Dahl: Is Minority Domination Inevitable?

In most of the sciences – whether human, social or natural – there is a symbiotic relationship between theoretical and quantitative approaches. Einstein would not have formulated the theory of special relativity had the Michelson-Morley experiment confirmed the existence of the aether wind. The academic study of politics, however, bucks this trend as theorists and political scientists rarely talk to each other. This is primarily because the term ‘political theory’ is generally preceded by the adjective ‘normative’, so a conversation between theorists and polsci professors might well be seen as a contravention of the naturalistic fallacy.

This is self-evidently the case in the field of social theory, dominated by the long shadow of Rawls and still dedicated to the study of ‘57 varieties of luck egalitarianism’ (Waldron, 2013, p. 21). But why should it apply to democratic theory? – common-sense would dictate this should be a combination of normative and descriptive work, as most modern poleis claim to be democracies. Yet the upgrade panel for my own PhD (on representation and sortition) advised me to choose between the theoretical and empirical literature and not to seek to reconcile the two. The recent thread on this blog discussing Gilens and Page’s claim to have disproved the median voter theorem is a good indication of the sharp divide between the two literatures.

The honourable exception to this intellectual apartheid is the work of Robert Dahl, so I would like to turn to Chapter 19 of Democracy and It’s Critics (1989), entitled ‘Is Minority Domination Inevitable?’ in order to attempt a reconciliation between these two warring tribes. In this chapter Dahl examines the ‘minority domination’ theory shared by Mosca, Marx, Lenin, Pareto, Michels and Gramsci (p.266), and concludes that it is a ‘distorted reflection of an important truth about human life’ (p.271). The ‘important truth’ is the quotidian observation that ‘even in ostensibly democratic organizations, it is the few who make the decisions and the many who do little more than go along’ (p. 267).[1] The ‘distorted reflection’ is a result of ignoring the ‘poly’ prefix of modern democratic governance (polyarchy) – Michels’ ‘Iron law’ of oligarchy was based on a study of the internal workings of a single political party (the German SDP). But when a number of equally oligarchic parties compete in the electoral marketplace then ‘competition among political elites makes it likely that the policies of the government will respond in time to the preferences of a majority of voters’ (p. 276). ‘Michels’s mentor Pareto, [when] writing as an economist, would never have said that these competing oligarchies produced monopolistic control over consumers and the market’ (ibid.), yet all the ‘minority domination’ theorists make the ‘elementary mistake’ of failing to apply the lesson of economics to competitive elections (ibid.).

Dahl notes that

For the most part the theorists of minority domination had little or no experience with systems of competitive parties in countries with broad suffrage or, certainly, with systematic analysis of competitive party systems. Marx, for example, did not live long enough to witness the operation of ‘mass democracy’ in Britain; and Lenin never really experienced it (even in exile in Switzerland). Pareto, Mosca, Michels, and Gramsci witnessed only its beginnings. (pp. 276-277)

In addition to the problem of anachronism, minority domination theories are, at root, normative or even apocalyptic. Although Marx claimed that his study of historical development was ‘scientific’, his vision of ‘true democracy’ was fundamentally

An ill-founded hope for an apocalyptic revolutionary transformation that will lead us into the promised land of perfect freedom, self-realization, and full acceptance of the worth of all human beings. . . democracy would be possible only under conditions that have hitherto not existed in all of recorded history and may well be out of reach of human efforts in the foreseeable future. (pp. 279, 272).

The theories of Marx, Lenin and Gramsci were ‘instrumental to their ideological ends’ (p. 269). John Gray (2007) and Anthony O’Hear (2000) confirm this view of Marxism as a secularized form of Christian millennial eschatology. Although Pareto, Mosca and Michels were drawn to the opposite pole of the political spectrum, Dahl argues that they were a ‘response’ to Marxism (p. 269), aimed to weaken and discredit it as a political force (i.e. their motive was normative and rhetorical rather than descriptive). He also claims that Gramsci’s modification of Marxist theory was ‘clearly influenced’ by Mosca’s Elementi di scienza politica (p.366, n.12).

