Overall I really enjoy this week's readings. Each of them explores the interplay of formal and informal institutions, which I agree is a very important yet long understudied topic, from different perspectives. I also learn a lot from these authors' integral use of quantitative and qualitative methods.
I want to start with Prof Frye's article. I have to admit that before reading this article I always intuitively assume private and public institutions surely complement each other, but the former can substitute the latter when the formal institutions are relatively weak and unreliable. My observation is based on China and my father's experience. Since the 1980s he has been doing business in China (i.e. he is among the first generation of foreign investors). When I was doing my research in Hong Kong, he kept telling me that how the business environment has been changed since the 1980s. In the past, he used to spend most of his effort establishing good relationships with other fellow merchants and the local party officials as everyone knows going to the public institutions and courts are not only costly but also unreliable. However, as the Communist Party of China started institutionalizing the business conducts since the 1990s, he is forced to deal with the formal institutions more often (but that does not mean they are better now). Therefore, I will tend to think private and public institutions can complement and substitute each other, depending on the specific context we are referring to. Maybe the question is: Why do informal (private?) institutions sometimes complement the formal (public?) ones, but in other cases substitute them? My speculation is it depends on how private institutions are established. For example, similar to Man Yan's previous inquiry, a lot of countries have displayed the importance of personal relations. As in Japan, the private institutions prevalent across the bureaucratic system have been long well-known, but they somehow do not weaken the Japanese state in providing public goods and services. Why is it the case? Maybe it is because to become bureaucrats you still have to pass the exams and master your professional expertise before others recognize you as part of the clientelistic network. Therefore, it seems like when informal institutions is built upon the formal ones, they will become much more compatible.
Before I move to Keefer et al., I have some more points for Prof Frye's piece. First, do business people avoid using courts to resolve their disputes only because the public institutions are unreliable? Or they choose not to appeal their cases to courts only because they are too costly? My impression is that no matter how reliable courts are people may still refuse to bring their disputes there because doing so is simply way too expensive and time-consuming. Second, while I think this survey should be conducted in China too and it will be very interesting to compare these two cases, why does this article choose to use "public" and "private" institutions, rather than "formal" and "informal" institutions?
Based on my understanding about the Japanese political system, I find Keeler et al.'s assumption quite problematic, and (like Cecelia said) I rarely have a strong sense of objection when reading an article - Why must building reputations through the investment of public infrastructure be the tradeoff of cultivating one's patronage network? Can't it be conditioned by the electoral system? I have these questions because the Japanese politicians did (not anymore now) manage to do both. Some may say Japan managed to achieve both of them because it is rich, but the thing is the Japanese politicians had been doing this way before the Japanese economy took off. In Japan, politicians' investment in their patron-client network did induce massive import in the building of public infrastructure while it has been so "privatized" that it is rather easy to tell who these collective goods are targeted. For instance, during the 1970s the Prime Minister Tanaka Kakuei proposed the plan of high speed railway, and interestingly he include several places his hometown Niigata Prefecture as the locations of train stations. Niigata used to be a very poor prefecture in Japan, so why they selected did arouse some concern. But since the plan covered the major metropolitan areas (Tokyo, Nagoya, and Osaka), no one paid attention until Tanaka's scandal broke out in the late 1970s. My point here is the dichotomy Keeler et al. suggest is not always the case, and I do not find a very compelling explanations why they think it must be the case. Even if it is really cheaper for politicians to control the patron rather than spend , then why must other voters support the politicians who do not take care of them but the patron? How do politicians resolve the agent problem? (In this sense, I will not say this approach will be any cheaper). I understand Keeler et al. try to find an alternative explanation since the traditional approach applied to distinguish different democratic systems fail to predict politicians' incentives to provide state spoils or public goods, but I do not really find their focus on "credibility" very impressive. They claim to endogenize it, but somehow they take it as given, but they do not explain what conditions a politician's "credibility."
My last comment regards Wilkinson's book. I really love his book, and the paper he presented at the comparative politics seminar is also very interesting. As a student of ethnic politics, I am really amazed how patient he is. The data he collects is surely thorough and he clearly understands his limitations but still tries his best to find other possible measurements. His review about the competing theories of ethnic conflicts is my favorite part. Overall, I do learn a lot from his book, and he beautifully justifies why readers should buy his political explanations. Three things I would like point out though. First of all, why do the Hindu-Muslim conflicts tend to benefit the former? Do these conflicts really help the Hindu candidates? Wilkinson mostly uses Hindu politicians' statements to illustrate this logic, but I wonder if there are other possible factors contributing to Hindu candidates' landslide victories. Second, if the Muslim understand the Hindu's provocative actions act as part of their electoral campaign, why do they still choose to fight back that exactly cater to the Hindu's expectations? (Or they actually do not see this?) Third, according Wilkinson, the Hindu-Muslim riots prevail mostly because the state government chooses not to intervene not because they are unable to do so, is there any way to differentiate these two situations?
