MVT and policy portfolios
> > - people spend an inordinate time satisfying extreme voters, even after > > winning a party nomination > > > They're trying to put together a winning coalition by targeting a variety > of market segments. You don't see General Motors or Altria/Philip Morris > targeting just the "median car buyer" or "median cigarette smoker." > > ~Alypius Well, this is not a prediction of the median voter theorem. If you have N policies, the MVT would predict that the candidate would gravitate to the center of each policy. What you are suggesting is that the candidate would go to the extreme position for each policy, as defined by some subpopulation who cares about the issue. Fabio
Re: MVT and policy portfolios
- Original Message - From: fabio guillermo rojas <[EMAIL PROTECTED]> To: alypius skinner <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Saturday, October 18, 2003 11:47 AM Subject: MVT and policy portfolios > > > > - people spend an inordinate time satisfying extreme voters, even after > > > winning a party nomination > > > > > They're trying to put together a winning coalition by targeting a variety > > of market segments. You don't see General Motors or Altria/Philip Morris > > targeting just the "median car buyer" or "median cigarette smoker." > > > > ~Alypius > > Well, this is not a prediction of the median voter theorem. If you have N > policies, the MVT would predict that the candidate would gravitate to the > center of each policy. What you are suggesting is that the candidate would > go to the extreme position for each policy, as defined by some > subpopulation who cares about the issue. > > Fabio > Yes, special interests--sometimes including the opinions of the rulers' own social class--are often more influential than the median voter preference. Furthermore, if a politician can put together a winning coalition--let's say one that reliably gives him about 55% of his constituents' votes-- he can usually ignore those market segments who are not part of his winning coalition. Sometimes we also see politicians neglecting part of their coalition--such as Democrats neglecting blacks or Republicans neglecting religious social conservatives--because, in a non-parliamentary system, they can be safely taken for granted. Thus, both President Bushes courted the homosexual lobby, because they knew that the Democrats would not nominate anyone the social conservatives, however unhappy, could vote for. (And I'm fairly sure, if David Duke could win the Democratic Party's presidential nomination, that he would carry the black vote in the general election.) The Democrats are enthralled to a coalition of special interests which would never allow someone acceptable to social conservatives to be nominated, just as the Republican coalition would never allow anyone to be nominated who was not acceptable to big business interests. As another example, notice how both Democrats and Republicans in Presidential elections always nominate someone whose views on abortion are to the left and the right respectively of the median voter, who, according to polls, prefers more restrictions than the Democratic nominee will endorse and fewer restrictions than the Republican nominee is willing to allow. So the two parties, rather than competing for the median voter, will compete for those market segments which would not drive away--or be driven away by--the other segments that make up the core of *either* party's coalitions. In a close election, this often means that both sides compete intensely for that segment of likely voters which is least informed, least consistent in its opinions, and most politically clueless. These people are often the "kingmakers" in democracies. Related to this is the question of whether there really is a median voter. Let's take 10 issues--abortion, gun control, gay rights, trade policy, tax rates, immigration, middle east policy, racial preferences, CO2/"global warming" policy, and SDI/"star wars" missile defense. What percentage of the electorate is in the middle quintile (if we could quantify these issues) on all 10? There also is the "weight" that each voter gives to each issue. For significant numbers of voters, abortion or support for Israel or support for Kyoto/"the environment" or gay rights positions or gun control or "affirmative action" policies will outweigh all other considerations. Much more often, even though one is not dealing with a true "single issue" voter, taking the "right" or "wrong" position on one issue may outweigh one's position on 2, 3, or more other issues that are a lower priority for a given voter. There is also the question of how committed an office seeker seems to be to a given issue. For example, among Republicans we often see the following tightrope being walked: the office seeker tries to be sufficiently supportive of traditional values that he endears himself to the social conservatives in his party--especially in the primaries--but not so supportive that socially liberal Republicans in the primaries (or independents in the general election) will think that he "really means it." On a weighted list of issues, there may not be enough "median voters" to bother with. Putting together a winning coalition of market segments is probably a surer path to victory. ~Alypius
[no subject]
PS--Fabio, this reply may not appear on the list. I think Bryan has been diverting my posts to some cyberfile 13. The last half dozen or so times I attempted to post, nothing ever appeared on the list, so I've pretty much given up trying. But I can still read everyone else's messages, so you can reply to me on the list if you wish to. Since the software upgrade, it seems like a post's author is no longer sent a copy, though I'm having trouble figuring out quite what's going on. Anyone who is having doubts can email me to confirm. -- Prof. Bryan Caplan Department of Economics George Mason University http://www.bcaplan.com [EMAIL PROTECTED] "But being alone he had begun to conceive thoughts of his own unlike those of his brethren." --J.R.R. Tolkien, *The Silmarillion*
Re: Median Voter Theorem, Part Deux
Anton Sherwood wrote: In a multi-dimensional sample, you can find a median along each of the axes, but does the intersection stay put with a different choice of axes? I would guess not, and therefore the Median Voter cannot be meaningfully defined in an election (as opposed to a referendum). The MVT of course does not hold in multidimensional space. But the fact that there are many issues does not show there are many dimensions. Empirical work on U.S. public opinion has found that it is surprisingly one-dimensional. This is true for Congress (Poole and Rosenthal). Many in social choice would be tempted to see this as an artificial "structure-induced equilibrium." But one-dimensional ideological measures on the left-right access are also highly predictive for the beliefs of the general public. I would also add that while the MVT does not hold for multi-dimensional space, taking the median position on all issues is probably a good rule of thumb. With two separate issues, you may know that the median is not necessarily best on either issue. But should you go above the median or below it, and by how much? Unless a politician has an amazing level of insight into your vote maximization function, sticking with the medians is probably a smart strategy. -- Anton Sherwood, http://www.ogre.nu/ -- Prof. Bryan Caplan Department of Economics George Mason University http://www.bcaplan.com [EMAIL PROTECTED] "But being alone he had begun to conceive thoughts of his own unlike those of his brethren." --J.R.R. Tolkien, *The Silmarillion*
Re: MVT and policy portfolios
alypius skinner wrote: Related to this is the question of whether there really is a median voter. Let's take 10 issues--abortion, gun control, gay rights, trade policy, tax rates, immigration, middle east policy, racial preferences, CO2/"global warming" policy, and SDI/"star wars" missile defense. What percentage of the electorate is in the middle quintile (if we could quantify these issues) on all 10? The General Social Survey is online and has a lot of information on public opinion, though not all of these exact topics. At least on a crude measure of "do you want more/less/the same" level of spending or regulation, the median position is usually "the same." -- Prof. Bryan Caplan Department of Economics George Mason University http://www.bcaplan.com [EMAIL PROTECTED] "But being alone he had begun to conceive thoughts of his own unlike those of his brethren." --J.R.R. Tolkien, *The Silmarillion*
Re: Median Voter Theorem, Part Deux
> I would also add that while the MVT does not hold for multi-dimensional > space, taking the median position on all issues is probably a good rule > of thumb. If we could find some way to estimate candidates' distances from the median position on a large number of number of public issues, and then average them, would the candidate who is closer to the median voter generally win? I'm not sure. However, I am sure of one thing--the median voter position in, for example, the South, is not the same on many issues as the median voter position in the northeast or the Pacific coast. The black median voter is not the same as the white median voter, the Arab-American median is not equal to the Jewish median, nor the Mormon median to the Unitarian median. Urban medians are different from rural medians, heavily populated states from less populated states, and men from women (and especially from unmarried women). Consistently taking the median position might leave one with lots of unmotivated voters on either side of the spectrum, while a "conviction politician" might strongly motivate certain groups of voters, but, so long as he is not generally perceived as dangerous, he will not automatically alienate an equal number of people in the middle. The optimum position may be far enough to the left or right to motivate the ideological "true believers" within certain political "market segments" but not so far as to shock the swing voters in the middle, who are accustomed to tolerating candidates who are a moderate distance to their left or right. In such a scenario, motivating the fringes might be the key to victory, so long as one can do it without frightening the center. BTW, does anyone know the correlation between election spending and votes received? Is there a point where additional campaign spending suffers from reduced utility? It seems like the RNC and DNC, as well as assorted special interests, would very much want to know the answer to this question, as it would help them to allocate political aid to better effect. ~Alypius