So why do ‘minority domination’ theorists (still the majority in political theory departments and the prevalent view on this blog) differ so markedly from their political science colleagues, for whom the median voter theorem is self-evidently true? In addition to the divergence between normative and empirical approaches to the study of politics, the theories are

presented at such a high level of generality that it is hard to determine what evidence could be brought to bear that would conclusively verify or disprove the central hypothesis of minority domination . . . an advocate’s commitment to a high-level theory is very likely to be far stronger than a rational decision would warrant [so] it would be fair to say that conceptual clarity and precision are not among the virtues of theories of minority domination. (p. 272)

In addition, and as marked in the critique of the Gilens/Page paper, the causal links are never specified whereby the minorities exercise their (alleged) control of the political process:

No theory or account contending that minority domination is a standard characteristic of countries governed by polyarchy has yet provided the evidence needed to verify the existence of such a chain of control. (p. 277)

The only mechanisms outlined are somewhat vague: ‘coercion’ (Marx/Lenin) or ‘social indoctrination, leading to false consciousness’ (Gramsci) (p. 274). More recently the hypothesized control factor has been wealth (Gilens and Page, 2014) but, given that the immediate currency of elections is votes rather than dollars, it is never explained exactly how the latter comes to trump the former in the minority of instances where the preferences of rich and median voters diverge. Correlation is not causation (especially when the level of statistical significance is marginal).

By contrast the chain of control in the median voter hypothesis is plain to see – votes are the immediate currency of electoral democracy so voters have ultimate control of political outcomes by obliging the competing oligarchies to reflect their preferences. Any party that fails to do this loses the election (or the next one), period.

Refs

Dahl, R. A. (1989). Democracy and Its Critics. New Haven: Yale University Press.

Gilens, M., & Page, B. I. (2014). Testing theories of American politics: elites, interest groups, and average citizens. Perspectives on Politics.

Gray, John (2007), Black Mass: Apocalyptic Religion and the Death of Utopia, NY: Farrar, Straus and Giroux

O’Hear, Anthony (2000), After Progress: Finding the Old Way Forward, London: Bloomsbury.

Waldron, J. (2013). Political political theory. Journal of Political Philosophy, 21(1), 1-23.

[1] Randomly-selected groups are no exception to this ‘important truth’, hence my claim that the speech acts of individual members will also lead to minority domination.

Advertisements

26 Responses

  1. >”Yet the upgrade panel for my own PhD (on representation and sortition) advised me to choose between the theoretical and empirical literature and not to seek to reconcile the two.”

    Wow. Coming from a physical science background that feels almost scandalous. The best work often comes from reconciling the theoretical and empirical when there is conflict between the two.

    >”but, given that the immediate currency of elections is votes rather than dollars, it is never explained exactly how the latter comes to trump the former in the minority of instances where the preferences of rich and median voters diverge.”

    In the American system it makes sense. You have several degrees of separation between the actions of the agents of the electorate and the actions of the government as a whole, which allows for an uncomfortable degree of agency loss. You also have oversized districts, weak parties, no public financing and little campaign finance regulation. It makes perfect sense for politicians to play a balancing act between the wishes of the masses and wishes of their sponsors (who are the ones who enable them to reach the masses). You also have extra veto gates, super-majority requirements, free-for-all primaries and extreme malapportionment, all of contribute to making the system relatively easy to game.

    This sort of thing ( http://thinkprogress.org/climate/2014/04/01/3421088/koch-brothers-tennessee/ ) is pretty common here. Even though Tennessee (where I live, actually) is a blood-red state, politicians still have to fear primary challengers. With enough money you can conjure up a primary challenger anywhere you’d like. Here’s an unusually overt example ( http://m.theweek.com/article/index/251041/the-tea-partys-newest-primary-challenger-big-business ) We’ve seen this stuff a hundred times before. Nobody acts surprised.

    Like

  2. Dahl doesn’t dispute the inequality introduced by the power of money but argues that “because it is easier to discover ways of reducing inequality than ways of achieving perfect equality (whatever that might mean), an advanced democratic country would focus on the reduction of the remediable causes of gross political equalities” (p.323) — this is in the same section as his proposal for a sortition-based minipopulus. Certainly the influence of money in British elections is far less than in the USA.

    Like

  3. There are some distinct questions here…1. Do the rulers serve the interests of the people (or “median voter” as a stand-in)? and 2. Does an elite class of people rule?

    The answer to the second question for the U.S. is clearly “yes.” After all a majority of members of Congress are millionaires, and nearly all the rest are culturally elite, if not wealthy.

    Whether these elite rulers serve the interests of the median voter any where near as well as they serve the interests of the wealthy seems obviously false to me, but it can be debated.

    Assuming sufficient examples and evidence is compiled of non-median-voter service, that still doesn’t prove causation, of course. But exactly WHY the elite rulers serve elites and not the median voter is an important question to answer. But it is NOT essential to answer THAT question in order to assert that the rule is not democratic.