Chao-yo, in regards to your comment that in China private institutions can act as substitutes to public ones, I don't think Professor Frye's article was necessarily to say that private institutions are wholly complementary; but that the evidence suggests that they tend to function best as complements. i think it seems somewhat obvious that private institutions tend to act as substitutes when formal institutions are weak; however, in that case, how truly useful are these private institutions? Without enforcement, one's reliance on reputation (or some other informal institution) is only as partially useful as it would be in a country with strong institutions. I think Krep's chapter on corporate culture attempts to theorize and model this behavior, with a certain degree of success.
Additionally, Frye's article (I think in the section about complements...I don't think this was Frye's original point) also brings up a salient point about how countries with strong institutions better disseminate information, thus reputations are useful because different organizations can access that information relatively cost free. In weak states, this information is less easy to access and corroborate.
I have only brief comments this week: first, I liked Wilkinson's book very much but I feel that in some ways it is unbalanced. Though he tests his first hypothesis that ethnic violence occurs in electoral districts that are highly competitive where politicians have an incentive to polarize ethnic communities, I feel that he less systematically analyzes the assertion that riots occur in states where the government is not using policing to stop violence because the political party in power is not dependent on minority communities to win elections. Wilkinson provides only brief anecdotal evidence to back up his claim that it is not a decline in state capacity that causes violence but the politicization of government resources. Im not sure how you would systematically test this theory but when contrasted to the tests he puts the first part of the argument through it looks somewhat deficient. My other critique is that this argument does not extend to other types of ethnic violence such as between caste violence or linguistic violence. Though he claims that Hindu-Muslim violence is common, he does not explain why politicians don't catalyze cross-caste violence when theoretically it could work in the same way as Hindu-Muslim violence.
The Susan Stokes piece on political clientalism is interesting to me in that it highlights something that common people usually suspect of politicians, but something, as the author acknowledges, that has not been thoroughly explored and analyzed in literature. The way the author takes up each portion of the puzzle and relates the literature associated is helpful in understanding the aspects of how, when, and why clientelism works. I am particularly intrigued by the idea of clientelism working to produce poverty and inequality, which in turn perpetuate it. If institutions such as the secret ballot and electoral rules work toward mitigating clientalism, is it only when economic growth has rendered the voters less amenable to temptation by politicians? While accepting that clientalism woks, and is detrimental, the author could have delved deeper into the relative effectiveness of different forms of clientelism: payoffs in the form of material goods (immediate but short term in effect) vs. patronage (slower but long term), protection vs. security etc.
I believe Stokes and Keefer&Vlaicu offer interesting and complementary perspective on the consequences of clientelism. They both (directly or indirectly) refer to two types of consequences: poverty/wefalre and policy. 1) For the poverty/welfare one, Stokes underlines the fact that clientelism may cause poverty, whether because the leaders voluntarily do not want income to grow or because of the declining relative productivity of monopolized-goods sectors. Keefer&Vlaicu, on the other hand, estimate that welfare may improve (for a short length). 2) On the broader policy front, Keefer&Vlaicu agree with Stokes that clientelism will bring lower provision of public goods compared to credible politician competitors with no patron. However, they add that compare to non-credible politicians, the ones with patrons may be better off.
Besides, Stokes seems to make the point, when she discusses the effectiveness of clientelism, that it is somehow surprising that political parties pursue clientelism even in cases (3 of the 4 possibilities of the game she exposes) that it is ineffective. She states that parties usually do not opt for uneffective strategies. I am not as surprised as her, because I have the impression she is analyzing parties and politicians as if they were ONLY seeking votes. I think it is not surprising that some politicians will, for example, try to get not necessarily votes, but also further benefits and possibilities (like a job once off politics) or to get personal immediate economic benefits. This would not fit in her restrictive definition of clientelism, I believe, since precisely they would not been looking for electoral support: but are politicians being driven by a definition of clientelism??
I found Wilkinson’s contention—that the degree of electoral competition influences whether or not ethnic violence occurs and the response of the state—interesting and well argued. I especially liked how he systematically tackled all the other arguments for why ethnic violence occurs chapter by chapter. Having recently read Varshney who argues civil society, especially formal associational networks that cross ethnic lines, prevents conflict, I would have liked Wilkinson to have addressed him in little more depth and to see if there was overlap between the arguments: if violence was low in areas that had both interethnic civic networks and low political competition characterized by intra-Majority party competition. If so, couldn’t both be a function of something else?
I guess I'm the contrarian this week, because the Stokes article struck me as a nearly Platonic fusion of trends I hate about American political science.
1. Her definition of clientalism makes no sense, which she acknowledges, because the exchange of votes for goods is plagued by uncertainty and available data suggests that rewards are targeted at regions and not just individual voters (this is the Snyder research on the US she references.)
2. She sets up a pointless game that demonstrates that, in a one-off scenario, both voters and clientalistic parties defect from their agreement. But then she points out that this doesn't comport to reality, and suggests information and reward size are important. Which further reinscribes my impression of game theory as a valuable literature that enables us all to say obvious things in complicated ways.
