Home > The Economics Profession > Krugman fails to take up the challenge

Krugman fails to take up the challenge

from Peter Radford

Sifting amongst the debris left after the recent collapse of economic orthodoxy is not easy. The big guns of the academy are clinging on for all they are worth to the idea that a few fixes and we are all back on track. They have the most to lose in reputation and standing, so I suppose this stubborn refusal to admit the extent of the defeat is understandable. But it is not acceptable. Things have to change in a more dramatic way for economics to rehabilitate itself as a credible body of thought upon which policies can be based. It is simply not enough to continue the old “freshwater” versus “saltwater” debate – that please recall is the argument between various grades of orthodoxy, the former ensconced in Chicago and other universities far from the coasts, and the latter based primarily on those two coasts.

Paul Krugman is a great example of the problem.

He presents a strong argument and  defense of a Keynesian approach to macroeconomics. This is the part of economics that deals with the big issues that occupy policy makers, especially in times of recession. In this he finds himself ranged against the more orthodox ideas of the Chicago style thinking that has invaded and taken over most of the high ground in the discipline. So he thinks of himself as an insurgent of “heterodox” thinker, and likes to do combat as a rebel.

This is curious since the rebellion is simply an attempt to reinstate the Keynesian theoretical constructs that were so successful in the immediate post war years until they were dethroned by monetarist and then hyper rationalist versions of what is called neoclassical economics. That neoclassical set of ideas was largely initiated by Paul Samuelson in 1948, when he published his enormously successful textbook and sought to incorporate Keynes within the pre-existing classical tradition.

The problem this melding produced as its legacy is that there is an inherent and implacable break between Keynesian and classical economics. Keynes spoke to aggregate issues: before him there was no notion of “aggregate demand” as a concept to be studied, managed, or fitted into policy. He argued that uncertainty was axiomatic and that the classical notions of perfect competition and so on were thus invalid. Unfortunately Keynes did not follow through and give us a full replacement for the old classical system. He was a practical and conservative person dealing only with the larger issues: indeed he invented what we now called “macroeconomics” so he could propose solutions to those larger issues, especially the devastation wrought by the Great Depression.

To put it bluntly: economics became a mess.

At the big picture level the emergence of Keynesian thinking split the subject apart. The saltwater versus freshwater fight is the result of this division.

But at the lower level much of the classical tradition survived untouched. The magic of markets; the workings of rational agents making perfectly rational decisions; the possibility of a “Pareto Optimum” – that mystical point at which an economy reaches its most efficient allocation of resources; and the other paraphernalia of classical “microeconomics” remained supreme.

So economics was cleft in two. The Keynesian constructs at policy level were based upon a worldview at odds with the micro analysis economists still relied on to understand the workings of the marketplace.

One response within economics was to replace Keynesian macro with a cobbled together version based upon classical micro. The mantra became “macro built upon a micro foundation”. This new classical tradition is the basis for freshwater thinking.

The other response was simply to ignore the contradictions between Keynesian macro and classical micro. This was the basis for saltwater macro.

Krugman falls into the second group. He voices little concern about the contradictions. Indeed he accepts them.

The obvious parallel is with physics, where the physics of very small things, known as generally as quantum mechanics, is incompatible with the physics of very large things, which is the relativity of Einstein. If physics can exist with a great divide, Krugman argues – I am putting words in his mouth – so too can economics.

This, frankly is rubbish.

Why?

For two reasons:

One, is that physics at each level is supported by empirical testing to an extraordinarily high degree of accuracy.

Two, is that physicists the world over are striving to connect to the two traditions. they do not find it acceptable to have deep contradictions within their thought, and so have set out to look for the necessary theoretical breakthroughs needed to reconcile them.

Not so in economics,according to Krugman. We, apparently will have to make do with the odd supposition that aggregate constructs are subject to inherent, and axiomatic, uncertainty. While our understanding of the deeper workings of the markets, firms, and consumer behavior from which those aggregates are composed are subject to no such thing. On the contrary those lower order objects behave supremely rationally in a world admitting of no uncertainty at all.

Talk about a muddle.

That this muddle persists is indefensible. It is time to eliminate the error left in the aftermath of the Samuelson synthesis. If Krugman won’t take up the challenge, then others will. Building a new micro based on the acceptance of uncertainty as its central axiom is the work we now need to undertake. Much has been done, much is yet to be done. One thing is clear: the challenge must be met.

  1. antonio garrido
    November 29, 2010 at 10:12 am

    Something have been done, however. See Marc Lavoie :L’economie postkeynesienne, Chapter nº2, where he put fordward an heterodox micro based in works of Simon, Lancaster,Galbraith,Pasinetti, Kalecki and Georgescu-Roegen.

    • Peter Radford
      November 29, 2010 at 5:07 pm

      I agree: I am a fan of Lavoie’s book. Much has been done. But we need to go further and work like this mainstream. Until the “establishment” is challenged to embrace it, we will linger in this twilight world where people like you know what’s going on, but the majority of our student’s still learn the broken, traditional micro.

      • Lucy Honeychurch
        December 3, 2010 at 2:20 am

        I hear that.

        When I asked these kinds of questions in my Graduate Micro classes at NYU in the 1990s, the response was – basically – swallow the ideas and regurgitate them on tests and you’ll get an A. Question the fundamentals – which really didn’t make sense to me – and you’ll risk a C grade.

        No wonder we’ve been stuck for years, if this is the way we educate our young up-n-comers.

  2. Paul Schächterle
    November 29, 2010 at 11:51 am

    Even without taking uncertainty into account i.m.h.o. standard micro economics is a mess. It is based on ridiculous assumptions (e.g. production function) and its founding idea (utility maximization) is either false or a tautology. So basing anything on standard micro will not produce sensible results. So I criticise Krugman for accepting neoclassical micro and just saying it is not important for macro economics, where he should outright reject it. But the question remains: Can you create reasonable macroeconomical models without the foundation of (reasonable) microeconomical model (which does not exist yet)? And in defence of Paul Krugman I would say that it is not excluded. The slogan “macro built upon a micro foundation” can be viewed as a mantra but also be viewed as a shibboleth on the part of neoclassical economists (ideologists).

    • Hannes
      December 1, 2010 at 8:36 pm

      2its founding idea (utility maximization) is either false or a tautology”

      Well I once thought that to be a valid criticism, too. But now it just seems like a misunderstanding of the concept of utility functions to me. Utility functions are a language of mathematically expressing preferences. Preferences say that you like some combinations of things more than others. Where’s the problem? If you say “1+1=2” that’s a way of mathematically expressing a relationship. You usually don’t criticize that to be a tautology.

      • Paul Schächterle
        December 2, 2010 at 3:51 pm

        Good point! What I actually meant to criticize were the assumptions about human behaviour that are loaded into the standard concept of utility maximisation. Without such assumptions about human behaviour *any* unforced decision by a person is a maximisation of her or his utility, including gifting things away and living in humility. Therefore you can’t construct a utility function in relation to quantities of goods on the concept of utility maximisation alone.

        Concerning aggregation:
        I am not at all convinced by Arrow’s impossibility theorem. There are methods of aggregation (e.g. Schulze-method) that fulfil all requirements of a rational group decision apart from the so called Condorcet-paradox, which is IMHO not a paradox but lies in the nature of groups.

        But … even if you accept all the – wrong – assumptions of standard micro about human behaviour there is absolutely no way to aggregate individual utility functions to make statements about collective supply and demand, hence the concept of a “representative agent” which itself is almost comically absurd.

  3. Keith Wilde
    November 29, 2010 at 12:15 pm

    Radford is admirably clear and on the mark here, as usual. He appears not to have noticed, however, that Krugman has been taken on publicly by Michael Hudson, who unfortunately does not seem to post contributions here. His argument with Krugman focused on criticism of Samuelson’s synthesis, and hence is very supportive of Radford’s complaint. A key distinction, however, is that Hudson distinguishes strongly between classical economics which came to an end in academic esteem with the rise of neoclassicism late in the 19th century. It is that neoclassicism that Keynes found defective, although he did term it “classical”. In his analysis of contemporary problems, Hudson affirms the classical distinction between earned and unearned income as one that needs to be embraced in response to the great disparity in wealth positions between the real economy and the financial sector that has emerged in the past three decades.

    As a specialist in international trade and payments (FX values), Hudson also has focused on the economic policy recommendations of American economists in the post-Civil War and Progressive era. These, he says, were a copy of the British practices in the days of its imperial ascendancy. That is, protect home industries and buy food and raw materials from colonies, abetted with the doctrine that the latter should practice “free trade”. Those practices, he says, should be re-introduced in order to deter further deterioration of America’s real economy. (Not to mention taxation of the “leisure class”.)

    Hudson’s theme is echoed firmly and eloquently in a 2009 book by Ian Fletcher, “Free Trade Doesn’t Work: Why America Needs a Tariff”. (Remarkably, although the book is super-abundantly researched and supported authoritatively, there is no reference to Hudson’s two books on the subjects of protectionism and trade individually, both of which are now in their second editions.)

  4. merijnknibbe
    November 29, 2010 at 1:42 pm

    @Keith Wilde: I agree with your first sentence. And with ‘the classical distinction between earned and unearned income as one that needs to be embraced in response’.

    @Peter:

    As much as I agree with you, the need for solutions is urgent (to say the least). I think I do have an usefull idea.

    Recently, I’ve been thinking about the question why monetary economists like Williamson and Sumner have large problems with a concept like ‘aggregate demand’ – a concept which is about the essence of Keynesian economics. It dawned upon me that answer is simple: montary economists are used to think about money and how it’s used, and ‘aggregate demand’ is indeed not a very coherent variable from a monetary point of view. Money is not only used for ‘aggregate demand’, but also for the ‘second hand trade’, like e-bay but also and, more important, existing houses and the ‘secondary market’, i.e. stocks and bonds. This might be fixed by changing the ‘Y+Im = C+I+G+Ex’ accounting identity of the National Accounts (i.e. production plus imports equals consumption plus government expenditures), which is the basis for Keynesian models of aggregate demand, into a monetized concept:

    Monetizing aggregate demand

    ‘Y + Im + SHS + i – imputed rent = C + I + G + Ex + SHB + i – imputed rent’,

    with
    * SHS is Second Hand goods Sold
    * SHB is Second Hand goods Bought,
    * ‘i’ is interest payed on mortgages (which is not included in conventional estimates of production or consumption)
    * ‘imputed rent’ is imputed rent on owner occupied dwellings, which is included in conventional estimates of production and consumption.

    This does not only couple monetary economics tighter to aggregate demand but it also enables one to incorporate for instance the housing market (existing houses) and the asset market into models of aggregate demand.

    To do this, we will have to redefine households (and companies and the government, but I will restrcit myself to households). Keynesian models mainly look at the ‘income and expenses’ account of households, monetary economists mainly look at the balance sheet and the cash flow statement. These can (and should) be combined (the recent Eggertsson and Krugman paper in fact does this, in a very rudimentary way and indeed, all of the so called micro-foundations can be left out). This will mean that aggregate demand indeed is influenced by liquidity constraints (including Minsky moments) and balance sheet preoccupations, like saving for your pension (or borrowing for your house or wedding).

    Detail: Steve Keen often includes ‘net borrowing’ into his definition of demand, which confuses many people used to the ‘conventional’ formula. Using and income and expenses accounts as well as an cash flow statement and a balance sheet in combination with the monetized macro identity formula above might solve this confusion problem.

    The good thing: balance sheets and income and expenses accounts of the sector households are already measured (National Accounts), just like cash flow statements (flow of funds).

    • Peter Radford
      November 30, 2010 at 6:21 pm

      Merijn: Have you worked this out more fully? I am aware of Steve Keen’s net borrowing approach, but haven’t seen it fully explicated. Besides, one of my ongoing gripes about economics is its obsession with accounting entities set in stone decades ago. What happened to the “knowledge revolution”, “the service economy”, or the “information age”? Sometimes I look at discussions between economists talking in terms of labor, capital, land etc and feel I am looking back in time because the terms sound so archaic. How can we solve modern problems when we are stuck using 19th century terminology?

  5. Jorge Buzaglo
    November 29, 2010 at 4:23 pm

    Most versions of Keynesianism, including of course Samuelson’s neoclassical synthesis, are examples of what Joan Robinson called “bastard Keynesianism.” If you want to get a taste of how genuine Keynesianism sounds, see for instance: “I expect to see the State, which is in a position to calculate the marginal efficiency of capital-goods on long views and on the basis of the general social advantage, taking an ever greater responsibility for directly organising investment; since it seems likely that the fluctuations in the market estimation or the marginal efficiency of different types of capital … will be too great to be offset by any practicable changes in the rate of interest.” (General Theory, p.164) Genuine Keynesianism is much closer to Kalecki than to academic economics.

  6. antonio garrido
    November 29, 2010 at 6:29 pm

    I see no point in finding the microfundations of macro (neither in what looks more interesting:the macrofundations of micro).What would be good is teach non-ortodox, critical economics.For that we need good textbooks, good teachers and universities that allow this tipe of education (and students attending).
    A little coordination among critical economist (some times more critical among themselves than against the Orthodox/Neoclassicals)would also help.
    For textbooks I would recomend the Lavoie one mentioned in my #1. I hope There is an English translation. An other good one,although a bit old and rather difficult for the kind of teaching american students are used to, is Robinson and Eatwell Introduction to Modern Economics

  7. December 1, 2010 at 11:50 am

    @Peter Radford
    “Sometimes I look at discussions between economists talking in terms of labor, capital, land etc and feel I am looking back in time because the terms sound so archaic.”

    What do consider is wrong with the classic definition of land together with Ricardo’s law of rent – apart from a bit of expansion?

    • Peter Radford
      December 1, 2010 at 12:12 pm

      Carol: I react that way because the terms are so imprecise and vague. The variations within “labor” for instance are huge. Is it labor as energy? Is it labor as skill? etc. The same can be said for capital. My own view is that we should start at a more basic level and define energy and physical resources [not just land] as the basic ingredients, then expand from there. This has the advantage of allowing us to build in a direct relationship with the environment immediately, and allows us to insert variables for the skill [i.e. knowledge] of the workforce, source, or type of capital etc.

      This is obviously preliminary, but I stand by my statement that the older terms reflect the issues and values of a bygone age. They need to be updated and refined, especially in our post-industrial context. Perhaps that’s the “expansion” you refer to?

  8. December 1, 2010 at 10:04 pm

    Dear Peter

    I am not an economist (though I have a degree in economics – from which I learnt almost nothing). I’m purely an amateur. But I know I’m not alone in feeling no compunction to redefine the factors of production. (I also know that you are not alone!)

    Labour is active, land is passive, capital is an intermediate product of labour on land. Energy is just capital – it requires extraction and processing of a natural resource by labour. I have just had an exchange with an accountant friend who tells me that land is ‘real capital’ – so accountants are to blame for the muddle! Economists constantly conflate land and capital which renders a great deal of their analysis defective. The main attribute of land is that it is in fixed supply. The definition of land includes natural resources contained within the location element of land until they are extracted (by labour) when they become capital. Natural resources are consumable. The location element of land can be used continuously and is unconsumable. When land can be used for production, housing, leisure, etc., yet stands idle that is a permanent loss which can never be reclaimed. This is market failure – land not allocated to best use. Correction is easy by collection of the unearned rent for public benefit. If you have to pay the rent you will use or sell to someone who will.

    The value of land is dependent on natural attributes, legal permissions, local public and private investment and the wider ‘state of the nation’. Good public investment feeds directly into land values, which could then be recycled via an annual land value tax, with frequent, regular assessments. Instead we get finance piling into unproductive land speculation and the accumulation of extreme wealth for which no labour has been expended.

    I don’t think the definition of land is vague, just ignored. Recognition of the importance of land in the economy (land appears to be invisible once it is built upon) would help us to build ‘a direct relationship with the environment’ and a far more just society.

  9. Merijn Knibbe
    December 2, 2010 at 1:04 pm

    @ hannes

    The Arrow paradox states that if you’ve got different groups in society with different preferences you can’t lump the individual prefrences together into a consistent, well behaved ‘utility based’ neo classical social welfare function anymore. The results become ‘chaotic’. I.e., the ‘representative consumer’ of microfoundations is not representative but undefined.

    On the tautology: it’s good to read the noble price lexture of Samuelson. He agrees with the tautology argument but states that measuring ‘revealed preference’ nails down preferences and therewith prevents neo classical economics from becoming a tautological science. To bad, however, that more than forty years after this lecture we still can’t measure ‘reavealed preferences’.

    • Hannes
      December 2, 2010 at 3:00 pm

      Paul’s statement that I was replying to was about microeconomic utility/preferences. I stilly think Utility functions are a useful tool for writing down which offered exchanges of bundles of goods a certain person would accept and which ones it would not accept.

      About the social utility function: I’d still accept to use it but not exactly with the social welfare/utility story told. More like: “If these are our (the guy checking the paper) preferences” or sometimes “if these are the market results without the certain friction we are analysing (not exactly of course because the utility function does not contain results), then how does the world with the friction analysed in the paper compare to this?”

      • December 2, 2010 at 4:27 pm

        I’m still thinkin’ things are worse than that. Keen’s first lecture on Behaviorial Finance:

        http://www.debtdeflation.com/blogs/lectures/

        One might still try to argue that Samuelson et al present a “normative” case, not a descriptive one, but ought should imply can, I thought, and Keen argues people CANNOT evaluate bundles of choices as even a normative interpretation suggests.

        As to Arrow: My favorite parts Footnote 1, second paragraph, page 1 of the 1963 edition of “Social Choice and Individual Values” and then Footnote 50, page 110.

        I guess that’s where the “Austrians” and the “Libertarians” just say “So what? Freedom/liberty is necessarily chaotic/irrational.”

        That may fly in America when only Pacific Islands and Bangladesh are sinking into the sea; but it’s starting to happen to Norfolk, Virginia … oh, yeah, New Orleans …

  10. omahkohkiaayo
    December 2, 2010 at 5:41 pm

    On the Debunking of Economics: Review of the Work of Steve Keen

    http://wwwthesixthestate.blogspot.com/2010/04/debunking-of-economics-review-of-work.html

  11. December 3, 2010 at 6:35 pm

    “I am not at all convinced by Arrow’s impossibility theorem.”

    Schultz-method violates independence of irrelevant alternatives. By my recollection, MANY of the proposed “solutions” succeed by dropping that condition, or making it weaker. Maybe that is the way to go: Surely, for a person, a choice ‘tween A and B should stand whether or not C’s available. So maybe, thinking that the “macro” solution should run as an analog to the “micro” solution is an ideological tenet…and maybe this point is redundant as I look back over the original post and comments.

    • Paul Schächterle
      December 3, 2010 at 7:01 pm

      “Schultz-method violates independence of irrelevant alternatives.”

      Sure, but AFAIK only because of the possibility of Condorcet-paradox triples (example: P1: a>b>c, P2: b>c>a, P3: c>b>a). If you remove one alternative, such a triple breaks up into a group preference (example: P2, P3: b>a; P1: a>b). So Arrow’s theorem is proven mathematically, ok, but IMHO does not at all give reason to believe that there are no rational ways of aggregating ordinal preference orders. And, of course Borda works if you accept cardinal judgments.

      • December 3, 2010 at 7:28 pm

        “..does not at all give reason to believe that there are no rational ways of aggregating ordinal preference orders.” Lord, this is pushing me way back into the recesses of things lernt long ago – I don’t want to say there are “no rational ways”, but rather, one is changing the criteria of rationality. And that may be what’s to be done. Hence my thinking I was maybe redundant, since a central point of the topic post maintained it was a mistake to “replace Keynesian macro with a cobbled together version based upon classical micro. The mantra became “macro built upon a micro foundation”.” I REMEMBER LEARNING THAT EXACT PHRASE even as an undergrad philo guy taking some econ classes.

        And suspecting that even Borda can be manipulated, I had to confirm that with a quick look at Wiki.

        Maybe those are unrealistic objections – I think I remember Tullock and Buchanan arguing in defense of simple majority rule on grounds that all the counter-examples were invented by people like me but better mathematicians out to destroy “democracy” – maybe not so polemically put. Surely, there HAS to be some better way to elect our President, as the election of 2000 had to show. But ranking “possible worlds” and then choosing the best of them? Me and Glenn Beck? Hoo boy …

        I think I am with you in hope, but lost in despair … Gonna go watch that new cartoon about QE2 now.

  12. Paul Schächterle
    December 3, 2010 at 8:00 pm

    “…one is changing the criteria of rationality.” That is correct. Maybe I should have said: I don’t accept Arrow’s definition of a “rational” aggregation. As for Borda: Maybe I shouldn’t have used the label “Borda”, but if you accept the idea that people are able to judge on a cardinal scale — again that is excluded by Arrow’s premises — then it is very easy to aggregate without the possibility of manipulation, simply by adding up the individual judgments. At least I think so. No liability assumed. ;)

    • merijnknibbe
      December 4, 2010 at 4:25 pm

      Your idea still seems to be that individual judgements are consistent, stable and transitive. That might not always be the case – people are prone to ‘herd behavior’ as recent Irish (and USA and Dutch and…) developments show. But that’s just my idea. There is whole science that looks at what people actually do, instead of just making assumptions about what they ought to do. And I do think that any economist thinking and writing about preferences and the like should now this science, one might consult for instance ‘Consumer behavior and marketing strategy’ by J. Paul Peter and J.C. Olson. A quote, from the introduction:

      “Consumer behavior is dynamic because thinking, feelings, and actions of individual consumers, targeted consumer groups, and society at large are constantly changing … Consumer behavior involves inteactions among people’s thinking, feelings, and actions, and the environment”.

      This is completely at odds with the thinking of people like Becker, Samuelson, Lucas and the like, who explicitely assume stable behavior, transitive choices and the like and do not bother to distinguish between thinking, feelings, and actions or to explain the asumptions. Authors who write about consumerbehavior research pay ample attention to the validity of their basic ideas about human behavior (150 pages in the book just mentioned).

      One nice difference between consumer science and the neo classical theory of exchange: neo classical theory assumes that utility is attained at the same time a purchase is made (an important building block of General Equilibrium theory). Consumer science pays ample attention to what happens after the purchase and how to manage buyers and their feelings after the purchase.

      Modern, scientific economists know about this science.

      • Paul Schächterle
        December 5, 2010 at 9:50 pm

        What I actually wanted to express is that you do not need Arrows impossibility theorem (which itself is based on premises that are at least problematic) to state that you can not aggregate individual demand curves even if you accept the basic premises of standard, i.e. neoclassical, micro.
        Personally, I do not believe in the neoclassical micro model even on an individual basis. But it does not matter since it is proven that you can not aggregate individual demand curves. And that does not depend on the Arrow theorem.

  1. No trackbacks yet.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.