For Geddes, I really like the way she sets her argument and her detailed descriptions for each case. To begin with, I find her conclusion very convincing, especially when she concludes that there exists no direct link between development and the improvement of bureaucracy (i.e. administrative reform). Even though reform surely does not guarantee the increase in government efficiency, this point more or less weakens Kurtz et al. argument that development breeds better governance, even though how to define governance remains a highly controversial issue. Second, to certain degree, her argument can be applied to other types of reform. For example, I wonder if electoral reform may be only possibly initiated when competing parties share relatively equal power in parliament. The cases that come to my mind are Japan and Taiwan. In the former case, the LDP stepped into the negotiation of electoral reform with opposition parties only after it lost for the very first time since 1955 in 1993. Even though the LDP may control most access to state spoils than opposition parties, facing great public demands of reform, the party head still made concession to start changing the Japanese electoral system (e.g. from multi-member SNTV to mixed single-member & PR system). For Taiwan, while the KMT had been in power for over five decades, it accepted to talk to the DPP regarding the reforms of electoral system when two parties possess relatively equal strength in the 2000s with one party holding the legislature and the other party controlling the executive branch. However, one thing I am not satisfied is Geddes' model actually does not consider party cohesion as one parameter but solely focuses on the share of patronage between two parities, which in turn induces a conceptual gap between her model and the case studies. Her theoretical story seems to contain two stages: The first is at the inter-party level, and the second regards the intra-party level (i.e. the degree of party cohesion). I think it would be better if she could mention this when she concludes the model section in her article (i.e. p. 377).
I also find Shleifer et al.'s article on corruption very interesting. Their reference to the perspective of organization into the analysis on bribing and government corruption is really amazing. Based on their findings, some points may be worthy of further discussion: Among the three scenarios, the level of bribes is lowest when there are competing public agencies which offer the same government goods, so the policy implication is the state should increase the competition between different public bureaus. However, there is one potential problem: Not every government service/good is mobile (e.g. transaction about land), so people may be restricted to one supplier of government goods. Also, to foster competition among public agencies, the state may need relatively rich resources to establish (e.g. the US), which is very unlikely when corruption is rampant within the government administration. Second, they seem to imply that democracies suffer less corruption due to the presence of stronger public pressure and more political competition. Yet is this really the case? Singapore may more or less counter this idea. Is Singapore an anomaly, or does the case of Singapore show that democracy may not be necessary to build a clean government? To echo Shleifer et al.'s argument, what matters seem to be state capacity, but I will focus more on if the state can effectively detect, rather than "coordinate," corruption. Finally, Shleifer et al. discuss the similarity between bribe and tax, and argue the former is different due to its illegality. Regarding the issue of legality, what is fee's legality then? My question is: Why do the government have to rights to collect fee and why will fees not become another source of corruption (with theft) and bribe?
My last comment regards Huber et al. and Moe's pieces. Before tacking the relationship (or even competition to certain degree) between the bureaucrats and the elected politicians, they all spend much effort to differentiate their preferences and roles within the political system, and how their interaction varies across different institutional settings (i.e. presidential system and parliamentary system). On the one hand, Moe talks about many things in his article, so sometimes I feel a little bit hard to tease out which groups (ordinary citizens, interest groups, politicians, and bureaucrats) he is referring to. But overall I find his analysis quite compelling. He especially does a very excellent job in distinguishing political transactions and economic transactions by highlighting that losers in the former are trapped in the structure constructed by the winners. On the other hand, while in Moe's words Huber et al.'s argument may be a conventional work of social choice to explore how politicians overpower bureaucrats, Huber et al. still provides some nuances to explain why in some countries the bureaucrats enjoy more discretion, though such discretion as Moe criticized is "delegated" from politicians. Also interestingly enough, while Huber et al.'s analysis obviously downplay the role of constituents, Moe tries very hard to include social actors into his analysis.
However even so, treating politician/bureaucrat relationship as a principal-agent relationship, I still find it very hard to say which one must be the principal. I believe that is exactly one of Moe's points. Very often researchers jump to the conclusion that no matter how much autonomous bureaucrats are they are still subject to the control of politicians. To be sure, it has been a central theme of the studies of Japanese post-war economic growth. Did the bureaucrats, dominating the political arena, lead Japan to prosperity as Chalmers Johnson argued? Or, is it the LDP politicians who dictated bureaucrats what to do? I have no intention to pull out the whole literature here, but there still one one nuance that I believe worthy of consideration: Do these two groups really exist exclusively to each other? I mean, before exploring which one controls the political realm, it may be necessary to see how they relate to each other. At least as shown by the Japanese situation, many politicians turn out to be senior/retired members of Japanese administration (so it is a source of political control over bureaucracy), but it is also the case that many bureaucrats and politicians go to the same university (e.g. University of Tokyo - Faculty of Law), which makes seniority a less powerful controlling mechanism as there are very complex personal relations between these two groups. To complicate the picture, some may even contend that it is neither politician nor the bureaucrats but the party, the LDP, that controlled the whole system through the Policy Research Committee, a party organ that convened party heads, Diet members, and bureaucrats. Therefore, I believe there are always more stories to tell about the contested relationship between the elected politicians and the bureaucrats as very often they may highly correlated to each other or may be even under the mandate of a third party (e.g. the majority party). In this sense, Geddes may be right for bringing the party system into the analysis.
