While the literature tends to define Western countries as wholly democratic, Paxton challenges the notion by noting that many Western countries did not grant universal suffrage until the 20 century. I found Paxton's article on women's suffrage to be a good reminder that democratization is a process, rather than an fully formed, concretized idea. When I begin to think about examples of consolidated democracies that didn't grant women's suffrage until the 20th century (France and the United States both came to mind immediately), it seems difficult to say with any certainty that these countries had inherently stable regimes. In retrospect, they may now appear to be consolidated democracies, but France during WW2 was decidedly undemocratic in many respects, and did not grant women's suffrage until 1946. Since universal suffrage is such an important part of our ideas of democracy, what does Paxton's work mean for our conceptualizations of democracy? And for our ideas about democratic transitions/consolidation?
Jones and Olken provide an interesting insight into (democratic) regime creation. Although I cannot speak to the statistical model employed, the idea that assassination of autocrats spark a move towards democracy seems a bit circumspect, since as Geddes (and others) have pointed out, not all autocratic regimes are the same. In thinkings of Geddes's explanations of regime breakdown, it seems unlikely that an assassination would be the impetus of authoritarian breakdown for military dictatorships or single-party regimes, especially since single-party regimes often have the mechanisms in place to replace leaders, whether they die by assassination, of natural causes or are replaced through elections (similar things could be said of military dictatorships, in that they at least follow a hierarchy of order, so if one ruler was assassinated, another one would be there to replace him).
Elaborated by Geddes, we have different analytical types of authoritarian regimes which she argued will help us understand better the diverse patterns of political transition across different countries. In a similar vein, there also exists a process for a country to move from the side of authoritarianism to become democratic, which some scholars may name as "liberalization." Putting both of them together, I wonder whether it is possible to incorporate them into the measurement of democratization. In other words, taking PL and JO's works as examples, what if we abandon the dichotomous categories of "democracy" and "non-democracy," and replace them with a continuous spectrum that is also implicitly suggested by Paxton? How will these changes in measurement alter our current theoretical understandings?
Also, regarding PL's research on modernization theory, they import new data for a long-standing unresolved theoretical puzzle. However, they chose to focus on the cases during the post-war period. Does this starting point affect their results? I am curious about this because Carles Boix and Susan Stokes' article "Endogenous democratization," by looking back to the 19th century, contended that modernization theory applies better to the pre-war period. Then we may need to ask: Why does modernization theory not fit the post-war democratization very well?
Third, I really like Geddes' article as she delineates different types of authoritarian regimes to help us understand why there are so many different patterns of regime changes that have puzzled political scientists for a long time. Yet her explanation on the interests of members of cliques is not very clear. For one thing, if personalist regimes do not need high economic growth to survive as the leaders only have to materially satisfy those in their clique, then why compared with other types of regimes the personalist ones are the least capable of weathering economic crisis (based on her data at p.135)? Also, why single-party regimes are the best in terms of their sustainability under economic crisis also lack a very compelling explanation.
Last but certainly not least, we have Boix's book this week! He indeed provides an elegant model to account for political transition and the choice of regime, with specified preferences and strategies of different actors involved, under different social and economic conditions. The synthesis of sociological accounts and the rational choice approach may constitute the most strongest part of his arguments. Nevertheless, sometimes I am not quite so sure if he is trying to elucidate democratization or democratic consolidation. Also, even though his model can explain the rise of democracies in the late 19th and the late 20th century, his theory cannot explain very well the post-war wave of democratization, during which we observed the strongest foreign impact (e.g. Germany and Japan). Even though Boix brings the international dimension back to his discussions in the end when he is making policy recommendations (by the way, his focus on immigration is definitely very interesting if not arguable), in his model the foreign impact may be seen only in the aspect of capital immobility. Most important of all, I do not understand the relationship between capital immobility, which he sometimes rephrases as "asset specificity," and the tax rate. While it is clear that the rich with less restriction for moving their assets abroad feel less uncomfortable about redistribution, which thus increases the chance of democracy, why do the voters prefer a lower tax rate under this situation? He argued "[a]s capital becomes more mobile, democratic governments must curb taxes" in order to prevent capital from overly moving abroad, but whose interests in the "democratic governments" does Boix try to explain? I need more clarification as his subsequent elaboration does not resolve my confusion very well.
What I find interesting about Geddes's article is that while the single-party regime and the personalist regime are structurally similar (distribution and benefits are closely aligned in her models of both regime types), the personalist regime collapses under economic pressure, while the single-party regime does not. Could this be because single-party regimes perhaps have more stable institutions (ie certain protocol dictating behavior), than those of personalist regimes (which are left to the whim of the autocrat)?
In the Hit or Miss article, there was a sentence about leaders being the cornerstone of authoritarian institutions, which I sort of can't stop thinking about. I think Geddes's article provides some more insight into that idea, by categorizing different authoritarian regimes.
I would first like to make a few comments in response to Kim's question on our conceptualizations of democracy and more generally on Paxton's paper.
I think Paxton has provided valuable insights on how the inclusion of women as political participants in our measures of democracy could change research outcomes and refute established "stylized facts" such as Huntington's notion of waves of democracy. However, it is precisely because our conceptualizations of democracy have been constantly evolving through time that a "benchmark" in terms of measurement has to be chosen by scholars in comparative studies(Paxton herself mentioned that these changing conceptualizations have resulted in a "moving target" for measurement.). Indeed, such a "benchmark" has emerged implicitly in scholarly literature on democratization--using male suffrage and excluding female participation. And the "norm" of using male suffrage instead of universal adult suffrage for measurement in research could simply be a reflection of general societal norms and views about women's limited role as political actors and participants. As Kim pointed out, even in the US, female suffrage was not granted at the outset and it was only granted in 1946 in France. To the extent that female suffrage has been a relatively new conceptualization of democracy vis-a-vis male suffrage, earlier scholars might have found it more convenient to work with a notion of democracy (as measured by male suffrage) that has existed for longer periods of time, which could then provide a greater sense of continuity and consistency for data collection.
Furthermore, rather than the issue of female suffrage, wouldn't the extension of the franchise to less privileged groups in society, e.g. those with lower levels of education or income, have greater implications on research outcomes? Even in the US, discriminatory voting practices persisted until the 1960s when the National Voting Rights Act was passed. Before that, could the US truly be considered a democracy when a large proportion of its population remained disenfranchised? Hence, although I think Paxton's arguments are valid, I also feel that there will often be gaps between our conceptualizations of democracy (or any other variable) and its measurements, and what is as important for comparative studies as choosing the right measure, is the adoption of a common benchmark for measurement so that scholars are comparing like with like.
Moreover, just as autocratic regimes are not all the same, democracies also differ substantially in the political institutions that they have in place, e.g. democracies with parliamentary systems versus those with presidential systems. Beyond the consideration of female suffrage, these complexities could also have further implications for research design and outcomes.
Last but not least, scholarly work on democratizations has focussed on the experiences of countries from the 19th and 20th centuries--because democratization was virtually unheard of before then. Without going too much into history, I thought it might be interesting to consider the ideological foundations for the concept of democracy, i.e. what led to the conceptualization of democracy itself, as most states in the world were governed by monarchies akin to dictatorships prior to this relatively modern democratization experience. Equally interesting to consider would be the factors that allowed these monarchies to persist for such long periods in history despite varying economic conditions such as economic growth and decline.
In light of Geddes' paper and our topic this week, I thought it might also be interesting to look at the case of North Korea. In her paper, Geddes considered the impact of exogenous shocks that could undermine the regimes, e.g. economic crises. Yet there was little mention of how exogenous/international factors e.g. North Korea's receipts of foreign aid or the nuclear issue, could possibly bolster the regime by creating a more supportive domestic environment for regime survival or succession. Generally, in the readings, the role of international factors has not been given much emphasis (since we are interested in Comparative Politics here), but these factors could also offer additional insights on democratization processes or the persistence of authoritarian regimes that are worth considering.
Jones & Olken’s colorfully titled “Hit or Miss” article makes impressive claims about the effect of successful assassinations on democratic transitions. However, upon closer examination their claim that the successful assassination of an autocrat increases the likelihood of a transition to democracy by 13 % points seems questionable. J & O point out that this 13 % increase is only the difference between a successful attempt (leader is assassinated) or an unsuccessful attempt (leader not assassinated). Disregarding the statistics, it is hardly surprising that in a case where the current autocratic leader dies it is more likely that a transition will occur than in a case where the current leader is still alive. Further, as Kim points out above it seems highly likely that the type of autocratic leader assassinated or not assassinated will impact the likelihood of a democratic transition. Both Geddes’ characterization of autocratic regimes and Linz’s in depth description of variation within totalitarian and authoritarian regimes point out the extent to which authoritarian regimes vary in regards to institutionalization and personalization. Highly bureaucratic autocracies might be relatively unaffected by the successful assassination of a leader if transitions are highly institutionalized. In general, autocracies are vulnerable in situations of leadership transition, whether due to death by natural causes or by some sort of institutionalized mechanism. In this sense, I wonder how different death by assassination is from death by natural causes? Ostensibly, an authoritarian regime with a sick or dying leader will have more time to prepare for the transition, but this is not necessarily always the case.
As noted above, I too am puzzled by some of the underlying logic of Boix’s argument. Boix argues that “excessive differences between the wealthy and the poor push the former to restrict the franchise to avoid the redistributive consequences if a fully democratic system, unless capital mobility restrains ability of poor to expropriate from the wealthy”. This seems to make sense, but it rests on the premise that the poor will tax when democracy is established. Is this assumption correct? If it is not, I am not at all sure that the logic will continue to hold. Additionally, is Boix claiming that inequality blocks democracy or that it jeopardizes the preexisting regime? At times it is unclear whether the mechanisms Boix describes are explaining democratic transitions or democratic consolidation.
I had a few comments and questions while reading Przeworski et al. First, their typology, while clearly defined, lacks substance. There are more features to democratic life than merely elections and transitions of power. While they acknowledge the minimalist nature of their definitions, their blatant dismissal of other democratic precepts-- accountability, transparency, government efficiency etc—severely limits their methodology. These concepts, contrary to their belief that they are “muddled”, can be defined and the governance literature of the last ten years has attempted to do just that. A state that meets the four conditions they propose but doesn’t fulfill these indicators of democratic governance cannot in my opinion be truly classified as a consolidated democracy. A typology that provides for nuances, such as Geddes' classification of different types of authoritarianism is more thorough, though I also realize the difficulty involved in doing this.
