The Intergovernmental Network of World Trade: IGO Connectedness, Governance, and Embeddedness1
Paul Ingram Columbia University Jeffrey Robinson New York University Marc L. Busch Georgetown University
Membership in certain intergovernmental organizations (IGOs), such as the World Trade Organization, has long been argued to stimulate trade. Yet, evidence linking IGOs to trade is mixed. The authors argue that identifying the inuence of IGOs requires attention not only to the institutions IGOs enact, but also to the network through which they enact them. This approach allows them to demonstrate that trade between two countries increases by an average of 58% with every doubling of the strength of IGO connection between the countries. They also contribute to debates regarding the mechanisms through which structural relationships inuence economic behavior by showing that substantial trade benets occur not only through economic IGOs, but also through IGOs that were formed for social and cultural purposes, and that connections through IGOs that are organizationally strong have more impact than those through minimalist IGOs. Institutions are the bedrock of commercial exchange. Dened as formal or informal sets of rules, norms, and decision-making principles, institutions help lower the uncertainty and risk inherent in transactions among
1 We are grateful to Pierre Azoulay, Charles Boehmer, J. B. Bonardi, John Freeman, Erik Gartzke, Wit Henisz, Guy Holburn, Ray Horton, Matt Kraatz, Ray Reagans, Ezra Zuckerman, and participants of seminars at Cornell University, Harvard University, New York University, the University of Chicago, the University of Pennsylvania, the Technion, and the University of Western Ontario, for comments on earlier
2005 by The University of Chicago. All rights reserved. 0002-9602/2005/11103-0005$10.00
AJS Volume 111 Number 3 (November 2005): 82458
IGOs and World Trade traders. In this light, institutions are widely viewed as a pillar of economic growth, bolstering incentives for commerce. Recently, attention has turned from whether institutions matter to questions about how they matter. In this article, we consider the link between the social structures in which institutions are embedded and their efcacy. Institutions are associated with social units (groups, networks, organizations, nations) that determine which actors are subject to the institutions, with which other actors they may more effectively trade, and what happens when they violate the institution. We examine the link between institution and social structure in the context of intergovernmental organizations (IGOs), a prominent institutional form aimed at promoting international trade and smoothing international interactions more generally. The signicance of social structure is readily apparent in international trade, where national and subnational borders can often act as substantial barriers, even when the social units they divide have comparable institutions (Frankel 2000). One of the most striking illustrations of the connection between social structure and institutional governance is the European Community (EC, now the European Union or EU), which is associated with an increase in intra-EC trade. The success of the EC is not so much a story of institutional innovation, at least with respect to the institutions that govern trade per se, but rather of the creation of an integrated, transnational society, which has helped to expand the set of actors that may effectively interact under the institutional umbrella (Fligstein and Stone Sweet 2002). Recent sociological analyses of the EC notwithstanding, most studies underemphasize the link between institutions and the social structures that host them. This is particularly clear in the literature on international institutions and trade, which has struggled to show a connection between IGOs and global commerce. IGOs are organizations that meet regularly, are formed by treaty, and have three or more states as members (Pevehouse et al. 2003). Prominent examples include the World Trade Organization (WTO) and the United Nations (UN). More representative of the more than 300 current IGOs, however, are organizations like the Andean Development Corporation or the Universal Postal Union. Researchers have sought for decades to identify the economic impact of these increasingly pervasive organizations, but have produced little evidence of any positive effect (e.g., Jacobson, Reisinger, and Mathers 1986; Rose 2004). Consequently, IGOs have been attributed only a marginal role in increasing

drafts. We are also grateful to Gueorgi Kossinets and Gokce Sargut for able research assistance, to Jon Pevehouse for help accessing the revised IGO data, and to the Chazen Institute for a summer research grant in support of this project. Direct correspondence to Paul Ingram, Graduate School of Business, Columbia University, 712 Uris Hall, New York, New York 10027-6902. E-mail: pi17@columbia.edu
American Journal of Sociology trade (Milner 1999). We contend that this mixed record is the result of a failure to account fully for the social structural implications of IGOs. More specically, IGOs create an intercountry network in which a large and interrelated set of trade-related institutions is embedded. Accounting for this broader network enables us to paint a very different picture of the inuence of IGOs on trade. The recognition that IGOs forge connections between countries makes relevant a large sociological literature that links interactor connections to exchange. This literature has shown that a range of formal and informal connections between actors smoothes exchange between them (Granovetter 1985; Uzzi 1996; DiMaggio and Louch 1998), and that the pattern of connections is a key determinant of competition (Burt 1992). We apply these ideas to help understand the inuence of connections through IGOs for bilateral trade and nd support for both the idea that more connections increase trade, and that the broader network affects competition, such that trade between two countries is less if they have similar patterns of connections to others. Our context also allows us to make a fresh contribution to the network approach to economic sociology by taking up two criticisms of that literature. The rst criticism is that network theorists reify social structure and underattend to issues regarding the origin and change of networks (Fligstein and Stone Sweet 2002). In our context it is clear that IGO connections are forged through IGOs themselves, and that these organizations form a link between trade-related institutions, on the one hand, and a broader network of countries, on the other. Furthermore, by tracking IGOs over an extended period (18851992), we produce dynamic measures of the network these IGOs help create. With sufcient dynamism in our network variables, we are able to overcome a related limitation of static analyses of network inuence: namely, that they may confound network measures with persistent attributes of the nodes or the dyads. We achieve this by including a variety of control variables and dyad xed effects, which permit us to show that IGO connections inuence trade independent of other notable factors, including physical distance, population and economy size, regime type, ethnic and colonial ties, and shared language. Our specications also include year xed effects, which control for global trends and events that may affect trade and/or international relations. The second criticism is that network theorists have underemphasized the institutional content of connections, treating social structure as an end in itself and failing to account for the fact that similar social structures can house various and sometimes opposing institutions (Salancik 1995; Nee and Ingram 1998). Essentially, this is the opposite of the criticism of the institutions literature that motivated us to consider social structure in 826

