In defence of Mankiw
When it comes to looking at policy, I started life fairly heavily left wing. When I started university at the age of 18, my first textbook was by Greg Mankiw. He was a Republican, while most of my economics reading at the time had been Marxist or a frustrated attempt at reading the General Theory by Keynes. I was immediately certain that I would hate the textbook, and that it had no value – at that point I was even more immature than I am now 😉
I was utterly and totally wrong – a situation I have become accustomed to. Mankiw’s first year textbook is clear, to the point, and is honest about what the economic method is and what it achieves. He “wears his assumptions on his sleeve” which I have learnt is the distinction of the best type of economist. His textbook, and his papers on macroeconomics and tax, have been insightful for me as a way of not just understanding economic ideas, but of understanding the economic method.
So I see he wrote a paper called “defending the one percent“. Undeniably it was titled that way to irritate people. And undeniably it succeeded. (Update: I’d note Cochrane states it is mistitled – and I believe to an economics auidence it is. I touch on why I think he gave it that title for his target audience below)
I was sent the paper by a reader on twitter (as I was reading it), and he was unhappy with it. We discussed it, we didn’t agree (that is not the purpose of discussion on normative issues), and that was good.
What isn’t good was when I saw Noah Smith decided to go to twitter insulting it. Noah is an insightful guy, with clear and consistent views on what economics is. And if he had read the paper more charitably I find it hard to believe he would have made a lame crack about it being any good at all.
Mankiw comes in, with standard economic welfare analysis, and applies well established principles of equity to try and articulate the impact of his assumed “cause” of changes in inequality – and the way that policy would respond given the fundamental equity-efficiency trade-off. This is basic “normative economics”. Note it wasn’t written to set the world of welfare analysis on fire – from what I can tell it is written to clearly articulate economic concepts around redistribution to the lay reader.
Now an economist might say “what is the point” – to which I’d note again, this paper is obviously not written for economists. The title is likely the way it is because there are people out there yelling “arg 1%” and he is saying, hey let’s think about this 1% a bit more and the idea of taxation.
All the stuff in this paper should be standard knowledge to virtually anyone who has studied economics – and I don’t mean at a high level. He is just trying to show the nature of the debate we can have around income inequality for non-economists – and illustrate that there are complicated issues, both in terms of getting the right “measurement” of things, and in terms of our normative assumptions around fairness. Any economist who has a modicum of modesty would be willing to accept just how danged hard this issue is – which is why it is something you don’t really see from bloggers (my fellow NZ blogging community excluded 😉 )
And his conclusion:
In the final analysis, we should not be surprised when opinions about income redistribution vary. Economists can turn to empirical methods to estimate key parameters, but no amount of applied econometrics can bridge this philosophical divide. I hope my ruminations in this essay have convinced some readers to see the situation from a new angle. But at the very least, I trust that these thoughts offer a vivid reminder that fundamentally normative conclusions cannot rest on positive economics alone.
He is writing to intelligent non-economists about tax and redistributive policy, and illustrating that there are trade-offs, and fundamental normative questions, that have to be faced before we can make a conclusion. Simply concluding “soak the rich” based on a single graph without context and analysis is moronic.
Seeing otherwise wildly intelligent economists say “MY POLICY IS THE GOOD POLICY, DO MY POLICY AS I AM ECONOMIST” is embarrassing – I accept all these guys are a lot smarter than me, but even with this intelligence they do not have any miraculous insight into the values of others. As a result, when it comes to these types of policy conclusions economists SHOULD (normative hat time) be making the trade-offs clear and concise, not grabbing hold of the economic levers and designing the society they want!
The counter to this is probably “but we do have the analysis, and we should make these changes based on it”. To which I’d answer “where”? Where the hell has someone actually done a god honest analysis of the fundamental trade-offs, then tried to determine the normative principles that society holds? I’ve seen Saez talk about the revenue maximising level of tax (fairly irrelevant to this), and I’ve seen Stiglitz just assume that the 1% are rent seekers (fairly bold claim). But where is the simulation work (eg behavioural microsimulation modeling), the work on implied labour demand responses, the estimates of where the tax burden actually falls (rather than where we place the tax)?