Chao-yo, Im not sure I understand with your criticism of Schleifer and Vishny's article in regards to mobile services. Land transactions are not mobile, but, for instance, citizens can go to one of many offices to deal with the bureaucracy of the transaction. Are there truly any government transactions that need to be dealt with in one locality? (the only one I can think of is customs, where transactions are limited to a point of entry) I do agree that establishing several offices for government transactions is costly, and that this may be a hinderance for less developed countries.
Additionally,if one is to believe TI scores, democratic countries tend to be less corrupt. Singapore is a consistent outlier in those rankings, and is generally in the top 10 least corrupt countries. Not being an expert on Singapore, I cannot say with any certainty why its seen as less corrupt, but perhaps there are other mechanisms involved, such as effective implementation of laws.
My own criticism of the S&V article is that I wonder if there is yet a third structure of corruption that is missing. S&V identify two structures, hierarchal and nonhierarchical, with the hierarchal structure being more conducive to growth. Im wondering if there are several "pyramids" of corruption, but they are not fully horizontal. Perhaps different agencies collude with each other to ensure one cost for a certain transaction (lets say, building permits and taxes), but due to competition between different ministries, not all the agencies work in collusion. For instance, a new resident is applying for a residency permit, and needs to show a work permit, a housing permit, a letter of invitation, and a visa; perhaps the individual is asked to bribe the visa agency and knows that this bribe will ensure she gets a labor permit and a housing permit, but still needs to pay a separate bribe for a visa. That sounds extremely convoluted and Im not sure my point is clear...And in thinking about it, this might have no real effect on growth rates (as its effects may be similar to that of nonhierarchical corruption), so I just wasted a lot of space trying to clarify this idea for no reason.
@ Chao-yo and Kim. I like your thoughts on the S&V paper. My own take on the relationship between corruption and democracy is that the large association between TI & Freedom House scores is because democracy does not develop successfully in countries where there is not a norm of universalism. Societies in which status is understood by distribution of power and where actors do not expect equal treatment by the state will be both more corrupt and less democratic. In this view Singapore is an outlier.
I appreciate Schleifer and Vishny’s article on corruption that attempts to derive a model of corruption and come up with some potential prescriptions for its reduction. This is a major issue as anticorruption is now a major industry. However in my opinion S&V overlook some key distinctions about corruption that call their remedies for corruption into question. It is very important to make the distinction about what the main spoils of corruption are and who is the principal agent of corruption. Is it the state (many post-communist societies) or private actors? In post-colonial societies the principal agent of corruption might be found elsewhere as the state is often only one of the actors competing for control. This distinction has great importance for both the treatment and the diagnosis of corruption. Another type of corruption overlooked by S&V but I think important in our discussions on institutions is electoral corruption. In many developing societies elites can exchange resources for votes so that voting becomes a commodity rather than a right or a choice. Though this type of corruption is overlooked in their article, this has serious ramifications for how the populace is tied to politics and perceptions about governance. In this case it is hard to see how their prescriptions for curbing corruption (increased economic and political competition) would be effective. Would increased political competition result in an increase in vote trading or would perceptions about what voting is actually change? Even increased political competition cannot stymie corruption: after the Orange Revolution one could but a seat in the Ukrainian parliament. It is clear that in many regimes popular perceptions of government corruption remain despite changes in government. Finally, V&S contend that better and increased accounting would help to reduce corruption levels. Here again a distinction on the nature of the principal actor is important: if it is the state, then increasing bureaucratic efforts to combat corruption seems implausible and counter-intuitive
@ Kim: Yeah, you are right. I did not have a very clear idea about an example when I was writing this point. But I will give it a try again. Regarding land, yes, citizens may go to different offices that provide the same service or government goods. But there may be one condition that citizens cannot do so. For example, some countries require lands to be valued before they can be used by private individuals for agricultural or industrial production. In other words, the land owners need to get permits from government to turn their land assets into profits, and clearly the industrial use of land is much more profitable in some countries (especially the small ones where land resource is more sparse). Under this condition, my imagination is it is very hard for citizens to have alternative choices and very often they need to bribe local officials to have their lands for more profitable use. If we use the case of visa application, some countries, if the bureaucrats are highly predatory, may restrict citizens' options by saying: People from a given state should go to apply for visa in a particular state city (I am just making this up).
Another example is resident registration (as you also mention). In some countries (e.g. China) resident identity (rural vs. urban) can be a very precious asset as different identity attached a totally different set of social privileges. Therefore, when citizens' choices are fixed to a specific locality, corruption may become prevalent. In this sense, I think my point is more like an elaboration of Shleifer et al.'s argument rather than a criticism?
Regarding your "pyramid" point, I think you are right that there may be simultaneous horizontal coordination of collection of bribes, though incomplete and unstable, within each level of administration. Actually, that is exactly the topic that many scholars of Chinese politics are tackling right now.