IGOs and World Trade the rst place. In our reading, it is not that network theorists ignore institutions, but rather that they are catholic as to the mechanisms through which connections may inuence exchange. This approach is empirically justiable, as most connections contain a diverse set of inuences on exchange. However, it would be theoretically useful to have more evidence that identies specic mechanisms through which connections affect exchange, especially in light of arguments by economists that the inuence of connections can be accounted for by nonsocial mechanisms (Gibbons 1999). We are able to produce such evidence by dividing IGO connections into those that arise through IGOs formed for economic purposes (EIGOs) and those formed for social and cultural purposes (SCIGOs). Consistent with a core principle of economic sociology, we nd that SCIGO connections bring substantial increases in bilateral trade. Furthermore, we nd that the magnitude of the effect of both EIGO and SCIGO connections depends on the organizational capacity of the IGOs that create them. This result sheds new light on the mechanisms of network inuence by linking the benets of association to more formal structures.
IGOS AND THE GOVERNANCE OF TRADE
The new institutional analysis of exchange relies on transaction costs, which arise because of the risk of malfeasance and uncertainty inherent in trading (Williamson 1975). In almost every exchange, there is a moment where one of the parties has control over all or most of the goods and must decide whether to follow through on the agreed-upon bargain, or make a grab for more. This problem is obvious in the simplest of exchanges, as where children swap toys on the playground. The risk of malfeasance increases substantially when the exchange is more complex, as in global commerce, where differences in law, physical distance, and language have all been found to impede trade (Frankel 2000). The second source of transaction costs, uncertainty, may be a more important inhibitor of international trade. The risk of malfeasance aside, exchange is fraught with difculties in recognizing opportunities for exchange, nding partners, measuring quantity and quality, and equating the value of goods that may be imperfectly divisible. Indeed, these factors are likely behind the so-called border effects (Helliwell 1998) that riddle trade, whereby commerce tends to ow more between subnational units of a country (i.e., provinces or states) than across national borders, controlling for physical distance and economy size. Here, political-cultural differences make communication and understanding more difcult, the upshot being that many opportunities for international exchange are doubtless missed (Evans 2003). 827
American Journal of Sociology Institutions are widely thought to moderate these transaction costs. Laws that enforce contracts at the domestic level enable exchange partners to credibly commit to future actions and reduce the risk of malfeasance (North 1990). When legal sanctions are ineffective or inaccessible, reputation and normative sanctions can create similar benets (Macaulay 1963; Greif 1994). An example of an IGO that promotes rules of fair exchange is the WTO, which sets out rights and obligations for trade based on the principles of nondiscrimination and reciprocity, and provides for a dispute resolution mechanism to adjudicate these rights and obligations (Busch and Reinhardt 2002). Other IGOs reduce uncertainty by promoting efforts at harmonization, like the International Organization for Standardization. Still others focus on specic issue areas, such as the World Intellectual Property Organisation, or on specic sectors, like the International Coffee Organization. For all the theoretical interest in IGOs, the fact remains that empirical studies have turned up results that are far from impressive. In the case of international trade, in particular, one could be forgiven for questioning all the attention to IGOs. The reason for this skepticism is that, despite persistent research efforts, there is little hard evidence that IGOs promote trade. Why this disconnect between theory and evidence? We argue that the literature has not given sufcient attention to the role of IGOs in affecting connections between their members. The very earliest efforts to identify the inuence of IGOs ignored membershipor at least the idea that membership brought specic countries under the umbrellas of IGOsand simply correlated counts of IGOs with international outcomes (e.g., Singer and Wallace 1970). Later efforts have partly overcome this problem by correlating outcomes for a specic country (i.e., levels of trade, GDP growth, participation in war, etc.) with the number of IGOs to which it belongs. Representative of this approach, Jacobson et al. (1986) nd mixed results regarding the link between IGO memberships and trade, noting that IGO memberships seem to matter only for developing countries and only in certain periods. While the count of memberships recognizes that countries must typically be part of IGOs to benet from their inuence, this approach misses the fact that IGO inuence often requires that both countries in a transaction be subject to the same IGOs. In other words, it is not just membership, but joint membership, that matters. This is most obvious with regard to the many IGOs that promote coordination. After all, what good is it to adopt a convention regarding measurement, data transmission, or accounting, when the parties one would like to transact with do not observe the same convention? The idea that IGO governance depends on connections created by joint membership has recently been applied in analyses of the likelihood of war and has reinvigorated that important 828

ORGANIZATIONAL CAPACITY OF IGOS
So far we have argued that IGOs inuence trade by forging a network of bilateral connections that hosts transaction-smoothing rules and affects 831
American Journal of Sociology intercountry sentiments. The next step in our theory development is to recognize that some IGOs have more impact than others beyond the structure they create and the institutions they overlay on that structure. This step is important substantively, because even casual observers of international organization realize that IGOs vary in their capacities to affect their members and achieve their goals, and that it would be a mistake to ignore the distinction between minimalist organizations (such as the International Wool Study Group) and more powerful ones (like the WTO). It also matters theoretically, as institutional arguments too often emphasize institutions of a given form (laws, organizational policies, social norms, etc.) while underattending to the interdependence between forms, which is so often fundamental to their impact (Nee and Ingram 1998). For example, the effectiveness of an organizational policy will depend on other institutions, such as the national law and culture within which the organization operates (e.g., Meyer and Rowan 1977; Dobbin and Sutton 1998) and the social norms held by the organizations participants (e.g., Heckathorn 1990). Our current claim is that the effectiveness of an IGO connection will depend on the organizational capacity of the IGO. The issue of whether an IGO has teeth is particularly salient, because the basis of IGO authority is voluntary association. IGOs bind their member countries through treaties, and if a country chooses to out a treaty, the only real sanctions are those imposed by other members, as there is no higher court (or third-party enforcer) that can compel members to follow through on their commitments. Just as the norms of a well-structured social group (e.g., Jewish diamond traders) can be expected to have more bite than those of a loosely structured group (e.g., passengers on a subway), the policies of IGOs with effective mechanisms of communication, coordination, dispute resolution, and enforcement should have more impact than those of minimalist IGOs. Gartzke (2002) demonstrates that the impact of IGOs depends on their organizational structures in an analysis of the determinants of war. As we do, he operationalizes connectedness between two countries as a function of the number of IGOs in which they share membership. He nds that connections through IGOs that were structured reduced the incidence of war, while connections through minimalist IGOs had no effect. According to Gartzke (p. 22), minimalist IGOs are without an extensive bureaucracy beyond research, planning, and information gathering, while structured IGOs contain structures of assembly, executive, and/or bureaucracy to implement policy, as well as formal procedures and rules. The emphasis here on a bureaucratic capacity to implement has a satisfying correspondence to arguments proffered by Weber (1946) and 832

2 Supplementary submission by the Central Districts Federated Clothing, Laundry and Allied Workers Union on the proposed Hong Kong free trade and investment agreement, at http://www.canterbury.cyberplace.org.nz.
IGOs and World Trade the EU, one fear being that Polands relatively close relations with nonmembers could well inspire a surge of transshipment into the lucrative European market. Straight transshipment may be the most obvious form of brokering in international trade, but a country spanning weakly connected others may also import raw materials or low-value inputs from one, transform them, and send value-added exports to the other.3 Alternatively, assembly industries may develop in a country because of preferential access to an export market. Volkswagen, for example, set up shop in Mexico to service the local market, but with trade liberalization sweeping that country in the lead-up to the North American Free Trade Agreement (NAFTA) and the completion of the Uruguay round of the GATT, Volkswagens Mexico facility emerged as a key exporter to the United States and Canada, its two NAFTA partners.4 The bases of all of these opportunities for ow-through trade are differences (inequivalences) in the strength of connectedness in the IGO network. If, for example, other countries established the same strong connection to China that Hong Kong hada development that would make them more structurally equivalent to Hong Kongthey could trade directly with China without relying on Hong Kong for transshipment. Trade between Hong Kong and its newly structurally equivalent alters would fall. Thus, we make the following prediction: Hypothesis 4.Trade between two countries will be negatively related to the structural equivalence between them in the IGO network.

ANALYSES OF TRADE

Model To test our hypotheses, we use the gravity model, which is the standard for analyses of bilateral trade. According to Rose (2004, p. 99), the gravity model is a completely conventional device used to estimate the effects of a variety of phenomena on international trade. The model uses a log3 We do not suggest that structural inequivalence in the IGO network is the only reason that countries occupy different positions in global production systems. Human capital, natural endowments, and industrial policy obviously affect which countries provide raw materials, labor, and technology in global production. Commodity-chain theory in sociology also highlights the importance of country power in determining the pattern of production (Geref and Korzeniewicz 1994). Our argument is consistent with these claims, and we would simply add to any of them that whatever determines which countries do what in globally distributed production, importing and exporting is required, and the IGO network smoothes those transactions. 4 See http://www.umich.edu/cibe/case_pdf/97-12.pdf and http://www.autonews.com/ news.cms?newsldp2709.