In fact, can you please tell me where this work is, as I would actually like to read it. I’ve seen bits and pieces in my travel, but a full objective analysis – and a full discussion around the normative principles that would be satisfied by counterfactual policy changes, given an estimate of a social welfare function – are the things the US needs before we can even try to say one sides policy prescriptions are “right or wrong” in any intelligible way. Even parts of that stuff – if you could just leave the papers in the comments I will send you a <3 . And if you add how you are using this analysis to form your views around policy you’ll get an extra <3 .
Economists will tell their first year students how amazingly important these concepts are … and then a surprising number will talk around them, give (inadvertently) misleading analogies, and/or avoid measurement. Mankiw comes out and lays his framework at people’s feet, and tries to articulate “why” the value judgments he’s applying lead to the policy … instead of just couching them in hidden terms, and demanding agreement. Can’t we have a bit more of that?
I like Mankiw’s essay. My person opinion is that a lot of the hate against the 1% is from people who have no understanding of what risk is and why capital allocation is very valuable activity. (John Cochrane in the Journal of Economic Perspectives finance issue has a good article on this stuff).
In terms of why a lot of academic work comes attached with a “lens”, I’d argue that many economists don’t see economics as a tool to use in obtaining some sort of better understanding about how the world works.
I think that the 80/20 principle applies – most economists stick to the basics, but there is a good proportion who ignore basic principles to push policy goals that are in contradiction with what most economists would think make society better off in the long run.
For example, I don’t think that Paul Krugman actually believes most of what he writes. I think he is clearly milking his NY Times column and the worship of fans for all that he can. His lecture notes and stuff on international trade is a whole world away from his blog posts and columns. From an economic perspective, perhaps he values public acclaim more than dispassionate analysis.
Some people I know who did well in 100 level econ but moved on to something else are better able to articulate why a certain policy would have welfare destroying trade-offs than some people in my tutorials at 300 levels!
From what I’ve read of the US economic scene, many economists seem to “pick a team” and work backward from there. An objective economist sadly seems to be like a purple squirrel. The marketable conclusions get the attention (Saez reckons 70% marginal tax rate wouldn’t be a terrible idea) and the polite alternative explanation proffered by someone “on the other team” gets dismissed without any inquiry.
That Noah dismissed Mankiw so quickly is just him signalling that he’s on a different team to his Twitter fans. Which is disappointing because Noahpinion is generally a very good economics blog.
I like the way he structures questions – and tells us exactly what he is assuming. People say they don’t agree with the assumptions, and then are rude about him … I don’t get this, isn’t the fact you can so easily disagree a positive thing?
I would love to be able to write that clearly, or to fit that many ideas into my mind and then articulate them in a way that makes them seem simple. Tis a skill!
Matt
Are you really pretty heavily Rawlsian? IIRC, you like to invoke decision making behind a veil of ignorance, using the idea of decision making under uncertainty, using expected utility maximisation as a route to utilitarian type thinking. If so, this makes you very much a follower of the economists Harsanyi and Vickrey (who did the whole veil thing before Rawls), but not of Rawls.
I know Vickrey, but I didn’t know Harsanyi.
Is the difference that Rawls comes with some associated value judgments on top of this schema – in that sense you would be right that I’m not strictly Rawlsian in that sense, and have been using the term (here included) more as short-hand for what you’ve described.
I know that, on a personal level, I find the idea of the “original position” compelling – akin to a social contract where concepts of justice are separated from your “realised” position in society. What I find difficult is asking, in that position what are the issues that I really care about in regards to equity?
I haven’t read Rawls since I was about 21/22, and my views on what constitutes economics, and the economic method, have developed significantly since then. Perhaps I should read some things on moral philosophy and try to make sure my inherent view isn’t contridictory 😉
If you have any suggested reading I’d appreciate it! I’ve spotted this book:
http://www.amazon.com/Justice-Political-Liberalism-Utilitarianism-ebook/dp/B001EWDFF0/ref=sr_1_1?s=digital-text&ie=UTF8&qid=1371775524&sr=1-1&keywords=John+Harsanyi
And I am a big fan of reading these types of series of essays
(Caveat – I’m armed with nothing beyond undergrad. phil)
Rawls’ veil of ignorance is a “thick” one, and this has attracted criticism. The whole purpose of the original position is to make us abstract away from morally irrelevant features (such as gender, ethnicity and the like) and focus only on what counts. But Rawls wants the veil to hide things like what people think the good life is, and the probabilities relevant to being born into a given station. We can question whether such a thick veil is justified, when something “thinner” (just veiling obviously morally irrelevant factors) would quite possibly lead to utilitarianism.