3. She then points out that some of the empirical research indicates that parties target swing voters, but others target core voters. No clear reason when they choose which.
4. Then she suggests clientalism may cause poverty. Or vice versa. ("it is probably both a cause and a consequence.")
So, in summary, there is a pattern of behavior in the world in which politicians buy votes with favors. These can be individual or collective. Not too sure who gets the favors or why. Seems to be mostly poor places, also not sure why. We should all probably do some more work to figure all this out, and it should probably involve a game somehow.
Maybe I'm just feeling cranky, but jumpin' Jeehosephat, that selection leaves me shaking my head.
I look forward to discussing it tomorrow to figure out what it is I'm missing here!
First I agree with Robert, I don't find the case by Stokes compelling at all. I must also say that I was already biased because after reading another article by Stokes I was less than impressed so I did not approach this with very high expectations.
Second, I found the concept of focal points outlined by Kreps both interesting and useful. In particular I believe it has considerable explanatory power to answer why some principle or rule is used naturally to select a certain behavior in view of many possible equilibrium behaviors (121). For someone who is interested in cooperation like me, this idea seems promising.
Third, according to Frye informal and formal institutions are complements. My question is if in more develop nations should we expect this complementarity to decrease given the increased certainty in formal mechanisms? Or, should we reevaluate how informal institutions also play a bigger role than we may think in developed countries?
I will start with Stokes since it is what Robert ended with. I am quite a fan of her work, especially regarding the role of trust and skepticism. I think her book Democracy and the Culture of Skepticism provides a good complement to Kreps and Frye in that reputation and skepticism are features of modern complex democracies. Trust, on the other hand, could be related to more primitive or smaller-scale societies/towns. That is, trust might make more sense when there is personal interaction, as in local politics. At the state and national level, the constituency is too large.
Since Robert already criticized the article, I will turn to what seems to me the main unresolved question of clientelism: its double commitment problem (p. 611), especially on the voter’s side. In the 21st century, even many semi-democratic regimes have secret ballot. When patronage, transfers, clientelism, etc. take place at the individual level, how can the politician rest assured they voted for him? Earlier, they would force workers to go to the polls, give people the ballot beforehand, etc. Such control is ever harder as the polling booths are more protected. The study of clientelism, thus, makes much more sense to me at the group level (ie, if province X votes for such party/candidate, they will receive transfers/infrastructure).
Naturally, I should continue with Keefer and Vlaicu. I believe their two main insights (building reputations and relying on patrons) are interesting. However, it is not clear to me that radical expansions of franchise maintains clientelism… it seems rather a coin-flip. The conclusions (less public goods, less development, less infrastructure and property rights) seem reasonable. They do not provide any test of their argument. Might the two cases by ideal types rather than in accordance with Mill’s methods (especially of agreement or difference)? It would be helpful since they lack a large-N analysis. Also, I got lost in the model, but the more models I read the more I see them as a precise explanation with more or less predictive power than as any kind of test to a theory.
The next pair of articles: Kreps and Frye. Since I read Frye’s first, I didn’t fully understand the emphasis on reputation. Still now, I don’t really understand the view of reputation and rule of law as substitutes… although the NYC Jewish example was fun to read. Even after reading Kreps, I wonder if Frye’s study could be “imitated” but switching from political economy to labor economics or civil law –or from Russia to elsewhere, for that matter. For example, why/when do tensions arise between employers and employees? In civil law, how are separations and divorces handled? In these two examples, do reputation of the employer and rule of law complement each other? It would seem so. From a methodological point of view, a test of differences in means and associated significance would help to determine how different are the numbers in the tables.
Given that Kreps is a review, all I want to point out is that it is a very helpful review. Besides, I taught me to appreciated the “hidden” side of transaction costs –ie costs that are not readily apparent- as well as the implications of repeated games in the theory of the firm. I will raise a couple of issues/critiques of Kreps and my take on Wilkinson tomorrow in class –I feel my comment is long enough… See you all then!
I have several comments/questions on Stokes' piece:
1. I am not very clear about the difference between "clientelism and other materially oriented political strategies", such as pork-barrel politics in the US, that Stokes has tried to distinguish. She argues that what sets clientelism apart from pork barrel politics is that the main criterion for distribution in the former is whether the client will vote for the patron whereas in the latter, the emphasis is on whether the client is living in the patron's district. As Robert mentioned, available data suggest that rewards are targeted at regions and not just individual voters, and this could become more prevalent given the increasing difficulty of monitoring individual votes with the adoption of the secret ballot in many countries.
What then sets clientelism apart from other materially oriented political strategies and how useful would this distinction be for research?
2. Stokes suggests that clientelism could cause poverty, for instance, because patrons are actively preventing growth in order to perpetuate clientelism and preserve the existing political machine. Indeed, as Robert pointed out, clientelism could be both a cause and consequence of clientelism (think endogeneity issues here). However, the puzzle that strikes me is this: clientelism is essentially a system based on the distribution of benefits to clients in exchange for votes or electoral benefits to the patron. In a vicious cycle where poverty and clientelism reinforce each other, wouldn't the economic pie eventually shrink to such a small size that the patron would no longer be able to distribute and credibly promise any substantively significant benefits for their clients? Would this suggest that patrons would only prevent growth to a certain extent, and that clientelism would only reduce growth up to a certain level that is compatible with the survival of the system?
3. On whether parties target swing voters or core voters, intuitively, I would think that if parties have sufficient resources, they would target both core and swing voters to maintain their stronghold in core areas and to gain extra votes in marginal districts. Instead of studying whether parties target either core or swing voters, would it also be helpful to look at when parties might target both, but perhaps offer smaller benefits to both groups (i.e. a catch-all versus a focused strategy)?
1. In deciding whether to use public institutions, e.g. courts, or private institutions, e.g. reputation, actors would most certainly weigh the benefits and costs of using each option (assuming it is available in the first place). One difference between public and private institutions is that before actors can resort to using private institutions, they often have to bear some direct, individual cost of establishing that institution in the first place (e.g. the cost of establishing good reputation through sustained good behavior), whereas the cost of establishing public institutions is often borne collectively, e.g. by the state.
This leads to my question: In studying whether public and private institutions are substitutes or complements, i.e. whether actors are inclined to use both or substitute one for the other, would it be helpful to distinguish between the different levels of costs that actors would face from having to establish private institutions within the context of different kinds of economic or industry structure? To illustrate my point, consider a small agricultural community, e.g. a fishing village, compared to a large, developed cosmopolitan economy or perhaps even the global market. Due to the small size of the former, every actor's/firm's actions are visible to other actors, so reputation and trust can be easily established by simply following a set of good behavioral guidelines dictated by the norms prevailing in that community. In a large economy, however, actors often do not have complete information about other actors' actions, and actors/firms might thus have to invest additional resources to portray a good reputation. A clear example would be the rampant use of advertising by the large multinationals, e.g. celebrity endorsements by Nike etc etc.
Would differences in the cost of establishing and thus of using private institutions such as reputation and trust in different kinds of economies and industries affect our understanding of whether public or private institutions are complements or substitutes?
2. In considering whether public and private institutions are substitutes or complements, would it also be useful to consider the level of institutions that we are analyzing and the purpose with which actors are using these institutions? In some cases, e.g. in developed economies like the US, we often see firms competing against each other to provide good customer service (money-back guarantees/refunds etc) for their customers and customers would usually resort to private institutions, e.g. firms' reputation, in choosing which firms to transact with and usually deal with the firms themselves when disputes arise. On the surface it might seem as though actors are substituting private institutions for public institutions, but in reality, actors are able to use private institutions precisely because of the set of strong public institutions underlying these private institutions. For instance, in my example, these public institutions could be strong anti-trust legislation and the strong courts that have allowed the market economy to flourish and market participants to act privately.
3. In terms of the data, Prof. Frye has concentrated on Russia. I am just wondering how comparable the findings would be in other countries and how far we can generalize the findings in this paper to all transition and developing economies.
4. On a sidenote, I really like the quote from "Othello" at the start, because it is quite hard to find something like this in a Political Science paper (and also because it was one of the core literature texts I studied in high school). Interestingly, that comment is made by Iago, who does not care about reputation per se, but understands its importance to other people such as Cassio, and uses the importance people attach to reputation as a weapon with which to manipulate them. This sort of suggests that reputation (and private institutions) could become a double-edged sword, where actors could be forced by someone with malicious intent to act suboptimally in order to uphold their reputation.
In many respects, I like the Keefer and Vlaicu article. The argument about the effect of patron-client relations in voting, and its relationship to credible commitments is interesting. Though I usually hate this kind of modeling exercises with any number of implausible assumptions, I do think K&V are able to isolate some important tendencies, in particular those reviewed in page 21 and 22.
However, as part of an explanatory framework meant to account for differences in public goods provision of democracies by age, I think it is methodologically unconvincing. Surely, there are any number of differences between older and younger democracies – political norms being but one, although an important one. The testable implications of a model should be able to distinguish between that model and at least those alternative models that immediately suggest themselves to the reader. As most other analysis of this nature, Keefer et al. lack any reference to pre-existing differences in the norms politicians hold or are forced to abide by in terms of honesty. However, if these are significant, cross country analysis is difficult because of fixed effect problems.
Finally, I would like to see how they would propose to look at this empirically. Voter credibility through patrons seems in many cases difficult to distinguish empirically from other forms of tribalism, i.e. that different parties have different credibility levels with some groups in the population. In some cases, even the conceptual distinction seems unclear to me. In many European countries in parts of the 20th century, where organized labor controlled labor parties, or where – in some countries – organized business was influential in conservative parties, this seems to have been the case.
Overall I really enjoy this week's readings. Each of them explores the interplay of formal and informal institutions, which I agree is a very important yet long understudied topic, from different perspectives. I also learn a lot from these authors' integral use of quantitative and qualitative methods.
ReplyDeleteI want to start with Prof Frye's article. I have to admit that before reading this article I always intuitively assume private and public institutions surely complement each other, but the former can substitute the latter when the formal institutions are relatively weak and unreliable. My observation is based on China and my father's experience. Since the 1980s he has been doing business in China (i.e. he is among the first generation of foreign investors). When I was doing my research in Hong Kong, he kept telling me that how the business environment has been changed since the 1980s. In the past, he used to spend most of his effort establishing good relationships with other fellow merchants and the local party officials as everyone knows going to the public institutions and courts are not only costly but also unreliable. However, as the Communist Party of China started institutionalizing the business conducts since the 1990s, he is forced to deal with the formal institutions more often (but that does not mean they are better now). Therefore, I will tend to think private and public institutions can complement and substitute each other, depending on the specific context we are referring to. Maybe the question is: Why do informal (private?) institutions sometimes complement the formal (public?) ones, but in other cases substitute them? My speculation is it depends on how private institutions are established. For example, similar to Man Yan's previous inquiry, a lot of countries have displayed the importance of personal relations. As in Japan, the private institutions prevalent across the bureaucratic system have been long well-known, but they somehow do not weaken the Japanese state in providing public goods and services. Why is it the case? Maybe it is because to become bureaucrats you still have to pass the exams and master your professional expertise before others recognize you as part of the clientelistic network. Therefore, it seems like when informal institutions is built upon the formal ones, they will become much more compatible.
Before I move to Keefer et al., I have some more points for Prof Frye's piece. First, do business people avoid using courts to resolve their disputes only because the public institutions are unreliable? Or they choose not to appeal their cases to courts only because they are too costly? My impression is that no matter how reliable courts are people may still refuse to bring their disputes there because doing so is simply way too expensive and time-consuming. Second, while I think this survey should be conducted in China too and it will be very interesting to compare these two cases, why does this article choose to use "public" and "private" institutions, rather than "formal" and "informal" institutions?
Based on my understanding about the Japanese political system, I find Keeler et al.'s assumption quite problematic, and (like Cecelia said) I rarely have a strong sense of objection when reading an article - Why must building reputations through the investment of public infrastructure be the tradeoff of cultivating one's patronage network? Can't it be conditioned by the electoral system? I have these questions because the Japanese politicians did (not anymore now) manage to do both. Some may say Japan managed to achieve both of them because it is rich, but the thing is the Japanese politicians had been doing this way before the Japanese economy took off. In Japan, politicians' investment in their patron-client network did induce massive import in the building of public infrastructure while it has been so "privatized" that it is rather easy to tell who these collective goods are targeted. For instance, during the 1970s the Prime Minister Tanaka Kakuei proposed the plan of high speed railway, and interestingly he include several places his hometown Niigata Prefecture as the locations of train stations. Niigata used to be a very poor prefecture in Japan, so why they selected did arouse some concern. But since the plan covered the major metropolitan areas (Tokyo, Nagoya, and Osaka), no one paid attention until Tanaka's scandal broke out in the late 1970s. My point here is the dichotomy Keeler et al. suggest is not always the case, and I do not find a very compelling explanations why they think it must be the case. Even if it is really cheaper for politicians to control the patron rather than spend , then why must other voters support the politicians who do not take care of them but the patron? How do politicians resolve the agent problem? (In this sense, I will not say this approach will be any cheaper). I understand Keeler et al. try to find an alternative explanation since the traditional approach applied to distinguish different democratic systems fail to predict politicians' incentives to provide state spoils or public goods, but I do not really find their focus on "credibility" very impressive. They claim to endogenize it, but somehow they take it as given, but they do not explain what conditions a politician's "credibility."
ReplyDeleteMy last comment regards Wilkinson's book. I really love his book, and the paper he presented at the comparative politics seminar is also very interesting. As a student of ethnic politics, I am really amazed how patient he is. The data he collects is surely thorough and he clearly understands his limitations but still tries his best to find other possible measurements. His review about the competing theories of ethnic conflicts is my favorite part. Overall, I do learn a lot from his book, and he beautifully justifies why readers should buy his political explanations. Three things I would like point out though. First of all, why do the Hindu-Muslim conflicts tend to benefit the former? Do these conflicts really help the Hindu candidates? Wilkinson mostly uses Hindu politicians' statements to illustrate this logic, but I wonder if there are other possible factors contributing to Hindu candidates' landslide victories. Second, if the Muslim understand the Hindu's provocative actions act as part of their electoral campaign, why do they still choose to fight back that exactly cater to the Hindu's expectations? (Or they actually do not see this?) Third, according Wilkinson, the Hindu-Muslim riots prevail mostly because the state government chooses not to intervene not because they are unable to do so, is there any way to differentiate these two situations?
