Overview

Are area studies and a cross-national approach really that different or they have more in common than we might expect? I think it is the latter, having three major similarities. First, these studies have increasingly merged to seek systematic explanations that cut through regions, which had been thought to be fundamentally different or exceptional (such as Latin America). As such, the second similarity is that they have also come to share some roles, such as confirming a theory. Third, the rise of mixed method approach that can combine area study and cross-national approaches further illuminates not only the second similarity but also a common and ultimate goal shared by the two approaches, which to expand our knowledge.

At the same time, how they go about playing similar roles and accomplishing the shared goal remains as a major difference. In other words, to answer the second question, each approach is better suited than another to answer particular types of questions. Area studies that are often in qualitative nature ask for conditions necessary or sufficient for particular outcomes to occur, while cross-national approach that tend to be in quantitative in nature is much suited for asking the average effect of an independent variable on such outcomes.

Similarities: Area Study Merging with Cross-National Approach

Once again, I think area studies and a cross-national approach have come much closer to each other, having three major similarities. The first major similarity is that both studies seem to have sought systematic explanations that cut through regions. While it is straightforward that a cross-national approach seeks such accounts, I argue that this similarity has emerged due to a change in seeing what area study should be.

Fundamentally, area study is a study that focuses on particular areas or regions of the world. In the mid 20th century, an area study approach had particularly been used as a “cookie cutter” strategy. That is, it ‘snips’ out regions or areas that do not conform to accepted ideas or particularly the ones that are studied through the scope of “ethnocentrism” (Wiarda 1993, 16). Thus, the study is driven by the idea—“what works in one context may not work in another” –and utilized to understand not just deviant or outlier but ‘exceptional’ cases (Wiarda 2005, 2). For instance, O’Donnell’s area studies on Argentina and Brazil (1973; 1976) caught grater attention in the 1970s; he challenged Lipset’s modernization theory (1959; 1960) that came out a decade ago based on Western states as a widely accepted idea displaying the positive relationship between economic development and democracy. Contrarily to Lipset, O’Donnell showed that a process of modernization actually yielded a bureaucratic-authoritarian regime in the ‘richest’ countries in the region. Dependency theorists, such as Frank (1969) and Dos Santos (1971) also elevated the importance of area studies by arguing that Latin American economies would not follow the path of Western states because the region was exploited as ‘satellite economies’ by the West and ended up contributing to the Western modernization. Recently, Mainwaring and Perez-Linan (2003) empirically demonstrated a non-linear relationship between economic development and democracy in Latin America; they concluded that “Latin American exceptionalism” existed in the mid to late 20th century because of distinctive economic policies (ISI) and a link between political elites.

Regardless of Mainwaring and Prez-Linan’s work, however, I see area study to increasingly become a ‘lesson-drawing’ approach (Wiarda 2005). Instead of pointing out regional distinction and exception as an end goal, these differences are used as lessons for building a ‘mega-theory’ or producing systematic accounts regarding comparative politics. This is in part because of the rise of other area studies focusing on the Middle East and East Asia, which show their paths towards democracy that are distinct from both Western states and Latin America. These studies, therefore, diminish Latin America exceptionalism. In addition, area studies, namely of O’Donnell’s (1973; 1976) have come under much attack for being ad hoc explanations, since Latin American turned to re-democratize in the 1980s. All of these factors have then called for a more systematic investigation for providing an account, which identifies common and different conditions contributing to such outcomes (Acemoglu and Robinson 2006).

For instance, Acemoglu and Robinson (2006) recognized Lipset’s modernization theory as one of four paths, rather than ‘the only path,’ and sought conditions that make democratization likely, using the cases of Argentina, Singapore, and South Africa.

O’Donnell’s later work with Schmitter (1986) on Latin America also merged with Przeworski (1991) and Haggard and Kaufman (1995), which utilized cases from various regions, such as the Philippines, South Africa, Bolivia, Uruguay, and Nepal. These studies have then provided a powerful account, suggesting the entrance and exit or authoritarianism to depend on a strategic bargaining between political (military) and economic elites in the wake of economic downturns.

Hence, the first major similarity is, once again, a tendency of both area studies and cross-national approaches to seek systematic accounts that cut through regions. While cross-national approach is essentially thought to have such a goal, area studies have come to understand the need of the goal, while no single region seems to stand as entirely exceptional or can be isolated from the rest of the world.

As such, the second similarity is that they have also come to share some roles. When area studies, at least some parts of the studies, have come close to cross-national approach theoretically, these studies can play a similar analytical role as well, namely the role in confirming a theory.

The theory of political activism may be a good example. Mainly based on Western states, including the United States, empirical cross-ational studies have suggested that well-established democracies have increasingly faced a “legitimacy crisis” or increase in “democratic deficits” (Norris 2011, 3-5). Using a wide range of indicators such as a declining civic engagement or voter turnout (Teixiera 1992; Putnam 2000), declining party loyalties (Aldrich 1995; Dalton et al. 1984), and surveys, they show dissatisfaction and decrease in confidence in national governments (Norris 2011). As such, Fung and Drakeley (2013) conducted an area study focusing on East Asia, ranging from South Korea to Indonesia and Cambodia, and confirmed that even in ‘transnational democracies’ face similar challenges with old democracies or what Norris (2011) calls ‘democratic deficits.’ The area study shows that East Asian states are remarkably similar with Western democracies in a sense that democratic regime may be ‘flawed’ but not ‘broken.’