Regarding contestation, after reading their explanation and description of the rule, I couldn’t stop reverting to the question of whether or not the type of contestation—e.g. electoral systems—matters? Different type of electoral system—PR vs plurality/majority—produce different outcomes. Isn’t possible that a government can fulfill their five rules and be classified as a democracy but use electoral engineering to skew the results? My initial inclination is to say yes. On a similar note, what if a country has free and fair elections for both the legislature and chief executive but has instituted a system of government that has devolved substantive powers to appointed regional councils (as is the case in a few countries in Africa)? Such a regime could pass the conditions set out but in reality be non-democratic in nature.
Given so many political transitions have happened in various geographical and temporal sites, it is not surprised to see divergent intellectual tastes of empirical analysis. Here, I’d like to pose some questions on methodological issues. First, as Epstein et al. correctly points out “leaving autocracy is not the same as entering democracy”, so there should be different mechanisms and variables which work in two stages of breaking down autocracy and establishing democracy. Also, the same variable may play contradictory roles before and after transition, which could be seen from practices of political elites.
Second question is about the concept and measurement of democracy. According to Epstein et al, hybrid regimes of autocracy and democracy, or so called “Partial Democracies”, dominate those new states in transition. But their identification method by Polity IV scales does not help us conceptually discriminate these countries from well-established democracies. I would suggest Bunce’s (2000) summarization that on common ground of uncertain results, fragile and mature democracies could be defined contrast with each other through uncertain versus certain procedures.
Valerie Bunce .2000. Comparative Democratization Big and Bounded Generalizations. Comparative Political Studies September vol. 33
1. I think how the piece of Geddes and the one of Jones & Olken talk to each other is very interesting and raises questions. Geddes' findings question J&O's ones: Is there any relevance for what type of autocratic leaders are assassined? Does this change the impact? Are these the cases where there is the most impact? At the same time J&O's perspective offers elements of discussion to Geddes: how are violent overthrow of personalist regimes likely to be assessinates?
2. I think Paxton proposes precise alternatives, and practical solutions to the problem of exclusion of women in measurement of democracy. I see it as a refreshing article, both convincing for its theoretical claims and empirically useful for its suggestions, in great contrast with Voigt's (non-) pragmatic approach we read last week.
3. Of course, an underlying theme of this week's readings is the modernization theory, and democracy and development debate. I must say in the Przeworski et al. VS Epstein et al. debate, I have been quite convinced by Epstein at al., and the relevance of allowing for partial democracies. However, more globally, Boix's approach of taking inequality instead of per capita income as a variable is a perspective that seems to me more connected to reality. I do however question his analysis, in particular on the extent to which actors are informed and can predict adequately the outcomes of democracy, as to align their strategy correspondingly.
So my comment is exceptionally short, but it is something that has always bothered me about Przeworski et al.
They spend an inordinate amount of time setting up and justifying a typology of authoritarian states. And then the report their results in a long table, indicating that they coded all these governments. And then the whole matter is dropped for the rest of the book without further mention -- either the typology doesn't matter and shouldn't be reported, or it does and they should explain why they didn't include those models. Everything is democratic/authoritarian.
Interestingly, Jennifer Gandhi, a Przeworski student, spends her entire book explaining that institutions in authoritarian regimes DO matter, which just makes the lack of that analysis in this work more confusing.
>>Furthermore, rather than the issue of female suffrage, wouldn't the extension of the franchise to less privileged groups in society, e.g. those with lower levels of education or income, have greater implications on research outcomes? Even in the US, discriminatory voting practices persisted until the 1960s when the National Voting Rights Act was passed.<<
This is precisely the issue that Paxton brings illuminates. One, were women not a less privileged group for most of the history (and arguably, still are)? Two, Paxton mentions in her article that researchers were willing to discount female suffrage by labeling countries democratic, but certain researchers (cant remember who off the top of my head) were willing to state that the US was not a democracy because it did not extend franchise to minorities. While extending the franchise to minorities is clearly important, she makes an interesting point that women can be excluded even though they make up about half of a nation and a country is still considered democratic...But if minority voting rights are nonexistent, that calls into question the state of a democracy? Why do we consider one a valid criterion but not the other?
I want to make a few contributions. (Those of Man Yan were well thought out, I believe.)
1) Anecdotically, I printed the some 400 dissertation pages of the student Robert mentions (Jennifer Gandhi) for my Professor back home, Josep Colomer. Of course, I only read the introduction but I agree with the point Robert made.
2) Interestingly, no one has pointed out an amazing aspect of this week's readings, maybe because Epstein's et al. article is technically complex and/or because Przeworski's "Modernization: Thories and Facts" is in "further reading". HOW RISKY CAN QUANTITATIVE WORK BE WHEN IT "ALL" DEPENDS ON INCORRECTLY CALCULATED STANDARD ERRORS? Also, although Przeworski is quite amazing, I think the autocracy vs. democracy dichotomy is favored, maybe among other reasons, because it results in a 2x2 Markov transition matrix. If he was using Freedom House, he would have to use an nxn so to speak, which would be highly impractical. Is he throwing away data for convenience? In any event, Epstein's et al. 3rd category seems to be needed.