Rawls isn’t defenseless against this charge that he has “rigged” (to use Hare’s accusation) the thought experiment to get his principles of justice – he argues the chosen principles must agree with our other ideas about justice in reflective equilibrium. So in this sense the original position was never alone intended to justify justice as fairness.
Whether this is a convincing defense is up to you. But it’s very possible to get utilitarianism from a thought experiment like the original position. (And I believe Hare argues that a better set up of the original position would lead to utilitarianism. In a sense the method may be less controversial than the specifics)
Undergrad Phil is more moral philosophy than I have 😉
“But Rawls wants the veil to hide things like what people think the good
life is, and the probabilities relevant to being born into a given
station”
That is a good point. I’d always felt that rather than being a direct thought experiment to abstract away everything, the goal from it should be to get an objective ranking of preferences that is not clouded by views and interests with regards to the current endowment of resources.
In this context, a hypothetical (although unobservable) unique original position should exist in the absense of value being “socially constructed” … which as you point out can be an issue.
I prefer using the idea of the original position to try to step back and ask “what sort of society, with given moral judgments, would have chosen the society we are in – or the society which would occur with X policy”. In this way, I use it more as an idea for moving backwards in though – rather than trying to solve for “optimal policy” which is a far harder question 😉
The question is, which Greg Mankiw do you like? Take a look at this post:
http://economistsview.typepad.com/economistsview/2013/03/reactions-to-mankiw-on-the-long-run-budget-path.html
The one that is clear with his assumptions, and writes about issues in a way a lay audience can understand. In the article they are criticising there he is informing people that in the long term the government should look at stabilising debt-GDP – that it isn’t about “surpluses” but it is about this. Then a whole bunch of people attacked him because “we have a demand shortfall now” … which wasn’t his point.
Mankiw seems to know that we should compartmentalize issues, and that this is an economists “strength”. He knows the strength of the method is in the transparency of assumptions (which makes his arguments an easy target – same with his book). But to be honest, that is the only method of understanding and teaching economics that is worth its salt.
I struggle to understand why so many of these left-wing blogs dislike his writing so much. I guess it is because his value judgments differ – hell they are different to mine as well. But he is one of the few economic writers out there that gives an argument for lay economists rather than using appeals to authority. Some people may think his analogies are misleading – and this would be a strong criticism – but I don’t agree for the most part.
Fair enough, Matt, but where you see clear assumptions, I see dissembling. Like Reinhart and Rogoff, he crafts an article that appears, at first reading, to support, say, the case for austerity, but embeds caveats so as be able to say, no, no, that’s not what I said.
All of which takes nothing away from the beauty of his textbooks. I’ll be with Perloff for Intermediate Microeconomics next semester, so I’ll be interested in comparing the two.
Fair call – but I would note that, by saying his assumptions so directly, he has made it very easy for people to quote them and say “this is stupid, he is stupid” … when he’s obviously not stupid. This clarity of assumptions is something all economists should do GIVEN THAT what they say is likely to be used as evidence for things!
Wow, only half way through the essay, but I found the quote:
“Our reluctance to apply utilitarianism at the global level should give us pause when applying it at the national level.”
Very thought provoking.
Just finished it, really liked the essay, thanks for posting.
Thanks for giving it a full read – the fact he links conclusions to the fact we have to make normative assumptions, in this case about moral judgments around distribution and fairness, is pretty cool. These are hard issues, and show the sort of care we should place in trying to analyse issues – interesting stuff!
A few thoughts (sorry this may be long):
1) It is a clearly written paper, and like you I applaud Mankiw for precisely laying out his moral assumptions. Mankiw also stresses that he’s not engaging in his field of expertise. I wish more economists would be clear and discuss the moral thinking that lies behind their policy proposals, because in some cases disagreements may be not about the way the world is, but about how it should be
2) What is this “normative economics” you speak of? I’ve never heard of it, just as I’ve never heard of normative physics. Both economics and physics can tell us the way the world is. To make normative judgements, we have to engage in moral philosophy.