I really enjoyed the banter between Kurtz and Schrank and the Kaufmann et al. The former make the case that current measures of governance are flawed and that economic growth might lead to better governance. Even before reading Kauffman et al.’s rebuttal, I found two points of contention. The first is the narrow definition of what constitutes good governance. As Kaufman et al note, Kurtz and Shrank’s focuses only on government effectiveness eschews other interdependent measures, subtypes and concepts of governance that are key to gauging the extent to which governance can be considered “good”. Political accountability (elections, and the degree to which the are free and fair), formal oversight institutions (independent legislature, judiciary, autonomous anti-corruption body), the role and functions of the media, civil society; decentralization (political, fiscal and administrative); and the role of the private sector (CSR initiatives, corporate governance; privatization of key industries) are other variables that affect public administration performance. Not taking these other measures of governance into account and only focusing on how well the public administration delivers services only gives one aspect of good governance. The other aspect that I kept shaking my head while reading was their tentative conclusion that economic development could be the impetus for better governance. The examples that most point to are South Korea and Singapore. Yet these are small outliers that buck a bigger trend towards the opposite conclusion. Here I thought Kaufmann et al really did a good job of putting the argument (governance -> economic growth) into its proper place by showing the overwhelming amount of work pointing in this direction. Reflecting on my experience working on governance initiatives for Transparency International in the Middle East, I have to say that my perspective is more in line with the Kaufmann et al camp. I have a feeling Kurtz and Schrank would say I’m too close to the trees to see the forest though.
I just posted my referee report on Geddes on courseworks... Aside what I have written there on it, I wonder how she would analyze the recent emphasis given by Pinera in Chile on the merit of civil servants. Maybe as the exception she finds in the Brazilian case? Whether her analysis can be extended beyond the cases she studies is a clear question that stand out of her article...
I believe Geddes's article can interestingly be seen as an analysis of the pre-period of Huber & Shipan's book. In a certain way (and although it remained limited in the Latin American cases studied), she looks at the decision of the politicians to (potentially) give more capacity to bureaucracies, whereas Huber and Shipan take bureacratic capacity as a requirement.
Specifically on Huber and Shipan: I found their argument very convincing, but one aspect of it still raises question for me: the definition of a bureaucrat. They define it as an “agent who plays a central vote during policy implementation after statutes are adopted” (p.82), and directly specify that a bureaucrat can also be a president, a governor, a cabinet minister, etc. How do they account for parliamentary systems when cabinet ministers are also legislators? How can they compare these systems with presidential ones? Do they consider cabinet ministers as bureaucrats in certain cases and not in others?
@ Osama: I would like to follow Osama's discussion on Kurtz et al. To summarize my point, I do not see South Korea and Singapore as outliers. They may seem to be so, as I argue below, is because they selectively choose a narrow portion of the notion of governance, and because they conflate dictators with the whole bureaucracy, which is not necessarily weak in both cases.
When I was reading this piece, while I was amazed by their insightful criticism regarding current data most political scientists use to measure governance, I have to say their concept of "governance," as Kaufman et al. clearly indicate, is rather ambiguous and selectively narrow. In the case of South Korea, I am not sure if it is appropriate to equate the military dictatorship to the whole bureaucracy. This applies to Taiwan as well, even though part of the administration might be predatory and caused much distortion as Kurtz et al. point out, the problem is we are not sure which part of administration/bureaucracy promotes economic growth. Maybe like Japan it is those technocrats who held the offices of the Ministry of Finance, rather than those who are in charge of foreign trade (or vice versa?), that contributed the rise of Four Tigers? To explain why Kurtz et al find the possibility of the reverse causal link between governance and development, one answer might be their inconsistent and unclear definition of "governance," and the assumption that treats bureaucracy as a single actor.
The first comment concerns the back and forth between Kurtz and Schrank (KS) and Kaufmann et al. There are so many articles that deal with economic growth and governance/democracy… AJR on settler mortality is a robust finding that institutions matter. Alternatively, Lipset’s modernization theory also has many supporters. Why can’t we simply accept that the two are intertwined and the feedback is bidirectional? The more articles I read, the more it seems causality attributions depend on data sets, time periods, controls, choice of instrumental variables, and what not! The World Bank governance indicators are regarded with suspicion even within the World Bank. My boss there this summer was reluctant to use them because of measurement error and operationalization problems… The biases in the World Bank (or Freedom House) should not be surprising if we look at the composition of the institution. Until recently, it was vastly dominated by economists who saw administration as interference and not as a guarantee of a fair process. That being said, I agree with Osama on his critique of KS. In the end, both articles should admit the only thing that is obvious: we don’t really know much thanks to the World Bank indicators because of period limitation, sensitivity to variables, etc.
Briefly about Geddes, I agree both with Chao-yo and Florence. the article is internally solid. However, is it externally valid? I honestly don’t know enough about Latin American administrative reforms. (Only if Huber and Shipan had focused on Latin America!)
Shleifer and Vishny: so good! I susbscribe Chao-yo’s summary rather than repeating it. An economic approach fits very well with the problem of corruption, especially when undertaken by individuals. In these circumstances, an expected utility approach seems sensible. The individual wants to know whether the monetary gains of corruption (g) minus the cost of being caught times its probability (p) is bigger than not doing anything (zero). That is, an individual will be corrupt if, for E[u(corruption)] = gains - p(caught)*cost|caught, E[u(c)] > 0. Simple, but powerful enough to give way to the three scenarios of corruption. The issue is a bit more complicated in organizational contexts, however. Imagine most around you are being corrupt like in the KGB or Stasi. Does the individual really make any calculations? Not really, if he was to be transparent and accountable to the Communist Party instead of his boss at KGB or Stasi he could be in big trouble. The point, then, is that when the individual is organizationally constrained other factors are more important than the personal monetary gains of selling public goods.
Few times have I found myself completely opposing an argument that an author presents. While I can appreciate the utility of rethinking concepts that we thought were good once and for all I was surprised by the proposal of Kurtz and Schrank. As Osama pointed out, is not only their narrow definition of good governance but their contention that it is in fact growth what causes good governance. I find that proposal lacking support for the following reasons: a)unless countries have institutions that can channel economic activity properly there will not be economic growth. b)economic growth requires secure property rights, a lawful environment, certainty,accountability, transparency etc etc that only derive from good governance. c)While there can be some economic growth without good governance this usually tends to be concentrated in the hands of few. Is that really the type of growth that we want? d)I find the policy implications of their paper quite dangerous. We do not know the mechanism but somehow growth will cause good governance.
I was a little disappointed by the defense of KKM. Again, while I can understand the value of methodology I do not believe that discussing halo effects is what will move the discipline to better theories. Yes, methodology matters but stats is just a tool that basically relies on qualitative classifications for its usefulness. In the defense of why good governance leads to growth, I expected a debate about ideas, not about tools.
@Chao-yo on Shleifer et al: The authors argue that the extent of bribery is lowest where there are competing public bureaus providing the same public good and your concern is that not every government good or service is "mobile", so taxpayers might be restricted to one particular bureau.
Instead of framing the counter-argument in terms of the "mobility" of goods or services, perhaps you could also consider other factors that might restrict the number of government bureaus/organizations in each area. For instance, public monopolies are often natural monopolies in industries (with huge fixed costs) that can only support one company, e.g. the standard example of utilities (water/electricity) providers. The argument that the industry cannot support more than one firm is often proposed in support of nationalizing the "natural monopoly" in this instance.
In addition, I think Shleifer et al's definition of corruption is rather narrow. They define "government corruption as the sale by government officials of government property for personal gain". Indeed, this aspect of corruption--where contracts might be awarded to less efficient or capable contractors in return for bribes--appears to have a direct, visible impact on development. But what about less visible forms of corruption that do not involve any exchange of property, for instance, when officials claim for expenses that they have incurred in a personal capacity, e.g. the corruption scandal in the UK cabinet earlier this year? This is only one example, but I am certain there are many different types of corruption in practice. Hence the interesting question is whether different types of corruption could have different effects on growth and development, or whether all types of corruption are equally detrimental to growth and development.
On Kaufmann et al’s rebuttal of KS’s critique, I agree very much with Kaufmann et al and I just have one small comment to make: Kaufmann et al mentioned that in looking for potential sources of bias in their data sources, they have examined whether the difference between the assessments of think-tanks and firm surveys was systematically correlated with the political orientation of the government in power in the countries being rated. They found that this was generally not the case and concluded that this additional source of bias is dubious. (pg 4) Although there is no correlation between the assessments of think-tanks and the political orientation of the government in power, could a correlation exist between the assessments of think-tanks and the ideologies/political orientation of the largest opposition party or of other smaller but nonetheless influential opposition parties? In autocratic regimes, simply investigating whether a correlation exists between the assessments of think-tanks and the political orientation of the government in power would probably suffice, but in democracies where an entire spectrum of political ideologies is often present and freely advocated by different groups, think-tanks might align themselves with a whole range of political ideologies, not necessarily that of the government in power, e.g. in the US. In this case, the ratings and recommendations of NGOs and think-tanks might still be coloured by different ideological orientations, though not necessarily reflecting those of the government in power.
I was positively surprised by the Geddes article. I think she combines empirical studies and the basic logic of game theoretical insights in a fruitful way; rather than relying on abstract mathematics and then testing quantiative results, she relies on what Alfred Marshall called "short chains", basic behavioral assumptions, the plausibility of which are part of the empirical testing.
I also liked Moe's piece, although it could probably have been shortened by 50 percent without losing content. I think his insights into something he says he knows little about - parliamentary systems, and their effect on governance and bureaucracy - are interesting, although he disregards the separation of powers that most parliamentary systems still have (supreme or constitutional courts in all European countries, House of Lords in the UK etc). However, these "checks" in European systems come into play only in fundamental questions, and not in nitty-gritty politics as usual.
As in the Geddes article, I like the way in which he employs rational choice theory very specifically - short chains - to explain a logic of a phenomenon (repeated games, uncertainty about outcomes etc.).