3) Geddes' article is well written and she provides a good summary. My question is: shouldn't we know many more "stylized facts" after so many publications? In other words, I am afraid we still know very little because data set A contradicts dataset B because they are both biased samples of the world's countries and years.
4) The following is a comment that Macartan Humphreys made on Boix's model this week in his class. Are inequality (X1) and asset mobility (X2) actually uncorrelated? If they were correlated, what problems does this bring into the model specification?
5) I just want to make sure there is a guy in the class who recognizes the efforts of Paxton. Unfortunately, Political Science has been a (maybe even unconsciously?) biased discipline.
@Kim: Indeed, women have been (and perhaps still are) a less privileged group in society.
I do think that Paxton's article is very insightful in highlighting that earlier research and theories on democratization have focussed (narrowly) on male suffrage and the findings should therefore not be generalized to include female suffrage. However, the point that I was really driving at (and I do not think it was clear from what I have written previously) is: scholars seem to be more consistent in excluding women in their measurements of democratization than in excluding other less privileged groups, e.g. those from lower income brackets/educational backgrounds. The sense that I got is that whereas most scholars have excluded women in their measures, they differ much more in deciding whether to include the other less privileged groups in their measurements, e.g. Huntington places the US as a democracy in 1828 when property requirements were relaxed; Muller claims the date is 1865 when universal male suffrage was instituted and Rueschemeyer classifies the US as a democracy only in 1965 when literacy requirements were eliminated. (Paxton)
Certainly, omitting women is an important measurement issue that has important implications for research outcomes. However, because most scholars have consistently omitted women in their measurements, we can at least say that earlier research findings are relevant as long as we keep in mind that they only apply to male suffrage. Conversely, if earlier works also differ in their measurements of democratization with regards to other less privileged groups, the extent to which these research findings are comparable would be much lower, even if we only intend to apply these studies to male suffrage. I feel that in comparative studies, it is important for scholars to standardize what and how they are measuring their variables of interest so that their results are comparable. Although we have a problem with the exclusion of women from measurements of democracy, at least scholars are consistent in making this error, compared to other measurement differences that are not even standardized.
To this end, Paxton has suggested using graded measures of democracy instead of dichotomous measures. This would certainly alleviate the problem but new problems with regard to differences in categorizing the scale of measurement would arise, e.g. how many stages of democratization should we include? The choice of these categories undoubtedly requires normative judgement and is thus debatable.
Another point is that I think it is worse in some respects to exclude other less privileged groups in society than to exclude women. If we may conceive of the nuclear family as the fundamental unit of which society is composed and each household is represented through voting by the male "head of the house", then even if women are excluded from voting, their voices could be indirectly represented. By contrast, if the male "head of the house" is not allowed to vote either, e.g. because of literacy/income requirements, then the entire household would not be represented at all. Women in these households would not even get an indirect say in political matters.
Of course, whether women who are not allowed to vote actually get any say through indirect representation depends on the dynamics of the family, and one can also argue that it is increasingly hard to conceive of the nuclear family as the basic unit of society today, e.g. with increasing divorce rates.
To conclude, I think there are always problems with different approaches of measurement, and there is certainly no perfect or right way to measure any variable. Perhaps what is more critical is for us to distinguish which of these problems are the most significant.
@ Man Yan & others: I do agree with you all Paxton's article is really interesting and there are a lot questions we can raise regarding her critique on the division between the conceptualization and the measurement of democracy. Also, her suggestion on a graded measurement of democracy is surely legitimate. However, what I am wondering this, (how) will the changes in the way we measure democracy alter fundamentally change our understandings on democracy and democratization? In the beginning of her article she is more critical with this regard, but in the end she becomes more soft and instead argues that the inclusion of female voters will enrich our current understandings about democracy and democratization (Actually, which one is she discussing?) to see if the factors that promote male suffrage will apply to the female constituencies or not. I personally think using female suffrage as the point when countries become democratic may be too harsh. Rather, the grant of voting rights to female voters should be seen as the point of democratic maturation (or consolidation even though I do not think these two things are identical because for one thing for democratic institutions to be accepted and stabilized may not require female suffrage). Depending on the definition Paxton has in her mind about the end of democratic transition, what Paxon envisions as a graded measurement of democracy may be not the one envisioned by other scholars since others tend to envision a continuous sequence between the initiation of democratic transition to democratic consolidation.
@ Joan: The question regarding the relation between inequality and asset specificity is very interesting, and I am really looking forward to hearing more if there is a theoretical story to connect these two dimensions though I think Boix made this quite clear that either of them will increase the chance of democracy in his case studies. However, while there seems to no very significant connection between these two, he more or less suggests that some factors will simultaneously affect both of them. For instance, on the one hand international trade will conditionally increase social equality, and on the other hand will also increase capital mobility, both of which in turn fosters the advent of democracy.