So when you say:
‘The counter to this is probably “but we do have the analysis, and we
should make these changes based on it”. To which I’d answer “where”?
Where the hell has someone actually done a god honest analysis of the
fundamental trade-offs, then tried to determine the normative principles
that society holds?’
we’re considering two completely different fields, right? The people doing the analysis of normative principles are moral philosophers in the philosophy journals, and the people doing the analysis of the economic trade-offs are the economists in the economic journals. I understand if your problem is economists making unjustified and hidden moral judgements. I think the solution is to say “Contingent on X being the right moral theory, it follows from the preceding economic analysis that Y is the right action to take”. It’s not the job of economists to really analyze moral theories: there are already a lot of really smart philosophers doing that, and wouldn’t it be best if economists don’t spread themselves too thin, and leave the analysis to the experts?
3) The problem with the paper is that we see a top economist doing bad moral philosophy. He starts his essay with an obvious Nozick analogy (Wilt Chamberlain) without mentioning Nozick. His objection to utilitarianism is bringing up actions he claims it would require that are not well-enough justified by him (would utilitarianism really require marginal tax rates on rich countries analogous to the rates on individuals, given questions over how much distribution would leak away in the transfer, and the lack of knowledge in development economics?). Mankiw also never considers the question of what to do if a moral theory conflicts with intuition. If we develop a moral theory that seems to take seriously the moral equality of all individuals, but it recommends something that goes against intuition, where do we go from there? How do we think about this dilemma? I see no nuance in how Mankiw addresses the question.
His dismissal of Rawls’ justice as fairness is similarly premature. Is it so obvious that those behind the veil would create institutions where forced donations occur, considering that they could well be the one involuntarily put under the knife? I think I’d rather choose a society set up in a way as to make voluntary donations as common and easy as possible (through nudges, advertisements, and perhaps organ markets).
I think if someone’s aim is to show that there’s little inequality of opportunity, that skill-biased technical can adequately explain recent trends, that utilitarianism is a fundamentally flawed moral theory, that Rawls’ thought experiment is a fundamentally flawed way of thinking about justice, and that a “just deserts” theory of justice is correct, all in 22 pages, then they’re going to have a bad time.
That being said, I’m in a way glad he wrote the paper because it’s important to discuss the topic.
Good points, cheers.
Normative economics is a name economists use for when they make ought statements – it is a field most economists are pretty keen to pretend they are avoiding but showing overconfidence that their result explains everything 😉
I agree with you that economists area of expertise should be describing the trade-offs, rather than trying to apply moral judgments to reach conclusions. Hence why I was so keen to defend Mankiw when he did point this out – compared to the situation where economists make implicit value judgments.
“It’s not the job of economists to really analyze moral theories: there
are already a lot of really smart philosophers doing that, and wouldn’t
it be best if economists don’t spread themselves too thin, and leave the
analysis to the experts?”
Indeed. Although I would also note the distinction between physcis and economics – namely that we have a significant number of unobserved variables that pop up when trying to model behaviour (expectations has to be the biggest). I think the difficulty this provides in terms of making statements about what is, and the fact that economists get pestered to give guidance, leads them to feel more comfortable saying what should be done then they … well … should.
Personally, I’m a big fan of interdisciplinary work and interdisciplinary papers. And I’ve noticed that since the 1980s social science disciplines are going hard in that direction, and a respect for philosophy has returned after a period of people trying to ignore it (not that I was around back then – just from reading papers, and chatting with economists who were in that period). I’m hoping interdisciplinary survey papers become the sort of things that influence policy – but that is just me being idealistic, and ignoring the type of conflict that is always inherent in groups 😉
“Normative economics is a name economists use for when they make ought statements…”
Then it’s just a description of a state of affairs, rather than a discipline or sub-discipline with its own claimed methodology? Because philosophy does have its own methodology, as unique as it may be: there’s a reliance on formal arguments, reasoning, thought experiments and perhaps reflective equilibrium (see here: http://plato.stanford.edu/entries/reflective-equilibrium/) I’m suspicious of the use of the term “normative economics”.
” Although I would also note the distinction between physcis and economics…”
Yep you’re very right here. And given that you say economists come under a lot of pressure to say what should be done, I need to take back what I said and accept that it is important that economists know some political and moral philosophy. I still think the best solution is to be explicit in saying “This policy advice is contingent on X being the right moral theory”, in which case familiarity with moral philosophy is primarily what is needed. I’m undecided about engagement – it’s great to see Mankiw debating Rawls and utilitarianism. I just don’t think he did it very well.