For Geddes, I really like the way she sets her argument and her detailed descriptions for each case. To begin with, I find her conclusion very convincing, especially when she concludes that there exists no direct link between development and the improvement of bureaucracy (i.e. administrative reform). Even though reform surely does not guarantee the increase in government efficiency, this point more or less weakens Kurtz et al. argument that development breeds better governance, even though how to define governance remains a highly controversial issue. Second, to certain degree, her argument can be applied to other types of reform. For example, I wonder if electoral reform may be only possibly initiated when competing parties share relatively equal power in parliament. The cases that come to my mind are Japan and Taiwan. In the former case, the LDP stepped into the negotiation of electoral reform with opposition parties only after it lost for the very first time since 1955 in 1993. Even though the LDP may control most access to state spoils than opposition parties, facing great public demands of reform, the party head still made concession to start changing the Japanese electoral system (e.g. from multi-member SNTV to mixed single-member & PR system). For Taiwan, while the KMT had been in power for over five decades, it accepted to talk to the DPP regarding the reforms of electoral system when two parties possess relatively equal strength in the 2000s with one party holding the legislature and the other party controlling the executive branch. However, one thing I am not satisfied is Geddes' model actually does not consider party cohesion as one parameter but solely focuses on the share of patronage between two parities, which in turn induces a conceptual gap between her model and the case studies. Her theoretical story seems to contain two stages: The first is at the inter-party level, and the second regards the intra-party level (i.e. the degree of party cohesion). I think it would be better if she could mention this when she concludes the model section in her article (i.e. p. 377).
ReplyDeleteI also find Shleifer et al.'s article on corruption very interesting. Their reference to the perspective of organization into the analysis on bribing and government corruption is really amazing. Based on their findings, some points may be worthy of further discussion: Among the three scenarios, the level of bribes is lowest when there are competing public agencies which offer the same government goods, so the policy implication is the state should increase the competition between different public bureaus. However, there is one potential problem: Not every government service/good is mobile (e.g. transaction about land), so people may be restricted to one supplier of government goods. Also, to foster competition among public agencies, the state may need relatively rich resources to establish (e.g. the US), which is very unlikely when corruption is rampant within the government administration. Second, they seem to imply that democracies suffer less corruption due to the presence of stronger public pressure and more political competition. Yet is this really the case? Singapore may more or less counter this idea. Is Singapore an anomaly, or does the case of Singapore show that democracy may not be necessary to build a clean government? To echo Shleifer et al.'s argument, what matters seem to be state capacity, but I will focus more on if the state can effectively detect, rather than "coordinate," corruption. Finally, Shleifer et al. discuss the similarity between bribe and tax, and argue the former is different due to its illegality. Regarding the issue of legality, what is fee's legality then? My question is: Why do the government have to rights to collect fee and why will fees not become another source of corruption (with theft) and bribe?
My last comment regards Huber et al. and Moe's pieces. Before tacking the relationship (or even competition to certain degree) between the bureaucrats and the elected politicians, they all spend much effort to differentiate their preferences and roles within the political system, and how their interaction varies across different institutional settings (i.e. presidential system and parliamentary system). On the one hand, Moe talks about many things in his article, so sometimes I feel a little bit hard to tease out which groups (ordinary citizens, interest groups, politicians, and bureaucrats) he is referring to. But overall I find his analysis quite compelling. He especially does a very excellent job in distinguishing political transactions and economic transactions by highlighting that losers in the former are trapped in the structure constructed by the winners. On the other hand, while in Moe's words Huber et al.'s argument may be a conventional work of social choice to explore how politicians overpower bureaucrats, Huber et al. still provides some nuances to explain why in some countries the bureaucrats enjoy more discretion, though such discretion as Moe criticized is "delegated" from politicians. Also interestingly enough, while Huber et al.'s analysis obviously downplay the role of constituents, Moe tries very hard to include social actors into his analysis.
ReplyDeleteHowever even so, treating politician/bureaucrat relationship as a principal-agent relationship, I still find it very hard to say which one must be the principal. I believe that is exactly one of Moe's points. Very often researchers jump to the conclusion that no matter how much autonomous bureaucrats are they are still subject to the control of politicians. To be sure, it has been a central theme of the studies of Japanese post-war economic growth. Did the bureaucrats, dominating the political arena, lead Japan to prosperity as Chalmers Johnson argued? Or, is it the LDP politicians who dictated bureaucrats what to do? I have no intention to pull out the whole literature here, but there still one one nuance that I believe worthy of consideration: Do these two groups really exist exclusively to each other? I mean, before exploring which one controls the political realm, it may be necessary to see how they relate to each other. At least as shown by the Japanese situation, many politicians turn out to be senior/retired members of Japanese administration (so it is a source of political control over bureaucracy), but it is also the case that many bureaucrats and politicians go to the same university (e.g. University of Tokyo - Faculty of Law), which makes seniority a less powerful controlling mechanism as there are very complex personal relations between these two groups. To complicate the picture, some may even contend that it is neither politician nor the bureaucrats but the party, the LDP, that controlled the whole system through the Policy Research Committee, a party organ that convened party heads, Diet members, and bureaucrats. Therefore, I believe there are always more stories to tell about the contested relationship between the elected politicians and the bureaucrats as very often they may highly correlated to each other or may be even under the mandate of a third party (e.g. the majority party). In this sense, Geddes may be right for bringing the party system into the analysis.