Chao-yo, in regards to your comment that in China private institutions can act as substitutes to public ones, I don't think Professor Frye's article was necessarily to say that private institutions are wholly complementary; but that the evidence suggests that they tend to function best as complements. i think it seems somewhat obvious that private institutions tend to act as substitutes when formal institutions are weak; however, in that case, how truly useful are these private institutions? Without enforcement, one's reliance on reputation (or some other informal institution) is only as partially useful as it would be in a country with strong institutions. I think Krep's chapter on corporate culture attempts to theorize and model this behavior, with a certain degree of success.
ReplyDeleteAdditionally, Frye's article (I think in the section about complements...I don't think this was Frye's original point) also brings up a salient point about how countries with strong institutions better disseminate information, thus reputations are useful because different organizations can access that information relatively cost free. In weak states, this information is less easy to access and corroborate.
I have only brief comments this week: first, I liked Wilkinson's book very much but I feel that in some ways it is unbalanced. Though he tests his first hypothesis that ethnic violence occurs in electoral districts that are highly competitive where politicians have an incentive to polarize ethnic communities, I feel that he less systematically analyzes the assertion that riots occur in states where the government is not using policing to stop violence because the political party in power is not dependent on minority communities to win elections. Wilkinson provides only brief anecdotal evidence to back up his claim that it is not a decline in state capacity that causes violence but the politicization of government resources. Im not sure how you would systematically test this theory but when contrasted to the tests he puts the first part of the argument through it looks somewhat deficient. My other critique is that this argument does not extend to other types of ethnic violence such as between caste violence or linguistic violence. Though he claims that Hindu-Muslim violence is common, he does not explain why politicians don't catalyze cross-caste violence when theoretically it could work in the same way as Hindu-Muslim violence.
ReplyDeleteThe Susan Stokes piece on political clientalism is interesting to me in that it highlights something that common people usually suspect of politicians, but something, as the author acknowledges, that has not been thoroughly explored and analyzed in literature. The way the author takes up each portion of the puzzle and relates the literature associated is helpful in understanding the aspects of how, when, and why clientelism works. I am particularly intrigued by the idea of clientelism working to produce poverty and inequality, which in turn perpetuate it. If institutions such as the secret ballot and electoral rules work toward mitigating clientalism, is it only when economic growth has rendered the voters less amenable to temptation by politicians? While accepting that clientalism woks, and is detrimental, the author could have delved deeper into the relative effectiveness of different forms of clientelism: payoffs in the form of material goods (immediate but short term in effect) vs. patronage (slower but long term), protection vs. security etc.
ReplyDeleteI believe Stokes and Keefer&Vlaicu offer interesting and complementary perspective on the consequences of clientelism. They both (directly or indirectly) refer to two types of consequences: poverty/wefalre and policy. 1) For the poverty/welfare one, Stokes underlines the fact that clientelism may cause poverty, whether because the leaders voluntarily do not want income to grow or because of the declining relative productivity of monopolized-goods sectors. Keefer&Vlaicu, on the other hand, estimate that welfare may improve (for a short length). 2) On the broader policy front, Keefer&Vlaicu agree with Stokes that clientelism will bring lower provision of public goods compared to credible politician competitors with no patron. However, they add that compare to non-credible politicians, the ones with patrons may be better off.
ReplyDeleteBesides, Stokes seems to make the point, when she discusses the effectiveness of clientelism, that it is somehow surprising that political parties pursue clientelism even in cases (3 of the 4 possibilities of the game she exposes) that it is ineffective. She states that parties usually do not opt for uneffective strategies. I am not as surprised as her, because I have the impression she is analyzing parties and politicians as if they were ONLY seeking votes. I think it is not surprising that some politicians will, for example, try to get not necessarily votes, but also further benefits and possibilities (like a job once off politics) or to get personal immediate economic benefits. This would not fit in her restrictive definition of clientelism, I believe, since precisely they would not been looking for electoral support: but are politicians being driven by a definition of clientelism??
I found Wilkinson’s contention—that the degree of electoral competition influences whether or not ethnic violence occurs and the response of the state—interesting and well argued. I especially liked how he systematically tackled all the other arguments for why ethnic violence occurs chapter by chapter. Having recently read Varshney who argues civil society, especially formal associational networks that cross ethnic lines, prevents conflict, I would have liked Wilkinson to have addressed him in little more depth and to see if there was overlap between the arguments: if violence was low in areas that had both interethnic civic networks and low political competition characterized by intra-Majority party competition. If so, couldn’t both be a function of something else?
ReplyDeleteI guess I'm the contrarian this week, because the Stokes article struck me as a nearly Platonic fusion of trends I hate about American political science.
ReplyDelete1. Her definition of clientalism makes no sense, which she acknowledges, because the exchange of votes for goods is plagued by uncertainty and available data suggests that rewards are targeted at regions and not just individual voters (this is the Snyder research on the US she references.)
2. She sets up a pointless game that demonstrates that, in a one-off scenario, both voters and clientalistic parties defect from their agreement. But then she points out that this doesn't comport to reality, and suggests information and reward size are important. Which further reinscribes my impression of game theory as a valuable literature that enables us all to say obvious things in complicated ways.