” And I’ve noticed that since the 1980s social science disciplines are
going hard in that direction, and a respect for philosophy has returned
after a period of people trying to ignore it”
Glad to hear!
“Then it’s just a description of a state of affairs”
Ha, sorry I was just being facetious about the use of the term! I’d say any demarcation economics does between “positive” and “normative” solely comes from the “equity-efficiency trade-off”. Most economics is followed on describing efficiency arguments and distributional consequences – when normative economics is when economists explicitly try to discuss policy considerations, and thereby state they are making moral statements.
There is a keen urge to make “positive moral statements” where possible, but I get a bit uncomfortable here, and just try to leave it be.
“I still think the best solution is to be explicit in saying “This policy
advice is contingent on X being the right moral theory”, in which case
familiarity with moral philosophy is primarily what is needed”
You’ll find the authors here agreeing with you on that. And economists who attempt policy conclusions tend to lean this way as well.
It may be an information problem though, the public can’t tell whether conclusions are determined on the basis of a consistent and appropriate moral framework and as a result reward economists on the basis of other loosely related, but observable, critera (such as confidence). This may lead to incentive problems for economists when it comes to “investing in a skill set”.
I’m not really sure – I just know from a personal perspective I have far too little knowledge of moral philosophy and try to stick to discussing trade-offs. Unless I’ve forgotten to have a coffee, in which case my prejudices no doubt slip out 😉
When economists say “efficiency,” 90% of the time they mean Kaldor-Hicks efficiency, 9% of the time they mean Pareto efficiency, and 1% of the time they have an explicit social welfare function that looks at distributional consequences. Over time I’ve come to appreciate the importance of Kaldor-Hicks efficiency improvements, because roughly speaking a society that pursues these will end up a lot richer over the long term with distributional consequences whose unpredictability looks vaguely like a veil of ignorance.
Indeed.
However I get a bit concerned about the way potential pareto improvements are treated – often the focus is tacit, based on an underlying statement about some factor like GDP – and often the “transfer” implied will in turn be seen as an uncomfortable sidenote. When economists do that, they are making implied moral judgments … a thought that is interesting.
GDP is in fact the operational approximation most often used in a Kaldor-Hicks evaluation. The problem with this use of GDP isn’t moral but operational–GDP gets some things systematically wrong in measuring welfare. Unlike many critics, I think the biggest problem is that it undervalues hedonic improvements of all kinds, but that’s a different topic.
What I meant about the “wisdom” of Kaldor-Hicks as a rough guide to policy is partly its implied morality–that material wealth is good on the margin but we don’t want to try to interpersonally compare marginal utilities. The combined moral/prudential judgment embedded in KH is that the dynamic consequences of trying to make those “potential” Pareto improvements “actual” would be dire. It would be like applying eminent-domain law to every transaction in the economy: Want to open up a new, more-popular restaurant? Can’t do it unless you cut in the existing, less-popular restaurants in on your profits. Gridlock is far too tame a metaphor for what would happen to the economy under a true Pareto-improvement standard.
“GDP is in fact the operational approximation most often used in a
Kaldor-Hicks evaluation. The problem with this use of GDP isn’t moral
but operational–GDP gets some things systematically wrong in measuring
welfare. Unlike many critics, I think the biggest problem is that it
undervalues hedonic improvements of all kinds, but that’s a different
topic.”
Indeed. Welfare measurements often move towards consumption instead as a result – and I think the “long-term” consumption view is generally a good way to start a policy debate. However, GDP or consumption will then often we taken as the end point also – something I don’t feel as comfortable about.
“Want to open up a new, more-popular restaurant? Can’t do it unless you
cut in the existing, less-popular restaurants in on your profits.
Gridlock is far too tame a metaphor for what would happen to the economy
under a true Pareto-improvement standard.”
Hehe, I like that description, very nice – I may have to steal it 🙂 . I’d probably make a distinction between the actions of government and the actions of all other agents. Government, due to its position as arbiter of the social contract, with the ability to set the rules of the game and redistribute through tax and transfers, has to take into consideration the distributional impact of policy. This involves taking things a step further when doing analysis of what should be done.
