Work blogging 2
As indicated the second paper on my reading list is not as easy to cover here as was the first one. It’s quite a bit more technical than the first paper was and so there’s a lot of stuff which is harder to cover. The paper covers many of the same themes the first paper does (it’s written by the same people and published around the same time), but it handles some of the aspects in far more detail. The modelling will probably be a bit hard to understand if you’ve never worked with economic models before but I’ve tried to outline in the post what at least some of the model-building stuff that’s going on is aiming at.
I have decided that I want to try to blog at least all of the material this specific course in question deals with; I haven’t yet figured out if it’ll make sense to try to ‘workblog’ the other stuff I’m doing this semester yet but it takes time to write these posts and so I can assure you that I’ll not try to cover everything I’m supposed to learn this semester. In the post I’ve decided to just write some relevant stuff about the various aspects of the models and -results presented in the paper and keep it relatively superficial (I’ve included nothing which relates to the stuff in the appendix) – hopefully you’ll understand a bit about what’s going on. Of course I mainly write these posts for myself – I know that I learn stuff from writing these posts – but please don’t forget that I’m actually also providing you guys a valuable service here; the last post I wrote was a condensed version of an almost 40 page paper which took me at least a few hours to read and prepare notes for which you could read in just, what, 5-10 minutes?
So, this new paper – what’s it all about? In the paper there’s some introductionary stuff which closely relates to the previous paper, there’s some theoretical model-work, and then there’s a part which handles some model simulations and numerical illustrations. I’ve spent most of my time with the model-work (both in the post and when working with the paper previously). A key policy challenge for decisionmakers is to find and settle for a ‘proper’ balance between incentives and insurance in the labour market, and part of what this paper does is to have a closer look at different aspects of workfare in order to figure out how workfare is likely to affect incentive structures in the labour market and thus labour market outcomes. If you haven’t read the first post, you should probably start there before going any further. As in the last paper, the (now no longer implicit) model operates with three groups: Employed, unemployed and people in activation. Again you have a threat effect, a lock-in effect and a wage effect. The paper disregards human capital considerations so the post-programme effect is absent and state dependence is not addressed. The modelling framework takes benefit levels as a given and then proceed to question whether workfare elements can change insurance/incentives-aspects of the model and improve labour market performance. In the previous paper I did not feel that it was completely clear how the different workfare dimensions worked and how they differed, but in the formalized presentation here it is made very explicit; the two main policy instruments are i) the probability that an unemployed person will be required to participate in an activation measure [P(au)] and ii) the activation work requirement [l(a)]. The latter refers to how much work you’re required to do while activated – the larger this is, the more time/effort you’re required to spend on activation. One way to think about it is that one is the probability of X and the other the effect size of X. Going away from a unidimensional workfare requirement isn’t just something they do ‘to add complexity to the model’; it is shown in the paper that the two variables can be expected to affect different groups in different ways and it is emphasized for this reason that the overall effects of various changes in the workfare requirements depend critically on the total policy package and the specific mixing of the two policy variables.
The utility functions are standard leisure-income specifications and the main variable of interest in the analysis is the search effort (and how this relates to unemployment). As already mentioned the model work illustrates (in more detail) the effects also covered in the previous paper, for example the threat effect [∂S(u)(∂l(a))>0, ∂S(u)/∂P(au)>0], and it also illustrates much more precisely the reasons why using a multidimensional specification of the workfare scheme is important when evaluating the effects of changes to the workfare requirements: ∂S(a)/∂P(au) is negative in the model, meaning that the effect on the job search effort of people in activation given a marginal increase in the workfare intensity is the opposite of the effect such a marginal increase would have on the job search effort of people who are unemployed (and not in activation).
Given the model specification, people in activation spend more time on the work requirement and job search combined than unemployed people spend on job search, but which group actually spends more time searching is ambiguous. Searching is of course only half of the story as there also needs to be some jobs that people who search can find. Unemployed search for jobs, firms have vacancies where people can get employed and the unsurprising equilibrium conditions are briefly outlined. The job finding rate (α) of people searching is decreasing in the wage rate. Wage determination takes place according to a Nash bargaining solution where the bargaining power is taken to be exogenous. A key variable when dealing with the matching aspect of the model is the labour market tightness, θ = v/s (where v is number of job vacancies available and s is the effective search volume).
How workfare affects job search incentives is important, but the main interest is of course rather the impact on (un)employment. The main fact to take away from that part of the formal analysis is that ‘things are complicated’. The net effect on the number of people who are unemployed, in activation and in employment from a given policy change “depends on the balance between counteracting effects”. The effective job finding rates (α*s) [job finding rate conditional on search times search volume] of the two main groups (unemployed and activated) are key, and their contribution can be decomposed into an indirect wage effect, which is unambiguously positive and will thus increase the effective job finding rate for both unemployed and activated, and a direct search effect the sign of which depends on the workfare dimensions and the groups in question. This is another reason why a general equilibrium framework is required to fully understand the effects involved (more below); as also mentioned in the previous paper, in analyses which do not include the indirect wage effect workfare elements will generally be perceived of as worse performing than they do in analyses which include the indirect effects.
The effects of a workfare policy change is not surprisingly dependent on the initial level of workfare introduced into the system; it is for example shown that if no workfare elements exists ex ante, a policy maker can decrease total unemployment by introducing workfare elements and holding the benefit level constant. The dynamics of the level-dependencies involved are made more explicit in the model simulations, where one of the conclusions is that at a low intensity of workfare intensity [P(au)] the threat and wage effects dominate (i.e. a marginal increase in the workfare intensity will impact employment positively) whereas when at a higher level the locking-in effect dominates (i.e. a marginal increase in the workfare intensity will impact employment negatively). On the other hand, when it comes to the work requirement [l(a)] total unemployment is unambiguously decreasing in the work requirement. Welfare goes down for all three groups analyzed, including the people in employment, when workfare is increased, although employers benefit from workfare because it impacts their profit share positively. As workfare can be thought of as to some extent effectively introducing slack into the budget constraint of the government, this party is of course another entity which gains from the introduction of the scheme.
The next paper in the series is Short-Run Equilibrium Dynamics of Unemployment, Vacancies, and Real Wages by Christopher Pissarides (who got the Nobel Prize two years ago). I’ve unfortunately not been able to find a non-gated version of this paper online.
No comments yet.