While the literature tends to define Western countries as wholly democratic, Paxton challenges the notion by noting that many Western countries did not grant universal suffrage until the 20 century. I found Paxton's article on women's suffrage to be a good reminder that democratization is a process, rather than an fully formed, concretized idea. When I begin to think about examples of consolidated democracies that didn't grant women's suffrage until the 20th century (France and the United States both came to mind immediately), it seems difficult to say with any certainty that these countries had inherently stable regimes. In retrospect, they may now appear to be consolidated democracies, but France during WW2 was decidedly undemocratic in many respects, and did not grant women's suffrage until 1946. Since universal suffrage is such an important part of our ideas of democracy, what does Paxton's work mean for our conceptualizations of democracy? And for our ideas about democratic transitions/consolidation?
ReplyDeleteJones and Olken provide an interesting insight into (democratic) regime creation. Although I cannot speak to the statistical model employed, the idea that assassination of autocrats spark a move towards democracy seems a bit circumspect, since as Geddes (and others) have pointed out, not all autocratic regimes are the same. In thinkings of Geddes's explanations of regime breakdown, it seems unlikely that an assassination would be the impetus of authoritarian breakdown for military dictatorships or single-party regimes, especially since single-party regimes often have the mechanisms in place to replace leaders, whether they die by assassination, of natural causes or are replaced through elections (similar things could be said of military dictatorships, in that they at least follow a hierarchy of order, so if one ruler was assassinated, another one would be there to replace him).
Elaborated by Geddes, we have different analytical types of authoritarian regimes which she argued will help us understand better the diverse patterns of political transition across different countries. In a similar vein, there also exists a process for a country to move from the side of authoritarianism to become democratic, which some scholars may name as "liberalization." Putting both of them together, I wonder whether it is possible to incorporate them into the measurement of democratization. In other words, taking PL and JO's works as examples, what if we abandon the dichotomous categories of "democracy" and "non-democracy," and replace them with a continuous spectrum that is also implicitly suggested by Paxton? How will these changes in measurement alter our current theoretical understandings?
ReplyDeleteAlso, regarding PL's research on modernization theory, they import new data for a long-standing unresolved theoretical puzzle. However, they chose to focus on the cases during the post-war period. Does this starting point affect their results? I am curious about this because Carles Boix and Susan Stokes' article "Endogenous democratization," by looking back to the 19th century, contended that modernization theory applies better to the pre-war period. Then we may need to ask: Why does modernization theory not fit the post-war democratization very well?
Third, I really like Geddes' article as she delineates different types of authoritarian regimes to help us understand why there are so many different patterns of regime changes that have puzzled political scientists for a long time. Yet her explanation on the interests of members of cliques is not very clear. For one thing, if personalist regimes do not need high economic growth to survive as the leaders only have to materially satisfy those in their clique, then why compared with other types of regimes the personalist ones are the least capable of weathering economic crisis (based on her data at p.135)? Also, why single-party regimes are the best in terms of their sustainability under economic crisis also lack a very compelling explanation.
Last but certainly not least, we have Boix's book this week! He indeed provides an elegant model to account for political transition and the choice of regime, with specified preferences and strategies of different actors involved, under different social and economic conditions. The synthesis of sociological accounts and the rational choice approach may constitute the most strongest part of his arguments. Nevertheless, sometimes I am not quite so sure if he is trying to elucidate democratization or democratic consolidation. Also, even though his model can explain the rise of democracies in the late 19th and the late 20th century, his theory cannot explain very well the post-war wave of democratization, during which we observed the strongest foreign impact (e.g. Germany and Japan). Even though Boix brings the international dimension back to his discussions in the end when he is making policy recommendations (by the way, his focus on immigration is definitely very interesting if not arguable), in his model the foreign impact may be seen only in the aspect of capital immobility. Most important of all, I do not understand the relationship between capital immobility, which he sometimes rephrases as "asset specificity," and the tax rate. While it is clear that the rich with less restriction for moving their assets abroad feel less uncomfortable about redistribution, which thus increases the chance of democracy, why do the voters prefer a lower tax rate under this situation? He argued "[a]s capital becomes more mobile, democratic governments must curb taxes" in order to prevent capital from overly moving abroad, but whose interests in the "democratic governments" does Boix try to explain? I need more clarification as his subsequent elaboration does not resolve my confusion very well.
What I find interesting about Geddes's article is that while the single-party regime and the personalist regime are structurally similar (distribution and benefits are closely aligned in her models of both regime types), the personalist regime collapses under economic pressure, while the single-party regime does not. Could this be because single-party regimes perhaps have more stable institutions (ie certain protocol dictating behavior), than those of personalist regimes (which are left to the whim of the autocrat)?
ReplyDeleteIn the Hit or Miss article, there was a sentence about leaders being the cornerstone of authoritarian institutions, which I sort of can't stop thinking about. I think Geddes's article provides some more insight into that idea, by categorizing different authoritarian regimes.
I would first like to make a few comments in response to Kim's question on our conceptualizations of democracy and more generally on Paxton's paper.