3. She then points out that some of the empirical research indicates that parties target swing voters, but others target core voters. No clear reason when they choose which.
4. Then she suggests clientalism may cause poverty. Or vice versa. ("it is probably both a cause and a consequence.")
So, in summary, there is a pattern of behavior in the world in which politicians buy votes with favors. These can be individual or collective. Not too sure who gets the favors or why. Seems to be mostly poor places, also not sure why. We should all probably do some more work to figure all this out, and it should probably involve a game somehow.
Maybe I'm just feeling cranky, but jumpin' Jeehosephat, that selection leaves me shaking my head.
I look forward to discussing it tomorrow to figure out what it is I'm missing here!
First I agree with Robert, I don't find the case by Stokes compelling at all. I must also say that I was already biased because after reading another article by Stokes I was less than impressed so I did not approach this with very high expectations.
ReplyDeleteSecond, I found the concept of focal points outlined by Kreps both interesting and useful. In particular I believe it has considerable explanatory power to answer why some principle or rule is used naturally to select a certain behavior in view of many possible equilibrium behaviors (121). For someone who is interested in cooperation like me, this idea seems promising.
Third, according to Frye informal and formal institutions are complements. My question is if in more develop nations should we expect this complementarity to decrease given the increased certainty in formal mechanisms? Or, should we reevaluate how informal institutions also play a bigger role than we may think in developed countries?
This comment has been removed by the author.
ReplyDeleteI will start with Stokes since it is what Robert ended with. I am quite a fan of her work, especially regarding the role of trust and skepticism. I think her book Democracy and the Culture of Skepticism provides a good complement to Kreps and Frye in that reputation and skepticism are features of modern complex democracies. Trust, on the other hand, could be related to more primitive or smaller-scale societies/towns. That is, trust might make more sense when there is personal interaction, as in local politics. At the state and national level, the constituency is too large.
ReplyDeleteSince Robert already criticized the article, I will turn to what seems to me the main unresolved question of clientelism: its double commitment problem (p. 611), especially on the voter’s side. In the 21st century, even many semi-democratic regimes have secret ballot. When patronage, transfers, clientelism, etc. take place at the individual level, how can the politician rest assured they voted for him? Earlier, they would force workers to go to the polls, give people the ballot beforehand, etc. Such control is ever harder as the polling booths are more protected. The study of clientelism, thus, makes much more sense to me at the group level (ie, if province X votes for such party/candidate, they will receive transfers/infrastructure).
Naturally, I should continue with Keefer and Vlaicu. I believe their two main insights (building reputations and relying on patrons) are interesting. However, it is not clear to me that radical expansions of franchise maintains clientelism… it seems rather a coin-flip. The conclusions (less public goods, less development, less infrastructure and property rights) seem reasonable. They do not provide any test of their argument. Might the two cases by ideal types rather than in accordance with Mill’s methods (especially of agreement or difference)? It would be helpful since they lack a large-N analysis. Also, I got lost in the model, but the more models I read the more I see them as a precise explanation with more or less predictive power than as any kind of test to a theory.
The next pair of articles: Kreps and Frye. Since I read Frye’s first, I didn’t fully understand the emphasis on reputation. Still now, I don’t really understand the view of reputation and rule of law as substitutes… although the NYC Jewish example was fun to read. Even after reading Kreps, I wonder if Frye’s study could be “imitated” but switching from political economy to labor economics or civil law –or from Russia to elsewhere, for that matter. For example, why/when do tensions arise between employers and employees? In civil law, how are separations and divorces handled? In these two examples, do reputation of the employer and rule of law complement each other? It would seem so. From a methodological point of view, a test of differences in means and associated significance would help to determine how different are the numbers in the tables.
Given that Kreps is a review, all I want to point out is that it is a very helpful review. Besides, I taught me to appreciated the “hidden” side of transaction costs –ie costs that are not readily apparent- as well as the implications of repeated games in the theory of the firm. I will raise a couple of issues/critiques of Kreps and my take on Wilkinson tomorrow in class –I feel my comment is long enough… See you all then!
I have several comments/questions on Stokes' piece:
ReplyDelete1. I am not very clear about the difference between "clientelism and other materially oriented political strategies", such as pork-barrel politics in the US, that Stokes has tried to distinguish. She argues that what sets clientelism apart from pork barrel politics is that the main criterion for distribution in the former is whether the client will vote for the patron whereas in the latter, the emphasis is on whether the client is living in the patron's district. As Robert mentioned, available data suggest that rewards are targeted at regions and not just individual voters, and this could become more prevalent given the increasing difficulty of monitoring individual votes with the adoption of the secret ballot in many countries.
What then sets clientelism apart from other materially oriented political strategies and how useful would this distinction be for research?