In turn, cross-national studies can also confirm a theory based on area studies. For instance, Lipset’s modernization theory on the basis on Western Europe has been reinforced with a growing number of empirical cross-national studies (Boix and Stokes 2003; Epstein et al 2006), although debatable (e.g. Przeworski and Limongi 1997; Kennedy 2010; Teorell 2010). Geddes (2003, 351-365) explicitly stated the literature on modernization theory has become much more “persuasive” because “large-n studies have begun to play a greater role in the comparative development fields.”

Third, the rise of mixed method approach that can combine area study and cross-national approaches further illuminates not only the second point but also a common and ultimate goal shared by the two approaches, which to expand our knowledge. For instance, Liberman (2005) recently suggested a mixed-method approach, called nested analysis, which is a research design employing both a Large-N statistical analysis and small-N case studies for in-depth investigation. In particular, this approach advocates the use of a large-N analysis as a guide to draw a subsequent small-case N analysis for two different purposes: a model-building tool for testing an outlier case and a model-testing tool for confirming an online case. Coppedge (2002) is a good example of the nested analysis; he developed a large-N study to determine the need of an area study on Venezuela, which appeared to have a large portion of residuals since the 1990s. Fish (2005) also employed a mixed approach, which conducted a large-n analysis and the Russian case study.

Similarly, King et al (1994) and Brady et al (2006) also suggest a mixed approach, which, though unlike Liberman, utilizes area studies to draw a large-n analysis. For instance, Krieckhaus (2006) briefly reviewed areas of Latin America, East Asia, and Sub-Saharan Africa to argue distinctive effects of democratic governance on economic growth. This area study thus confirms not only a null relationship between the two variables in cross-national studies but also the positive and negative relationships that appear when empirical analyses are conducted separately.

In short, these mixed approaches show that area studies and cross-national studies can reinforce their finding or give a valid reason for each to be conducted; and ultimately, these mutual roles highlight the most important similarity –both studies contribute to enhance our knowledge in comparative politics (Walt 1999).

Different Questions and Approaches

At the same time, how they go about playing similar roles and accomplishing the shared goal remains as a major difference. In particular, area studies are usually qualitative in nature, with some exceptions (e.g. Mainwaring and Prez-Pinan 2003).

This means that, as I mentioned sometimes, area studies are a small-n or case study, which intensively examine particular events with careful attention to historical and cultural contexts. King et al (1994) similarly argue that a small-n study is better at conducting a descriptive inference, which is the “process of understanding an observed phenomenon on the basis of a set of observations” (55).

As such, area studies are particularly suited for asking two questions. The first one is, “what are conditions necessary or sufficient for a particular event to arise?” Returning to the works by O’Donnell and Schmitter (1986), Przeworski (1991), and Haggard and Kaufman (1995), they essentially found economic downturn and ‘authoritarian bargaining’ between political and economic elites as crucial and interactive conditions that change the likelihood of a regime change. Acemoglu and Robinson (2006) as well as Boix (2003) also constructed a model where the level of income inequality and capital mobility to interactively alter the probability of democratization, as they contribute to power relations between political elites and mass citizens.

Area studies are also suited for questions that identify important actors. The identification of domestic actors is crucial because they, according to Mahoney (2011, 115), “create…structures, which in turn shape subsequent actor behaviors, which in turn lead to the development of institutional structural patterns” It is also important, as Walt (1999, 12) points out that the main task of political science research is to produce “useful knowledge about human social behavior.” As such, the above studies are also praised for identifying important actors, such as political elites, business actors, and military, which are “black boxed” (Rueschemeyr et al. 1992, 29) in Lipset’s modernization theory and subsequent empirical studies that focus on the relationship between economic development and democracy (e.g. Prezeworksi and LImongi 1997; Epstein et al 2006; Boix and Stokes 2003; Kennedy 2010; Teorell 2010). Ziblatt (2006:

322) commented, “their accounts improve upon the agentless structural functionalism implicit in modernization theory by reasserting the primacy of collective actors’ resources, preferences, and strategies.” Teorell (2010, 151) also argues, “The key theoretical virtue of this novel approach is that it integrates the previous …traditions by providing structural conditions explaining preference and actions of ordinary citizens, in turn affecting the strategic choices made by political elites.”

In turn, cross-national studies are naturally equipped with a larger sample size and conducted through statistical or quantitative analyses. As such, they are better suited for asking, “what is the average effect on an independent variable on the same or similar outcome seen across the world?” (Mahoney 2011; King et al 1994). Put differently, King et al (1994) argues that, while area studies tend to be good at descriptive inference, large-n studies are better suited for causal inference—that is—to “demonstrate the causal status of each potential linkage in such a posited mechanism the investigator would have to define and then estimate the causal effect underlying it” (86). For instance, Boix and Stokes (2003, 531), building on Lipset, specifically concluded, “ A simulation of the results shows that for low and medium levels of development, the probability of a transition to democracy grows by about 2 percent for each $1,000 increase in per capita income.” Similarly, Kennedy (2010, 797) notes “a 1% increase in per-capita GDP above the country mean” increases the probability of democratic transition. These specific numbers would not come out of area studies; for instance, although Haggard and Kaufman (1992) identified economic downturn as a crucial condition for a regime change, they do not specify exactly how bad the economic situation has to be; it was rather relative judgment in comparing cases.