Home > Uncategorized > Putting the baby in the tub: unemployment in a neoclassical (?) macro model

Putting the baby in the tub: unemployment in a neoclassical (?) macro model

Is it possible to model unemployment in neoclassical ‘DSGE’ macro-economic models ? I’m occupied with a project which compares neoclassical macro concepts with statistical macro concepts. One of the glaring differences between the statistics and the models: we measure unemployment as a matter of routine but DSGE-models do not conceptualize or define, let alone operationalize it. When you model our society as a one person ‘Robinson Crusoë’ ‘society’ you will have somebody who works a little less or more but who will never be entirely unemployed. Models with heterogeneous agents have problems with this, too. On twitter, Lukas Freund however pointed me however to the 2018 article ‘Unemployment (fears) and deflationary spiral’ by Wouter J. den Haan, Pontus Rendahl and Markus Riegler, which does model unemployment.

Does this article require me to rewrite my stuff on ‘labor’? Nope. To be able to model unemployment they do not (yet) put a silver bullet through the heart of the models. But they do cut out its liver.

What do I mean? Central to DSGE models is the ‘Euler equation’ which relates present consumption to consumption yet to come. The basic intuition behind this formula (not the formula itself) is:

U = C-N

Or: Social Utility is Utility of total consumption minus Disutility of total labor. Consumption is gain, labor is pain. The idea is rather vaguely defined. Social Utility is not well-defined (if at all), consumption as a rule excludes government products and services like education and law and order and bridges and often, hours and persons are not well distinguished or a vague , sub-scientific description like ‘labor services’ is used for labor. Also, consumption and labor are recalculated into ‘utility’ without clear-cut deep, sound, fundamental parameters like Avogadro’s number. Never mind. Even in this fuzzy environment it is difficult to model unemployment, as increasing unemployment is supposed to diminish pain. While it, as we all know, it increases pain. There is a momentous literature about this, even when people like Thomas Sargent and Robert Lucas and Edward Prescott and Christopher Pissarides, al winners of the Sveriges Riksbank Prize in Economic Sciences in Honor of Alfred Nobel, all use the word ‘leisure’ when they mean ‘unemployment’.

So, what do Den Haan, Rendahl and Riegler do? Brilliant. They cut out N. The basic formula at the heart of all DSGE models is fundamentally altered. This is a paradigm shift. As they state it: ‘there is neither home production not any disutility from working’. In a sense they are following Keynes with this and unlike labor markets in ‘normal’ DSGE models their labor market does not fit Keynes’ description of the theory  of ‘classical’ labor market models in chapter 2 of the General Theory (it’s not a new paradigm shift) and is much closer to a Keynesian labor market as well as unemployment as we measure it. In the project I state that ‘labor’ and to be precise wage labor is the core of the ‘macro battles’. This corroborates this idea. Thanks!

How about the rest of the article? Some remarks:

The model is not about a modern economy. Part of the trick used to model unemployment is assuming: ‘there is one worker attached to each active firm’. As, every period, some firms go bust and as it takes time to establish a new one, there will, hence, also be unemployment. And even cyclical swings in unemployment when for whatever reasons in one period more firms go bust than usual and you put in a positive feedback loop between the amount of firms going bust and whatever parameter you have which influences the amount of firms which going bust. Unemployment will go down as new firms will be established. To estimate the number of new firms created they use ‘search theory’: the less firms there are, the easier it gets to establish one and the more unemployed there are, the faster this will happen, even when the model has the useful addition that creating a firm requires capital, too. This, however, is not a description of our present day wage labor economy but a nostalgic but modernist  description of the handicraft economy of the pre-industrial village or, less benign and closer to the model (which assumes that there is only one product) the mass production of textiles or other consumer goods in poor and often destitute households in the countryside in eighteenth and nineteenth century Europe. From Engels to Marx to Veblen to Means to Keynes to Coase to Robinson to Galbraith to Lee to Keen economists have been shouting at the top of their lungs to neoclassicals: ‘Hey guys, even when own account workers are still very important we’re basically having a wage labor economy now with monopsonic labor markets and large companies, centralized control and with administered prices defined in long-term value chain contracts as well as positive economies of scale… Live with it’. Assuming everybody is self-employed, as the article does, is not an apt description of our international value chains Walmart administered prices economy.

Graph 1. Number of workers world wide, divided into three classes. Source: ILO.

