Tuesday, June 12, 2007

Motives and public policy views

The question I posed to this week’s Blogocracy-organised group blog (Joshua Gans, Tim Dunlop, Ken Parish, Kim, Robert Merkel Andrew Bartlett and Tigtog) was:


How relevant are motives in assessing the public policy stance of a politician or commentator?
I had early posted on this topic here. To summarize my tentative conclusions I argued that if someone was asserting a statement of fact that the observer cannot evaluate ('I wasn’t there at the time', 'I didn’t steal the potato chips') then motives are important. This is obviously the case in a court of law where the culpability of a defendant is asserted.

But elsewhere, for example, if you are offering a desired political policy, they are unimportant. If you are going to evaluate policies sensibly you should ignore motives and concentrate on the consequences of policies. Indeed to do other than this is to commit a ‘Motives Fallacy’.

Logicians identify the 'Motives Fallacy’ as the belief that exposing the motives behind an expressed opinion shows that the opinion is false. This is a fallacy since it appeals to ulterior (extralogical motives) rather than the truth of the issue at hand.

The problem with falling prey to this fallacy in a political debate is that attention is turned away from the analysis of policy consequences. Policies just become part of a political game that seeks to establish who might win or lose. The specific effects of policies remain unanalyzed by the person who says 'X is only just saying that because of Y' where Y has nothing to do with the effects of the policy.

Excellent discussions of this (and other) fallacies is here and the Jamie Whyte book.

In Australian political debates motive fallacies are so repeatedly and commonly used – particularly, but not exclusively*, by the political left - that serious discussion is almost nonexistent. John Howard supports WorkChoices not to improve the efficiency of labour markets but because he is opposed to trade unions. John Howard supports the war in Iraq not as a defense against the expansion of terrorism but because he wants to win favor with the Americans, because he hates Arabs. Ayaan Hirsan Ali sees Islam as oppressive towards women because she had her genitals cut out with a pair of scissors.

Other favourites of mine include claiming that teachers are lying about workloads because they just want lighter teaching loads or that university academics are lying about university salaries being uncompetitive because they want more money.

To be clear the asserted motives might be correct here – John Howard may hate trade unions and have developed WorkChoices to attack them but this has zero to do with rejecting the claim that WorkChoices may improve labour market efficiency. Similarlly AHA might be making sound judgments about Islam irrespective of her history. Maybe teachers are overworked and universities are not recruiting the best staff they can.

Arnold Kling points out that even respected economists such as Paul Krugman and Brad de Long are often in the grips of a motive fallacy.

Motives are often attributed to people because we are too lazy or too time-constrained to explore the implications of an argument. They can also be used to reflect our attitudes to an opinion or viewpoint that someone expresses. But this has naught to do with expressing judgments on the value of political positions adopted.

In my view we should try to get rid of motive-based arguments in our own thinking and to try to recognize – and criticize – this type of flawed argument when it arises in the media and in blog discussions. It will be a long process in bringing about this type of cultural change and I am not exclusively pointing the finger here.

In short, we should try to reason logically.

Anyway I will respond in turn to the various arguments by my co-bloggers.

Joshua Gans’ response to the question seems to me to ignore the point. He is concerned with whether or not you can identify the motives which from the perspective of policy evaluation is irrelevant. The argument itself still needs to be considered regardless of the motives for advancing it. Joshua consider’s Telstra’s claim:


that there is insufficient return to investing in fibre to the node infrastructure and that they will only do it with a subsidy or protection of competition.
Since they have a stated interest in taking care of their clients one might be sceptical as Joshua suggests but you cannot disprove their claim on the basis of motives. Nor does it matter if Telstra is stating something that ran against its interests or whether Telstra was operating under the constraint that it could not directly lie. The Milgrom-Roberts piece on truth-telling he refers to assumes that people will lie if it is in their self-interest to do so. But that is not correct. Claims regarding the case for revisions of media ownership laws might be correct even if made by proprietors and sound environmental claims might be made by poor as well as wealthy environmental ministers.

In all cases Joshua is committing a motives fallacy by supposing you infer something about the truth of Telstra’s claims by positing motives for them.