Chao-yo, Im not sure I understand with your criticism of Schleifer and Vishny's article in regards to mobile services. Land transactions are not mobile, but, for instance, citizens can go to one of many offices to deal with the bureaucracy of the transaction. Are there truly any government transactions that need to be dealt with in one locality? (the only one I can think of is customs, where transactions are limited to a point of entry) I do agree that establishing several offices for government transactions is costly, and that this may be a hinderance for less developed countries.
ReplyDeleteAdditionally,if one is to believe TI scores, democratic countries tend to be less corrupt. Singapore is a consistent outlier in those rankings, and is generally in the top 10 least corrupt countries. Not being an expert on Singapore, I cannot say with any certainty why its seen as less corrupt, but perhaps there are other mechanisms involved, such as effective implementation of laws.
My own criticism of the S&V article is that I wonder if there is yet a third structure of corruption that is missing. S&V identify two structures, hierarchal and nonhierarchical, with the hierarchal structure being more conducive to growth. Im wondering if there are several "pyramids" of corruption, but they are not fully horizontal. Perhaps different agencies collude with each other to ensure one cost for a certain transaction (lets say, building permits and taxes), but due to competition between different ministries, not all the agencies work in collusion. For instance, a new resident is applying for a residency permit, and needs to show a work permit, a housing permit, a letter of invitation, and a visa; perhaps the individual is asked to bribe the visa agency and knows that this bribe will ensure she gets a labor permit and a housing permit, but still needs to pay a separate bribe for a visa. That sounds extremely convoluted and Im not sure my point is clear...And in thinking about it, this might have no real effect on growth rates (as its effects may be similar to that of nonhierarchical corruption), so I just wasted a lot of space trying to clarify this idea for no reason.
@ Chao-yo and Kim. I like your thoughts on the S&V paper. My own take on the relationship between corruption and democracy is that the large association between TI & Freedom House scores is because democracy does not develop successfully in countries where there is not a norm of universalism. Societies in which status is understood by distribution of power and where actors do not expect equal treatment by the state will be both more corrupt and less democratic. In this view Singapore is an outlier.
ReplyDeleteI appreciate Schleifer and Vishny’s article on corruption that attempts to derive a model of corruption and come up with some potential prescriptions for its reduction. This is a major issue as anticorruption is now a major industry. However in my opinion S&V overlook some key distinctions about corruption that call their remedies for corruption into question.
It is very important to make the distinction about what the main spoils of corruption are and who is the principal agent of corruption. Is it the state (many post-communist societies) or private actors? In post-colonial societies the principal agent of corruption might be found elsewhere as the state is often only one of the actors competing for control. This distinction has great importance for both the treatment and the diagnosis of corruption.
Another type of corruption overlooked by S&V but I think important in our discussions on institutions is electoral corruption. In many developing societies elites can exchange resources for votes so that voting becomes a commodity rather than a right or a choice. Though this type of corruption is overlooked in their article, this has serious ramifications for how the populace is tied to politics and perceptions about governance. In this case it is hard to see how their prescriptions for curbing corruption (increased economic and political competition) would be effective. Would increased political competition result in an increase in vote trading or would perceptions about what voting is actually change? Even increased political competition cannot stymie corruption: after the Orange Revolution one could but a seat in the Ukrainian parliament. It is clear that in many regimes popular perceptions of government corruption remain despite changes in government.
Finally, V&S contend that better and increased accounting would help to reduce corruption levels. Here again a distinction on the nature of the principal actor is important: if it is the state, then increasing bureaucratic efforts to combat corruption seems implausible and counter-intuitive
@ Kim: Yeah, you are right. I did not have a very clear idea about an example when I was writing this point. But I will give it a try again. Regarding land, yes, citizens may go to different offices that provide the same service or government goods. But there may be one condition that citizens cannot do so. For example, some countries require lands to be valued before they can be used by private individuals for agricultural or industrial production. In other words, the land owners need to get permits from government to turn their land assets into profits, and clearly the industrial use of land is much more profitable in some countries (especially the small ones where land resource is more sparse). Under this condition, my imagination is it is very hard for citizens to have alternative choices and very often they need to bribe local officials to have their lands for more profitable use. If we use the case of visa application, some countries, if the bureaucrats are highly predatory, may restrict citizens' options by saying: People from a given state should go to apply for visa in a particular state city (I am just making this up).
ReplyDeleteAnother example is resident registration (as you also mention). In some countries (e.g. China) resident identity (rural vs. urban) can be a very precious asset as different identity attached a totally different set of social privileges. Therefore, when citizens' choices are fixed to a specific locality, corruption may become prevalent. In this sense, I think my point is more like an elaboration of Shleifer et al.'s argument rather than a criticism?
Regarding your "pyramid" point, I think you are right that there may be simultaneous horizontal coordination of collection of bribes, though incomplete and unstable, within each level of administration. Actually, that is exactly the topic that many scholars of Chinese politics are tackling right now.
I really enjoyed the banter between Kurtz and Schrank and the Kaufmann et al. The former make the case that current measures of governance are flawed and that economic growth might lead to better governance. Even before reading Kauffman et al.’s rebuttal, I found two points of contention. The first is the narrow definition of what constitutes good governance. As Kaufman et al note, Kurtz and Shrank’s focuses only on government effectiveness eschews other interdependent measures, subtypes and concepts of governance that are key to gauging the extent to which governance can be considered “good”. Political accountability (elections, and the degree to which the are free and fair), formal oversight institutions (independent legislature, judiciary, autonomous anti-corruption body), the role and functions of the media, civil society; decentralization (political, fiscal and administrative); and the role of the private sector (CSR initiatives, corporate governance; privatization of key industries) are other variables that affect public administration performance. Not taking these other measures of governance into account and only focusing on how well the public administration delivers services only gives one aspect of good governance. The other aspect that I kept shaking my head while reading was their tentative conclusion that economic development could be the impetus for better governance. The examples that most point to are South Korea and Singapore. Yet these are small outliers that buck a bigger trend towards the opposite conclusion. Here I thought Kaufmann et al really did a good job of putting the argument (governance -> economic growth) into its proper place by showing the overwhelming amount of work pointing in this direction. Reflecting on my experience working on governance initiatives for Transparency International in the Middle East, I have to say that my perspective is more in line with the Kaufmann et al camp. I have a feeling Kurtz and Schrank would say I’m too close to the trees to see the forest though.
ReplyDeleteI just posted my referee report on Geddes on courseworks... Aside what I have written there on it, I wonder how she would analyze the recent emphasis given by Pinera in Chile on the merit of civil servants. Maybe as the exception she finds in the Brazilian case? Whether her analysis can be extended beyond the cases she studies is a clear question that stand out of her article...
ReplyDeleteI believe Geddes's article can interestingly be seen as an analysis of the pre-period of Huber & Shipan's book. In a certain way (and although it remained limited in the Latin American cases studied), she looks at the decision of the politicians to (potentially) give more capacity to bureaucracies, whereas Huber and Shipan take bureacratic capacity as a requirement.
Specifically on Huber and Shipan: I found their argument very convincing, but one aspect of it still raises question for me: the definition of a bureaucrat. They define it as an “agent who plays a central vote during policy implementation after statutes are adopted” (p.82), and directly specify that a bureaucrat can also be a president, a governor, a cabinet minister, etc. How do they account for parliamentary systems when cabinet ministers are also legislators? How can they compare these systems with presidential ones? Do they consider cabinet ministers as bureaucrats in certain cases and not in others?
@ Osama: I would like to follow Osama's discussion on Kurtz et al. To summarize my point, I do not see South Korea and Singapore as outliers. They may seem to be so, as I argue below, is because they selectively choose a narrow portion of the notion of governance, and because they conflate dictators with the whole bureaucracy, which is not necessarily weak in both cases.
ReplyDeleteWhen I was reading this piece, while I was amazed by their insightful criticism regarding current data most political scientists use to measure governance, I have to say their concept of "governance," as Kaufman et al. clearly indicate, is rather ambiguous and selectively narrow. In the case of South Korea, I am not sure if it is appropriate to equate the military dictatorship to the whole bureaucracy. This applies to Taiwan as well, even though part of the administration might be predatory and caused much distortion as Kurtz et al. point out, the problem is we are not sure which part of administration/bureaucracy promotes economic growth. Maybe like Japan it is those technocrats who held the offices of the Ministry of Finance, rather than those who are in charge of foreign trade (or vice versa?), that contributed the rise of Four Tigers? To explain why Kurtz et al find the possibility of the reverse causal link between governance and development, one answer might be their inconsistent and unclear definition of "governance," and the assumption that treats bureaucracy as a single actor.
The first comment concerns the back and forth between Kurtz and Schrank (KS) and Kaufmann et al. There are so many articles that deal with economic growth and governance/democracy… AJR on settler mortality is a robust finding that institutions matter. Alternatively, Lipset’s modernization theory also has many supporters. Why can’t we simply accept that the two are intertwined and the feedback is bidirectional? The more articles I read, the more it seems causality attributions depend on data sets, time periods, controls, choice of instrumental variables, and what not! The World Bank governance indicators are regarded with suspicion even within the World Bank. My boss there this summer was reluctant to use them because of measurement error and operationalization problems… The biases in the World Bank (or Freedom House) should not be surprising if we look at the composition of the institution. Until recently, it was vastly dominated by economists who saw administration as interference and not as a guarantee of a fair process. That being said, I agree with Osama on his critique of KS. In the end, both articles should admit the only thing that is obvious: we don’t really know much thanks to the World Bank indicators because of period limitation, sensitivity to variables, etc.
ReplyDeleteBriefly about Geddes, I agree both with Chao-yo and Florence. the article is internally solid. However, is it externally valid? I honestly don’t know enough about Latin American administrative reforms. (Only if Huber and Shipan had focused on Latin America!)