Polanyi graph2

Aside: two very good studies about the establishment of seventeenth, eighteenth and nineteenth cottage industries are Ogilvie,  Küpker, and Maegraith (2012), “Household Debt in Early Modern Germany: Evidence from Personal Inventories”,  and Pfister (2007), “Rural land and credit markets. The permanent income hypothesis and proto-industry: evidence from early modern Zurich” and for the Netherlands: Zuijderduijn, Moor and Van Zanden,  “Small is beautiful. On the efficiency of capital markets in late medieval Holland”. All these studies show that institutions evolved which guarantee that households could lend to set up shop or save/invest for old age. Fun fact: the Ogilvie e.a. households with a milk cow had hoarded less cash than households without one, which gives the phrase ‘cash cow’ its original meaning back.The model suffers from a common failures of DSGE models: ‘there is no financial intermediation’ i.e. no banking sector. This, too, is representative of eighteenth and nineteenth country villages. Also: the monetary asset ‘serves as the unit of account’. No. A unit of account is attached to a monetary asset but you can’t equate it with it. In a sense, as there is a central bank the sector banks is included in the sector households.

There are too many neoclassical platitudes like: “Karsrud e.a. report that the amount of consumption that is financed out of an increase in debt actually decreases following a displacement [getting unemployed, M.K.]”. Yeah. When somebody cuts of my leg I can’t walk as fast anymore… In the neoclassical world, this seems to be a surprise as people are infinitely lived and will, eventually, be able to pay off the debt. ‘Where I find myself pointing out to economists that people do not live forever’.

The definition of unemployment used is interesting. It is not equal to the definition as used by statistical institutes *all over the world* which is remarkable as they do use this official rate for comparison. The official definition is based on individual persons, not on households. The definition of the model is however based upon households (highly interesting!). Which means that they have to exclude quite a chunk from the economy. A large EU household survey is used (good!) but they exclude quite some households:

Excluded Households. To ensure that the data are used resemble the households of our model, we exclude types on agents that are not present in our model. In particular, we apply the following filters:

  1. Always exclude the entire household if any household member is (according to PE0100a)
    • #2 (sick/maternity leave),
    • #5 retiree or early retiree,
    • #6 permanently disabled,
    • #7 compulsary military service, and
    • #9 other not working for pay.
  2. Always exclude the entire household if nobody in the household is
    • either #1 (employed) or #3 (unemployed).
  3. Exclude the entire household if there are more than two people either employed or unemployed.
  4. Exclude the entire household if any in the household is aged 65 years or higher.
  5. Exclude the entire household if nobody is aged 20 years or older.
  6. Exclude individuals without reported labor force status.

I would have loved to see a more thorough comparison of these data with the official statistics.

They add the unemployment rates of France, the UK, Germany and Spain together, even when labor markets in these countries behaved differently which might lead to fallacies of composition while they also do not pay attention to differences between changes in hours per job and changes in unemployed persons; after 2008 average hours per person working declined quite a bit in Germany, which must have mitigated the increase in unemployment with somewhere between 2,5 and 3.5percentage points, but not in Spain. Using search theory requires one to look at labor market flows which forces one to include people outside the labor market as net new entrants are quite important when it comes to changes in unemployment (even when changes in the total number of jobs are hardly affected). This is not part of the model. For the elephant in the room: see the graph. How smart is it to take an average of these countries?

Graph 2. Unemployment in Spain and Germany. Source: Eurostat.

employment rate

A general equilibrium approach is used which means that, in the end, unemployment will be zero (aside from friction unemployment) even when what Keynes called ‘involuntary unemployment’ seems to exist (basically, unemployment which is not solved by lower real wages). Important is however that they realize that increases in employment during expansions are slower than reductions during crises (graph). In reality, however, job finding rates of the unemployed do not always go down during crises even when job finding rates of new entrants and, later, long-term unemployed do. Meaning that during a crisis, unemployment and inactivity themselves become a friction, hampering the decline of unemployment. Even then, job finding rates do not explain unemployment, changes in the amount of jobs relative to the total labor force do. I’m also not too happy with their idea that “a reduction in the price level that actually goes together with an increase in expected inflation and substantial decrease in the [expected?, M.K.] real interest rate”. In a general equilibrium world, a temporary decrease of the price level will be counteracted by higher inflation tomorrow. This is not what happened in Spain (or Greece). Price levels decreased, but inflation is still lower than it used to be even when the price level is still way below the German level. By the way: Greece and Spain are examples of involuntary unemployment in the sense of Keynes in chapter 2 of the General Theory.

Graph 3. HICP inflation and constant tax rates. Source: Eurostat.


Some other quibbles:

The authors use the GDP deflator. As this is influenced by the terms of trade (as I show here) they should have used the domestic demand plus export deflator (graph). 2013-2016 clearly was a low inflation period, unlike 2017-2018, total difference over this period is about 3 to 4% which is – a lot (as can be seen, inflation has rebounded after the 2012-2016 ‘double dip’, something not entirely unimportant but not visible in the GDP deflator data). I leave it to the authors to do the arithmetic, but as domestic demand plus export prices are more volatile than GDP prices, they might enable better testing of the model.

Graph 4. Different macro-deflators. Source: Eurostat.

Domestic demand deflator

The authors use RULC: real unit labor costs or nominal labor costs per real unit of output. On the macro level, this variable is prone to a whole array of fallacies of composition, as I show here. Basically: you can’t assume that GDP, a historical aggregate monetary variable, is one unchanging physical product.

The authors state that there are only relatively short (OECD) series for employment, starting in 1980. Finding and comparing sources is part of craft of being a macro-economists. Longer series exist, for all the countries mentioned, here only one (graph) (Wow: never realized that it was only after 1985 that the UK really surpassed WWII employment… Another wow: the female 60-64 employment rate in France is higher than the male employment rate). 

Graph 5. Long term employment, UK. Source: Bank of England.


Graph 6. Employment rates in France. source: INSEE.


Long term data for Germany here. Data for Spain on ‘trading economics’ already show longer series than back to 1980. This is the direction macro is going, join the pack.

A HP filter is used to calculate trends. This method has been discredited.

As stated: cutting out N is wonderful – the baby can finally be put in the tub.  But there still is some bathwater which has to be thrown out. Introducing wage labor and replacing social utility with transactions based variables might be the next steps. A famous economist, what’s his name again, once did this already…



  1. F
    April 23, 2019 at 3:08 pm

    The answer to your initial question, “Is it possible to model unemployment in neoclassical ‘DSGE’ macro-economic models?”, is simple. You could follow some algorithmic process through to its conclusion, but this would be meaningless. The neoclassical postulates are not those of the real world (Zambelli 2018) so the conclusions will not be those of the real world. Therefore it would be an exercise in futility as will any manipulations around the neocalssical assumptions.


    Stefano Zambelli, The aggregate production function is NOT neoclassical,
    Cambridge Journal of Economics 2018, 42, 383–426. doi:10.1093/cje/bex011

    • Frank Salter
      April 23, 2019 at 3:09 pm

      The name should be attached as this

  2. April 24, 2019 at 12:32 am

    Really, why do you spend time on models that assume equilibrium? Real modern economies are far from equilibrium, all the time. Instabilities everywhere.

  3. Helen Sakho
    April 24, 2019 at 1:37 am

    I just had a look at the website above, and I look forward to reading you book Geoff.

    • April 26, 2019 at 1:21 am

      Good Helen. You can get some of the gist from the Little Green Economics Book, available on the site.

      Now if I could just get a response from Editor when my book might become available … :-)

  4. Mike Ryan
    April 24, 2019 at 5:03 am

    neoclassical Econ is a false science meant to deceive the working class. Stop even talking about this stuff. For $1.00 you can learn the false mathematics and the psychological tricks used to fool the public,

    Not so sure?? The University of Chicago was founded by the Rockefeller family in the 1890’s. The Rockefeller family rescued the school in the early 30’s from the great depression. They hated unions – recall the Ludlow massacre where 21 people were killed.

    They wanted capitalism painted in a fair light and they wanted to teach this to every high school kid in America before they were smart enough to know better. Along comes Paul Samuelson – he heard their request and delivered the most popular High School textbook of all times.

    Why is it the University of Chicago has so many Economics Nobel Prize recipients. (Remember, this prize is given out by bankers – not scientists) Bankers love theories that glamorize unrestrained capitalism.

    That is how the rich expand their influence,,,, they get people to work for them and write papers that paint capitalism as the best thing since sliced bread. ala think tanks. Think again…

  5. Frank Salter
    April 24, 2019 at 7:54 am

    To Geoff, Helen, Mike at all: the only theoretically justifiable quantitative analysis in time is my paper, Transient Development, RWER-81, 2017. The analysis passes every test of the quality calculus and describes the nature of manufacturing projects and industries. It is not falsified by the empirical evidence, Popper, and it a Lakatosian progressive research programme.

  6. Ken Zimmerman
    April 29, 2019 at 11:08 am

    In his book, ‘23 Things They Don’t Tell You About Capitalism’ Ha-Joon Chang of Cambridge University serves up devastating assessments of both economics and capitalism. And these are from an economist who favors capitalism. Economics is failing and its final failure is its fiasco to even anticipate the 2008 crisis or to offer any worthwhile recommendations to fix it. In light of such multiple and colossal failures why would anyone of good sense listen to economists about anything, particularly about economics? Ha-Joon Chang writes: “Economists are not some innocent technicians who did a decent job within the narrow confines of their expertise until they were collectively wrong-footed by a once-in-a-century disaster that no one could have predicted.” Economics is anything but an inward-looking, hermetic discipline. Quite the contrary, economics has been a hugely powerful – and profitable – enterprise, shaping the policies of governments and companies throughout much of the world. The results have been little short of disastrous. As Chang puts it: “Economics, as it has been practised in the last three decades, has been positively harmful for most people.” With wit and appropriate subversion, and ordinary commonsense, Chang unwraps economics and with it capitalism. For example, we learn that free market policies rarely make poor countries richer; global companies without national roots belong in the realm of myth; the US does not have the highest living standards in the world; the washing machine changed the world more than the internet; more education does not of itself make countries richer; financial markets need to become less, not more efficient; and – perhaps most shocking to Chang’s colleagues – good economic policy does not require good, or any other kind of economists.

    If capitalism is to survive and if humans are to survive capitalism, capitalism must be reformed at its core. Making markets more transparent is not enough. “If we are really serious about preventing another crisis like the 2008 meltdown, we should simply ban complex financial instruments, unless they can be unambiguously shown to benefit society in the long run.” Chang argues this is not really a radical step. It is no different from those that have been enforced on other dangerous products. “This is what we do all the time with other products – drugs, cars, electrical products and many others.” This is an excellent recommendation but in making it we need to accept that powerful (very rich) groups oppose it, particularly in the USA. Then there’s the issue of how we make such changes on a world-wide basis, and the related issue of the changing make up of capitalism as it grows in India, China, Europe, etc. Chang’s reforms may sound impractical, even impossible. But his account of where we find ourselves today is strikingly accurate. For anyone who wants to understand capitalism not as economists or politicians tell us, their marks, it is, but as it operates with peoples and nations, this book is invaluable.

    Listing the 23 things they don’t tell you about capitalism rounds out Chang’s message. Some of these have been discussed on this blog. What’s not discussed enough is ridding the world of economists and establishing new principles to guide resource creation, acquisition, and distribution. Chang gives us 23 principles to consider.
    1. There is no such thing as a free market
    2. Companies should not be run in the interest of their owners
    3. Most people in rich countries are paid more than they should be
    4. The washing machine has changed the world more than the internet has
    5. Assume the worst about people and you get the worst
    6. Greater macroeconomic stability has not made the world economy more stable
    7. Free-market policies rarely make poor countries rich
    8. Capital has a nationality
    9. We do not live in a post-industrial age
    10. The US does not have the highest living standard in the world
    11. Africa is not destined for underdevelopment
    12. Governments can pick winners
    13. Making rich people richer doesn’t make the rest of us richer
    14. US managers are over-priced
    15. People in poor countries are more entrepreneurial than people in rich countries
    16. We are not smart enough to leave things to the market
    17. More education in itself is not going to make a country richer
    18. What is good for General Motors is not necessarily good for the United States
    19. Despite the fall of communism, we are still living in planned economies
    20. Equality of opportunity may not be fair
    21. Big government makes people more open to change
    22. Financial markets need to become less, not more, efficient
    23. Good economic policy does not require good economists

  7. April 30, 2019 at 2:25 am

    Or you can try my Sack the Economists (and disband their departments).

    • Rob
      April 30, 2019 at 5:11 am

      Sounds like a plan!

    • Ken Zimmerman
      April 30, 2019 at 10:59 am

      Geoff, your book is great. I recommend it often. But Chang’s a flashier writer and besides he’s an economist at Cambridge. The perfect critic for both capitalism and economics. Plus, his books tend toward shrewdness rather than sarcasm.

    • Craig
      April 30, 2019 at 6:42 pm

      Our we could base economic systems firmly on an integration of profit and the unimpeachable concept of grace as in Monetary Gifting with the following policies:

      1) A 50% discount to every consumer product at the point of retail sale that is rebated back to the enterprise gifting it to consumers by a monetary authority mandated to do so.

      2) A $1000/mo. universal dividend distributed to everyone 18 and older for life.

      3) End the idiotic charter of private money creation which violates Occam’s Razor and Lord Acton’s dictum that absolute power corrupts, and replace it with a non-profit publicly administered national banking and monetary system based on the same integratively wise, benevolent and yet sovereign concept as the economy.

      4) As a non-profit national banking and financial system does not need to make a profit the point of note signing could become the new terminal ending point of the economic process and an additional 50% discount to the consumer could be implemented at that point making a $300k house reduced to $150k at retail sale become a $75k note at 0% interest, a $50k lithium battery auto $12.5k

      5) Incentivize saving which in a true money economy is no longer a macro-economic vice like it is in a Debt Only based money system by making any savings used for ecologically sane consumption additionally multiplied by 2. Thus a totally off the grid solar etc. powered $300k home would be further reduced by $50k if $25k of savings were used at note signing making the final note $25k. This policy could be used to incentivize any consumer purchase, capital outlay or mandated anti pollution/green house gas reduction regulation.

  1. No trackbacks yet.

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

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