Joshua refers to the case of a courtroom but here establishing motives is important for establishing culpability because claims concern issues of fact that a juror cannot often make a direct determination on. This has nothing to do however with making a case for or against some public policy proposal on the basis on the motives of the politician or commentator.

I’ll comment on other responses to my question as they come in and possibly revise my appraisal of Joshua's views if he makes an indignant, convincing response.

Robert Merkel responds with a fair amount of sympathy for my position. He also recognises, as I do, that we attribute motives as a shortcut in dealing with arguments. This won’t however - I would emmphasise - invariably lead to sound judgments.

It seems to me that backing expert judgment is a sensible approach that involves no motives issues. If you want to form a reasoned view on climate change and don’t have background expertise to evaluate a proposition then listen to the consensus view of knowledgeable climate scientists. That is more-or-less what I do.

On the other points that Robert suggests that motives-related heuristics can be justified I disagree. Repeating something often enough does not make it true and motives need not determine how policy will be modified given exogenous shocks.

Finally Robert (like Joshua) makes a case for using motives to assess an argument when there is imperfect information. Well that is true if you cannot make a judgement about the situation yourself – the case where expertise comes in or in a testimony in a criminal trial. I agree with this.

Tim Dunlop. I agree with some of Tim’s remarks but not the main thrust which as with Joshua I think constitute another motives fallacy.

Tim agrees with me that motives are important in assessing court evidence. But he questions my view that motives are irrelevant in assessing political debates. He argues that motives are relevant in assessing the believability of a person making a claim and that is what is important ‘what we are assessing is not the outcome of the policy, but the believability of the person making the claim’.

But his argument here reverts to the case for believing evidence at a trial. He claims it might be very difficult to assess arguments, such as the case for WorkChoices that I mention, so it is important to assess the believability of the claims which depend on motives.

If motives are driven by considerations other than the argument then it is reasonable to assume that the case he is putting forward is being influenced by those motives. His view might wrongly reflect these motives so it is reasonable to take them into account.

But again now Tim reverts to making a motives fallacy. He is forced now to assess the status of the motives (whether or not they enhance believability) which the ‘motives fallacy’ makes clear are completely irrelevant if you are in a situation where you can think through the implications of a policy proposal. If you cannot think through these implications then you are back in the courtroom situation.

You cannot dismiss an argument on the grounds that a person has low credibility. You can say you don’t have time to go into it or that their track record might be poor but not that the argument is unsound. You need to put in some work to reach the latter conclusion.

It strikes me that all sides of politics commit these fallacies so that, for example, much blog discussion is people shouting past each other and not even trying to hear. How many times have I heard advanced the proposition that John Howard or George Bush have proven themselves to be a total liar and that therefore their policy prescription on X (X= global warming, immigration, labour market reform, Iraq policy etc etc) is wrong. Normally I do not agree with the ‘liar’ description but defending these individuals on that basis is to commit a motives fallacy myself that parallels that of the lazy critics. The best defense is to try to force these critics to come to grips with the specific proposal and to examine and critique it. That is what political discussion should be about.

Relying on motive fallacies produces weird conclusions. Mr Rudd and the Labor Party recently condemned the Coalitions ‘cap and trade’ policy for addressing global warming on the grounds that the policy was purely poll-driven. This is a pure motives fallacy since the proposal might have been a good one even if these motives were accurate. You need to think about the effects of the policy and whether a carbon tax would work better, whether agriculture and waste generation should have been included and so on. This type of analysis gives the proposal fair treatment and helps develop better policy.

Kim Jameson, Correctly identifies attempts to avoid the motives fallacy as an attempt to make political debate rational. Her view is that rhetoric and appeals to emotion are just as important as logic. Kim disagrees with the proposition that political debate should be primarily on the merits of policy when in fact we vote for packages of policies identified as ‘left’ or ‘right’, for example, so that ‘general orientation’ is as important as specific policies. If you agree with this approach to politics then Kim is right. You should consider the general motives of politicians not the specific policies they espouse.