Shleifer and Vishny: so good! I susbscribe Chao-yo’s summary rather than repeating it. An economic approach fits very well with the problem of corruption, especially when undertaken by individuals. In these circumstances, an expected utility approach seems sensible. The individual wants to know whether the monetary gains of corruption (g) minus the cost of being caught times its probability (p) is bigger than not doing anything (zero). That is, an individual will be corrupt if, for E[u(corruption)] = gains - p(caught)*cost|caught, E[u(c)] > 0. Simple, but powerful enough to give way to the three scenarios of corruption. The issue is a bit more complicated in organizational contexts, however. Imagine most around you are being corrupt like in the KGB or Stasi. Does the individual really make any calculations? Not really, if he was to be transparent and accountable to the Communist Party instead of his boss at KGB or Stasi he could be in big trouble. The point, then, is that when the individual is organizationally constrained other factors are more important than the personal monetary gains of selling public goods.
Few times have I found myself completely opposing an argument that an author presents. While I can appreciate the utility of rethinking concepts that we thought were good once and for all I was surprised by the proposal of Kurtz and Schrank. As Osama pointed out, is not only their narrow definition of good governance but their contention that it is in fact growth what causes good governance. I find that proposal lacking support for the following reasons:
ReplyDeletea)unless countries have institutions that can channel economic activity properly there will not be economic growth.
b)economic growth requires secure property rights, a lawful environment, certainty,accountability, transparency etc etc that only derive from good governance.
c)While there can be some economic growth without good governance this usually tends to be concentrated in the hands of few. Is that really the type of growth that we want?
d)I find the policy implications of their paper quite dangerous. We do not know the mechanism but somehow growth will cause good governance.
I was a little disappointed by the defense of KKM. Again, while I can understand the value of methodology I do not believe that discussing halo effects is what will move the discipline to better theories. Yes, methodology matters but stats is just a tool that basically relies on qualitative classifications for its usefulness. In the defense of why good governance leads to growth, I expected a debate about ideas, not about tools.
@Chao-yo on Shleifer et al: The authors argue that the extent of bribery is lowest where there are competing public bureaus providing the same public good and your concern is that not every government good or service is "mobile", so taxpayers might be restricted to one particular bureau.
ReplyDeleteInstead of framing the counter-argument in terms of the "mobility" of goods or services, perhaps you could also consider other factors that might restrict the number of government bureaus/organizations in each area. For instance, public monopolies are often natural monopolies in industries (with huge fixed costs) that can only support one company, e.g. the standard example of utilities (water/electricity) providers. The argument that the industry cannot support more than one firm is often proposed in support of nationalizing the "natural monopoly" in this instance.
In addition, I think Shleifer et al's definition of corruption is rather narrow. They define "government corruption as the sale by government officials of government property for personal gain". Indeed, this aspect of corruption--where contracts might be awarded to less efficient or capable contractors in return for bribes--appears to have a direct, visible impact on development. But what about less visible forms of corruption that do not involve any exchange of property, for instance, when officials claim for expenses that they have incurred in a personal capacity, e.g. the corruption scandal in the UK cabinet earlier this year? This is only one example, but I am certain there are many different types of corruption in practice. Hence the interesting question is whether different types of corruption could have different effects on growth and development, or whether all types of corruption are equally detrimental to growth and development.
On Kaufmann et al’s rebuttal of KS’s critique, I agree very much with Kaufmann et al and I just have one small comment to make: Kaufmann et al mentioned that in looking for potential sources of bias in their data sources, they have examined whether the difference between the assessments of think-tanks and firm surveys was systematically correlated with the political orientation of the government in power in the countries being rated. They found that this was generally not the case and concluded that this additional source of bias is dubious. (pg 4) Although there is no correlation between the assessments of think-tanks and the political orientation of the government in power, could a correlation exist between the assessments of think-tanks and the ideologies/political orientation of the largest opposition party or of other smaller but nonetheless influential opposition parties? In autocratic regimes, simply investigating whether a correlation exists between the assessments of think-tanks and the political orientation of the government in power would probably suffice, but in democracies where an entire spectrum of political ideologies is often present and freely advocated by different groups, think-tanks might align themselves with a whole range of political ideologies, not necessarily that of the government in power, e.g. in the US. In this case, the ratings and recommendations of NGOs and think-tanks might still be coloured by different ideological orientations, though not necessarily reflecting those of the government in power.
This comment has been removed by the author.
ReplyDeleteI was positively surprised by the Geddes article. I think she combines empirical studies and the basic logic of game theoretical insights in a fruitful way; rather than relying on abstract mathematics and then testing quantiative results, she relies on what Alfred Marshall called "short chains", basic behavioral assumptions, the plausibility of which are part of the empirical testing.
ReplyDeleteI also liked Moe's piece, although it could probably have been shortened by 50 percent without losing content. I think his insights into something he says he knows little about - parliamentary systems, and their effect on governance and bureaucracy - are interesting, although he disregards the separation of powers that most parliamentary systems still have (supreme or constitutional courts in all European countries, House of Lords in the UK etc). However, these "checks" in European systems come into play only in fundamental questions, and not in nitty-gritty politics as usual.
As in the Geddes article, I like the way in which he employs rational choice theory very specifically - short chains - to explain a logic of a phenomenon (repeated games, uncertainty about outcomes etc.).