ReplyDeleteI think Paxton has provided valuable insights on how the inclusion of women as political participants in our measures of democracy could change research outcomes and refute established "stylized facts" such as Huntington's notion of waves of democracy. However, it is precisely because our conceptualizations of democracy have been constantly evolving through time that a "benchmark" in terms of measurement has to be chosen by scholars in comparative studies(Paxton herself mentioned that these changing conceptualizations have resulted in a "moving target" for measurement.). Indeed, such a "benchmark" has emerged implicitly in scholarly literature on democratization--using male suffrage and excluding female participation. And the "norm" of using male suffrage instead of universal adult suffrage for measurement in research could simply be a reflection of general societal norms and views about women's limited role as political actors and participants. As Kim pointed out, even in the US, female suffrage was not granted at the outset and it was only granted in 1946 in France. To the extent that female suffrage has been a relatively new conceptualization of democracy vis-a-vis male suffrage, earlier scholars might have found it more convenient to work with a notion of democracy (as measured by male suffrage) that has existed for longer periods of time, which could then provide a greater sense of continuity and consistency for data collection.
Furthermore, rather than the issue of female suffrage, wouldn't the extension of the franchise to less privileged groups in society, e.g. those with lower levels of education or income, have greater implications on research outcomes? Even in the US, discriminatory voting practices persisted until the 1960s when the National Voting Rights Act was passed. Before that, could the US truly be considered a democracy when a large proportion of its population remained disenfranchised? Hence, although I think Paxton's arguments are valid, I also feel that there will often be gaps between our conceptualizations of democracy (or any other variable) and its measurements, and what is as important for comparative studies as choosing the right measure, is the adoption of a common benchmark for measurement so that scholars are comparing like with like.
Moreover, just as autocratic regimes are not all the same, democracies also differ substantially in the political institutions that they have in place, e.g. democracies with parliamentary systems versus those with presidential systems. Beyond the consideration of female suffrage, these complexities could also have further implications for research design and outcomes.
Last but not least, scholarly work on democratizations has focussed on the experiences of countries from the 19th and 20th centuries--because democratization was virtually unheard of before then. Without going too much into history, I thought it might be interesting to consider the ideological foundations for the concept of democracy, i.e. what led to the conceptualization of democracy itself, as most states in the world were governed by monarchies akin to dictatorships prior to this relatively modern democratization experience. Equally interesting to consider would be the factors that allowed these monarchies to persist for such long periods in history despite varying economic conditions such as economic growth and decline.
In light of Geddes' paper and our topic this week, I thought it might also be interesting to look at the case of North Korea. In her paper, Geddes considered the impact of exogenous shocks that could undermine the regimes, e.g. economic crises. Yet there was little mention of how exogenous/international factors e.g. North Korea's receipts of foreign aid or the nuclear issue, could possibly bolster the regime by creating a more supportive domestic environment for regime survival or succession. Generally, in the readings, the role of international factors has not been given much emphasis (since we are interested in Comparative Politics here), but these factors could also offer additional insights on democratization processes or the persistence of authoritarian regimes that are worth considering.
ReplyDeleteJones & Olken’s colorfully titled “Hit or Miss” article makes impressive claims about the effect of successful assassinations on democratic transitions. However, upon closer examination their claim that the successful assassination of an autocrat increases the likelihood of a transition to democracy by 13 % points seems questionable. J & O point out that this 13 % increase is only the difference between a successful attempt (leader is assassinated) or an unsuccessful attempt (leader not assassinated). Disregarding the statistics, it is hardly surprising that in a case where the current autocratic leader dies it is more likely that a transition will occur than in a case where the current leader is still alive. Further, as Kim points out above it seems highly likely that the type of autocratic leader assassinated or not assassinated will impact the likelihood of a democratic transition. Both Geddes’ characterization of autocratic regimes and Linz’s in depth description of variation within totalitarian and authoritarian regimes point out the extent to which authoritarian regimes vary in regards to institutionalization and personalization. Highly bureaucratic autocracies might be relatively unaffected by the successful assassination of a leader if transitions are highly institutionalized. In general, autocracies are vulnerable in situations of leadership transition, whether due to death by natural causes or by some sort of institutionalized mechanism. In this sense, I wonder how different death by assassination is from death by natural causes? Ostensibly, an authoritarian regime with a sick or dying leader will have more time to prepare for the transition, but this is not necessarily always the case.
ReplyDeleteAs noted above, I too am puzzled by some of the underlying logic of Boix’s argument. Boix argues that “excessive differences between the wealthy and the poor push the former to restrict the franchise to avoid the redistributive consequences if a fully democratic system, unless capital mobility restrains ability of poor to expropriate from the wealthy”. This seems to make sense, but it rests on the premise that the poor will tax when democracy is established. Is this assumption correct? If it is not, I am not at all sure that the logic will continue to hold. Additionally, is Boix claiming that inequality blocks democracy or that it jeopardizes the preexisting regime? At times it is unclear whether the mechanisms Boix describes are explaining democratic transitions or democratic consolidation.
I had a few comments and questions while reading Przeworski et al. First, their typology, while clearly defined, lacks substance. There are more features to democratic life than merely elections and transitions of power. While they acknowledge the minimalist nature of their definitions, their blatant dismissal of other democratic precepts-- accountability, transparency, government efficiency etc—severely limits their methodology. These concepts, contrary to their belief that they are “muddled”, can be defined and the governance literature of the last ten years has attempted to do just that. A state that meets the four conditions they propose but doesn’t fulfill these indicators of democratic governance cannot in my opinion be truly classified as a consolidated democracy. A typology that provides for nuances, such as Geddes' classification of different types of authoritarianism is more thorough, though I also realize the difficulty involved in doing this.
