable achievement. There are veryfew other programs targeted at the poor that havemanaged to reach so many people. However, thestructure of the program, which is the source of itssuccess in lending to the poor, is such that we can-not count on it to be a stepping-stone for largerbusinesses to be created and financed.

Miao Lei, with a good degree and a reasonablebusiness model, had to gamble to survive. NarayanMurthy and Nandan Nilekani, despite their degreesfrom the ultra-prestigious Indian Institute of Tech-nology, could not get a loan to start the firm Infosysbecause the banker objected that the bank could seeno inventory to lend against. Infosys today is one ofthe largest software firms in the world. It is hard notto assume that there are a lot more people like thesethree, but who just couldn’t make it because theydidn’t get the right financing at the right time

Microfinance gives itsclients every incentive to play it safe, so it is not wellsuited to discover who has an appetite for risk tak-ing.

The necessary focus on repayment discipline im-plies that microfinance is not the natural or bestway to finance entrepreneurs who want to gobeyond micro-enterprises.

This time, Spandana and Share, one of its competit-ors, were accused of being the reason a number offarmers had committed suicide. According to a newseries of articles in the press, loan officers hadpushed the clients to overborrow, then put unfairpressure on them to repay. The MFIs obviouslydenied the charges, but before anything could be re-solved, the district commissioner of Krishna (theadministrative head of the district) decreed that re-paying one’s loan to Spandana or Share was . . . il-legal.

that microfinance is auseful financial product. But that apparently is notenough. Trapped by decades of overpromising,many of the leading players in the microfinanceworld have apparently decided they would ratherrely on the power of denial than take stock, regroup,and admit that microfinance is only one of the pos-sible arrows in the fight against poverty

In our minds, mi-crocredit has earned its rightful place as one of thekey instruments in the fight against poverty

People in the Spandana neighborhoods were morelikely to have started a business and more likely tohave purchased large durable goods, such as bi-cycles, refrigerators, or televisions. Households thatdid not start a new business were consuming morein these neighborhoods, but those who had started anew business were actually consuming less, tighten-ing their belts to make the most of the new oppor-tunity.

All of these innovations contribute to reducingthe administrative costs of lending, which, as we ar-gued above, get blown up by the multiplier effectand make lending to the poor so very expensive.This is how most MFIs in South Asia manage tomake money by lending to the poor at interest ratesof around 25 percent per year,

From its modest beginnings with the BangladeshRehabilitation Assistance Committee (universallyknown as BRAC) and the Grameen Bank in themid-1970s in Bangladesh, microcredit is now aglobal phenomenon.

he Indian branch of Citibank gotinto serious trouble when it was discovered that itwas using “goondas” (local hooligans) to threatenborrowers who did not repay vehicle loans. And thecourts are not really an option, either. In 1988, theLaw Commission of India reported that 40 percentof the cases for asset liquidation (of bankrupt bor-rowers) were more than eight years pending

Kabuliwala kills someone who was trying tocheat him. The same logic also explains why theMob in the United States was for many people the“lender of last resort.”

Thereport describes debt collectors in India making useof the old social prejudice against eunuchs to collectfrom long-standing defaulters. Because people be-lieve that seeing a eunuch’s genitals brings bad luck,the eunuchs were instructed to show up at the de-faulter’s house and threaten a “showing” if they con-tinued to be uncooperative

To make matters worse, this creates what eco-nomists call a multiplier effect. When the interestrate goes up, the borrower has more reason to try tofind a way not to repay the loan. That means theborrower needs to be monitored and screened morecarefully, which adds to the cost of lending. Thispushes the interest rate up even further, which ne-cessitates more scrutiny, and so on.