I don’t agree with this view and don’t think it gets you far in analysing political issues. It was the Labor Party that introduced enterprise bargaining and cut tariffs not the Liberals. The Liberal Party conversely has proven itself to be a high immigration, big spending government of the type Labor might have been expected to be. You need to look at specific policies and analyse their consequences – ‘general orientation’ gets you nowhere.

Andrew Bartlett is a Senator in the Parliament so I was very interested in his views. Andrew agrees with other commentators that we assign motives as a shortcut to assessing policies and notes this says nothing about the accuracy of claims. He tries to consider arguments on their merits but does consider motives particularly if a proposer has a bad track record of propounding poor arguments. He claims that motives give a guide as to whether a politician will keep promises when implementing policy.

Ken Parish. I disagree with Ken's views - my reasons are at his blogsite. These views are the reason I posed this question. Like many on the left Ken seems to be prepared to reject policy views entirely if they come from politicians who have what are, in his view, poor motivation. It puts the emphasis on the reasons behind a proposal rather than the proposal itself. My blog posting on Tony Blair's comments on the media here shows these attitudes to be among Blair's concerns as well. It seems to me this approach almost guarantees poor public policy outcomes for the reasons Blair discusses.

Tigtog. This came in late and I was running out of steam. Its similart to Ken Parish's views but has greater support for my position that what matters are the consequences of policies not their motivational rationale. To Tigtog however motivation is however important for judging 'trustworthiness'. Of course as I argued above - at length - my view is that trustworthiness, deceit, honor, honesty, union-hating, jew-hating, believing in Santa Claus motivations have bugger-all to do with assessing whether a policy is good policy.

(*) Using the terms ‘left-wing’ and ‘right-wing’ can support motives fallacies. Of course while we should be aware of fallacies when people are arguing with us we should also watch out for our own fallacies.



The other discussants are Robert Merkel, Kim Jameson, Ken Parish, Tim Dunlop, Joshua Gans, Andrew Bartlett and Tigtog. My response to their points is over the fold.

11 comments:

Joshua Gans said...

You slams me for engaging in a “motives fallacy.” You argue that my discounting of Telstra’s views doesn’t mean that they are wrong. Harry misreads me. I didn’t say that they were wrong I just said that if they are the only one’s now talking, I’m not listening.

And, yes, Harry I think people will lie if it is in their interest to do so (I thought most economists thought that). That doesn’t mean what interested parties say is always a lie. You commit that fallacy. What it means is that we can’t tell whether they are lying or not. Find me someone without a special interest who says the same thing and I will assess that truthiness of that more favourably. You need to apply equilibrium thinking here. The point of Milgrom and Roberts is that when we treat interested party’s views with skepticism it generates conditions from which the truth might emerge. So I agree that a statement by an interested party does not mean it is automatically wrong and the reverse true, just that you cannot often get information — true or otherwise — from those parties. You need a better information source.

hc said...

Joshua,

You need to listen to Telstra and evaluate their arguments. You cannot dismiss them simply because they have a motive justifying their position.

I don't believe people will always lie if it is in their interests to do so. The assumption of self-interested behaviour is just that - an assumption. Not even Adam Smith believed it as he made clear in his Theory of Moral Sentiments .

If you can't tell if a policy claim is true or false you can't make a decision about accepting it without considering the implications of the claim. Does it make sense or not. An appeal to motives tells you nothing.

It seems towards the end of your comment that you agree with the point that you cannot draw any inference from motives. That is exactly my claim.

Why then your discussion of whether motives support a claim or not? Why the discussion about whether the claim runs against their interests? Why the claims about disregarding the views of those with a stake in the media industry or (by inference) for accepting the views of wealthy environmental advocates?

Anonymous said...

This is the problem with most of the left, Harry.

There is a short answer to that question. Motives don't matter one bit, in terms of government motives, results do. Period.

Joshua Gans said...

Harry, how do you explain that when Telstra are in the entrant position (as they are in NZ) they argue EXACTLY the opposite?

Anyhow, being skeptical does not mean usually dismissing a whole argument but discounting it. It is just with Telstra that such complete discounting appears to be warranted by past behaviour.

Anonymous said...

In economics we understand the exclusive role of motivation and constraints of strategic interaction. I don't know if logicians take into account motives for advocacy, which is usually costly, or if they take into account the understanding that it is costly for the opposition to refute advocacy. Let alone the fact that in refuting advocacy one takes into account the costs of counter refutes.

I'm inclined to think that advocacy can't be understood outside the economic framework.

Nevertheless, the typical examples I've seen have been ones where the policy advocated is first shown to be incoherent (usually quickly) and then, despite all, those advocating the nonsensical positions persist unchanged with their advocacy. Then one is lead to ask why are these otherwise intelligent people persist? They must be motivated by secondary factors. (To be clear, this is not my point. My point is that one cannot begin to understand advocacy outside an economic setting).

Take for instance those who advocated war on civilization and that persisted with their war drums despite all. They persist to this day. What motivates them? (Well we see from the 'Scooter' Libby trial that they see themselves as soldiers in some mythical army; Onward Christian Soldiers, as it were.)

hc said...

Joshua, I cannot see the relevance of your point.

Rabee, I think logicians are concerned with deriving rules of thinking that suggest what is the case.

A lot of people are concerned with truth rather than self-interest. Hence the rules of logic apply to their endeavors.

Anonymous said...

It seems to me absurd to think that a person who has a postive incentive to lie and who knows that his lie is impossible to detect would instead tell the truth.

It simply is not economics. Might as well chuck out the rest of economic reasoning if we are willing to accept this type of thinking.

hc said...

Rabee, I am not sure where this is leading. Are you saying that motives need to be assessed in considering the case for a political proposal on say climate change?

Or should you not just consider the implications of the proposal?

How does understanding motives help you decide on the value of the policy proposal?

Anonymous said...

We can't abstract from motives. People are influenced by self interest. It all depends on the framework within which people are interacting.

Now there are mechanisms that seek to guarantee truth telling. There are also institutions that seek to do this. For instance you can be confident that an academic scientist tries to tell the truth in her research work. The reason is not a natural propensity to tell the truth, but is that she is part of an institution whose principal function is to elicit truth.

There are also separate crude mechanisms that try to do this too. For instance the conflict of interest rules in various government institutions.

There are almost universal norms and understandings that try to mitigate the effects of motive inspired lying, for instance we all agree that politicians are liars.

In most situations there are institutions that guarantee that being caught lying is costly. So people advocating positions typically go to great lengths to avoid being shown to be liars. Some for instance frame their advocacy in general and imprecise language. In such situations, establishing whether the person advocating is telling the truth usually means figuring out the institutional framework in which the discourse is taking place.

In short, I don't think that we can learn much about motives from informal meta-logic.

hc said...

Rabee,

Of course motives influence peoples claims about the world. The 'motives fallacy' just says

Proposition: you cannot sensibly falsify a proposition by identifying the motives that you believe drive a person to make their claims.

Proof: An agent may stupidly misconstrue the implications of hisd motives and propose something very sensible even though it is opposed to his self-interest.

If you want to falsify the claim you need to consider the implications of the claim itself. There care no shortcuts.

That's all I am saying. Its not a deep proposition. Its not attempting to derive implications about motives from 'meta-logic'.

tigtog said...

Of course as I argued above - at length - my view is that trustworthiness, deceit, honor, honesty, union-hating, jew-hating, believing in Santa Claus motivations have bugger-all to do with assessing whether a policy is good policy.

I have some sympathy with that view, but I think "bugger-all" is way too simplistic. Motivations matter in how rigorously a proposed policy is likely to be followed through, and also how a policy proposer is likely to respond to unintended consequences of their policies.

That's where we want to trust in their motives, because if we trust their motives we trust them to truly push for the policy rather than just lipservice it, and we trust them to sort out unforeseen problems according to those perceived motives.

e.g. It makes a real difference in how one believes that unforeseen problems in Industrial Relations will be sorted out if one perceives those with power over policy to be pro-union or pro-bosses. The two parties could have identical policies on paper and those motivations would still be relevant to assessing the likely impact of those IR policies.