    Like

  4. Terry,

    “Democracy” means that the people have power, not that the people rule in person (that would be demarchy). Dahl introduced the word “polyarchy” (originally coined by the political theorist Ernest Barker in 1913) into modern political science in 1953 as it was clear to him that the people did not rule. It may seem “obvious” to you (and most political theorists) that the polyarchs rule in their own interests, but the overwhelming majority of political scientists claim (as Gilens and Page acknowledge) that the evidence supports the median voter theorem. The real scandal, as Naomi points out, is that the two groups don’t even talk to each other, so Gilens and Page should be praised for seeking to bridge this divide, even though their analysis is so clearly flawed. Dahl’s observation that an “advocate’s commitment to an [outmoded] high-level theory is very likely to be far stronger than a rational decision would warrant” would suggest psychological and ideological mechanisms. To this I would want to add the desire of political theorists to protect their discipline from infiltration by public-choice economics (the irony being that both Marx and Pareto were economists).

    For the people to rule in person in large complex poleis would contravene the laws of arithmetic, although we all agree that the aggregate judgment of a statistically-representative sample would be a more reliable way of discerning the informed preferences of the median voter.

    Like

  5. P.S., Dahl would most likely agree as the “minipopulus” would constitute “an attentive public that represents the informed judgment of the demos itself” (p. 340). Can anyone recall whether Fishkin acknowledges Dahl as the godfather of the Deliberative Poll?

    Like

  6. Terry,

    > But exactly WHY the elite rulers serve elites and not the median voter is an important question to answer.

    One might as well ask why any person is primarily busy with their own interests rather than those of their neighbor.

    In terms of theory, the real question is the opposite one: is there a plausible reason to think that elite rulers would serve the population at large?

    Like

  7. Yoram:> is there a plausible reason to think that elite rulers would serve the population at large?

    Not just plausible, demonstrable. In Dahl’s own words:

    ‘competition among political elites makes it likely that the policies of the government will respond in time to the preferences of a majority of voters’ (p. 276).

    I append below a selection of references that confirm this (by correlating straw-poll preferences with policy outcomes):

    Arnold, 1990; Bartels, 1991; Burstein, 1998; Clingermayer & Wood, 1995; Erikson, 1978; Erikson, Mackuen, & Stimson, 2002; Erikson, Wright, & McIver, 1993; Fiorina, 1973; Geer, 1996; Grogan, 1994; Hartley & Russett, 1992; Hinckley, 1992; Hobolt & Klemmemsen, 2005; Jacobs, 1993; Jacobs & Shapiro, 1992, 1994b; Jones, 1994; King, 1997b; McDonough, 1992; Miller & Stokes, 1963; Monroe, 1979, 1998; Page & Shapiro, 1983, 1992; Page, Shapiro, Gronke, & Rosenberg, 1984; Quirk, 2009; Quirk & Hinchliffe, 1998; Schuman, Steeh, & Bobo, 1985; Simon, 2003; Stimson, 1999; Stimson, et al., 1995; Taggert & Winn, 1993; Weaver, 2000.

    Note that the proposed causal mechanism is that elite politicians are motivated (in part) by re-election.

    Like

  8. > ‘competition among political elites makes it likely that the policies of the government will respond in time to the preferences of a majority of voters’ (p. 276).

    That’s apparently a reference to the rewards-based mechanism, which is not credible.

    As for the empirical evidence: I am not sure where you lifted your long list of references from. (Unless you created this list yourself, reproducing it here without attribution is impolite, to put it mildly).

    In any case, this list is a poor substitute for real evidence. I doubt that any of them show what you claim they show. If you actually read any of them, which I doubt, I have no reason, based on your track record, to assume that you understood any of the papers that you did read.

    A minute’s search turned up what Bartels says in 2005 about the Bartels 1991 paper which you cite (and about other papers on your [for some definition of “your”] list):

    Meanwhile, statistical studies of political representation dating back to the classic analysis of Miller and Stokes (1963) have found strong connections between constituents’ policy preferences and their representatives’ policy choices (for example, Page and Shapiro 1983; Bartels 1991; Stimson, MackKuen, and Erikson 1995). However, those studies have almost invariably treated constituents in an undifferentiated way, using simple averages of opinions in a given district, on a given issue, or at a given time to account for representatives’ policy choices. Thus, they shed little or no light on the fundamental issue of political equality.

    Like

  9. Keith, I think you press a sensible challenge, at least insofar as I would agree that all political philosophers and theorists need to have empirical support for the empirical premises in their arguments. (Many, if not most, arguments in political philosophy have empirical premises.)