2. Stokes suggests that clientelism could cause poverty, for instance, because patrons are actively preventing growth in order to perpetuate clientelism and preserve the existing political machine. Indeed, as Robert pointed out, clientelism could be both a cause and consequence of clientelism (think endogeneity issues here). However, the puzzle that strikes me is this: clientelism is essentially a system based on the distribution of benefits to clients in exchange for votes or electoral benefits to the patron. In a vicious cycle where poverty and clientelism reinforce each other, wouldn't the economic pie eventually shrink to such a small size that the patron would no longer be able to distribute and credibly promise any substantively significant benefits for their clients? Would this suggest that patrons would only prevent growth to a certain extent, and that clientelism would only reduce growth up to a certain level that is compatible with the survival of the system?
3. On whether parties target swing voters or core voters, intuitively, I would think that if parties have sufficient resources, they would target both core and swing voters to maintain their stronghold in core areas and to gain extra votes in marginal districts. Instead of studying whether parties target either core or swing voters, would it also be helpful to look at when parties might target both, but perhaps offer smaller benefits to both groups (i.e. a catch-all versus a focused strategy)?
Some thoughts on Prof. Frye's paper:
ReplyDelete1. In deciding whether to use public institutions, e.g. courts, or private institutions, e.g. reputation, actors would most certainly weigh the benefits and costs of using each option (assuming it is available in the first place). One difference between public and private institutions is that before actors can resort to using private institutions, they often have to bear some direct, individual cost of establishing that institution in the first place (e.g. the cost of establishing good reputation through sustained good behavior), whereas the cost of establishing public institutions is often borne collectively, e.g. by the state.
This leads to my question: In studying whether public and private institutions are substitutes or complements, i.e. whether actors are inclined to use both or substitute one for the other, would it be helpful to distinguish between the different levels of costs that actors would face from having to establish private institutions within the context of different kinds of economic or industry structure? To illustrate my point, consider a small agricultural community, e.g. a fishing village, compared to a large, developed cosmopolitan economy or perhaps even the global market. Due to the small size of the former, every actor's/firm's actions are visible to other actors, so reputation and trust can be easily established by simply following a set of good behavioral guidelines dictated by the norms prevailing in that community. In a large economy, however, actors often do not have complete information about other actors' actions, and actors/firms might thus have to invest additional resources to portray a good reputation. A clear example would be the rampant use of advertising by the large multinationals, e.g. celebrity endorsements by Nike etc etc.
Would differences in the cost of establishing and thus of using private institutions such as reputation and trust in different kinds of economies and industries affect our understanding of whether public or private institutions are complements or substitutes?
2. In considering whether public and private institutions are substitutes or complements, would it also be useful to consider the level of institutions that we are analyzing and the purpose with which actors are using these institutions? In some cases, e.g. in developed economies like the US, we often see firms competing against each other to provide good customer service (money-back guarantees/refunds etc) for their customers and customers would usually resort to private institutions, e.g. firms' reputation, in choosing which firms to transact with and usually deal with the firms themselves when disputes arise. On the surface it might seem as though actors are substituting private institutions for public institutions, but in reality, actors are able to use private institutions precisely because of the set of strong public institutions underlying these private institutions. For instance, in my example, these public institutions could be strong anti-trust legislation and the strong courts that have allowed the market economy to flourish and market participants to act privately.
3. In terms of the data, Prof. Frye has concentrated on Russia. I am just wondering how comparable the findings would be in other countries and how far we can generalize the findings in this paper to all transition and developing economies.
ReplyDelete4. On a sidenote, I really like the quote from "Othello" at the start, because it is quite hard to find something like this in a Political Science paper (and also because it was one of the core literature texts I studied in high school). Interestingly, that comment is made by Iago, who does not care about reputation per se, but understands its importance to other people such as Cassio, and uses the importance people attach to reputation as a weapon with which to manipulate them. This sort of suggests that reputation (and private institutions) could become a double-edged sword, where actors could be forced by someone with malicious intent to act suboptimally in order to uphold their reputation.
In many respects, I like the Keefer and Vlaicu article. The argument about the effect of patron-client relations in voting, and its relationship to credible commitments is interesting. Though I usually hate this kind of modeling exercises with any number of implausible assumptions, I do think K&V are able to isolate some important tendencies, in particular those reviewed in page 21 and 22.
ReplyDeleteHowever, as part of an explanatory framework meant to account for differences in public goods provision of democracies by age, I think it is methodologically unconvincing. Surely, there are any number of differences between older and younger democracies – political norms being but one, although an important one. The testable implications of a model should be able to distinguish between that model and at least those alternative models that immediately suggest themselves to the reader. As most other analysis of this nature, Keefer et al. lack any reference to pre-existing differences in the norms politicians hold or are forced to abide by in terms of honesty. However, if these are significant, cross country analysis is difficult because of fixed effect problems.
Finally, I would like to see how they would propose to look at this empirically. Voter credibility through patrons seems in many cases difficult to distinguish empirically from other forms of tribalism, i.e. that different parties have different credibility levels with some groups in the population. In some cases, even the conceptual distinction seems unclear to me. In many European countries in parts of the 20th century, where organized labor controlled labor parties, or where – in some countries – organized business was influential in conservative parties, this seems to have been the case.