Work blogging

I thought I’d try this and see how it goes. When starting a semester there’s always some easy overview stuff that should not cause people outside the field any problems and I thought I’d start with that. The current post will be based on the paper Flexicurity – labour market performance in Denmark by Andersen and Svarer. Monday I printed approximately 40 papers (un)like this to be read during the semester, and I’m not sure I’m going to be blogging all of them but we’ll see how it goes.

The article is as mentioned just an overview article that can be read by pretty much anyone who understands English. It’s not hard, it’s just stuff I need to know. I filled one A4 paper with notes related to the paper and my blogpost will be based on those notes, rather than the text; I assume this approach will be useful in terms of preparing for the exam because at that point I will not have time to reread the paper. Some of my remarks may not be from the paper but instead related to stuff covered during the first lecture.

First off they talk a bit about the (‘Danish’) flexicurity model, which is based on a combination of a relatively flexible labour market and relatively high social transfers providing social insurance. This model has often been argued to be a major factor behind the Danish (and in other contexts, Scandinavian) economic performance. In the paper they argue that the flexicurity model has been ‘around’ to a significant extent since the 70’es and given the economic performance of Denmark in the 70’es and 80’es the flexicurity model is probably ‘not the whole story’. They argue in the paper that a third factor, active labour market policies, has been crucial for the relative success of the model.

They mention and talk a bit about – and I believe they also misspell – the Ghent system (Gent in the text), which relates to how the Danish UI (unemployment insurance) benefit payments scheme works. Other noteworthy features (in this context) of the Danish labour market: Many small firms, and people who are temporarily laid off constitute a substantial number of the unemployed at any given point in time.

They talk a bit about EPL [employment protection legislation] and argue that there’s a distinction to be made between ‘job security’ (strict EPL) and ’employment security’ (lax EPL). Denmark has relatively lax EPL.

Denmark has a high replacement rate (UI benefits are relatively high compared to wages of people in employment) especially for low-wage workers. So low wage-workers generally confront the highest marginal tax rates related to the state transition from unemployment to employment.

Reforms in the 90’es had three main effects: i) Shorter duration of benefits, ii) changed rules regarding eligibility – getting a job basically became a requirement for ‘resetting the clock’ regarding benefits; it was/is no longer enough to participate in a job training program, ii) workfare. Youth unemployment was dealt with by lowering benefits (to the level of study grants) for young people and by implementing stricter activation requirements for this population segment. Wage formation has become less centralized over time.

Activation measures generally last about 6 months. Workfare affects both employed and unemployed people. Unemployed people in the active labour market programmes are subject to a lock-in effect which means that the activation requirement may crowd out job search. They are also subject to a positive effect, the post-programme effect, which deals with the fact that an activation programme may increase human capital. (though it’s worth noting here that even if human capital goes up, job search efforts may still be impacted negatively by the programme, e.g. by more narrow job-search post-activation). Unemployed people who are not in an activation programme may increase search efforts prior to being faced with activation measures, as activation measures are generally unenjoyable (in the literature they are often modelled as a tax on leisure). This threat/motivation effect has been shown in a Danish context to be both real and significant. People who are employed are also impacted by the workfare requirements of people who are unemployed, because they make the outside option (other jobs) less attractive, which means that wage demands of people employed will be impacted by the policies. This is because from the point of view of a person who’s already employed workfare can be considered a tax on job searching and/or an increase in search costs. They argue in the paper that active labour market policies have impacted wage formation in Denmark during the 90’es. The wage effect is an indirect effect which is hard to observe and it illustrates how a general equilibrium framework is necessary to evaluate costs and benefits of labour market policies.

The time profile of the UI scheme has changed since the reforms were first implemented, as compensation is now falling with the duration of unemployment (in the 80’es it basically wasn’t). In a long time this fall was caused by both the jump from UI-benefits to kontanthjælp after the UI-benefits had been exhausted and by the implicit tax on leisure which hit unemployed people who had received UI benefits for some time and thus became subject to workfare requirements. Today unemployed people face workfare requirements from day one, but as the UI benefits duration has shortened even further (to 2 years in 2010, not in text) the time profile aspects of the system are still very important.

Noteworthy is the fact that workfare requirements introduce a screening element to the benefits system, as benefits are arguably better targeted to people who ‘really need them’ (and thus are willing to be subject to the workfare requirements). Also noteworthy is that how one perceives workfare requirements can impact the effects they can be expected to have; for example, one might choose to perceive of workfare requirements as ‘an option to prolong the benefits period’ rather than a ‘condition to get benefits’, and a recipient of UI-benefits might start to ‘think of workfare as a job option’ – such perceptions would be expected to cause workfare to crowd out job search.

Workfare seems to be popular politically, compared with lowering benefits. Most voters care more about income distributions which can be measured than utility functions which can’t.

Empirically, the lock-in effect is more significant  in the short run than the post-programme effect, and (as already mentioned?) the threat effect is real and significant. The wage effect is hard to measure but given current estimates it’s probably quite significant. In the long run the post-programme effect is likely to be larger than in the short run; this again relates to the extent of hysteresis/state dependence. When evaluating the costs and benefits of workfare, it’s important to deal with this aspect. In general, the reforms of the 90’es have improved cost effectiveness, but this is still an issue. Denmark is in the absolute top of most measures of spending on active labour market policies.

Sanctions, which are imposed on people who are subject to workfare requirements but do not meet the requirements, have increased over time. Arguably males are more responsive to sanctions than females. Workfare may be improved through better targeting of programmes; for example supplementary education, the most common activation measure, is more likely to be cost-effective for people with low education than for people with a high education. In the public sector the use of matching groups have been implemented to improve the efficiency of the programmes.

Any kind of feedback is most welcome.


September 5, 2012 - Posted by | Economics, Papers


  1. I read through, and found it very informative – thanks. My takeaway is that Denmark managed to transform a system that was beginning to fail into a system that works by decreasing not the amount of (monthly?) unemployment benefits, but their average duration by building incentives into the eligibility requirements for continued participation. This is a very legitimate public choice, as long as there is public support for it – good job, guys. I think the insight here is not really from the POV of economics (the total benefits are still decreased, and the outcome is what an economist expect), but of political science. Structure matters, aggregate data can lead to bad policy – and these guys dis-aggregate the data. Good stuff.

    My question is: is the new system going to handle the challenges ahead? Now, understandably, yet regrettably, my specific knowledge of Danish demographics is lacking, so I hope you (US) can shine some light for us foreign readers. My impression is that every single country in the EU is facing sub-replacement fertility – there are less that 2.1 children per woman, and hence population, net of immigration, is aging and declining. Can the Danish system handle the resulting decreasing labor force participation? Can it handle the outflow of young educated Danes who take full advantage of the excellent (and costly!) education, and then leave for greener (less taxed) pastures? If Denmark is (as most of Europe) importing refugees from dramatically different cultures (Muslim, Eastern European), will the social support for relatively generous redistribution last when the pressures to redistribute even more are ratcheted up?

    I do not mean to sound as if I do not appreciate the benefits of immigration – I am a libertarian. However, open borders and generous redistribution do not seem to mix well, at least at the intellectual level. On the other hand, without open borders, would Denmark be choosing the Japanese model – stagnation, due to demographic pressure, and thus slow, but virtually irreversible decline?

    Finally, a disclaimer. I find it hard to comment on macroeconomic topics because I have steep reservations about the value of macroeconomic analysis. Its “hit rate” has a hard time matching that of a coin flip. When it’s right, it has beautiful stories to explain why; when it misses, there are always convenient variables that were not considered. When you have so few data points (even on a quarterly basis), and so many variables… how do you say “overfitting” in Danish? 🙂 I also find in macroeconomics, at lease as practiced currently, an uncanny manifestation of Hempel’s paradox – the same observations are interpreted as confirmations of mutually exclusive points of view. It is glorified tea-leaf reading, with advanced math and statistics thrown in for good measure to deter common sense critics who (damn ignoramuses!) claim that you should take your foot off the aggregation pedal, and put your thinking cap on. Rant over.

    Thanks again for the insight. I, personally, and I know I am in a minority, but what the hell, would love to see more of your work in micro, econometrics (the statistical aspect of it), markets and market infrastructure, etc.

    Comment by Plamus | September 8, 2012 | Reply

  2. It’s very nice to know that at least someone will probably be reading along – I hadn’t forgotten that a few of my readers have a background in economics, but of course I was hoping that even people who don’t might benefit from the posts. That’s probably aiming too high, but…

    The course I’m covering has a reasonably micro-oriented focus which is a big part of why I’m taking it. So you can expect more micro stuff to come. A prerequisite for the course is a stats course called microeconometrics, which itself has econometrics as a requisite (and to have had econometrics you need regression analysis, linear algebra, etc.) – so it’s not unlikely that there’ll also be some econometrics stuff to cover later on.

    I could have said the same things you did about macro (I think I may have done just that in the past) – I’m generally not a big fan. The good thing about quite a few of the macro-related courses that are on offer here though is that the treatment of the subjects is relatively formalized and model-based, which makes it much ‘easier to think about matters’ for me; to be able to think in terms of models will be a much more useful ability to possess when entering the job market than will knowing the main results of the Ramsey-Cass-Koopman model. The math applied in the models is not always used solely to obfuscate matters.

    The questions you ask are not, I believe, questions which will be addressed to any significant extent in this course. I have had a course in the past which dealt with these issues though; the course (‘The Economics of the Welfare State’) was incidentally taught by Torben Andersen, one of the authors of the two first papers. If asked in a conversation AFK I could probably answer the questions without getting a lot of things wrong, but I’ll go have a look at my notes from the course before I answer this anyway to get the details right.

    Comment by US | September 8, 2012 | Reply

  3. Thanks again, and please use your valuable time wisely – if my request is much effort, ignore it or put it on hold, I’ll surely understand.

    As you can probably see, I have a tendency to look at macro with a very long-term perspective – basically, let the noise cancel out, and the supposed benefits of aggregation kick in. This, of course runs the risk of looking like Bill Gates with his infamous “640K ought to be enough for anybody” statement – the rate of technological change for everything has increased dramatically in the last 2 decades or so, and, contra Tyler Cowen, I do not think it’ll be slowing down, unless of course governments are dumb enough to kill it. Still, for me, trying to glean insight about the future of the cell phone market from watching the original “Wall Street” is more useful that trying to shave a tenth of a percent off the margin of error for your estimate of economic growth for Q4 of 2014.

    Comment by Plamus | September 8, 2012 | Reply

    • I generally don’t spend my time wisely and so this comment is highly unlikely to make much of a difference; if I was actually spending my time wisely I’d be either working or looking for a girlfriend right now, instead of writing a random and forgettable online blog comment. Anyway I’d much rather be doing this right now than to actually be doing ‘something useful’.

      Back to your questions: I’ve actually written quite a bit about Danish immigration stuff and the economic dynamics of the model currently in place before, but you may have missed those posts. There’s a lot of stuff and a lot of data at the links, including stuff on fertility rates of both Danes and immigrants. The Danish TFR is ~1,9 and so it is below replacement level.

      Anyway, some general remarks:
      i) “The demographic shifts in the Scandinavian countries are roughly of the same order of magnitude. The demographic old age dependency ratio will thus increase from about 0.25 to 0.4 within the next 3-4 decades.” (a quote from Torben’s lecture notes.) The Ministry of Finance made a projection of the public finances for Denmark last year where they presented a structural budget balance calculated for unchanged welfare arrangements, but taking into account demographic shifts. The structural budget balance will reach -4% of GDP in 2050 given the current political arrangements. “The order of magnitude of the financial problems is substantial, and deteriorating public finances will lead to systematic deficits.” (another quote).

      ii) The ‘Danish model’ is very sensitive to changes in labour force participation rates. As also mentioned briefly in the paper, the model basically is only sustainable as long as labour force participation can remain high. This is because the Danish government finances are hit much harder when a person leaves the labour force or goes from employment to unemployment than is the case for most other countries, due to the combination of the high tax rates of people in employment and the high replacement rate.

      iii) Private sector productivity gains can not be expected to contribute much to solving the financial problems caused by changing demographics and increasing health care expenditures over time. This is because the growth rate of the Danish transfer payments are more closely linked to the private sector wage growth than is the case for most other countries (they are not perfectly linked, but it’s close).

      iv) Most economists who’ve looked closely at the numbers will (…I hope?) tell you that changes in immigration policy is unlikely to be ‘the answer’ to the problems the model is facing; for some details you can see part 3 and 4 of my immigration posts above. For any politically realistic immigration model in the future, total immigration will remain a cost driver. Emigrants are on average net contributors, immigrants are on average net beneficiaries and this adverse selection aspect of Danish migration patterns basically hasn’t changed in 30 years. If you look at the numbers the origin countries with the worst performing immigrants are almost exclusively Muslim. Note that changing the composition of immigrants by increasing migration from more developed countries on its own will not make much of a difference when it comes to fiscal sustainability; an estimate Torben Andersen presented in the course I mentioned gave a net effect on the sustainability indicator from a net increase of 5000 immigrants from more developed countries of -0,03 – i.e. basically no effect either way. “Better (labour market) integration” should in my view be considered a cost-containment measure, not a way to solve the long run fiscal sustainability problem. Do note here that the Danish employment gap (the gap between the employment rate of Danes and the employment rate of immigrants) is one of the highest, if not the highest, in the EU. Eurostats ‘Employment gap of immigrants’ statistic is ‘under preparation’, but data from that source presented last year by Torben Andersen gave numbers in the neighbourhood of ~16% for males (highest of 15 countries included) and ~26% for females (second highest – Netherlands had a gap of ~32%). The high employment gaps of Scandinavian countries (the numbers for Sweden and Norway are relatively high as well) are partly explained by the relatively high labour force participation rates of the native population, but these numbers underline how difficult it will be to use immigration to solve problems with the fiscal balance given the model’s sensitivity to the labour force participation rate.

      v) I don’t know if you’re familiar with it or not, but Baumol’s law […related to yet another course I’ve taken but haven’t blogged] states that there’s no long run steady state in a two-sector economy with different productivity growth rates. So if you’re thinking about the very long run, it’s clear that the Danish model is not sustainable.

      Comment by US | September 8, 2012 | Reply

  4. Indeed, I am aware of Baumol’s cost disease, but the paper you linked is an excellent demonstration – I saved it among my links. “The welfare state has no steady state” is brilliant.

    I started, but have not made my way through your previous posts on Danish demographics, which I had indeed missed, and they are so far excellent – thanks for the links. I should have been smart enough to look under the demographics category, but… old age takes its toll [begins to mumble about kids these days and how when we did not blog, but hunted dinosaurs].

    I’ll just note that I find it mildly amusing and puzzling that Poland is considered a Western country, and ex-Yugoslavia not. The two share the same geographical longitude, and culturally a Pole probably has a lot more in common with a Slovenian than with a Brit, but I guess you have to draw a line somewhere – just maybe not a very useful line.

    Comment by Plamus | September 9, 2012 | Reply

Leave a Reply

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

You are commenting using your 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.

%d bloggers like this: