Quiggin vs Williamson: The home game

A while ago, my stoush with US economist Stephen Williamson over his attack on Zombie Economics (in some blog posts and what was presented as a review, for the Journal of Economic Literature) attracted a fair bit of attention around the Intertubes. Now Williamson’s longer review, to which I briefly responded here, has turned up in Agenda, published by the ANU School of Economics.

There’s a history here. Way back before blogging was born, the current editor of Agenda, William Coleman co-authored a book, Exasperating Calculators, published by Keith Windschuttle’s MacLeay Press in which I got a brief but critical mention, along with lots of others. I wrote a fairly scathing review (over the fold, also with a review of a book by Wolfgang Kasper) and I think there may have been one or two more rounds.

So, I wasn’t all that surprised to see Williamson’s piece appearing in Agenda, although I do feel (given Williamson’s putdowns of me as an Aussie yokel, and member of the “farm team”) that they could have tried for an Australian, instead of an import on this occasion. I don’t have time for a full-length response at the moment, except to say that I don’t think Williamson really engages with my argument at any point.

Quiggin, J. (2001), ‘The economic rationalists strike out’, Australian Financial Review, 6 April. Review of, Building Prosperity: Australia’s Future as a Global Player by Wolfgang Kasper (2000) and Exasperating Calculators: The Rage over Economic Rationalism and the Campaign against Australian Economists by William Coleman and Alf Hagger (2001).

It is a brave economist who revisits the predictions they made twenty years ago, particularly if those predictions were accompanied by policy advice, and that advice has been followed. If your advice has been rejected, it is usually easy enough to find some subsequent disaster and say ‘I told you so’. But economic outcomes are so hard to predict and control, and effective advocacy requires such a high level of optimism, that the successful advocate is almost bound to be disappointed by the outcomes.

Wolfgang Kasper is not deterred by such considerations. In 1980, along with four other eminent Australian economists, he published a book entitled Australia at the Crossroads, advocating a program of radical microeconomic reform. The book contrasted two possible paths for Australia, referred to as ‘mercantilist’ and ‘libertarian’. The mercantilist path was seen as a continuation of the policies of the post-war period and was described as involving:
• protection against import competition
• protection against changes wrought by new technologies
• maintenance of restrictions on capital inflows and on free competition in capital markets
• defence of a fairly rigid system of relative occupational wages and of real wages irrespective of market forces
• continuation of an extensive government role as benevolent provider of many basic services including education, health and welfare
• government by lobbying and associated control by producer groups with short sighted policies aimed essentially at winning the next election
• consumerism and environmentalism supported by bureaucratic regulation

The libertarian alternative path involved
• free international trade
• acceptance of the structural changes wrought by new technology and the removal of protection
• elimination of restrictions on international capital flows and on free competition in the domestic capital markets
• resolute application of anti-monopoly and restrictive trade practices legislation
• deregulation of many markets and other activities, especially in the area of entry by persons and firms that want to compete
• greater variation of occupational wages and of real wages in response to market forces
• reduction of the government’s role as a producer of many basic services, including education, health and welfare
• expansion of the government’s role as a provider of income maintenance

Using scenarios devised by Kasper, the Crossroads group estimated that the mercantilist approach would yield annual growth in income per person of 1.7 per cent, while the libertarian approach would yield growth of 3.8 per cent. Over 25 years, the resulting income gap is around 25 per cent.
With the wisdom of hindsight, it is easy to see that nearly all the items on the libertarian agenda have been implemented. Kasper appears satisfied with the result. Over the period since 1975 income per capita has grown by 1.8 per cent annually, a result which Kasper describes as ‘an acceleration’ and ‘not overwhelming but not disastrous’. In order to achieve this rate, as he says, Australians have been forced to ‘work harder and to compete’.

Alert readers will have noticed a problem. Isn’t growth of 1.8 per cent what we were supposed to get under the dismal Mercantilist scenario? And at least in that scenario we got to be ‘relaxed and comfortable’ instead of hard-driven and competitive. Is it any wonder that Australians are less than thrilled with radical microeconomic reform ?

Kasper has two answers to this objection. First, he says the reforms were delayed and only partially implemented. Second, he relies on Productivity Commission research, drawing on ABS data, which shows that productivity accelerated dramatically towards the end of the century, from a historical average rate of 1.2 per cent to a 2.4 per cent for the period from 1993-94 to 1997-8.

On the first point, it is worth noting that in many areas reform went further than the Crossroads group dreamed possible. For example, they did not seriously discuss privatisation and they envisaged a more extensive role for the Arbitration Commission than is involved in the system of enterprise bargaining introduced in the early 1990s.

On the second point, Kasper, like many other economists, failed to note the publication, without much fanfare, of revised statistics in 1999. These lower the productivity growth rate for the for the period from 1993-94 to 1997-8 to 1.7 per cent, and showed productivity decelerating in the late 1990s. The long-run average was also revised downwards, but only to 1.1 per cent. When proper account is taken of cyclical factors (productivity always grows faster in expansions) and other data problems, the revised data give little or no support to the idea of a productivity ‘miracle’ or ‘new economy’.

Whatever the problems of past predictions, Kasper is happy to offer new ones. This time he offers a ‘cricket’ scenario based on backsliding into regulation, and a ‘bee’ scenario based on more extensive economic liberalism. The cricket scenario, he claims, will yield annual growth in income per person of 1.25 per cent over the period from 2000 to 2025, while the libertarian approach would yield growth of 3.0 per cent.

These predictions are basically the same as those in Crossroads, except that both are a bit more pessimistic and probably therefore more realistic. More importantly, by 2025, both Kasper and his critics will have long since been replaced by other economists debating other issues.

By contrast with Crossroads, Kasper’s policy proposals this time around are surprisingly ill-defined. He has abandoned belief in ‘expansion of the government’s role as a provider of income maintenance’ (actually, I suspect this was a compromise forced on him by his co-authors), and now denounces the welfare state. He wants to limit public spending to 25 per cent of GDP. But nothing is spelt out, and most of the advocacy in the book deals with tired ideas for constitutional reform (community initiated referenda, term limits, a Hayekian ‘Third Chamber’ and so on).

A final comment relates to Kasper’s subtitle ‘Australia’s future as a global player’. I can only assume that this nationalistic phrase, quite inconsistent with Kasper’s libertarian stance, was added for marketing reasons. Readers looking for any discussion of strategic industry policy, the ‘branch office economy’ debate, or anything else that might be suggested by this subtitle, will be sorely disappointed.

Another look at old debates is provided by Coleman and Hagger, who revisit the debate launched by Michael Pusey’s Economic Rationalism in Canberra in 1991. Their book, Exasperating Calculators*, is indeed exasperating. Whenever the authors make a definite statement about their own beliefs they appear eminently reasonable. Using the method of selective quotation favored by the authors, it is possible to produce the following summary of their argument:
(1) Economic rationalism is highly unpopular in Australia, indeed the Economic Rationalist is a ‘folk devil’ (p 299)
(2) This is also true in the rest of the world ‘By the mid-1990s, Economic Rationalism – Rogernomics, Thatcherism, Reaganomics– was politically spent’ (p 261)
(3) ‘In the mind of the Economic Irrationalists [Pusey, Manne and others], it is economists who have brought about Economic Rationalism’ (p 289)
(4) But in reality, ‘pure neoclassical theory cannot be identified with Economic Rationalism’. The majority of Australian economists ‘strongly disagreed with the proposition that government outlays should be reduced as a percentage of GDP’ and ‘the stereotyping of the Professoriate as uniformly Rationalist is wrong’ (pp 206-207)
(5) Moreover, there is little evidence to back the strong claims made by Economic Rationalists ‘We do not mean to suggest, for example, that the fact that industrial production grew by a smidgen more under Thatcher than under her predecessor vindicates her policies’ (p 82)
(6) Economists should educate the public about the points on which they agree, in particular the point that there is an extensive role for government intervention, including, but not limited to, externalities, public goods and macroeconomic stabilisation. In particular, economics does not support a presumption in favour of small government: ‘as far as economics is concerned, the size of government doesn’t really matter’ (p 259)

From this summary, it would be reasonable to conclude that the authors endorse the judgement they correctly attribute to this reviewer ‘[Quiggin] agrees with the Economic Irrationalists about the [adverse] impact of Economic Rationalism. But he disagrees with their remedy: get rid of economists’.
Unfortunately, the quotations presented above give a highly misleading picture of the tone of the book as a whole, which is more accurately summarized by the use of the term ‘Economic Irrationalist’ as a pejorative label for opponents of ‘Rogernomics, Thatcherism and Reaganomics’. In a continuous series of sly asides, pointed footnotes and pedantic quibbles, it is made crystal clear that the Economic Rationalists are the good guys and that anyone who opposes them (or even fails to defend them with sufficient ferocity) is an enemy of reason and progress.

Unfortunately, as the authors have observed, the majority of Australian economists fall into the latter class. As a result, although the authors claim to be responding to a ‘campaign against Australian economists’, their book contains more personal attacks on Australian economists, living and dead, of all schools and persuasions, than any other volume I have read. Those denounced include H.C. Coombs (‘elderly’ and ‘nostalgic’), Russel Mathews (‘frenzied’), Geoffrey Brennan (an ‘appeaser’), Stephen King and Peter Lloyd (‘indefensible’), Clive Hamilton (‘florid irrationalism’), Ted Wheelwright (‘insignificant’) and even Wolfgang Kasper, among many others. (The present reviewer gets off relatively lightly, as a ‘distinguished economic theorist’, who is prone to ‘foolishness’ in matters of policy).

A couple of professional economists of the free-market persuasion (John Freebairn and Ian Harper) are favorably mentioned in passing. But the real defenders of Australian economics, it seems, are three writers who have abandoned the economics profession for right-wing politics: John Hewson, P.P McGuinness and John Stone. With friends like these, those concerned with the public image of the economics profession may well ask, who needs the enemies attacked by Coleman and Hagger?
Coleman and Hagger claim to promote values of academic integrity, trashed by the Economic Irrationalists, but conspicuously fail their own tests. They criticise the ‘Irrationalists’ for offering various and inconsistent definitions of Economic Rationalism, then fail to offer any definition of their own. This does not stop them from making throwaway statements to the effect that particular policies and ideas are, or are not, consistent with Economic Rationalism.

They criticise statistical claims made by non-economists like Manne and Pusey, but play fast and loose with the data themselves. A typical example is the following (p 113) ‘Pusey in February 1992 diagnosed the New Zealand economy as dead in the water. Over the next five years, New Zealand’s real GDP rose by 22 per cent’. Why, one might ask, does this refutation refer to a five-year period? Coleman and Hagger know, but unwary readers may not, that after a five-year recovery, the New Zealand economy relapsed into recession in 1997, and was once again ‘dead in the water’ by the late 1990s.
More trivially, Coleman and Hagger make repeated fun of Pusey’s misspelling of proper names. Yet they refer to their prominent colleague, Wolfgang Kasper, as ‘Kaspar’. Fellow-economist Brian Dollery comes off even worse, appearing as ‘Brean Dollery’ in the text and ‘Brean Dolery’ in the index. It is hard to avoid such errors, but those who live in glass houses shouldn’t throw stones.

Connoisseurs of vituperation, a field in which Australians have long excelled, will find this book a worthy addition to their shelves. Those looking for a balanced view of economic rationalism and its critics would do better to seek out the ‘indefensible’ volume edited by King and Lloyd.

Wolfgang Kasper (2000), Building Prosperity: Australia’s Future as a Global Player, Centre for Independent Studies, St Leonards NSW, 118 +xxv pp, $27.45.
William Coleman and Alf Hagger (2001), Exasperating Calculators: The Rage over Economic Rationalism and the Campaign against Australian Economists,MacLeay Press, Paddington NSW , 336pp, $24.20

see also
King, S. and Lloyd, P. (ed.), (1993) Economic Rationalism: Dead End or Way Forward?, Allen and Unwin, St. Leonards, NSW.

* The title is drawn from Hancock’s Australia which has lots of neat quotes about Australian attitudes to economists. This was a fairly obscure work by the 1990s, but I cited it on exactly this point in my 1996 book Great Expectations, a fact not mentioned by Coleman and Hagger. Just sayin’.

104 thoughts on “Quiggin vs Williamson: The home game

  1. @Chris Warren

    In the wierd and wonderful world of General Equilibrium analysis, it is possible to have upward sloping demand curves for factors of production. But this has to do with the factor intensity of production of the goods demanded as a result of the increase in income of the people whose income goes up, versus that of the people whose income goes down.

    It has nothing to do with the marginal propensity to comsume. At an equilibrium real interest rate, all leakages and balanced by injections at the equilibrium level out output. Less savings by some causes a higher rate of interest and more savings by others, less investment and higher capital inflows (1). This effect is temporary until the higher savings by others makes interest rate and the growth path converge again.

    It is logically possible that low skilled workers have amazingly high income elasticities of demand, and low price elasticities of demand, for the goods that they produce, but I don’t see any particular reason to assume it. I remember Milton Friedman’s admonition to focus on the way the economy works in the real world, rather than enumerate logical possibilities.

    (1) Now if the economy worked like a real business cycle model, the higher EQM interest rates would be temporarily expansionary, as the rest of the workforce worked longer hours for lower wages and saved by raising their income rather than lowering their consumption. However, I don’t think that this is the way the economy works either.

  2. @Chris Warren
    China has a ticking debt time bomb in the form of the debts of its publically owned manufaturing industries. These run at a loss, and the government strong arms the banks into lending them money to cover their losses in order to try and cushion employment (China has a fairly large structural unemployment problem).

    The whole thing is a ponzi game, and one day may explode in the face of the Chinese government.

  3. Tim Peterson :
    also, productivity and output have (common) stochastic trends, so a production function regression in levels will be spurious.

    Now that doesn’t sound right. So if two variables are co-integrated then a regression in levels will be spurious, is that what your saying? Are you sure?

  4. @Tim Peterson

    Except that you didn’t take into account of the Chinese Government is a dictatorship. It might pay back the money to the bank; but if they don’t want to, they have the ultimate decision whether to pay them back or not. The power to control the economy the way the government want is something that a democratic society could not do; however whether if that is a good thing or a bad thing is not easy to conclude. When it comes to the Chinese Government at least they did not provide much of the economic growth to the public but allowed them to flow through corporate and political corruption.

  5. @Tim Peterson

    For a technical critique of the concept of NAIRU please read papers and books written by Bill Mitchell such as “Full Employment Abandoned” written with Joan Muysken. I think that this blog is not the right place for such a debate, if in your opinion that critique of NAIRU is invalid, you can challenge Bill on his blog. I skimmed through that book and I agree with Bill’s arguments.

    “The reason that the natural rate of unemployment exceeds frictional unemployment is mostly mismatch between skills supplied and demanded. The minimum wage and wage rigidity in general is another reason. Show me a valid regression that shows that the NAIRYU is 2%!”

    I think that this reasoning is deeply flawed as I am supposed to agree with the explanatory value of the NAIRU theory before discussing possible causes of unemployment and remedies – or give up, admitting my ignorance. But the whole concept of NAIRU hinges on a pseudo-scientific theory that a certain level of involuntary unemployment is a natural feature of every possible market-based economy and if attempts are made to push the unemployment below that boundary the inflation will accelerate out of control. So I am supposed to agree that I will not argue for full employment before I can approach this topic and start putting arguments how full employment could possibly be achieved. All the supply-side “full employability”-type strategies only aim at mitigating the problem. Of course training people is very important but in the end if we have 50 vacancies and 100 unemployed no matter how hard these people study, the end result will be the same. And we cannot simply re-train unemployed brick layers over a 6 months period to get dentists or lawyers.

    So what am I supposed to estimate?
    “The NAIRU can be measured by decomposing unemployment in two parts. One is the NAIRU and the other the ‘unemployment gap’ (gap between actual unemployment and the NAIRU). Two disturbances are assumed to affect fluctuations in unemployment: the NAIRU disturbance and the gap disturbance. In this approach the NAIRU is defined as the part of unemployment that is inflation neutral in the long run. It is mathematically derived by setting the long-run effect of the gap disturbance on the NAIRU to zero.” ([1] Derived Measurement in Macroeconomics: Two Approaches for Measuring the NAIRU Considered, Peter Rodenburg)

    There is a certain circularity in this concept. What was observed in the 1970s was a period of high inflation and high unemployment. This was also a result of the economic policies of that period and external factors such as the oil shock. Instead of trying to understand and mathematically describe the root causes of the problems (there were obviously such attempts outside of the neoclassical stream), a quasi-empirical model was cobbled together. From that time the NAIRU model was also used to determine the policies – what included using the monetary policy for inflation-targeting as a tool of choice. Unemployment was accepted as a natural state. These policies which could have eliminated involuntary unemployment were actually thoroughly eliminated. Monetary policy is not one of them. Then I am supposed to look at the state of the system after all these political experiments were completed and run a regression to extract the parameters of the model. What if I ask whether the state of the system is a result of the policies inspired by the NAIRU theory? It is not that we only have a market, employees and employers. We also have policies. We should have Job Guarantee and active fiscal demand-management policies and also supply-side policies instead of the promotion of asset bubbles and unchecked grow of the consumer debt. As a result we also have an overgrown and parasitic finance sector. This all leads to a massive misallocation of resources. Does anyone think I don’t know what I am talking about? I am talking about the US where the construction sector was growing unchecked during the housing bubble. When the bubble burst, the unemployment in that sector reached something about 25% (in some states it was higher). This is where the NAIRU sits. It is obvious that increasing the aggregate demand up to the point when all these workers are absorbed would lead either to inflation of the prices of assets or CPI inflation. This is what also leads to the skills mismatch. It was the neoclassical economics what was used as an excuse for all the fraud in the lead up to the crisis – the worship of the “market”. And I am still supposed to believe that behind the thick veneer of intimidatingly looking maths lies true science in the sense of seeking the truth not just “science on demand” as served by Milton Friedman, the inventor of NRU.

    And this is what we get:
    “As a consequence of the difficulty of identifying parameters, the 95% confidence interval for all computations covers a very wide interval for the NAIRU, usually somewhere between 4 percent and 8 percent. Staiger et al. (1997b: 34) ague: “The most striking feature of these estimates is their lack of precision. For example, the 95 percent confidence interval for the current value of the NAIRU based on the GDP deflator is 4.3 percent to 7.3 percent. In fact, our 95 percent confidence intervals for the NAIRU are commonly so wide that the unemployment [of the USA] has only been below them for a brief periods over the last 20 years.” Other studies find similar wide confidence intervals.
    This empirical problem arises from not knowing the parameters of the model concerned. A wide range of values is consistent with the empirical evidence.” [1]

    Look I am an engineer not a “scientist”. You know what I do when I get a measurement or estimation with that level of confidence and precision. This is not a search for a Higgs boson. There is a rubbish bin next to my desk. This is where all this pseudo-science belongs to, not to the meeting rooms of cabinet ministers and central bankers.

    If you think that implementing Job Guarantee that is a very specific intervention of the government on the market with a fixed price of the labour ($15.51/hour in Australia) – not pushing on the whole market – will lead to accelerating inflation please give arguments what is wrong in this approach, preferably on Bill’s blog. There is no skills mismatch – everyone will get a job. There is no competition with the private sector – except at the minimum wage. There is no crowding out of the demand – except for very narrowly defined services such as environmental remediation. The only relevant factor is that the workers will be slightly less afraid of losing jobs, knowing that they cannot be thrown into abyss by their bosses. But this issue also has certain moral aspects. Do we want to live in a society where people are treated like animals? This may not be the case at the unemployment rate of 5% but may develop at 10% or 15% and it is very common at 20%, I witnessed that with my own eyes. I reckon the only reason Job Guarantee is rejected is purely political. Michal Kalecki wrote about it in 1943. I think that he has the last word. I am done with this topic and I cannot add much, sorry about that.

  6. @Freelander

    Running a regression of levels without lags on cointegrated variables will give you correct parameters but incorrect t-statistics and R^2. Running one with lags will give you spurious parameters.

  7. @Adam (ak)

    “So what am I supposed to estimate?”

    Wage growth as a function of unemployment, lagged inflation and expected inflation.

    “These policies which could have eliminated involuntary unemployment were actually thoroughly eliminated.”

    What politicies were they?

  8. @Adam (ak)

    “From that time [the 1970s] the NAIRU model was also used to determine the policies”

    From that time, in Germany, Japan and Switzerland, the NAIRU was used to determine policy. As a result, the great inflation came to an end early in these countries.

    In the English speaking world, policy makers like Fed Chairman Arthur Burns believed that inflation case caused by exoginous cost push factors. They believed that output/employment gaps would have no effect on inflation in the presence of strong unions and monopolistic firms, and that monetary policy should accomodate inflation to sustain output and employment. They regarded price and income controls as the only viable way to control inflation.

    As a result we got double digit inflation in those countries, which did not end until the early 80s, when policies based on the NAIRU/natural rate were brought in.

  9. @Tim Peterson

    You want me to estimate “Wage growth as a function of unemployment, lagged inflation and expected inflation.”

    So you are making an assumption that you can describe/ estimate growth of wages as dw(t)/dt =F(u(t), dw(t-tau)/dt, dw_e(t)/dt)) where w_e(t)/dt is the expected inflation.

    What I am telling you is that people who are asking the question in the way you are doing it have already got the answer. That equation which “shows” the functional dependency of the wage inflation on certain parameters ignores all the other relevant parameters. It employs the usual “ceteris paribus” method of intellectual fraud which is endemic to economics and not used in engineering or proper science – we first has to prove that a parameter has no influence on the result if we want to remove it as an exogenous variable from the analysis.

    By writing that equation (function definition) we restated that we do not care nor do not want to think about any possibilities that the government can directly employ all the workers who wish to work for the minimum wage – because there is no explicit parameter “Job Guarantee participation”.

    Obviously then a discussion shifts to what’s hidden in the constant parameters of the equation, about bringing down the fake “NAIRU” constant. What I am telling you is that the equation you want me to write cannot describe the reality correctly. One cannot lump together everything – the changes in the share of salaries of bankers and the share of wages of the street cleaners in the GDP, the changes in the manufacturing processes and the effects of globalisation, asset bubbles, various changing paradigms in monetary and fiscal policies – into changes in just one parameter “u*”. Or if we do it we’ll get a 100% percent precision of the estimation of the parameters and we will blind ourselves to the possible solutions of the social problems.

    If we do it we will get GIGO. So there’s nothing to estimate, sorry about that.

    BTW I know perfectly well why the expected rate has been introduced. Because if you have dw(t)/dt=F(u(t)) you may have got a static Philips curve and a certain trade-off between inflation and unemployment can emerge, one may say I don’t care let the inflation be 12% but I want to have the unemployment set at 3%. If you just introduce the lagged inflation as a parameter you’ll get a lagged differential equation with (most likely) a stable state solution.

    But if you introduce the expectations you effectively end up with an equation showing d2w(t)/dt2 = F(u(t)) that is an equation which “proves” that below a certain level of unemployment the inflation will keep accelerating forever (there will be a price-wage spiral) and there is no trade-off. So it has been “empirically proven” that we must accept a certain level of involuntary unemployment because nobody wants to live with hyperinflation.

    But what are these forward-looking expectations?
    Let’s look at [1] again:

    “In the most general form, the expectations augmented
    Phillips-curve relation is formulated as:
    ? t = ? te ? ? (ut ? u* ) + ? X t + ?t (4)
    where, ? te is expected price inflation, X t a regressor included to control supply shocks, ? a
    parameter, and ?t an error term. When expectations about inflation are realized, that is,
    when expected inflation coincides with realized inflation, thus when ? t ? ? te = 0 , and in the
    absence of supply shocks ( X t = 0), the NAIRU u * will coincide with the actual
    unemployment rate ut”

    And because there will be a constant mismatch between the expected rate of inflation and realised inflation, the realised inflation will keep accelerating. This argument itself looks dodgy to me but let’s leave the discussion about it for now. So we got what we wanted – the long-run Philips curve as a vertical line drawn at (u*).

    Or if we move to the pseudo-Keynesian framework:

    p ? we = ? 0 ? ?1ut + ? 2 X 1t (7)
    w ? p e = ? 0 ? ? 1ut + ? 2 X 1t + ? 3 X 2t (8)

    and then we may conclude that:

    ” stable inflation will now occur when expectations about future prices and wages are realized, thus when the identities p = p e and w = we hold. When expectations are not fully realized, a wage-price spiral will occur, as wage-setters try to regain the losses imposed on them by price setters, and vice versa, and inflation will start to accelerate. Therefore, equilibrium, and hence non-increasing inflation, exists at the intersection of the wage-setting and price-setting curve, creating a NAIRU-
    level of unemployment of u*”

    The interpretation of this dubious theory meets the definition of intellectual fraud. So the essence of the mainstream interpretation of the results of the analysis is that in the short term the bargaining power of the trade unions needs to be suppressed by bringing their bargaining power w ? p e determined by (1-u) by pushing it below (1-u*) and the long-term solution to bring down the NAIRU u* is to destroy the bargaining power of the unions and shift the wage-setting curve downwards.

    But I have a better solution. If you consider this model as valid (I don’t) and don’t like the Job Guarantee (which doesn’t fit into the model) – what about destroying the bargaining power of the firms and shifting the price settings curve upwards? Come on, we can cut some fat from the “1%”, nothing bad will happen because of that.


    “These policies which could have eliminated involuntary unemployment were actually thoroughly eliminated.”
    For example any large-scale interventions on the job market similar to New Deal or better, Job Guarantee-like programs. Also – more active fiscal policy such as public infrastructure spending (what was outsourced to public-private partnerships). But to debate that we must move to Sandpit or to Bill’s blog.

    Please be aware that I actually lived in a country which had 20% unemployment rate as a result of the implementation of the monetary and fiscal tightening policy (disinflation) based on the NAIRU theory. Actually I left that country because the society has been made seriously sick and I didn’t want my kids to grow up there. That’s why I am so personally enraged when I hear about this idiotic theory.

  10. @Tim Peterson

    “As a result we got double digit inflation in those countries, which did not end until the early 80s, when policies based on the NAIRU/natural rate were brought in.”

    … what links this issue with the next one – the GFC and the possible demise of the Western capitalism in the longer run (unless the current delusional economic paradigm is changed). The apparent success in bringing back the inflation under control came at the price of reducing the share of wages in GDP – by shifting down the wage setting curve I was talking about in the previous comment. These political settings allowed for the private debt to take off – to finance consumption so that the aggregate demand gap is closed (the governments refused to finance the gap because of the return to “sound money” doctrine). This in turn brought about the Great Recession.

    What is really interesting is that the income inequality in the US is now greater than it used to be in Roman Empire. (source: “persquaremile” blog, “Income inequality in the Roman Empire” by Tim De Chant). We all know how they ended up – if not, Michael Hudson wrote a lot about income inequalities and debt.

  11. @Adam (ak)

    You restate the expectations augmented Phillips curve then diss it out of hand without refuting it. Are you suggesting that wages don’t respond to inflation and unemployment, or that the long run Phillips curve isn’t vertical?

    As for ommited variables, Layard Jackman and Nichols book on unemployment suggests the following: long term unemployment, hiring and firing costs, and the level of unemployment benefits.

    A job creation programme that targeted the long term unemployed could reduce the natural rate of unemployment, since the long term unemployed are seperated from the labour market and do not exert much downward pressure on unemployment.

    As for unemployment in transitional economies:

    1) These started out with massive amounts of supressed inflation as a result of decades of forced saving to pay for investment and the military

    2) The new goverments did not have a great deal of credibility in fighting inflation, so inflationary expectations were not easy to budge, and

    3) Shifts in the structure of production from central planning to market forces generated a lot of structural unemployment.

    (1) and (2) meant that unemployment had to go way above the natural rate for some time. (3) meant that the natural rate itself was high.

  12. @Tim Peterson

    As I stated before the formal “refutation” or rather critique of the vertical “expectations augmented Philips curve” can be found in in papers written by Bill Mitchell and Joan Muysken. A more lightweight version can be found on Bill’s blog “The dreaded NAIRU is still about!”

    If you really want me to go after the model – I know where the flaw is. Yes I think that the line is not vertical. You won’t get hyperinflation just because of low unemployment. Other conditions must be met.

    To me the idea of forward looking rational expectations is just violating the laws of physics about the impossibility of travel in time. There’s nothing to prove I am afraid and the burden of proof is not on my side. All the expectations are adaptive. Then we may look into the way people can adapt to accelerating inflation and build a proper dynamic model.

    The effect of Job Guarantee is not in shifting the NAIRU. It is in overriding the system. Then one may re-evaluate that model and realise that something has changed.

    What really happened in Eastern Europe was much more fundamental. You essentially did not have the market mechanism in place at all because of the central setting of the product prices and rationing (as in Poland or USSR) and this resulted in forced savings. There is an old book “From Marx to the Market” which I recommend – Brus and Laski knew the system very well. NB Laski predicted the outcome of the shock reforms (almost 20% drop in GDP and 16% unemployment) – but there were other external political factors in 1989/1990 so back then, there possibly had been no other choice. Killing hyperinflation in 1990 was actually achieved by applying a very restrictive fiscal policy – there was an universally hated tax on wage increases above the inflation level (“Popiwek”). The credibility and expectations were much less important.

  13. @Tim Peterson

    You obviously have not graphed the ABS data.

    Without this, your statements are Ptolemaic.

    The notion that there was suppressed inflation is fanciful.

    Market forces did not create unemployment – the greed of oligarchs did. A perfectly free market guarantees a basic wage to everyone who wants to, and is able to, participate.

    There is no such thing as a natural rate of unemployment unless you are talking about capitalism.

  14. @Adam (ak)

    Rational expectations does not mean perfect knowledge of the future. It means that the future is forecast using all relevant available data and knowledge of the structure of the economy. The second condition seems too strong to me and many economists, but there is a large/growing literature on learning that copes with this problem.

    In practice one might use instrumental variables to proxy inflationary expectations (regress future inflation on all the lags in the model and use the fitted value) or one might use estimates of inflation derived from financial markets (the break even point between indexed and nominal bonds).

    I fail to see how the job guarantee scheme over rides the private sectors responses.

  15. @Chris Warren

    Graphed the ABS data on what?

    Repressed inflation in the USSR, Poland etc, not here! Can anyone deny that there were chronic shortages of consumer goods in these countries under Communism? These shortages were the symptom of repressed inflation.

  16. @Tim Peterson

    It helps if you pay attention. What was wrong with the data I cited earlier in this thread?

    Huh?

    Do you know how to set up a scatter plot?

    There are no shortages of goods under communism unless the economy is impacted by a World War, a Cold War, economic warfare and in the case of Poland and Yugoslavia, IMF conditional funding. Such politically induced shortages are not symptoms of anything but the hatred of capitalists.

  17. @Chris Warren

    The only ABS data you cite is on per-capita GDP; I don’t see what that has to do with the NAIRU/natural rate. Maybe one of your posts got lost?

    Communism didn’t just create shortages of consumer goods to fund cold war military expenditures, it also created them to fund high levels of investment. These resources were not very well utilized; by the early 80s GDP in the USSR was growing by 1%-2% p/a while the investment share was around 40% of GDP. And all, or more than all, of this growth was coming from the oil and gas sector of the economy. Economic efficiency was even worse when you consider the very high levels of investment in human capital (the USSR excelled in training mathematicians and physicists).

  18. @Adam (ak)

    I have read Mitchells paper “The dreaded NAIRU is still about!” and I am not very impressed. Your own point about omitted variables applies to his bivariate analysis; he should include the labour mkt variables mentioned above and other variables like oil prices that can drive a wedge between unit labour cost growth and inflation. The point about nonlinear employment dynamics is irrelevant. The point about the range of inflation outcomes consistent with the differing NAIRUs in TRYM is spectacularly irrelevant.

    He misdefines the natural rate of unemployment as being voluntary unemployment. Up till now, I have been using NAIRU and natural rate interchangably. This is not quite correct. The NAIRU stems from models with adaptive expectations, where the lags on inflation add up to unity, generating a vertical Phillips curve from the regression.

    The natural rate is a more general concept of a vertical long run phillips curve. If the real process driving wages responds to forecast inflation, and one estimates instead lags of inflation, then the coefficients on the lags depend on univariate proporties of the process for inflation. Only if inflation is a random walk will you get the coefficients on the lags adding up to unity. Otherwise, if inflation is a stationary AR(n) process the lags will add up to less than one. But that doesn’t mean that the LR Phillips curve is not vertical! Change the policy regime to target less than NAIRU unemployment and the ‘shallow’ parameters in the model will change driving the estimated Phillips curve to the right. This is known as the Lucas critique, and has fundamentally revolutionized macroeconomics since the 1970s.

    So what if the natural rate varies over time? It does not invalidate the concept of the natural rate. Its great if we can track these changes with explanatory variables, but if we can’t, then level shift dummy variables can be used.

    Most bivariate relationships in labour economics are unstable. Just look at the Okun curve that was supposed to relate changes in unemployment to economic growth. Also, especially in light of the unstable Okun curve, rember than unemployment is only one measure of economic slack. Particularly in non-unionized industries, under employment may have an effect on wage setting. As a result, many economists use the output gap instead of the unemployment gap in their Phillips curves, which results in more stable models and better out of sample performance.

  19. @Adam (ak)

    Also, I just had another look at Mitchell’s paper, and the table relating unemployment to change in lagged inflation is truly bizarre!

  20. So called Phillips Curve…
    June 1978 – SEp 2011

    Ue [ABS data A163165V] 3 month Averages
    CPI [ABS data A2325846C] Quarterly

    If you copy this data into your clipboard, you can simply click a single cell in Excel 2007 and paste – and it will automatically flow into two columns.

    It is a simple matter to then create a scatter graph, which is your Phillips curve.

    ============================================

    Ue CPI
    6.2 7.6
    6.1 8.0
    6.1 8.1
    6.9 8.8
    6.2 9.6
    6.0 9.9
    6.0 10.5
    6.6 11.2
    6.2 10.9
    5.9 10.4
    5.8 9.7
    6.3 9.7
    5.5 9.0
    5.6 8.8
    5.8 10.9
    6.8 10.0
    6.5 11.0
    6.9 12.9
    8.5 11.9
    10.4 12.2
    10.1 11.2
    10.0 8.7
    9.3 7.6
    10.1 4.9
    9.1 2.9
    8.5 2.8
    8.2 2.3
    9.2 4.3
    8.4 6.5
    7.9 7.5
    7.5 8.3
    8.6 9.2
    7.8 8.9
    8.0 9.1
    8.0 9.7
    9.0 9.4
    8.1 9.2
    7.8 8.5
    7.6 7.2
    8.2 7.5
    7.6 7.3
    6.8 8.3
    6.4 9.2
    7.1 6.9
    6.1 8.0
    5.9 8.1
    5.6 7.4
    6.7 9.1
    6.4 8.1
    7.0 5.9
    7.6 6.4
    9.4 4.8
    9.5 2.8
    9.7 2.8
    9.9 1.5
    11.2 1.2
    10.6 1.0
    10.6 0.8
    10.7 0.3
    11.8 1.1
    10.7 1.8
    10.6 1.7
    10.5 1.3
    11.2 0.8
    9.8 1.5
    9.2 2.1
    8.8 2.8
    9.5 4.2
    8.2 4.9
    8.1 5.7
    8.1 5.8
    9.1 4.7
    8.3 3.9
    8.4 2.5
    8.3 1.8
    9.4 1.3
    8.5 0.3
    8.2 -0.3
    7.7 -0.2
    8.5 0.1
    7.6 1.0
    7.5 1.8
    7.2 1.9
    7.8 1.6
    6.9 1.3
    6.7 1.8
    6.4 1.9
    7.0 2.6
    6.3 3.3
    5.8 6.0
    6.0 6.0
    7.0 6.5
    6.8 6.3
    6.6 2.9
    6.6 3.3
    7.2 2.9
    6.4 2.8
    6.0 3.1
    5.9 2.8
    6.6 3.0
    6.0 2.4
    5.7 2.0
    5.4 2.3
    6.0 2.0
    5.4 2.3
    5.3 2.7
    4.9 2.6
    5.6 2.2
    5.1 2.4
    4.8 2.9
    4.8 2.5
    5.5 2.7
    4.9 3.8
    4.5 3.7
    4.3 3.2
    5.0 2.2
    4.3 1.7
    4.1 1.3
    4.1 2.4
    4.5 3.9
    4.3 4.3
    4.0 4.9
    4.2 3.8
    5.8 2.4
    5.7 1.3
    5.5 1.3
    5.3 2.2
    5.8 3.0
    5.3 2.9
    5.0 2.6
    4.9 2.4
    5.4 3.2
    4.9 3.8
    5.0 3.7

  21. And if that fails, as the blog wipes-out tabs, you will have to save the data set to a temp word file, and convert to a table using either a space as a separator (previous post data) or vertical bar for the following dataset.

    Then a subsequent cut and paste to Excel will work.

    The Pearson coefficient is -0.06.

    Ue|CPI
    6.2|7.6
    6.1|8
    6.1|8.1
    6.9|8.8
    6.2|9.6
    6|9.9
    6|10.5
    6.6|11.2
    6.2|10.9
    5.9|10.4
    5.8|9.7
    6.3|9.7
    5.5|9
    5.6|8.8
    5.8|10.9
    6.8|10
    6.5|11
    6.9|12.9
    8.5|11.9
    10.4|12.2
    10.1|11.2
    10|8.7
    9.3|7.6
    10.1|4.9
    9.1|2.9
    8.5|2.8
    8.2|2.3
    9.2|4.3
    8.4|6.5
    7.9|7.5
    7.5|8.3
    8.6|9.2
    7.8|8.9
    8|9.1
    8|9.7
    9|9.4
    8.1|9.2
    7.8|8.5
    7.6|7.2
    8.2|7.5
    7.6|7.3
    6.8|8.3
    6.4|9.2
    7.1|6.9
    6.1|8
    5.9|8.1
    5.6|7.4
    6.7|9.1
    6.4|8.1
    7|5.9
    7.6|6.4
    9.4|4.8
    9.5|2.8
    9.7|2.8
    9.9|1.5
    11.2|1.2
    10.6|1
    10.6|0.8
    10.7|0.3
    11.8|1.1
    10.7|1.8
    10.6|1.7
    10.5|1.3
    11.2|0.8
    9.8|1.5
    9.2|2.1
    8.8|2.8
    9.5|4.2
    8.2|4.9
    8.1|5.7
    8.1|5.8
    9.1|4.7
    8.3|3.9
    8.4|2.5
    8.3|1.8
    9.4|1.3
    8.5|0.3
    8.2|-0.3
    7.7|-0.2
    8.5|0.1
    7.6|1
    7.5|1.8
    7.2|1.9
    7.8|1.6
    6.9|1.3
    6.7|1.8
    6.4|1.9
    7|2.6
    6.3|3.3
    5.8|6
    6|6
    7|6.5
    6.8|6.3
    6.6|2.9
    6.6|3.3
    7.2|2.9
    6.4|2.8
    6|3.1
    5.9|2.8
    6.6|3
    6|2.4
    5.7|2
    5.4|2.3
    6|2
    5.4|2.3
    5.3|2.7
    4.9|2.6
    5.6|2.2
    5.1|2.4
    4.8|2.9
    4.8|2.5
    5.5|2.7
    4.9|3.8
    4.5|3.7
    4.3|3.2
    5|2.2
    4.3|1.7
    4.1|1.3
    4.1|2.4
    4.5|3.9
    4.3|4.3
    4|4.9
    4.2|3.8
    5.8|2.4
    5.7|1.3
    5.5|1.3
    5.3|2.2
    5.8|3
    5.3|2.9
    5|2.6
    4.9|2.4
    5.4|3.2
    4.9|3.8
    5|3.7

  22. @Tim Peterson

    You haven’t responded to the majority of Bill Mitchell points or the majority of the points which I have raised. If any of them sticks then the hypothesis stemming from the NRU/NAIRU theory interpreted as the inability to bring down unemployment to the frictional level without experiencing accelerating inflation within the framework of a capitalist or mixed system does not hold water.

    Let me address the separate issues you’ve raised.
    “He [Bill] misdefines the natural rate of unemployment as being voluntary unemployment.”
    Where? I have read the Bill’s blog entry several times and I cannot find such an elementary error.

    “table relating unemployment to change in lagged inflation is truly bizarre”
    This is easy to explain. If there is a causal relationship between unemployment rate and inflation (not the other way around) then we would see such correlation on lagged data.

    “he should include the labour mkt variables mentioned above and other variables like oil prices that can drive a wedge between unit labour cost growth and inflation” – not if his goal was to debunk the NAIRU model where such variables were not included

    Now about maybe not the best worded point I was trying to make about the expectations.

    We can build a filter or learning system fed with past and current period data which is supposed to predict the future behaviour of the system. In that sense “expectations about the future” make more sense to me – but we are still dealing with an adaptive system. However this does not mean that we should automatically assume (as in the original NRU theory) that there is no systematic error in estimation. When I looked at the graphs from Australia and the UK I spotted that the so-called inflationary expectations were usually higher than realised inflation. This would affect the estimation of the parameters of the model specifically the “beta” parameter in the original model (see below) that is the weight with which expectations affect the realised rate. Can it be the case?

    When trawling the Internet for the arguments debunking your view that the long-run Philips curve has to be vertical and we must have involuntary unemployment in excess of the frictional level or face accelerating inflation I came across the following document:

    “Modelling Inflation in Australia David Norman and Anthony Richards RDP 2010-03 Reserve Bank of
    Australia”

    They have found that the long-run curve indeed is not vertical.

    They tried to fit the data to several models. The model which offered the best fit was the standard Philips curve model with expectations.
    pi_t = c + beta*E_t_1_pi + ksi*(1/ur_t_1) + WSOCOI + WSOCIIP
    where
    pi represents inflation
    ur unemployment rate
    WSOCOI weighted sum of changes in unemployment rate
    WSOCIIP weighted sum of changes in import prices
    E_t_1_pi expectations of inflation over the next s periods, formed in period t–1

    The coefficient beta related to the was found to be significantly below 1 (about 0.4). The coefficient beta in the NRU model I previously commented on is 1 (? t = ? te + …) . This means that in the RBA model the trade-off between inflation and unemployment does exist. (Please notice that I am not advocating exploiting it)

    From the article:
    “A Vertical Long-run Phillips Curve Restriction

    The concept of a vertical long-run Phillips curve – that is, that there is no trade-off in the long run between unemployment and inflation – is a cornerstone of most inflation models and has a long history in the literature (beginning with the rational expectations revolution in the 1970s). Indeed, this idea underlies the pursuit of price stability by central banks in most industrial countries. However, neither our standard Phillips curve, the mark-up model nor the OLS NKPC incorporate this standard feature. This reflects the fact that the sum of the coefficients on the right-hand-side nominal variables (inflation expectations, and growth in unit labour costs and import prices) in each of these models is significantly less than unity. This result survives when we simplify the models to represent inflation as being driven solely by inflation expectations and the output gap or unemployment rate; in that case, the coefficient on inflation expectations is significantly different from 1 for samples beginning in or after 1987. The rejection of this restriction is, however, not unique to our work; for example, both Anderson and Wascher (2000) and Williams (2006) have found such a result using US data. There are several possible explanations for this result. One relates to econometric issues, including from the relatively small sample used in estimation, or the
    possibility of a bias to our coefficients from measurement errors in our regressors. For example, our measure of expectations of inflation over the next 10 years is likely to be imprecisely measured (as are other measures of expectations) and only imperfectly correlated with the expectation that influences price-setting (which is likely to be for a one- or two-year horizon). However, it is hard to assess how important these effects are, and the long-run restriction is also obtained if we use a
    measure of shorter-term (one- or two-year) inflation expectations derived from the term structure of bond yields. Another possible explanation is suggested by Akerlof, Dickens and Perry (2000), namely that it may be ‘near rational’ for agents to devote limited attention to inflation when it is at low levels, resulting in the long-run Phillips curve appearing to be non-vertical within a range of low inflation outcomes.”

    I understand that RBA paper cannot just openly state that they have discovered that NRU/NAIRU hypothesis is either incorrect or a fraud.

    But as I stated before to me the whole empirical study does not prove anything in regards to Job Guarantee and using other non-standard means to reduce unemployment as it will not use the aggregate demand channel and will not significantly affect resource utilisation in the private sector.

    The study only shows that pumping up aggregate demand to reduce unemployment may result in either higher or even accelerating inflation. Fluctuations in aggregate demand did occur during the examined period of time and were driving the unemployment rate.

    Nobody has proven that we must persistently have a few % of the workforce unemployed to scare the rest of the workers to stop demanding higher wages what is the real implication of the vertical long-run Philips curve. Nobody has proven to me that government policies cannot be changed so that full employment is achieved while price stability (not necessarily defined as 2..3% CPI but certainly not 25%) is still maintained. If this means more Government intervention in the market processes – this is exactly what we have the Government for.

  23. @Tim Peterson Yes. The regression isn’t spurious but the statistics will be biased. Spurious regression typically refers to regression where the non-stationary variables are not co-integrated.

  24. @Adam (ak)

    The RBA article that you mention found that an old fashioned accelerationist Phillips curve (ie vertical long run Phillips curve) fitted the data better than the model mentioned above that has an apparent tradeoff with inflation. Why ignore this evidence?

    Modelling inflationary expectations using financial market variables (like the above study) is tricky because nominal bonds have a time varying inflation risk premia and indexed bonds are less liquid.

  25. @Tim Peterson

    Please do not misinterpret the results of the RBA study. It is not “ie vertical long run Philips curve”. It is the opposite.

    The authors of the study departed from the straight and narrow path of the neoclassical – Neo-Keynesian orthodoxy just a bit – and see what they found!

    The study was obviously limited – as an attempt to create an empirical model of the current system with its political constraints and within is unique historic context. The authors correctly stated that the model may not be valid if for example inflation jumps to 15% for any reason. I understand perfectly well why they had to pay lip service to NAIRU.

    What I want to say is that if the policy changed to accommodate full employment – the empirical model would change as well.

  26. I can’t seem to download the paper but the abstract says:

    “We find that traditional models, such as the expectations-augmented standard Phillips curve or mark-up models, outperform the more micro-founded New-Keynesian Phillips curve (NKPC) in explaining trimmed mean inflation, both in terms of in-sample fit and significance of coefficients.”

    Now the model you quote above is an NK Phillips curve. Expectations augmented Phillips curves are accelerationist.

    The time varying inflation risk premnia on nominal bonds, which is positively correlated with inflation (higher inflation is more variable) means there is measurement error in any crude attempts to measure inflationary expectations from financial markets by the difference between nominal and indexed bonds.

  27. It is available as a pdf file on rba publications.

    They also tried other sources of inflationary expectations as well, not only bonds.

  28. Tim Peterson :The reason that the natural rate of unemployment exceeds frictional unemployment is mostly mismatch between skills supplied and demanded. The minimum wage and wage rigidity in general is another reason. Show me a valid regression that shows that the NAIRYU is 2%!

    Actually most studies of the Australian NAIRU in the 1940s-early 1970s that I have seen seem to suggest that the NAIRU in that period was around 2%, which was also the government-mandated tolerated rate of unemployment; which suggests that the NAIRU is something of a post-hoc rationalisation for “whatever the unemployment rate was at the time”.
    On Phillips curves, Paul Ormerod showed there is a strong correlation between change in inflation and change in unemployment, that holds over the whole post wwii period; in other words, the underlying base level of inflation/unemployment are indeterminate with respect to each other, but change in one produces change in the other.

  29. PS the Lucas critique may have “fundamentally revolutionised Macroeconomics”, but that doesn’t say anything about its applicability to the real world. As JQ points out in his book, most concepts stemming from Lucas and his school such as Real Business Cycle theory, Ricardian equivalence, and a rigid relationship between the monetary base and inflation have all failed the test of reality dismally, and his microfoundations critique of macro should be tossed in the same “fail” basket.

  30. @Adam (ak)

    I have been reading up on the NK Phillips curve.