The way lenders all over the world protect them-selves against the different forms of willful default isby asking for a down payment, some collateral, orwhat is sometimes called the promoter’s contribu-tion, which is the part of the firm’s capital thatcomes from the entrepreneur’s pocket. If the bor-rower defaults on the loan, the lender can punishher by seizing the collateral

But second, the interest rates charged byinformal sources tend to be higher for the poor thanfor the less poor. The average interest rate from aninformal source drops by 0.4 percent per month foreach additional hectare of land owned by the persontaking out the loan

and only 6.4 percent from a formalsource.

According to Padmaja, the ragpicker ex-plained that the bank would not lend to someonelike her. She could have gotten a loan from amoneylender, but the rates would have been so highthat it would not have been worth it. In the end,Padmaja decided to give her a loan. The ragpickerreimbursed it faithfully and flourished. Soon after,people were lining up at Padmaja’s doorstep forloans, and she decided to quit her job to startSpandana. Thirteen years later, in July 2010,Spandana had 4.2 million loan clients, with an out-standing portfolio of 42 billion rupees.

Padmaja Reddy

This in-terest payment is 4.69 percent per day.1 To seewhat this means, try the following calculation: Ifyou borrowed 100 rupees ($5.10 USD PPP) todayand kept it until tomorrow, you would need to repay104.69 rupees. If you kept this amount a furthertwenty-four hours and repaid it the following day,you would need to repay 109.6 rupees. After thirtydays, you would owe almost 400 rupees, and after ayear, 1,842,459,409 rupees ($93.5 million USDPPP). So the equivalent of a $5 loan, if it goes unre-paid for a year, leaves a debt of nearly $100 million

Children who are completely lost inthe regular school system (their test scores are waybehind those of other children when they enter

This message of scaling down expectations, focus-ing on the core competencies, and using technologyto complement, or if necessary substitute for, teach-ers, does not sit well with some education experts

Despite the fact that they only got to play fortwo hours a week, the gains from this program interms of math scores were as large as those of someof the most successful education interventions thathave been tried in various contexts over the years,and this was true across the board—the strongestchildren did better, and so did the weakest children.This highlights what is particularly good about thecomputer as a learning tool: Each child is able to sethis or her own pace through the program. 4

A pro-gram in Kenya that offered a $20 USD PPP scholar-ship for the next year to girls who scored in the top15 percent on an exam not only got the girls to domuch better, but it also put pressure on the teachersto work harder (to help the girls), which meant thatboys did better, too, even though there was noscholarship for them. 42 I

allow children to learn at their own pace, and inparticular to make sure the children who are laggingbehind can focus on the basics.

Infosys, one of India’s IT giants,has set up testing centers where people, includingthose without much formal qualification, can walkin and take a test that focuses on intelligence andanalytical skills rather than textbook learning.Those who do well get to become trainees, and suc-cessful trainees get a job.

It should now be clear why private schools do notdo better at educating the average child—their en-tire point is to prepare the best-performing childrenfor some difficult public exam that is the stepping-stone toward greater things, which requires power-ing ahead and covering a broad syllabus.

World Ban

But this behavior creates a poverty trap even wherenone exists in the first place. If they give up, theywill never find out that perhaps the child could havemade it. And in contrast, families that assume thattheir children can make it, or families that don’twant to accept that a child of theirs will remain un-educated, which tend to be, for obvious historicalreasons, more elite families, end up confirmed intheir “high” hopes.

he effects ofthe treatment on schooling are large and striking; there are fewexamples of policies or interventions that result in a 0.20- to 0.35-year increase in schooling, much less interventions that are aspotentially inexpensive as this one.

This finding suggests a possible inefficiency andmay even a reflect a potential development trap, as the relativeskill composition demanded by the labor market is not transmit-ted to youths in the form of greater perceived returns, resultingin an undersupply of skilled labor, which in turn inhibits the de-velopment of domestic skill–intensive industries or the ability toattract foreign direct investment.

Overall, the effects for the least poor students are large andstriking. The magnitudes compare favorably with large-scale pro-grams implemented elsewhere, such as Mexico’s PROGRESA,which provided direct cash incentives to increase school atten-dance. 29 And many of these other programs are extremely expen-sive, 30 whereas in the present case, information could potentiallybe provided at low cost.