Personally, I think information requirements imply that most of the time, merely using Kaldor-Hicks to come to a conclusion can be best – but I’d be remiss if I did not say that I felt the argument needed to be made 🙂
Louis Kaplow’s textbook is nice. He uses a ‘distribution neutral’ approach to tax reform packages. He covers quite a range of instruments, includes serious discussions of moral theories and also considers a wide degree of expectation formation mechanisms and preferences.
I give the dynamic Mirrleesian models a bit more weight than Mankiw does. Sargent and Ljunqvist, Salanie or Kocherlakota are probably good places to start. It is difficult to talk about transfers outside of a Mirrleesian model, although imo there is a lot more work to be done before we can take the numbers seriously. I think though that this literature is moving very very quickly right now in the very capable hands of Farhi, Werning, Golosov, Kocherlakota, Albanesi and others…
Anyway, these models are really useful when we want to analyse issues where we do think that the ‘first best’ could look pretty utilitarian, but for some information asymmetry issues (for example unemployment insurance and disability/accident insurance, but not necessarily welfare in general). They are also useful for political economy formulations, where we can be a lot more “positive”.
As an aside, it is very important when reading mechanism design approaches to tax and welfare problems to keep private markets in mind. Kiwi Melanie Morten has a really nice job market paper that considers the interaction of a government program with informal insurance networks in village India: http://pantheon.yale.edu/~mem226/Site/Research_files/JMP_Morten_Website.pdf
Hi,
That book sounds great – is it this one here: http://press.princeton.edu/titles/8656.html
“I give the dynamic Mirrleesian models a bit more weight than Mankiw does”
Indeed, I was actually pretty surprised to see some of the transfer principles he moved against – I’ve always found incentive compatability as on of the more consistent reasons for the existence off the current tax system. But he was clear about why, and that’s all I was really after. Furthermore, I do really strong agree with him in terms of labour demand/GE effects – I’ve seen more of the public policy journals fiddling around on this margin as well.
“As an aside, it is very important when reading mechanism design
approaches to tax and welfare problems to keep private markets in mind.
Kiwi Melanie Morten has a really nice job market paper that considers
the interaction of a government program with informal insurance networks
in village India: http://pantheon.yale.edu/~mem2…”
That is very cool.
That’s the book. I just picked it up a few months ago and I love it. Kaplow is a law and econ professor at Harvard and he publishes in top journals in both fields. He has all the technical ability to present frontier economic models, yet his discussions of assumptions are very rich, and his explanations of the economics are very intuitive.
That’s it, I’m buying it. I really don’t have time to read it … but I’ll find a way.
I apologize for offering a half-remembered tidbit (if I remembered it better, the last fifteen minutes I spent looking for a source might have proved more fruitful), but I recall reading an assertion that, for a certain set of data including information one might expect to predict a teenager’s adult earnings (parents’ income, test scores, etc.), the best predictor was how much the teenager said s/he wanted to earn. (Actually, now that I think about it, I might have seen it cited by one of the papers Mankiw cites on heterogeneity of preferences.)
For most normative analysis, I would say Saez’s “revenue maximising level of tax” is relevant insofar as it ought to provide an upper bound on tax levels; people who knowingly advocate taxes above that level are tipping their hand that, no, they aren’t worried simply about the poor, they are envy-driven who, like the Puritans, can’t stand the thought that someone might be happy.
Heya,
That is a cool anecdote, thanks. It is interesting how much the heterogeniety of preferences, and the value of leisure as the opportunity cost of earning income, gets ignored in public. It is a neat issue to have an idea about, as it doesn’t just impact on concepts of fairness – but it also influences the argument around the elasticity of labour supply at different income levels.
Good point – I was a bit loose calling it the Saez paper irrelevant to the question, it does indeed place an “upper bound”. I suspect that cynicism came from talking with people who confused upper bound with optimal level over the past few months 😉
Matt — just wondering, which Mankiw textbook were you referring to?
This fellow:
http://www.amazon.com/Principles-Economics-N-Gregory-Mankiw/dp/0030259517/ref=sr_1_3?ie=UTF8&qid=1372271497&sr=8-3&keywords=principles+of+economics+mankiw
I’m pretty sure that was the edition we were using when I was 18.