    Say you have (a) an inflation targetting central bank and (b) a Calvo process for price setting; a fixed proportion of firms vary their prices each period and then hold their prices constant for several periods. Now consider the NK Phillips curve:

    Pi(t)=alpha + beta E(t) Pi(t+1) + gamma (y(t)-y*(t))

    where Pi is inflation, E(t) Pi(t+1) is forecast inflation and y(t)-y*(t)) is the output gap

    because the firms that vary their prices this period are concerned with conditions this period as well as future periods, by* and alpha will increase, rendering the Phillips curve vertical.

    Yet another case of the Lucas critique.

  31. @James Haughton

    Yep in the good old days the natural rate of unemployment was 2%. But so what: it is higher now.

    An inverse relationship between inflation and unemployment is what one would expect with an inflation targetting central bank or treasury. Plot the 70s alone (when policy accomodated inflation) and you will get a positive relationship.

  32. @Adam (ak)

    I meant to say:

    because the firms that vary their prices this period are concerned with conditions this period as well as future periods, beta will be less that 1. But that doesn’t mean that the long run Phillip’s curve isn’t vertical, because the alpha parameter is ‘shallow’. If the central bank targets y-prime>y* then alpha will increase, generating a vertical long run Phillips curve.

  33. @James Haughton

    Why throw the baby (The Lucas critique) out with the bathwater (all that other stuff)? About 2 years ago in a discussion here, Prof Q agreed with me that there was no going back to pre Lucas critique Keynesianism.

  34. @Tim Peterson
    If the only explanation you can give for the NAIRU changing so drastically over time is “it is higher now”, a dormitive virtue argument if ever I heard one, you have no cause to argue that raising the minimum wage and hence demand would cause inflation because it would push unemployment below the NAIRU, because you can’t show that the NAIRU itself wouldn’t change – and if it does change all the time, then it becomes a vacuous concept of no use to policy.

  35. sorry, that wasn’t very well phrased: which aspects of the Lucas Critique do you think are left and still have validity when all the “other stuff” is thrown out? The days of “sniggering at the back of the class” when someone taught classical Keynesianism are gone.

  36. @James Haughton

    “Its higher now” is not the only explanation I can give. Slower productivity growth, higher energy prices, shifts in sectoral demand interacting with relative wage rigidity, frictions like hiring and firing costs. The list goes on.

    As far as the good bits of Lucas: emphasis on forward looking variables, and the variability of regression model parameters when confronted with regime change.

    Also, partial Ricardian equivalency; something like Mankiw’s spenders/savers model. This is one of the reason that multipliers are small.

  37. @Tim Peterson
    That list sounds like something of a grab-bag. Particularly since I would guess “relative wage rigidity” and “hiring and firing costs” and other such frictions were much higher in the heavily-unionised, full employment post war era and measures of productivity growth seem to vary from economist to economist. All the microeconomic deregulation Australia and the anglosphere has indulged in since the 1980s doesn’t seem to have done tuppence to affect our productivity growth, which IIRC still hasn’t equalled that of much of the full employment era; which to me suggests productivity is a function of unemployment rather than a cause of it.

    Both Lucas’ forward looking and any Ricardian equivalence seem to depend on the notion of a predictable future, or at least a consistent set of assumptions among actors about what the future will be like (Ricardian equivalence also depends on the loanable funds doctrine when it comes to government spending). I’d go with the Post-Keynesians in contrast in stressing the “Knightean” uncertainty of the economic path of the future.

  38. @James Haughton

    Note that I said sectoral demand shifts _interacting_ with relative wage rigidity. If the system started out with relative wages that were not out of whack due to the demand shifts, it would not start out with a high natural rate of unemployment. To sectoral demand shifts one could add shifts in demand for factors due to ‘biased’ growth in technology.

    Thatcherism caused an increase in UK productivity growth, particularly in manufacturing.

    Ricardian equivalence has nothing to do with loanable funds theory; an increase in G reduces C before there is any chance of crowing out I.

  39. @Tim Peterson
    “an increase in G reduces C”

    Has anyone ever seen it or is this another “shallow” variable?

    “O’Brien held up the fingers of his left hand, with the thumb concealed.
    ‘There are five fingers there. Do you see five fingers?'”

    (George Orwell 1984 Part 3, Chapter 2)

    I’m done.

  40. To quote “UNDERSTANDING THE EFFECTS OF GOVERNMENT SPENDING ON CONSUMPTION
    by Jordi Galí, J. David López-Salido and Javier Vallés at the European central bank – chosen solely because it’s the first paper that pops up when you google “effects of government spending on consumption”:

    “What does the existing empirical evidence say regarding the consumption effects
    of changes in government purchases? Like several other authors that preceded us, we
    find that a government spending leads to a significant increase in consumption, while
    investment either falls or does not respond significantly. Thus, our evidence seems to
    be consistent with the predictions of IS-LM type models, and hard to reconcile with
    those of the neoclassical paradigm.”

    On the face of it, the idea that consumers would cut back their spending NOW, in response to a government spend which may (or may not) have to be “paid for” at some point in the indefinitely far future, is ridiculous. It’s concepts like this that led to the call for a “post-autistic” economics.

  41. I didn’t say that Ricardian equivalence was completely true, just that there was some truth in it (some agents act like that, and this reduces the size of the multiplier). Jordi Galí, J. David López-Salido and Javier Vallés may find large multipliers but many researchers do not.

    My point was that, if Richardian equivalence was correct, it would not rely in crowding out.

    Note that with life-cycle consumers we can get a non-linear response to deficits. If debt ratios are small, repayment of the debt can be put off until the next generation, so government debt is viewed as net wealth. If debt ratios are large, big primarary surpluses loom in the near future so the debt is not viewed as net wealth.

    I suspect that the population is made up of a mixture of dynastic (‘infinitely lived’) life-cycle and rule of thumb consumers.

  42. @James Haughton

    This makes no sense.

    On Phillips curves, Paul Ormerod showed there is a strong correlation between change in inflation and change in unemployment, that holds over the whole post wwii period; in other words, the underlying base level of inflation/unemployment are indeterminate with respect to each other, but change in one produces change in the other.

    The fact is Ormerod says the exact opposite;

      The analysis shows that reliance on any kind of trade off between inflation and unemployment for policy purposes is entirely misplaced.
     

    See: http://www.paulormerod.com/pdf/curve.pdf

  43. @Tim Peterson

    Plot the 70s alone (when policy accomodated inflation) and you will get a positive relationship.

    But you would have to be a cherry-picking fanatic to do that, wouldn’t you?

    Oh well – this is the basis of philistine capo-theory.

  44. @Chris Warren

    My reference for Ormerod on the Phillips curve is his older book “the death of economics”. I don’t think the two are inconsistent; the Death of Economics paper is all bout the first derivatives of unemployment and inflation and acknowledges that the relationship between the base variables is indeterminate.

    @ Tim Peterson

    A few comments ago you were claiming that C(onsumption) goes down when Government spending goes up, or so it seemed. Or were you saying that that’s what Ricardian equivalence claims, but you don’t subscribe to it personally?

Leave a Reply

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

WordPress.com Logo

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

Google photo

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

Twitter picture

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

Facebook photo

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

Connecting to %s