For the poorest households, the effect of the treatment is ex-tremely small and not statistically significant for all three mea-sures of schooling. This is despite the fact that in column (8),the treatment appears to have had a large effect on perceivedreturns to schooling for these students. By contrast, for youthsfrom wealthier (though still quite poor) households, the effectsare large, and statistically significant at the 10% level or bet-ter for all three education measures (though the effect for finishedsecondary is not statistically significant without the additional co-variates (Table A.6 in the Online Appendix), and only marginallysignificant with them).

As noted above, within the standard human capital frame-work, demand is not always sufficient for schooling. For someyouths, even if they wanted to attend school, a combination ofcosts, low family income, and credit constraints will limit theeffectiveness of the intervention.

The table showsthat the treatment group was four percentage points (7%) morelikely to have returned to school the following year, though thedifference is only marginally statistically significant (p-value of.091). They also achieved on average 0.18 more years of schoolingover the next four years. Finally, the difference in the likelihood ofcompleting secondary school is positive, but small (two percentagepoints) and not statistically significant.

Overall, these results are further confirmation thatsurvey measures of perceptions are useful predictors of schoolingoutcomes, supporting the conclusions of Kaufmann (2008) andAttanassio and Kaufmann (2008).

The fact that students have such low perceived returns toschooling raises the possibility that providing information on thehigher measured returns may improve schooling.

The means here are lower than own expected earnings forboth levels of education, consistent with a general optimism bias

Of course, these regressions may be plagued by omitted vari-ables bias (e.g., those with low perceived returns may attendlower-quality schools) or reverse causality (e.g., a “sour grapes”effect whereby those who want to go to school but are constrainedfrom doing so by poor grades or low income report low returns). Al-though we would not want to attach a strong causal interpretationto these results, they provide an initial impression that measuredperceptions do have some predictive value (though it should benoted that the perception measure alone can account for only1%–2% of the total variation in the various schooling outcomes

The average householdincome is approximately RD$3,500 per month, and 38% of youthshave fathers who finished high school. At baseline, students ex-pect earnings at age 30–40 of RD$3,516 if they only finish primaryschool and RD$3,845 if they finish secondary school, both of whichare slightly greater on average than what they believe currentworkers aged 30–40 with those levels of education earn

ssign-ment of the treatment was done at the school level rather thanfor individual students within schools because students in thesame school are likely to communicate, which would contaminatethe control group.

The two sets of questions can therefore be used to deter-mine in part whether students’ perceptions differ from measuredreturns because (a) they have poor or inaccurate information onprevailing wages in the labor market (as captured by the secondset of questions), or (b) they have information or beliefs aboutthemselves (correct or incorrect) that influence what they expectearnings will be for them personally (as captured by the first setof questions).

Overall, we were able to obtain follow-up informa-tion in the October 2001 survey directly from 93% of youths, with2% from relatives and 5% from neighbors. By the 2005 survey,this had changed to 89% from youths, 4% from relatives, and7% from neighbors. In all cases, we attempted to verify educa-tional attainment by contacting the school that students were re-ported to be attending or had attended.

Using data from a panel survey of boys in the Dominican Re-public in the eighth grade, the last year of compulsory schoolingand the point at which most students terminate their education,we find that perceptions of the returns to secondary schooling areextremely low for most students, especially relative to returnsmeasured with earnings data. Although many factors may affector limit school attendance, such as poverty and credit constraints,these results raise the possibility that for at least some youths,school dropout may be the result of low demand due to low per-ceived returns

only about 25%–30% completesecondary school (Oficina Nacional de Estad ́ıstica, Rep ́ublica Do-minicana 2002).

Manski(1993), it is the returns perceived by students and/or their parentsthat will influence actual schooling decisions.

Though many studies estimate these returns with earningsdata, it is the perceived returns that affect schooling decisions, and these percep-tions may be inaccurate.