ReplyDeleteRegarding contestation, after reading their explanation and description of the rule, I couldn’t stop reverting to the question of whether or not the type of contestation—e.g. electoral systems—matters? Different type of electoral system—PR vs plurality/majority—produce different outcomes. Isn’t possible that a government can fulfill their five rules and be classified as a democracy but use electoral engineering to skew the results? My initial inclination is to say yes. On a similar note, what if a country has free and fair elections for both the legislature and chief executive but has instituted a system of government that has devolved substantive powers to appointed regional councils (as is the case in a few countries in Africa)? Such a regime could pass the conditions set out but in reality be non-democratic in nature.
Given so many political transitions have happened in various geographical and temporal sites, it is not surprised to see divergent intellectual tastes of empirical analysis. Here, I’d like to pose some questions on methodological issues.
ReplyDeleteFirst, as Epstein et al. correctly points out “leaving autocracy is not the same as entering democracy”, so there should be different mechanisms and variables which work in two stages of breaking down autocracy and establishing democracy. Also, the same variable may play contradictory roles before and after transition, which could be seen from practices of political elites.
Second question is about the concept and measurement of democracy. According to Epstein et al, hybrid regimes of autocracy and democracy, or so called “Partial Democracies”, dominate those new states in transition. But their identification method by Polity IV scales does not help us conceptually discriminate these countries from well-established democracies. I would suggest Bunce’s (2000) summarization that on common ground of uncertain results, fragile and mature democracies could be defined contrast with each other through uncertain versus certain procedures.
Valerie Bunce .2000. Comparative Democratization Big and Bounded Generalizations. Comparative Political Studies September vol. 33
I would like to make three short comments.
ReplyDelete1. I think how the piece of Geddes and the one of Jones & Olken talk to each other is very interesting and raises questions. Geddes' findings question J&O's ones: Is there any relevance for what type of autocratic leaders are assassined? Does this change the impact? Are these the cases where there is the most impact? At the same time J&O's perspective offers elements of discussion to Geddes: how are violent overthrow of personalist regimes likely to be assessinates?
2. I think Paxton proposes precise alternatives, and practical solutions to the problem of exclusion of women in measurement of democracy. I see it as a refreshing article, both convincing for its theoretical claims and empirically useful for its suggestions, in great contrast with Voigt's (non-) pragmatic approach we read last week.
3. Of course, an underlying theme of this week's readings is the modernization theory, and democracy and development debate. I must say in the Przeworski et al. VS Epstein et al. debate, I have been quite convinced by Epstein at al., and the relevance of allowing for partial democracies. However, more globally, Boix's approach of taking inequality instead of per capita income as a variable is a perspective that seems to me more connected to reality. I do however question his analysis, in particular on the extent to which actors are informed and can predict adequately the outcomes of democracy, as to align their strategy correspondingly.
So my comment is exceptionally short, but it is something that has always bothered me about Przeworski et al.
ReplyDeleteThey spend an inordinate amount of time setting up and justifying a typology of authoritarian states. And then the report their results in a long table, indicating that they coded all these governments. And then the whole matter is dropped for the rest of the book without further mention -- either the typology doesn't matter and shouldn't be reported, or it does and they should explain why they didn't include those models. Everything is democratic/authoritarian.
Interestingly, Jennifer Gandhi, a Przeworski student, spends her entire book explaining that institutions in authoritarian regimes DO matter, which just makes the lack of that analysis in this work more confusing.
>>Furthermore, rather than the issue of female suffrage, wouldn't the extension of the franchise to less privileged groups in society, e.g. those with lower levels of education or income, have greater implications on research outcomes? Even in the US, discriminatory voting practices persisted until the 1960s when the National Voting Rights Act was passed.<<
ReplyDeleteThis is precisely the issue that Paxton brings illuminates. One, were women not a less privileged group for most of the history (and arguably, still are)? Two, Paxton mentions in her article that researchers were willing to discount female suffrage by labeling countries democratic, but certain researchers (cant remember who off the top of my head) were willing to state that the US was not a democracy because it did not extend franchise to minorities. While extending the franchise to minorities is clearly important, she makes an interesting point that women can be excluded even though they make up about half of a nation and a country is still considered democratic...But if minority voting rights are nonexistent, that calls into question the state of a democracy? Why do we consider one a valid criterion but not the other?
I want to make a few contributions. (Those of Man Yan were well thought out, I believe.)
ReplyDelete1) Anecdotically, I printed the some 400 dissertation pages of the student Robert mentions (Jennifer Gandhi) for my Professor back home, Josep Colomer. Of course, I only read the introduction but I agree with the point Robert made.
2) Interestingly, no one has pointed out an amazing aspect of this week's readings, maybe because Epstein's et al. article is technically complex and/or because Przeworski's "Modernization: Thories and Facts" is in "further reading". HOW RISKY CAN QUANTITATIVE WORK BE WHEN IT "ALL" DEPENDS ON INCORRECTLY CALCULATED STANDARD ERRORS? Also, although Przeworski is quite amazing, I think the autocracy vs. democracy dichotomy is favored, maybe among other reasons, because it results in a 2x2 Markov transition matrix. If he was using Freedom House, he would have to use an nxn so to speak, which would be highly impractical. Is he throwing away data for convenience? In any event, Epstein's et al. 3rd category seems to be needed.