American Journal of Sociology log specication to explain trade between two countries as a function of their joint income, asserting trade ows will be proportional to the product of their GDPs.5 We use the bench-line specication of the gravity model described by Rose (2004, p. 100) with control variables suggested by Oneal and Russett (2001) and the variables we have created to represent the IGO network ln (Tradeijt ) p b0 b1 ln (GDPi GDPj )t b2 ln (GDPi GDPj /Popi Popj )t b3 ln (IGOCONijt ) b4 StrucEquivijt b 5 Democ ijt b6 Allyijt

fij Dij

a tYt ijt ,
where i and j are the countries in a dyad, t denotes time, and the variables are 1. Tradeijt is the real value of bilateral trade between i and j in year t. 2. GDP is real GDP. 3. Pop is population. 4. IGOCONijt, which tests hypothesis 1, is IGO connectedness, the number of IGOs that i and j are simultaneously members of in year t. IGOCON is replaced by subcomponents representing connectedness through economic and social/cultural IGOs to test hypothesis 2. Those variables are in turn replaced by their subcomponents representing connectedness through minimalist and structured IGOs to test hypothesis 3. 5. StrucEquivijt tests hypothesis 4 and is the Pearson product-moment correlation between the vectors of i and js IGO connections to other countries in year t. 6. Democijt is the minimum of the democracy/autocracy scores (taken from the Polity III Database) of i and j in year t. This control is included because democracies are expected both to trade more and join more IGOs with each other. 7. Allyijt is an indicator variable coded one if i and j have a military alliance in year t (taken from the Correlates of War Data Set). This control is included because allies are expected both to trade more and to join more IGOs with each other. 8. {D} is a set of dyad-level xed effects. 9. {Y} is a set of year xed effects.
5 The typical gravity model also includes the log of the distance between the two countries, which we cannot include because we use dyad xed effects. We show below that our results are robust in a random-effects model that includes the log of distance.
IGOs and World Trade The dyad and year xed effects are important in this specication (Green, Kim, and Yoon 2001; Rose 2004). First, they account for the nonindependence of observations in our data. Second, they effectively control for all stable dyadic and time-varying global inuences on trade. Examples of relevant dyad-level inuences include the distance between the two countries, whether they share language, a border, religion, or colonial history. The dyad xed effects take all of these stable inuences out of the mix. Similarly, the year xed effects account for historical inuences that affect all dyads, including discrete events of global import such as the Great Depression, the world wars, and the fall of state socialism, as well as trends such as the legitimacy of international relations or international trade. In other words, the xed effects control for all inuences on trade except those that vary both within a dyad and across time. With the xed effects in place, the coefcients indicate the expected change in bilateral trade of a one-unit change in an independent variable for a given dyad in a given year, not correlations between variables across dyads or time. This is the best way to test our hypotheses directly. Data and Variable Construction We take our trade, GDP, and population data directly from the data used in Oneal and Russett (2001). These data are particularly appealing for our purposes because they cover the period 1885 to 1992, whereas other data sets focus on the most recent 50 years, when trade data have been more readily accessible. The period before World War II represents substantial change in the network of IGO connections, so this longer time series is important for testing our hypotheses, though as we show below, our results are comparable when we restrict our analysis to the postwar period. Furthermore, given that some analyses of bilateral trade have shown sensitivity to certain key coding decisionsnotably what to do about zero trade values before taking their natural logarithmit is convenient to use Oneal and Russetts data, given that their codings are the result of a series of methodological debates (though here too, we show below that our results are robust to various popular treatments of the relevant variables). Oneal and Russett obtained trade data from (a) the IMF for the post World War II era, (b) the League of Nations for the interwar period, and (c) annual editions of The Statemans Yearbook (e.g., Epstein 1913) for preWorld War I. They also relied on other archival sources in their effort to compile the data and check its reliability and robustness (Russett and Oneal 2001, pp. 13940). They converted current values of trade and GDP to real U.S. dollars (1990 p 100), listed bilateral trade in millions of dollars and added $100,000 before taking the log (to allow the logs of 837

American Journal of Sociology dyads with zero trade), and listed real GDP in millions of dollars and population in thousands before logging. We take their data in these forms, so the basic data inputs to our gravity models, including the dependent variable, are the same as theirs. Table 1 lists the 135 countries in the data set and the time frame over which each country is observed. The basis of our operationalizations of the IGO network is the timevarying listing of IGOs and their members, from 1816 to 2001, provided by Pevehouse et al. (2003). According to their denition, an IGO must 1. include three or more members of the Correlates of Wardened state system, 2. hold regular plenary sessions at least once every 10 years, and 3. possess a permanent secretariat and corresponding headquarters. IGOs may be formed directly by the states themselves or may be emanations formed by another IGO. Pevehouse et al. list all of the IGOs formed directly by states, but exclude emanations. This treatment is appropriate for our purposes because emanations are not independent from their parent IGOs and do not therefore represent independent sources of IGO connections. Pevehouse et al. identify 497 IGOs that existed at some point in history. In the rst year of our analysis, 1885, there were 14 IGOs operating, 43 in 1914, 65 in 1938, and 314 in 1992, which is the last year of our analysis. We used the IGO member listing to create a time-varying afliation matrix of connectedness between two countries. The afliation matrix is produced by multiplying Xt, a country by IGO matrix, with the cells indicating whether a country is a member of a given IGO at time t, by its transpose: Ct p Xt 7 XtT. Thus, Ct is a symmetric country-by-country matrix where the cell cijt indicates the number of IGOs that country i and country j share joint membership in at time t. To test hypotheses 2 and 3 we followed the same procedure to create afliation matrices of connections through EIGOs and SCIGOs, and the minimalist and structured representatives of each of those types. This required that we code IGOs as to their function and structure. We followed available coding schemes to do so and describe the process in the appendix. We take the natural logarithms of the IGO connectedness variables, both for consistency with the treatment of other variables in the gravity model, and because exploratory analysis indicated that this functional form best represented the impact of IGO connectedness in our models (our results are qualitatively similar when the IGO connectedness variables are not logged). We added 0.1 to all measures of IGO connectedness before taking the natural logarithms. Structural equivalence, which tests hypothesis 4, is simply the Pearson product-moment correlation between the vectors that represent i and js 838

IGOs and World Trade IGO connections to all other countries (Wasserman and Faust 1994, p. 368). This measure captures the degree of similarity between two countries IGO connections to others, to get at our argument that similarities in relations to others represent competition, while dissimilarities represent brokering opportunities. Some readers have wondered why we use this continuous measure of structural equivalence rather than a measure based on pure structural holes (complete disconnects) in the IGO network. The reason is that brokering opportunities in the IGO network arise from differences in the strength of IGO connections, not from complete disconnects which were relatively rare, especially in later years. For example, currently, all countries have some connection to China in the IGO network (there are no complete disconnects), yet much trade to and from China ows through Hong Kong because its IGO connection to China is stronger than that of most other countries. Descriptive statistics for the variables are presented in table 2. Generally, the correlations between variables that appear in the same models (e.g., not comparing IGO connectedness to its economic and social/cultural components) are modest, although there are some correlations greater than.50 among the network variables. We therefore conducted a number of investigations to insure that our estimations were not compromised by multicollinearity. We estimated hierarchically nested regression models and used F-tests to indicate the joint signicance of more highly correlated variables (Kmenta 1971, p. 371). The nested models (presented below) showed that the coefcients of correlated variables were robust to various model specications, and the results of the F-tests were consistent with the tests of individual signicance, indicating that their standard errors were not inated. Additionally, we estimated our models on random subsamples of the data, obtaining results comparable in all ways to those we report below (Greene 1997). We also estimated models with the network variables entered singularly, and again, the results were consistent with those we report here. Thus, there is no evidence that multicollinearity compromised our estimations. Results Table 3 presents the results of xed-effects gravity models. Model 1 includes the control variables. Model 2 adds structural equivalence and is a signicant improvement over model 1 (F 1,143278 p 352.22 , P !.001). The coefcient on that variable indicates that there is less trade between two countries when they are more structurally equivalent, in support of hypothesis 4. Model 3 adds the aggregate IGO connectedness measure and improves on model 2 (F 1,143277 p 4231, P !.001). The positive coefcient for IGO connectedness is as predicted by hypothesis 1: as two countries 839

American Journal of Sociology become more connected to each other through joint membership in IGOs, the trade between them increases. When IGO connectedness between two countries doubles, the level of trade between them is expected to increase by 58% (20.656 1). Since IGO connectedness is based on the full set of IGOs, this result indicates the average impact of IGO connections on trade. Model 4 replaces the aggregate IGO connectedness measure with one that reects connections only through EIGOs and SCIGOs, excluding IGOs that had general or military/political functions.6 As expected, connectedness through EIGOs and SCIGOs (which make up more than 80% of all IGOs) brings a large increase in trade. Model 5 breaks out the separate effects of EIGO and SCIGO connections. A test of joint signicance indicates that the inclusion of these measures improves on model 2 (F 2,143276 p 1939 , P !.001; model 2 is the appropriate comparison because the logging of the IGO connectedness measures means that model 5 is not nested in models 3 or 4). Consistent with hypothesis 2, the positive coefcient of SCIGO connectedness indicates that afliations through these social and cultural organizations do increase bilateral trade. The nal model in table 3 breaks EIGO and SCIGO connectedness into that which comes from minimalist and structured IGOs. The four connectedness measures that result are jointly signicant (F 4,143274 p 486.5, P !.001). As hypothesis 3 predicted, connections through structured IGOs do more to increase trade than connections through minimalist IGOs. This is true for both EIGOs (F 1,143274 p 45.99, P !.001) and SCIGOs (F p 615.65, P !.001). 1,143274 The coefcients in model 6 suggest that doubling the level of connection through minimalist and structured EIGOs is associated with increases in trade of 7.8% and 12.2%, respectively. Corresponding gures for SCIGOs are 1.1% for minimalist and 19.6% for structured. While we predicted that SCIGOs would increase trade, we were surprised by the magnitude of the effect of connections through structured SCIGOs, which is even larger than that of structured EIGOs (F 1,143274 p 65.13, P !.001). The control variables in all of the models in table 3 behave as expected. Richer countries, as indicated by GDP and GDP per capita, trade more. Trade is also higher as a function of the minimum level of democracy in

IGOs and World Trade which trade in a dyad leads to specic changes in the IGO network. A single IGO connection between two countries emerges through membership in an IGO that includes at least one other country-member, and usually many more, and is therefore coupled to IGO connections to all of them. A given country cannot target an IGO connection to another country in response to trade in the dyad, because any attempt to do so would have repercussions throughout the network. IGO connectedness may be a signal of goodwill, not causally related to trade, but just something countries do to indicate that they are open to trade.This claim is inconsistent with the fact that structured IGOs matter more than minimalist IGOs. If IGO connections were merely a signal of goodwill or openness, then there is little reason why their impact should depend on the organizational capabilities of the IGOs. If endogeneity does exist, coefcient estimates for IGO connectedness will be inconsistent.The above arguments lead us to believe that the causal relationship we specify is most consistent with the results of the analysis. Nevertheless, it would be foolish to deny that IGO connectedness and trade may have some reciprocal relationship. If they do, the consistency of our coefcient estimate for IGO connectedness would be compromised. Instrumental-variable estimation is an increasingly popular method for adjusting for endogeneity (Greene 1997; for recent sociological applications see Ingram and Roberts 2000; Burris 2004). This technique involves creating proxies for the endogenous variable by using variables other than the dependent variable of the regression. In other words, we need a model of IGO connectedness that does not rely on past levels of trade. To build this model, we relied on the literature on the causes of war, with the logic that peace and IGO connections are two types of bilateral relations that may be explained by similar factors (Russett and Oneal 2001). Specically, we used lagged values of the following variables to predict the IGO connectedness between two countries:8 whether they share a border; the distance between them; whether they began a militarized dispute in either of the two previous years; the time since their last militarized dispute; whether either was a major power; whether they were military allies; the similarity of the countries levels of democracy; the
We applied the instrumental variables procedure only for the aggregate IGO connectedness variable and not its social, economic, minimalist, and structured subcomponents. Applying instrumental variables to the subcomponents would result in the awkward specication of including in the same model two or more covariates that rely on substantially the same instruments. In supplementary analyses, we reestimated four versions of our full model, substituting instrumented versions of the four subcategories of IGO connectedness one at a time. Results of those regressions were comparable to those reported in model 13.

American Journal of Sociology IGOs, which are often assessed to members using various formulas (i.e., based on GDP). These direct costs may typically pale in comparison to the benets of increased trade, but they are not always trivial, as evidenced by the ongoing battle between the United States and the UN over dues to that organization. The second cost is the risk that IGOs may be diverted from their original purposes, or the will of their members, by powerful bureaucrats. Michels iron law represents a threat not only to the effectiveness of IGOs, but also to the very autonomy of their member states (Strange 1996). Cox and Jacobson (1973) present case studies of decision making in eight IGOs. They identify a trend to bureaucratization, and citing UNESCO and the International Labor Organization as specic examples, claim, The existence of a large organization is itself a potentiality and a pressure for the expansion of tasks (p. 424). Indeed, goal displacement and unjustied budgetary growth were among the criticisms the United States made when withdrawing from UNESCO in 1984. Cox and Jacobson begin the process of identifying features of an IGOs structure and mandate that affect whether it is likely to be more subject to the inuence of the individual participants (bureaucrats, consultants, member representatives) or of its member states. This distinction is an important one for extending our research and fully specifying its policy implications. A clear understanding of what preserves member inuence in IGOs would be useful for (1) identifying which IGOs are most useful for promoting trade and other desired outcomes, (2) helping countries decide which IGOs to join, and (3) guiding the designers and managers of IGOs. Another contributing factor to the U.S. decision to withdraw from UNESCO, that organizations perceived anti-Westernism and anti-Semitism, is useful for illustrating the third, and perhaps greatest, cost of IGO connections. The sociological literature on embeddedness makes clear that there is a dark side to relational constraints (Uzzi 1996). They bind related parties for better or for worse. To this point, we have concentrated on the advantages of relational constraints to smooth trade. In the IGO context, relational constraints may also subject states to unwanted economic, political, and ideological dictates. It is not possible to mitigate this risk fully through careful design of the structure and scope of IGOsany relationship from which the parties derive benet opens the door to normative inuence on a range of issues (Homans 1950). These potentialities suggest that a given IGO connection may be a panacea or a devils compact, depending on the IGOs structure and mandate, and the cultures, histories, economies, and polities of the connected countries. At the same time, we do not want to slight the benets to trade of IGO connectedness merely because they are only part of the equation of benets and costs. The gains to trade from IGO membership 852

IGOs and World Trade are substantial, and their pattern sheds important light on the interdependence between economy and society.

APPENDIX

Coding of IGO Function and Structure Information on the functions and structures of IGOs comes mainly from the listing for each IGO that appears in the Yearbook of International Organization, braced by various other sources. A research assistant who was unfamiliar with our hypotheses performed the coding. It was not practical to have multiple coders because the coding effort required extensive archival work, which involved a learning curve for nding information on IGOs that were sometimes obscure. For both function and structure, the coder initially applied a ne-grained coding scheme. We then collapsed ne-grained subcategories into the categories we used in the actual analysis. This process allowed us to be more precise about the exact nature of each IGO and therefore more condent in the aggregate categories that we use for our analyses. The coder also identied her condence in each coding, based on the quality of the evidence that supported it. We used those condence measures in supplementary analyses to insure our results were robust to the data quality supporting the coding.
IGO Function IGOs have specic functions that are outlined in their mandates. We began with the four-category coding of IGO functions (general purpose, military, economic, and social/cultural) that Jacobson (1996) provides for IGOs in 1981 and 1992. By comparing Jacobsons coding to the available information on each IGO, we identied the criteria for each category. We then broke down the criteria for economic and social/cultural codings into subcategories to produce the nine-category scheme presented in appendix table A1, which we applied to all IGOs in our data. In the analysis, IGO connectedness is calculated using IGOs from all nine categories. EIGO connectedness is calculated using the 241 IGOs categorized under EIGO in table A1. SCIGO connectedness is calculated using the 167 IGOs categorized under SCIGO in table A1. We also examined in preliminary analysis the effects of the economic and social/cultural subcategories of IGOs. These were comparable to those using the aggregate categories, although IGO connectedness measures using the subcategories tended to be highly correlated. 853
TABLE A1 Coding Scheme for IGO Function
Function General purpose (59/497 IGOs) Description of Organizations Umbrella organizations that focus on communication between and the administration of governments; perform multiple functions of standardizing, harmonizing, monitoring, and administering international agreements. Regional political or military alliances; any organization created for military alliance/defense/security purposes. Examples UN, Nordic Council, African Civil Service Observatory, Organization of American States
Military/political (30/497 IGOs)

American Journal of Sociology

REFERENCES

Angell, Robert C. 1950. UNESCO and Social Science Research. American Sociological Review 15:28287. Bilkey, Warren J., and Erik Nes. 1982. Country-of-Origin Effects on Product Evaluations. Journal of International Business Studies 13:89100. Boli, John, and George M. Thomas, eds. 1999. Constructing World Culture: International Nongovernmental Organizations since 1875. Stanford, Calif.: Stanford University Press. Burris, Val. 2004. The Academic Caste System: Prestige Hierarchies in Ph.D. Exchange Networks. American Sociological Review 69:23964. Burt, Ronald S. 1992. Structural Holes: The Social Structure of Competition. Cambridge, Mass.: Harvard University Press. Busch, Mark L., and Eric Reinhardt. 2002. Testing International Trade Law: Empirical Studies of GATT/WTO Dispute Settlement. Pp. 45781 in The Political Economy of International Trade Law: Essays in Honor of Robert E. Hudec, edited by Daniel L. M. Kennedy and James D. Southwick. Cambridge: Cambridge University Press. Cox, Robert W., and Harold K. Jacobson. 1973. Anatomy of Inuence. New Haven, Conn.: Yale University Press. Crawford, John C., and Charles W. Lamb. 1982. Effect of Worldmindedness among Professional Buyers upon Willingness to Buy Foreign Products. Psychological Reports 50:85962. DiMaggio, Paul J., and Hugh Louch. 1998. Socially Embedded Consumer Transactions: For What Kinds of Purchases Do People Use Networks Most? American Sociological Review 63:61937. Dobbin, Frank, and John R. Sutton. 1998. The Strength of a Weak State: The Rights Revolution and the Rise of Human Resource Management Divisions. American Journal of Sociology 104:44176. Ellickson, Robert C. 1991. Order without Law. Cambridge, Mass.: Harvard University Press. Epstein, Martin, ed. 1913. The Statesmans Yearbook, 1913. London: Macmillan. Erev, Ido, Gary Bornstein, and Rachely Galili. 1993. Constructive Intergroup Competition as a Solution to the Free-Rider Problem: A Field Experiment. Journal of Experimental Social Psychology 29:46378. Evans, Carolyn L. 2003. The Economic Signicance of National Border Effects. American Economic Review 93:12911312. Evans, Peter B., Dietrich Rueschemeyer, and Theda Skocpol. 1985. Bringing the State Back In. Cambridge: Cambridge University Press. Fligstein, Neil, and Iona Mara-Drita. 1996. How to Make a Market: Reections on the Attempt to Create a Single Market in the European Union. American Journal of Sociology 102:133. Fligstein, Neil, and Alex Stone Sweet. 2002. Constructing Polities and Markets: An Institutionalist Account of European Integration. American Journal of Sociology 107:120643. Frankel, Jeffrey A. 2000. Globalization of the Economy. Working Paper no. W7858. National Bureau of Economic Research, Cambridge, Mass. Gartzke, Erik. 2000. Preferences and the Democratic Peace. International Studies Quarterly 44:191212. 2002. Do Interorganizational Organizations Promote Peace? Manuscript. Columbia University, Department of Political Science. Geref, Gary, and Miguel Korzeniewicz, eds. 1994. Commodity Chains and Global Capitalism. Westport, Conn.: Greenwood.

In order to make the screen examples match what you see on your PDA, most screenshots are shown in three versions: portrait, landscape and square. iGO can also take full advantage of PDAs with VGA (480 by 640 pixels) display. This setting is also automatic. Should you need the standard quarter-VGA (240 by 320 pixels) resolution for any reason, just turn on the Force QVGA switch in Advanced Settings (5.6.1.1). Most of the screens have a Return ( ) button in the top left corner. This arrow returns to the previous screen or directly to one of the map screens. ) button in the top right corner. This will show Settings screens also have a Help ( a detailed description of the current settings screen.

Operating iGO (Controls)

iGO is designed for easy operation. All controls are operable by fingertips. Wherever possible, pushbuttons and lists are provided to make accessing functions or changing settings as easy as possible.

3.1 Hardware buttons

Hardware buttons are mostly used on map screens to zoom, tilt and rotate the map, or to open other screens. In addition you can use the direction buttons (5-8) and Enter (9) on list screens to make your selection. Where it is applicable, the Recorder button (not shown on the illustration) usually marked by an audio cassette symbol can also be used to operate iGO. Some PDAs may have another additional button (often marked by a camera symbol). iGO uses this button as a quick way to minimise the program, and return to the operating system. Your PDA may also have a wheel or slider control, mainly used for setting the volume of the device. As you will find later in this document, the voice settings of iGO are independent of the PDA settings, but this wheel is programmed to adjust the master volume in iGO, too. When you use one of these buttons or the wheel, the actual master volume level is displayed in the centre of the screen.
On PDA telephones the pick up (green) and hang up (red) buttons of the device are neither used nor reserved by iGO. This means that you will be able to pick up incoming calls while navigating, or minimise iGO with the help of these buttons.
Function Tilts down (towards flat view) Tilts up (towards 2D view) Opens Route Information screen Toggles between Map and Cockpit modes Rotates map left, moves left in lists Rotates map right, moves right in lists Zooms in maps, moves up in lists Zooms out maps, moves down in lists Re-enables Lock-to-Position or announces current voice instruction, makes selection in lists Places a Pin on the map Adjusts volume up and down Minimises iGO

3.1.3 Rotate left and right
Track-up map orientation (4.5.8) will always rotate the map so its top faces your direction of travel when your GPS position is available. Use these hardware buttons if you need to check something on your left or right. If you rotate the map during navigation, Smart Zoom will still zoom and tilt the map but the automatic rotation (Track-up) will be deactivated, so the map orientation will remain as you set it. You need to press Enter (No. 9), or tap the Lock button (4.5.4) to reactivate the automatic map rotation.
3.1.4 Voice instruction announcement
When using either the Map or the Cockpit screen, you can repeat the latest voice instruction by pressing the Enter button (No. 9). Tip: Use this button any time during your journey if you need the distance and type of the next route event. Note: In order to receive voice instructions, sound should not be muted (4.5.11), voice guidance must be enabled (5.3.2), and a route must be active. Note: If you have disabled Smart Zoom by using the functions above or moved the map during navigation (i.e. the Lock button appears on the screen), Enter will first reenable normal navigation (Lock-to-Position and Smart Zoom), and will say the instruction only after being pushed again.
3.1.5 The Route Information screen
To give easy access to the most important data screen during navigation, a hardware button is reserved for opening the Route Information screen. Alternatively, tap the bottom section of the Cockpit screen (4.5.14) or the Info button (4.7.3.6) in the Route menu) to display this screen.
3.1.6 Toggling between the Map and Cockpit screens
You may want to temporarily switch between the Cockpit screen and the Map screen during navigation. For example you may want to use the Map screen to have a better view of the map, or to save a POI. The Cockpit screen can come in handy if you wish to quickly check the name of the street you are on. This button will let you toggle between the two screens. Alternatively, you can return to the Main menu, and select the other screen from there.
3.1.7 Sticking a Pin into the map
The quickest way to record a reminder on the map is to press the Record button of the PDA (if applicable). Whenever your GPS position is available and Lock-toPosition (4.5.4) is enabled, iGO will stick a drawing pin into the map at the current position. If you have disabled Lock-to-Position by selecting a point on the map (4.4.7), the pin will appear at the position of the cursor. If your PDA has no Record button, as an alternative action you can open the Cursor menu (4.5.13) and tap the Pin button.

Note: When the GPS connection cannot be established with the last used parameters (protocol, port and speed), iGO keeps on trying to connect. To save the battery power in cases when you use iGO without a GPS, iGO first tries to open the port in every few seconds, but later only twice every minute. If you later switch on the GPS, the connection is still established without any user action but the repeated attempts will not run the battery down.
4.3.3 GPS data quality indicator
In the top left corner there is a satellite dish to show the quality of the GPS position. Different colours represent different signal quality: - black with a red cross means there is no connection with the GPS device, - red means the GPS is connected but no GPS position is available, - yellow means 2D reception. A GPS position has been aquired, iGO is ready for navigation, but the GPS is using enough satellites for calculating the horizontal position only. Elevation data is not provided, and the position error may be significant. - green means 3D reception. The GPS receiver has enough satellites to calculate altitude. Position is generally correct (yet it can still be inaccurate due to different environmental factors). IGO is ready for navigation.
4.3.4 GPS configuration (Detect and Config.)
You have two black buttons in this screen. These buttons open the automatic (Detect) and the manual (Config.) GPS Setup screens. For information on how to set up the GPS connection refer to the iGO Quick Start Guide and 5.7 in this manual.
4.3.5 Time synchronization
In the top right corner of the screen you have another button that leads to a new screen where you can synchronize the clock of your PDA to the very accurate time provided by the connected GPS.
Turn on the Auto Correction switch to let iGO frequently check and correct the PDA time with the GPS time.
Below that button you will see the current values of the GPS and the PDA clocks. You can check here whether any correction is needed. Tap the manually synchronize the time. button to
Below the PDA time you have and controls to manually correct the time with or without a valid GPS time. It also gives you the chance to correct the time after synchronization if your PDA does not support time zones or daylight saving time. You can also use these buttons if you prefer to set the PDA time from this screen instead of using the time setup features of the operating system.

4.4 The map

The line of the route is displayed on the driving side of the road for two-way and on the axis in case of one-way streets. When the map is zoomed in and the line is wide enough, small arrows show the direction of the route. This can be useful if you preview the route before starting the journey or when entering a complex junction.
4.4.10.4 Inactive legs of the route Future sections of a route are inactive. They are also shown on the map with the same colour but a darker tint than the active one. An inactive route section becomes active as soon as you reach its starting via point.
4.4.10.5 Roads in the route excluded by your preferences Although you can choose whether to include or avoid some road types in Route parameter settings (5.4.4), sometimes they are impossible to avoid near the starting point, via points or the destination. If so, iGO will display those segments of the route with an alternate colour. They are shown in red with daylight colours, and yellow with the night colour scheme, both in active and inactive legs of the route.

4.5 Screens with map

Having explained the contents of the map, the description of the other parts of the map screens follows. There are two map screens: the Map screen and the Cockpit screen. The way they show the map is the same but their look and controls are optimised for different purposes. The Map screen is to be used mainly without a GPS, to browse the map, create user POI items, or to plan your route based on map points. The Map screen is designed to give you the maximum map area. This screen is usually used in 2D North-up mode (you can set iGO so it always opens the Map screen like that see 5.6.1.2).
The Cockpit screen is for driving purposes. Besides showing the map, it contains some additional travel information if you are just cruising (speed, current street your are driving in, speed limit for the current street), and some more route data if you are navigating (e.g. next street in your route, distance to travel, type of the next route event). This screen is typically used in 3D Track-up mode (you can make iGO always open the Cockpit screen like that see 5.6.1.3). There are several controls that function in a similar fashion on the two screens. They are described on the following pages. Map screen contents:

Cockpit screen contents:

5.2.1 Daylight / Night colour profile
iGO comes with different colour schemes for both daylight and night use. There is always one selected daytime scheme and one selected night-time scheme. iGO uses these when switching from day to night and back. Tap the appropriate button and select a new scheme from the list.
5.2.2 Cockpit / Map mode map details
The less detailed the map, the faster iGO can draw and move it. You will find that iGO has a very fast graphic engine that is capable of moving even detailed maps at 66
high speed. Yet you may experience that the performance of some old PDA models can be enhanced by decreasing the amount of detail on the map. This setting has three values: More detailed, Normal, Less detailed (fast). They influence when iGO hides and shows certain map features when zooming in or out. Setting a higher detail level will result in objects appearing sooner when zooming in. Although the map visualisation of iGO is basically the same in Map and Cockpit modes, you can set a different map detail for each one, so the map is better suited for the purposes of that screen.
5.2.3 Alternative Road Names
Some highways have an international name or number for foreign travelers. You can decide whether to show only the local name or both.

5.2.4 Show Street Labels

You can set whether or not to see the names of the streets and the POI icons on the map when driving. Based on the current zoom and tilt levels, street names are displayed either aligned with the street or on signposts stuck into the streets (4.4.4). If you switch these signs on, it will be easier to recognise your location on the map, if you turn them off, it will be easier to see the road geometry. Note: You can only disable street names and POI icons if iGO follows your position. As soon as you move the map and Lock-to-Position (4.5.4) is disabled, street names and icons become visible again. Press Enter or tap Lock to re-enable Lock-toPosition and have the labels and icons disappear again.

5.2.5 Textured Polygons

Enable textured polygons, and have rivers, lakes, forests and other large objects displayed in a more realistic and attractive way on the map. Disabling it will lead to uniform areas but better performance for old PDA devices with a slow processor. See 4.4.5.

5.3 Sound settings

Settings on this page determine the way iGO sounds.
5.3.1 Master sound volume/switch
The loudness of iGO is independent of your PDA settings. When you run iGO, the position of this fader determines the device volume level. When you exit the program, the device settings return. The left part of this control works as a mute button. Tap to mute all iGO sounds. This works in tandem with the Mute button (4.5.11) on the map screens. Switching one of them will change the status of the other.
5.3.2 Voice guidance volume/switch
The switch on the left can turn on or mute iGO's audible guidance. When turned on, the slider on the right will adjust the loudness of voice prompts. In its leftmost position the voice guidance is suppressed, in its rightmost position the master volume applies.

5.3.3 Key sound volume/switch
The switch on the left can turn the key sounds on or off. Key sounds are audible confirmations of either pressing hardware buttons or tapping the touch screen. When key sounds are turned on, the slider on the right will adjust their loudness. In its leftmost position the key sounds are suppressed, in its rightmost position they are played at the master volume level. Note: The sound effects of iGO are context sensitive. They will be different if, for example you open or close a window, or you enable or disable a setting. They will even let you know if you have entered enough letters of a street name to have a short list of matches on the screen.

5.3.4 Dynamic Volume

When driving at high speed, the noise in the car may be too loud to clearly understand the voice guidance messages and to perceive the key sounds. Using Dynamic Volume you can instruct iGO to increase the volume when your speed exceeds a certain minimum, and reach its highest volume (determined by the Master volume slider in Sound settings) at the given maximum speed. 68
Tap the Dynamic Volume button to turn on the feature. This will also open the screen where you can set the minimum and maximum speeds.

5.3.5 Attention Tone

If this is set to Disabled, the voice guidance will sound without a preceding attention tone. Setting it to Single Tone will initiate a single attention tone before the instructions, while Double Tone will use a double tone. Note: The Mute function (4.5.11) accessible from the map screens overrides the settings on this screen. When iGO is muted, all sounds disappear. These settings will not be changed; only the output will be muted temporarily.
5.4 Route parameter settings
This is a very important page. Settings here determine how routes will be calculated. This screen is directly accessible from the Route Information screen (4.6).

5.4.1 Method

Use the handle to set the speed of the route calculation. In the leftmost position the route may not be optimal but the calculation will be very fast. Sliding it to the right will result in more accurate routing in an increased amount of time.
Note: Since iGO calculates routes very quickly, the position of this slide is only used for long routes. Short routes are always calculated to give you the optimal result independent of the slider.

5.4.2 Route

Here you can choose from three different route types. 5.4.2.1 Shortest
Choosing Shortest will result in a route that has the smallest total distance of all possible routes between the given points. This is usually preferred by pedestrians, cyclists or slow vehicles. 5.4.2.2 Fastest

5.6.1.1

Force QVGA mode
Even if you have a PDA capable of VGA resolution display (480 by 640 pixels) you may want to set iGO to display things in the less detailed QVGA mode (240 by 320 pixels). Use this switch to do so. You need to restart iGO to change the resolution. iGO will ask you whether you want to do this immediately. If you answer no, iGO will turn the switch on, but the resolution will only be changed the next time iGO is started. 5.6.1.2 2D in Map mode (and north-up orientation)
The normal use of the Map mode is to browse the map and look for different places on it. It is usually done in a top down view having north towards the top of the map. By default iGO uses the same look for the map in both Map and Cockpit modes. Use this switch to instruct iGO to always open the Map mode in 2D with North-up orientation for map browsing purposes. Note: You will still have the possibility to rotate and tilt the map, but the map will return to 2D mode whenever Map mode is started. 5.6.1.3 3D in Cockpit mode (and track-up orientation)
The normal use of the Cockpit mode is cruising or navigating, when the road lying in front of the driver is the most important part of the map. It is usually done in a 3D view with the current direction towards the top of the map. By default iGO uses the same look of the map in both Map and Cockpit modes. Use this switch to instruct iGO to always open the Cockpit mode in 3D with Track-up (heading up) orientation for driving purposes. Note: you will still have the possibility to switch the map to 2D view or north-up orientation, but the map will return to 2D mode whenever Cockpit mode is started. 5.6.1.4 Zoom in after find
When this switch is turned off, iGO will centre the map to the location selected in Find but will not change the zoom level. If you turn this switch on, iGO will also zoom in to the selected point.
The zoom level in this case depends on the type of object you have searched for. For example in the case of a city search, the zoom level will show you the whole settlement or at least a significant part of it around its centre. Selecting a POI, a junction or an exact address will give you a closer look with only a few streets on the display. 5.6.1.5 Coordinate display format
Positions are sometimes displayed with their addresses, and sometimes with their coordinates. This setting lets you choose between displaying coordinates in degrees (dd.dddddd); degrees and minutes (dd mm.mmm); or degrees, minutes and seconds (dd mm ss.s). Note: the coordinate display mode is independent from the coordinate input in Find menu. You can use all three formats freely on the coordinate input screen. 5.6.1.6 Left / Middle / Right field

As described at the hardware buttons and the map screen functions, you can move, rotate, tilt and zoom the map during navigation. In these cases a Lock icon (4.5.4) appears on the screen. When you rotate, tilt or zoom the map, that part of Smart Zoom is automatically deactivated, the map will still follow your position, but will not change the view parameter you have modified. You can return to the full Smart Zoom mode by tapping the Lock button. If you move the map, all parts of Smart Zoom become deactivated, and the view freezes. Tapping the Lock button makes iGO follow your position again (Lock-toPosition) and also re-enables Smart Zoom. You can make iGO push the Lock button for you automatically after a few seconds of inactivity. Restore Lock-to-Position Turn this switch on if you want iGO to return to your current GPS position after you have moved the map during navigation. This is useful if you have moved the map accidentally, or if you have moved it to quickly check something near your route. After a certain timeout set below, the Lock button disappears, and the map immediately jumps back to show your current position.
Restore Smart Zoom Use this switch if you want iGO to re-enable Smart zoom after zooming, tilting or rotating the map during navigation. This is useful if you have changed the view accidentally, or you modified it to quickly check something near your route. After a certain timeout set below, the Lock button disappears, then Smart Zoom changes the view back for navigation. Unlike with Restore Lock-to-Position, the view changes smoothly. Note: Smart Zoom can only be re-enabled when it is activated at the top of this page. With Smart Zoom turned off, neither does the Lock button appear, nor will the view return to its previous state if you zoom or tilt the map. Delay before restoring You can set the timeout for both Restore Lock-to-Position and Restore Smart Zoom here. Shorter delays are best if you tend to accidentally change the map display, but you may prefer a longer delay if you often look for things around your position while driving. Remember to only look at the display if it is absolutely safe to do so. Note: If you push the Lock button before the automatic Lock comes into effect, Smart Zoom and Lock-to-Position will be re-enabled instantly.

5.6.4 Route options

You can set the basic route parameters on the Route parameter settings screen described in (5.4). On this screen you have some more ways to influence route planning and route recalculation.

5.6.4.1

Off-route sensitivity and Recalculation delay
Depending on the quality of your GPS device, the GPS antenna location in the car and the environment you are driving in, route recalculation can behave differently. iGO may think you have departed the proposed route and perform a recalculation even if you have not done so.
When the GPS reception is poor (e.g. driving in an urban environment with a low sensitivity GPS), occasional wandering (called position fluctuation) of the calculated GPS position is likely. Although iGO employs a sophisticated Lock-on-Road system that will mostly suppress these position errors by aligning your position to the recommended route and the road network of the map, sometimes the errors are too large to correct. To reduce the effect of large errors, you can increase the recalculation tolerance two ways. Off-route sensitivity This is a range of relative values from 0 to 10, telling iGO how far the GPS position should be from the recommended route before the program decides to recalculate. Lower values make iGO insensitive to position errors; higher values will result in quicker reactions. Recalculation delay This is a setting that helps suppress the effects of position fluctuation. With a few seconds delay in recalculation even large position jumps can be survived without a need for route recalculation. 5.6.4.2 U-turn penalty

5.6.5.3

Remove Pins
Normally pins can be deleted one by one. Since they are shown at all zoom levels, and you may end up having too many of them, this button lets you delete all of them together. iGO will warn you that you are about to lose all your pins. 5.6.5.4 Clear Data
This button will delete all user data. It is a reset to factory settings. Tapping this button means you lose all your saved data and customised settings. iGO will warn you about this. 5.6.5.5 Reset Advanced Settings
There are a multitude of Advanced settings in iGO. Some changes may cause iGO to behave in an unsatisfying manner. Tap this button to restore the default settings.
5.6.6 Startup & minimised
Here you can set how iGO should behave when minimised, and how it should integrate with the operating system.

5.6.6.1

Autorun enabled
When this switch is on, iGO will start automatically if you insert the SD card into the PDA. Turn it off if you do not need this service. 5.6.6.2 Today screen entry
An iGO strip appears on the Today screen of the operating system in order to let you start the program easily whenever you like. You can set whether to hide that entry, display it as a single Today strip, or in a double height version that is easy to tap with a fingertip.

5.6.6.3

GPS is active while iGO is minimised
Sometimes you need to minimise iGO during navigation. Here you can decide whether to keep the GPS connection active during that period or not. If you like saving track logs, you should consider turning this feature on. 5.6.6.4 Voice Guidance is active while iGO minimised
You can choose whether to keep the voice navigation active or not when the GPS connection is kept active during background operation. If you have a PDA-telephone, you may wish to switch this function off so as not to bother you during your phone calls, alternatively you may find it useful to be alerted about the turns even if you are not watching iGO on the screen.
5.7 Manual GPS configuration
This screen is separated from the other settings as it is normally used only once, when you set up iGO for the first time. It can be accessed from the GPS Data screen (4.3.4) by pushing the Config. button. There are three values to set on this screen (as you may have already read in the Quick Start Guide).