    If I offer an argument that has a premise that in electoral conditions like the United States, the rich and powerful will come to control what outcomes are generated, that claim would, ideally, be bolstered by substantial empirical evidence. Of course, that evidence can be hard to come by, given the small ‘n’ problem, and the many possible confounding variables. The evidence will be suggestive, perhaps, but rarely decisive.

    Another difficulty is that it might be true that political outcomes will be those most preferred by the median voter, while also being true that the rich and powerful control what outcomes are generated. How? By the rich and powerful controlling what the median voter most prefers. So these hypotheses do not conflict. Indeed, I would think those who think that corporate controlled media and unregulated campaign finance are problems for “real” democracy (rather than “nominal” democracy) would predict that something like the median voter hypothesis would be true. (Something like social indoctrination via wealth. This is not a crazy hypothesis in the United States, certainly, when one thinks of things like defense spending, insurance regulation, finance regulation, health care and pharmaceutical industry regulation, and so on.)

    Of course, there is now a difficult question: if the fact that a system does what the median voter most prefers is not enough to establish that the system is a “real” democracy, what is enough? That is a hard question to answer, but that is where much of the action has been with deliberative democracy, deliberative polling, reflective democracy, and discussion of “phantom” opinions and the like.

    A final small note. I like this blog, and when I visit there are usually many very interesting points made and discussions had, but the constant sniping back and forth and somewhat overheated argumentative tone sometimes seems counterproductive. Why not see the blog in a ‘big tent’ way, where we can develop many different arguments about and on behalf of the use of lottery in political institutions? It’s not as if there are so many people interested in defending sortition! So, some of us might think that the key premises to motivate the use of sortition involve elite capture and domination of the electoral process. Others might think that the key premises are something else. But these are all potentially interesting and useful arguments, given that different people might accept different premises, and might be moved to consider sortition from different starting points. Indeed, some of these arguments needn’t even compete, since they might conclude in conditional claims: “If X is a problem for a society, then we should use sortition for reasons A, B, and C.” “If Y is a problem for a society, then we should use sortition for reasons C, D, and E.” Or whatever. Part of the discussion can then concern the support that can be (or has been) offered for various premises, including the empirical premises, without people being quite so dismissive, and with efforts made at constructive engagement and development of the arguments. [Ok, climbing down off the soap box…]

    Like

  10. Alex,

    I agree with your point about the big tent and apologise for any sniping that I may have been guilty of, especially if this puts off people like yourself from participating in the debate. I guess I feel in something of a beleaguered minority as most people working in this field do believe the problem is elite capture, whereas I’m of the view that the problem is that the median voter is poorly informed. Most people side with the theorists, whereas I side with the political scientists. That’s why the DP methodology is my favoured option.

    >the rich and powerful controlling what the median voter most prefers

    That’s why I’m with Dahl that we need to ameliorate the problems of media monopolies and the domination of the political process by money (especially in America), alongside experiments with sortition.

    Yoram,

    >I am not sure where you lifted your long list of references from.

    The bibliography for my draft PhD. It’s too long to post here, so I’ll email it to you.

    >However, those studies have almost invariably treated constituents in an undifferentiated way . . . Thus, they shed little or no light on the fundamental issue of political equality.

    That’s undeniably true. Unfortunately the only study to attempt to remedy this is deeply flawed, for all the reasons that we have discussed on this blog (and Dahl predicted).

    Like

  11. Alex,

    If I read you correctly, you seem to assume that electoral systems are responsive to the views of the median voter. Why? The evidence is quite clearly pointing in the other way.

    If the public thought that Congress is responsive to its views, why would the approval ratings of Congress be so low? If such responsiveness existed, why would congressional salaries be at a level that the overwhelming majority of voters thinks is too high? And how would you explain the findings of Gilens and Page which are that given elite opinion, policy is uncorrelated with the average opinion?

    I do agree, BTW, that due to the elite’s power to set the agenda and affect public opinion even if there was correlation between average opinion and policy, the causation could quite possibly be different from the presumed “opinion determines policy” mechanism. Such correlation, however, doesn’t seem to exist.

    Like

  12. Yoram, I think the evidence is mixed regarding the responsiveness of electoral systems to the median voter. For one thing, there is huge variation in the details of electoral systems. I think that for many issues in the US, for example, Gilens and Page are clearly correct: if the interests of the elites depart from the interests of the median voter, the outcome is what the elites prefer. (Indeed, I rely heavily on this claim in my forthcoming paper, Against Elections: The Lottocratic Alternative, which I post a link to below–an early version; the final version will come out in Philosophy and Public Affairs soon.) But that doesn’t undermine the median voter hypothesis for all electoral systems everywhere… I do think the US is particularly bad in this regard, but it’s also the only system I know very well.

    A perhaps bigger problem (but not an unrelated problem, I think), is that the median voter is poorly informed, or even has no real views at all on most policy issues. I think this *leads to* capture by elites, by making capture considerably easier. That’s what I argue for in the first part of the above referenced article. So maybe this is a way of bringing Keith’s views and Yoram’s views closer together: ignorance leads to, or makes possible, capture…

    As for the low public opinion of Congress, and the high salaries of Congress, it’s hard to know what evidential value to place on those, given that most people cannot even name 5 members of Congress, or 3 things Congress has done in the past year. But that is a bit of a quick rejoinder. I think the Gilens book is really remarkable, and I cite several other things as well in this paper:

    http://www.alexguerrero.org/storage/PAPA_Against_Elections_Final_Web.pdf

    Like

  13. Yoram: > “If the public thought that Congress is responsive to its views, why would the approval ratings of Congress be so low?”

    Congress is extremely unresponsive. It, and the American government as a whole, has many features which promote agency loss, as I discussed above. You can’t just say, “Congress is unresponsive. Congress is an elected institution. Therefore, elected institutions are unresponsive.” No. If it is in the interest of politicians to win election/reelection, then it is in their interest to act in a fashion consistent with winning election/reelection. This is self-evident. It is also self-evident that if you change the rules of a game then you also change the way the game is played. Different rules, different strategies.

    > “If such responsiveness existed, why would congressional salaries be at a level that the overwhelming majority of voters thinks is too high?”

    This goes back to why I favor PR/coalitions. If you only have two serious competitors it’s all too easy for them to pick and choose what they want to compete on. Salaries are a lose-lose thing to compete on. Maybe a party could gain a slight, temporary, advantage if it made an issue out of it. But then everyone would end up agreeing (as popular opinion is so one-sided) and everyone would end up with a lower salary. No one wants a lower salary. Why would politicians make an issue of it? They have little incentive to do so. Things are a bit different if you have a half-dozen or more viable parties. Smaller parties are always looking for ways of distinguishing themselves.

    Like

  14. Alex,

    >Gilens and Page are clearly correct: if the interests of the elites depart from the interests of the median voter, the outcome is what the elites prefer.

    Really? There is a searching critique of their (unpublished) paper on this blog at https://equalitybylot.wordpress.com/2014/04/17/commentary-on-gilens-and-page-average-citizens-have-no-political-influence/ See especially the comments of the statistician HH.

    >maybe this is a way of bringing Keith’s views and Yoram’s views closer together: ignorance leads to, or makes possible, capture…

    Yes, I think that would be a good compromise. But, in order to accept it, all participants in this debate would have to abandon the claim that elections will necessarily (ex hypothesi) lead to elite capture of the political system. A combination of election, sortition and appointment can lead to government in the public interest (demokratia), assuming reform of the electoral procedures (as Naomi proposes), limits on campaign funding and additional safeguards against media monopolies. Demarchy (in the etymologically literal sense) is not possible in large modern states, even allowing for the use of sortition, so the achievement of demokratia will require a mixed system of governance.

    This also has the pragmatic advantage that elected politicians will be less hostile to the implementation of reforms that include an ongoing role for themselves. And if it is the case (as post-Gramsci Marxists and other critical theorists claim) that popular consciousness has been corrupted by elite indoctrination, this rules out a popular uprising against “electoralism”, hence the need to adopt a piecemeal and incremental approach to political reforms. Unfortunately there can be no heroic vanguard Kleroterian Party to orchestrate the mass uprising, as it would have to appoint its own leaders by lot!

    PS Do you know anything about the Perspectives on Politics journal that has accepted the Gilens/Page paper and what their policy is on peer review? The word “perspectives” in the title suggests a more inclusive editorial approach than that adopted by mainstream politics journals. Is the editorial policy just to let 1,000 flowers bloom, rather than to eliminate the weeds prior to publication? I appreciate that both authors are distinguished and well-published scholars (Page was himself “surprised” by Gilens’s findings) and it would be hard to conceal their identity via a double-blind review process. I publish a few academic journals (including History of Political Thought) and we would take a dim view of such an embargo breach (the paper is not due for publication until this Autumn).

    Like

  15. Alex, Naomi,

    > I do think the US is particularly bad in this regard, but it’s also the only system I know very well.

    > This goes back to why I favor PR/coalitions.