3) Geddes' article is well written and she provides a good summary. My question is: shouldn't we know many more "stylized facts" after so many publications? In other words, I am afraid we still know very little because data set A contradicts dataset B because they are both biased samples of the world's countries and years.
4) The following is a comment that Macartan Humphreys made on Boix's model this week in his class. Are inequality (X1) and asset mobility (X2) actually uncorrelated? If they were correlated, what problems does this bring into the model specification?
5) I just want to make sure there is a guy in the class who recognizes the efforts of Paxton. Unfortunately, Political Science has been a (maybe even unconsciously?) biased discipline.
See you tomorrow :)
@Kim: Indeed, women have been (and perhaps still are) a less privileged group in society.
ReplyDeleteI do think that Paxton's article is very insightful in highlighting that earlier research and theories on democratization have focussed (narrowly) on male suffrage and the findings should therefore not be generalized to include female suffrage. However, the point that I was really driving at (and I do not think it was clear from what I have written previously) is: scholars seem to be more consistent in excluding women in their measurements of democratization than in excluding other less privileged groups, e.g. those from lower income brackets/educational backgrounds. The sense that I got is that whereas most scholars have excluded women in their measures, they differ much more in deciding whether to include the other less privileged groups in their measurements, e.g. Huntington places the US as a democracy in 1828 when property requirements were relaxed; Muller claims the date is 1865 when universal male suffrage was instituted and Rueschemeyer classifies the US as a democracy only in 1965 when literacy requirements were eliminated. (Paxton)
Certainly, omitting women is an important measurement issue that has important implications for research outcomes. However, because most scholars have consistently omitted women in their measurements, we can at least say that earlier research findings are relevant as long as we keep in mind that they only apply to male suffrage. Conversely, if earlier works also differ in their measurements of democratization with regards to other less privileged groups, the extent to which these research findings are comparable would be much lower, even if we only intend to apply these studies to male suffrage. I feel that in comparative studies, it is important for scholars to standardize what and how they are measuring their variables of interest so that their results are comparable. Although we have a problem with the exclusion of women from measurements of democracy, at least scholars are consistent in making this error, compared to other measurement differences that are not even standardized.
To this end, Paxton has suggested using graded measures of democracy instead of dichotomous measures. This would certainly alleviate the problem but new problems with regard to differences in categorizing the scale of measurement would arise, e.g. how many stages of democratization should we include? The choice of these categories undoubtedly requires normative judgement and is thus debatable.
Another point is that I think it is worse in some respects to exclude other less privileged groups in society than to exclude women. If we may conceive of the nuclear family as the fundamental unit of which society is composed and each household is represented through voting by the male "head of the house", then even if women are excluded from voting, their voices could be indirectly represented. By contrast, if the male "head of the house" is not allowed to vote either, e.g. because of literacy/income requirements, then the entire household would not be represented at all. Women in these households would not even get an indirect say in political matters.
Of course, whether women who are not allowed to vote actually get any say through indirect representation depends on the dynamics of the family, and one can also argue that it is increasingly hard to conceive of the nuclear family as the basic unit of society today, e.g. with increasing divorce rates.
To conclude, I think there are always problems with different approaches of measurement, and there is certainly no perfect or right way to measure any variable. Perhaps what is more critical is for us to distinguish which of these problems are the most significant.
@ Man Yan & others: I do agree with you all Paxton's article is really interesting and there are a lot questions we can raise regarding her critique on the division between the conceptualization and the measurement of democracy. Also, her suggestion on a graded measurement of democracy is surely legitimate. However, what I am wondering this, (how) will the changes in the way we measure democracy alter fundamentally change our understandings on democracy and democratization? In the beginning of her article she is more critical with this regard, but in the end she becomes more soft and instead argues that the inclusion of female voters will enrich our current understandings about democracy and democratization (Actually, which one is she discussing?) to see if the factors that promote male suffrage will apply to the female constituencies or not. I personally think using female suffrage as the point when countries become democratic may be too harsh. Rather, the grant of voting rights to female voters should be seen as the point of democratic maturation (or consolidation even though I do not think these two things are identical because for one thing for democratic institutions to be accepted and stabilized may not require female suffrage). Depending on the definition Paxton has in her mind about the end of democratic transition, what Paxon envisions as a graded measurement of democracy may be not the one envisioned by other scholars since others tend to envision a continuous sequence between the initiation of democratic transition to democratic consolidation.
ReplyDelete@ Joan: The question regarding the relation between inequality and asset specificity is very interesting, and I am really looking forward to hearing more if there is a theoretical story to connect these two dimensions though I think Boix made this quite clear that either of them will increase the chance of democracy in his case studies. However, while there seems to no very significant connection between these two, he more or less suggests that some factors will simultaneously affect both of them. For instance, on the one hand international trade will conditionally increase social equality, and on the other hand will also increase capital mobility, both of which in turn fosters the advent of democracy.