5.7.1 Protocol

This is the language the GPS and iGO communicate in. The default NMEA is the worldwide standard that practically all GPS devices are able to use. We suggest you keep this setting if you have no special reason to change it. If you have a GPS device using a SiRF chipset, you can change this setting to SiRF. Note: If you use SiRF protocol in iGO, you need to change the protocol of your GPS to SiRF as well. This cannot be done in iGO. You need to quit iGO and use the application provided with your GPS.

5.7.2 Port

There are different wired and virtual serial ports on your PDA. You need to know which of them your GPS is connected to, and set it here.

5.7.3 Baud rate

You also need to set the speed your GPS is communicating at. The higher the speed, the smaller the position delay there will be (there is always a small delay between the real position and the position shown in navigation systems due to the time consuming processing and communication). Tip: Unless you are comfortable with these settings, or you have a special configuration that demands manual set-up, you should use the automatic GPS detection feature of iGO to configure your GPS.
One of the most frequently used functions of iGO is selecting the destination. Once you have the destination, you can start navigating. Getting to this point should be as fast as possible. iGO provides you with a versatile search engine designed to find your chosen destination after only a few taps on the screen. Note: If you select a location in any part of the Find system, you will normally drop back to the map screen, where you will have several possible actions (set as start, set as destination, add as via, continue route with, mark with a pin, or add as a POI item). However, if you enter the Find system using the Find & GO button, iGO will immediately put you into Cockpit mode and start navigation.
6.1 Find & GO (Main menu)
As just mentioned, the fastest way to find your destination and start navigating is to tap Find & GO in the Main Menu (4.1). This will put you on the Find screen, and as soon as you pick a destination, iGO will immediately show the Cockpit screen and start navigation. This means that if you plan your route to one of your Favourite destinations, you only need two taps to start navigating.
6.2 Selection by tapping the map
It is also very easy to set your destination using the map. Just browse to your desired destination on the map, tap it, and the Cursor menu with the possible actions will then open automatically. Note: When appearing automatically, the Cursor menu remains open for a few seconds only. If you decide not to choose any of the actions listed, the menu closes automatically. When you open it manually, it will remain until you close it or switch to another screen.

Just pick any of the recent locations as your destination. Here you have no possibility to reorder the list or filter it by name, but the Next and Previous buttons let you browse through the complete list to find your preferred point. Tip: If you will need a location later but you do not want to save it as a POI, just mark it with a pin, and remember its colour to find it easily in the History list.

6.3.3 Find Coordinates

iGO also lets you enter your destination by map coordinates. The coordinates need to be in latitude/longitude format and, based on the WGS84 earth model (the one used by most GPS devices).
When you enter this page, the coordinates of the current GPS position (or the selected map point (Cursor) if Lock-to-Position is inactive) are shown at the top of the display. The coordinates are always shown in the format configured in Advanced settings Display options (5.6.1.5), but you can enter the coordinates in any of the three formats. You can even enter the latitude and longitude in different formats. Entering a latitude/longitude pair is easy. The left field contains the latitude. It starts with an N (North) or S (South) letter. This tells iGO whether the point is in the Northern or the Southern hemisphere. Use the button to change the hemisphere. Enter numbers for the latitude. Use the decimal point if the degrees, minutes or / / button (the label depends on the seconds are not integers. Use the current cursor position inside the latitude) to start entering minutes after degrees or seconds after minutes. When finished, tap the longitude on the right, and enter it as you did latitude. This time the hemisphere-changing button tells iGO whether the point is located East or West from the meridian crossing Greenwich, UK. Once you have entered both numbers, press Done to make the selection. Tip: The quickest way to tell the coordinates of a point is to select it by tapping on the map or by using Find, and then come to this page and read the coordinates. Tip: If you need to reformat coordinates to the format selected in iGO at 5.6.1.5, enter the coordinates in the format you have, press Done to show it in the map, then come back here to see the same location in the chosen coordinate display format.

6.3.4 Find a POI

You can select your destination from the thousands of POIs included with iGO or from the ones you have previously created. This screen helps you find the one you 91

Tap the Manoeuvre button to return to the Route Information screen with the recalculated route. As you can see iGO has reconsidered the whole route, not only the part starting from the manoeuvre you banned. The first route event leads left and not right as it did before.
As you start driving along the newly planned part of the route, you can zoom out (3.1.2) using a hardware button, that Safety Mode leaves operational, to see the whole remaining part of the route (the Lock button automatically appears on the screen). As you concentrate on the destination and the map, you miss the next turn and go straight on to the next junction. There is no need to worry, iGO waits for a few seconds to be sure you have really left the route, and recalculates it in an instant replacing the missed right turn with three left turns. You just need to keep on following the instructions. iGO will remember the previously avoided turn for the whole journey, so it will not be considered as a route event even if recalculation becomes necessary. These restrictions will be remembered even after arriving at your destination. To make iGO forget them, delete the route manually (4.7.3.2), or exit iGO. After the third left turn you approach the junction where you made the mistake. On the left you can see the point from which the route has been recalculated.
Now the destination is near but as you stop at a red light, you see a nice shop on the right. As the car is stopped, you are allowed to use the touch screen. Just tap the map where the shop is. The cursor (red dot with radiating red circles) appears, and if Popup Info (4.7.2.5) is enabled in the Quick Menu, you can also see the name of the street and the house number. If you tap Pin in the Cursor menu (4.5.13) that appeared automatically as you selected the point on the map, a drawing-pin will appear there. Remember the colour that iGO selected for the pin, and later you can easily find it in the History list.
Now the checkered flag shows the next event is your destination. Smart Zoom will zoom and tilt as with any other route event, and when you reach the given point, the route will be deleted.
The only thing left is to create the POI item later when you are not driving. Now tap MAP in the Main menu then History (6.3.2) in the Menu screen. You will find the drawing-pin on the list. Tap it to have it displayed on the map.
The Cursor menu appears automatically. Tap the Add POI button. You are asked to give a name for the newly created point first. Then you can see its details.

The blue (yellow at night) arrow shows my location, but I cannot see the route (green or red line), and no voice instructions are announced.
I cannot see the Lock button on the Look for a small, red N letter on the screen, yet the map is not rotated during compass icon (4.5.8) or an aeroplane icon instead of it. It is likely that you had driving. initiated the North-up map orientation or the Overview mode unintentionally. Tap this icon to revert to Track-up mode with automatic map rotation. Creating a multi-point route I have tapped the Route To button once for each destination but only the last one appears in the list, all previous points have disappeared. The Route To button is for starting a new route only. For a single route you tap this when the destination is selected. Multi-point routes can be created after you have established a single route. Add points to the single route by using the Add Via and Continue buttons. Applying Route To again will delete the whole route. In your case only single routes existed before, so they were deleted without a warning message. If you already have a multi-point route, iGO warns you before deleting the whole route.
The speaker icon shows iGO is not The muting available from both map muted yet I hear no voice guidance. screens is just a quick silencer that cancels the sound output. Voice guidance and key sounds need to be enabled and volume must be set at the Sound Settings screen (5.3).
I enabled the speed warning as soon as I bought the product but I have just been fined for speeding because iGO failed to warn me.
For an accurate speed warning the actual speed limits of each street and road must be present correctly on the map. This is a relatively new feature of digital maps, so it may not be available in some countries, and its accuracy still has much room for improvement (ask your local dealer for the details in your region). This feature can help you in several cases but it cannot be considered as serious speed control. That is why it can be turned on and off separately from the reliable voice guidance (5.1.4).