    AFAIK the US system is better documented and studied than any other system. It is therefore convenient (for supporters of electoralism) to claim that it performs worse than the norm. I don’t think there is good evidence that this is indeed the case. Are you aware of any?

    Just as an anecdotal example: Israel, where I live, uses a unicameral parliamentarian proportional representation system – institutionally, about as far from the US system as electoral systems go. Yet, over two thirds of Israelis believe that Israeli politicians put their own interests ahead of those of the public. (Here, page 49.)

    > A perhaps bigger problem (but not an unrelated problem, I think), is that the median voter is poorly informed, or even has no real views at all on most policy issues.

    This is largely true, but if anything it means that responsiveness, if achieved, would not be a valuable achievement. But given that there is no responsiveness, this is a secondary issue. As things stand, even when voters are informed and do have definite opinions (and the matter of MP salary is an example of such a situation) there is no responsiveness. So why should voters bother to become better informed (on those matters where this is even feasible – and many matters are not)?

    Like

  16. Hi Keith, just on the point about Gilens/Page: I was actually thinking more about the Gilens project over the last decade, including the book Affluence and Influence. It’s relevant, too, the “for many issues in the US” caveat that came before the part you quoted. With those qualifications, I think the point is well demonstrated.

    I haven’t read the newer work closely, but Perspectives on Politics is an excellent, peer-reviewed forum for original articles (it’s easy to google and find their submission and review policy online), and Gilens and Page are both very well regarded political scientists (well trained in formal methods, among other things). Having read your critique, I think some of your points they would take on board, particularly in terms of the scope and limitations of the data. One will never be able to empirically demonstrate an unqualified claim like “elections will necessarily, under every possible circumstance, lead to elite capture of the political system.” (I’m not sure why anyone would believe that, stated in such an unqualified way.) And of course they don’t establish anything like that.

    In general, there are always serious limitations on the empirical study of political institutions–lots of hard to account for variables, seriously limited data, few comparable systems, difficult to rule out competing hypotheses, and so on. So there are always limitations to the work, and most people who do empirical political science seem to appreciate (and remain appropriately modest about) the limited scope of their conclusions. Of course, people can then get carried away in talking about what the work shows or establishes, and this often happens, both in the popular press, and in the work of non-political scientists in other fields (including political theory and political philosophy). I have seen many people defend electoral democracy across the board (for every issue, for every context) simply based on Amartya Sen’s empirical work on the apparent strength of electoral democratic responses to famine prevention, for example.

    Like

  17. Yoram, I agree with much of what you say. The US seems clearly worse with respect to allowing money to run wild in politics. But it’s not clear that this is the biggest or only factor. In the paper I link above (sorry to keep referencing it, but it’s the main place where I set out my thinking) I try to identify the general factors that contribute to elite capture. Some of these are worse in the US, but I think many of them are present everywhere. Some of what I argue gets ahead of the empirical evidence, but there isn’t evidence that contradicts what I say (at least not to my knowledge).

    And I agree with your points both about the importance of responsiveness, and about the rationality of not becoming better informed, given the current circumstances in many places.

    I think PR systems are better for some things, but they are no panacea. They introduce a host of complications. They allow more fine-grained political parties, and more issue-driven competition, but this increases the information burden on the electorate, without increasing the incentives to become informed (or at least not in a particularly significant way). And they can lead to relatively fringe parties becoming unduly powerful, because of the importance for forming governing coalitions. So it’s a bit of ‘pick your poison’…

    Like

  18. Yoram:> given that there is no responsiveness [to median voter preferences, emphasis added]

    This certainly bears out Dahl’s observation that “an advocate’s commitment to a high-level theory is very likely to be far stronger than a rational decision would warrant” (p.272). I would have hoped that our extensive discussion of Gilens’s dataset might have moderated your absolute rejection of median-voter theory. It makes me wonder how deliberative democracy is ever going to work if participants are so resistant to shifting their views on the basis of reasoned argument.

    Alex:> “elections will necessarily, under every possible circumstance, lead to elite capture of the political system.” (I’m not sure why anyone would believe that, stated in such an unqualified way.)

    Unfortunately there are people on this blog who do believe precisely that.

    >Of course, people can then get carried away in talking about what the work shows or establishes, and this often happens, both in the popular press, and in the work of non-political scientists in other fields (including political theory and political philosophy).

    Gilens and Page have been dining out on this unpublished paper and have made no attempt to rein in their media acolytes.

    Looking forward to reading your forthcoming piece and will respond in due course.

    Like

  19. Keith,
    Perhaps the problem you have with the Gilens/Page article really comes down to the way the popular media has presented it as proof of a hypothesis that the rich decide things in the American electoral democracy. But that was NOT the hypothesis their paper tests, or proves. They were very clear that they had instead succeeded in falsifying the widespread median voter hypothesis. IF government policy was causally responsive to the median voter, then even for those issues where the median voter preference differed from the preferences of the richest 10%, those policy outcomes should more closely track with the median voter than with the preferences of the highest income decile. They showed that just the opposite is true. You have complained that there could be OTHER reasons for this than that the wealthy control policy, that they never examined, and proposed no theory of causation (there are countless other variables that may be more significant than wealth)…but they were not proposing a NEW hypothesis for why median voters do not have any significant independent influence on policy when their preferences diverge from the wealthy…they were testing, and successfully disproving a long-standing hypothesis.

    Like

  20. > Gilens and Page have been dining out on this unpublished paper and have made no attempt to rein in their media acolytes.

    Sutheland is up to his usual trolling. As Alex pointed out just a couple of comments before, (and as I have informed him multiple times in the past) the findings under discussion (i.e., the analysis of the influence of opinion by income group) have been published by Gilens years ago in a paper and in an extended format in a book. (The recent paper only adds the analysis of influence of interest groups.)

    Again, I really can’t understand why anyone would bother trying to have a substantial conversation with Sutherland. He is obviously impervious to any rational argumentation. If a matter of simple fact such as the publication date of a particular finding cannot be established, what chance is there of having a useful discussion of matters of any complexity?

    Like

  21. Keith,

    I also recalled reading something about the lack of “median voter” responsiveness in a recent book by Bartels (whose 1991 paper you cite). In “Unequal Democracy: The Political Economy of the New Gilded Age” on page 287 he concludes “Whatever elections may be doing, they are NOT forcing elected officials to cater to the policy preferences of the ‘median voter.”

    Like

  22. >”Unfortunately there can be no heroic vanguard Kleroterian Party to orchestrate the mass uprising, as it would have to appoint its own leaders by lot!”

    Reminds me of the “suicide squads” that were elected with the sole purpose of abolishing their institutions. That’s how a few upper houses were gotten rid of.

    Yoram,
    The points I’ve been making are generally conventional wisdom in comparative institutional circles. I would recommend reading Lijphart’s “Patterns of Democracy,” Shugart and Carey’s “Presidents and Assemblies,” and Linz’s “The Failure of Presidential Democracy.” I’m largely parroting points I read in those (and similar) books.

    Actually, the bulk of *comparative institutional* research seems to have been done on various European parliamentary governments. If my former university’s political science section is representative, that is. If you crack open a random old book on comparative party structures or something like that it’s generally a given that the context is European parliamentarianism. You’d be surprised how often the authors don’t even bother to mention that bit, as if it just goes without saying. This is especially true of the older, pre-WWII, research. The only good example of a developed, stable presidential regime that has remained stable and developed over generations is the United States, and one data point isn’t worth much. There are many presidential governments, sure, but they have a distressing tendency to experience disfunction and collapse. Requiring consensus to function yet providing institutional rewards that encourage the taking of an adversarial stance is patently absurd. It’s a recipe for absurd behavior. See most of Linz’s work.

    Israel is the opposite extreme from the US. As I said in the other thread, you need enough parties to where there is enough overlap to where most voters can shift their vote from one party to another without shooting themselves in the foot (so they can punish misbehavior) BUT not so many that the effect of a person’s vote is unpredictable. Having twelve or thirteen parties in parliament is absurd. When there are too many effective actors, blame gets spread too thinly. Fear of blame becomes a less effective deterrent. Having too many small parties is the same sort of problem as having highly undisciplined parties. Coalitions end up with too many actors within them. You start to see the same committee-style accountability problems. The Israeli electoral system fails at encouraging a healthy INTRAparty dynamic, instead it channels differences that would better be expressed as competition between members of the same party into the INTERparty dynamic.

    Like

  23. Terry>: the popular media has presented it as proof of a hypothesis that the rich decide things in the American electoral democracy. But that was NOT the hypothesis their paper tests, or proves. They were very clear that they had instead succeeded in falsifying the widespread median voter hypothesis.

    Not so. The paper set out to test four competing hypotheses and came to the conclusion that three of them were false. If the authors were strict Popperians they would have resisted the claim that their results favoured the remaining hypothesis: “the preferences of economic elites . . . have far more independent impact upon policy change than the preferences of average citizens do” (m/s p.22). The word “impact” implies a causal mechanism, whereas the paper only supplies evidence of a correlation. In their public interviews and TV appearances the authors made no effort to resist the strong interpretation of their correlation data. This is unsurprising, as the strong claims were in the paper itself.

    >Perhaps the problem you have with the Gilens/Page article really comes down to the way the popular media has presented it

    No, my problem is with lead author, Martin Gilens (Page has acknowledged that his only input was the four competing hypotheses and that he was surprised by the results), and with the journal’s editors and referees for authorising the publication of unfounded claims. We have dissected these arguments at length elsewhere, so I will only summarise them here:

    1. The portrayal of “affluent” and median-voter preferences in binary terms, whereas the dataset indicates that, in the minority of cases when they do diverge, the variance is of marginal statistical significance (I rely here on HH’s analysis of Gilens’s dataset).

    2. The presentation of correlation evidence as supporting one of the hypotheses (see above) without an intervening causal mechanism. The median-voter theorem would suggest that on the rare occasions when policy outcomes diverge from median preferences that these will be low-salience issues. If this were the case then the conclusions could be presented as “policy outcomes follow the preferences of the median voter except in issue areas that they deem to be unimportant”.

    3. The findings are presented in terms of the “impact” of different sets of voters (again without any suggestion as to how this “impact” will trump the immediate currency of electoral politics (votes). In their more cautious moments the authors precede the words “impact” and “influence” with the adjective “independent”, but their overall claim: “the preferences of the average American appear to have only a minuscule, near-zero, statistically non-significant impact upon public policy” (p.21) is unqualified. Evidence of differential “impact” when preferences diverge tells us nothing about the impact of voters’ preferences when they align, and the fact that the immediate currency of elections is votes, not dollars, would suggest that the preferences of 90% of voters would count for more than the preferences of 10%.

    Yoram,

    I imagine remarks like “Sutherland is up to his usual trolling” and “I really can’t understand why anyone would bother trying to have a substantial conversation with Sutherland” are examples of the “constant sniping back and forth” and “overheated argumentative tone” that Alex felt marred the debate on this blog.

    Like

  24. Naomi,

    > The points I’ve been making are generally conventional wisdom in comparative institutional circles.

    What is the kind of empirical evidence they use? Since you are familiar with this literature, I invite you to write a post laying out the evidence they present.

    > Coalitions end up with too many actors within them.

    The current Israeli government relies on a coalition of four parties. Is that too many? So the sweet spot is either 2 or 3? (1 being too few and 4 being too many?) Hoping that this kind of fine tuning would make a material difference seems like clutching at straws to me.

    Like

  25. Keith,

    I was focusing on the one theory of democracy that we were discussing (;median voter responsiveness”… but yes they used multivariate analysis to see if the evidence supported or falsified four distinct theories. They showed that their data set was consistent with two of the theories and inconsistent with the other two (one of which was the median voter theory). As they write in their abstract “The results provide substantial support for theories of Economic Elite Domination and for theories of Biased Pluralism, but not for theories of Majoritarian Electoral Democracy or Majoritarian Pluralism.”

    The theories they evaluated assert CAUSATION. This is not something the authors invented or attached…they simply tested whether the evidence fitted the theory of causation…not whether this association or lack of it was in fact CAUSED as each theory asserts.

    Finally, You should stop implying that only Gilens holds these findings valid, and that Page doesn’t support the findings of the paper he co-authored. Page may have been “surprised,” but fully endorses the findings. I have heard the two authors speak together on TV, and they are in agreement.

    Like

  26. Terry,

    I think it would be more accurate to say “their interpretation of their dataset” — HH and others took issue with their methodology. The focus on data points, without investigation of causal hypotheses is the chosen methodology of parapsychologists — so long as correlations exist then they claim that causal hypotheses are unnecessary. Unfortunately human and social sciences don’t really work in this way (“lies, damn lies, and statistics”) as the choice of intervening variables is so great. HH suggested that it would be more fruitful to examine the salience factor as this might better explain how politicians can successfully ignore median-voter preferences in the minority of cases when preferences differ from the top economic decile. Although they claim not to offer a causal connection, the clear implication (which they did not seek to deny in their media appearances) is that the rich are buying the policies that are in their interests (hence the use of words like “impact” and “influence”) — if they really were only interested in demonstrating correlations then they should not have used this kind of language.

    I saw the (somewhat nauseating) TV chat shows. I was not trying to suggest that Page was distancing himself from the results, merely that this is Gilens’s paper which has benefited from co-authorship with a well-known researcher who is normally associated with a more neutral position than his own.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: