Are there DSGE models with agents with different skill levels (and according wage differences) who can develop or lose their skills and with employees with dynamic/changing skill levels and skill domains? Can such models be used for developing and validation policies that are designed according to Amartya Sen Capability Approach theory? Capability Approach is the leading approach for government programmes that invest in human capital development and poverty eradication, e.g. Germany's policies are based on it. In such a way we could use DSGE for validating and forecasting and design of wider policies, no just monetary or fiscal policies. If there are no such DSGE models, are there any other tools for modelling human capital development. Why not to use DSGE as the common framework for all the major policies?
1 Answer
Most DSGE models don’t model heterogeneity in individual labor market. Although I suppose you could do that if you would really want to but, it would be incredibly data intensive to actually estimate.
You can have some DSGE models with endogenous growth where human capital accumulation plays a role but most of them are based on just exogenous technological growth.
Well you can in principle use DSGE to evaluate policy I would not use it for evaluating some microeconomic policies. There are several reason for this:
Even simple DSGE models are incredibly data intensive. Introducing heterogeneity in individual labor markets by let’s say assuming there is some probability distribution over human capital and further possibility for heterogenous human capital accumulation would push this even further. I am not sure you would find dataset to do this even if you would have full access to all available data.
DSGE models are built for describing and forecasting macroeconomy or macroeconomic policy analysis. While in principle I can imagine you could try to use them for evaluating some microeconomic policy by just changing parameters of one market and then examining the impulse response functions I don’t think you should and I never seen it done before. DSGE is not build for that kind of task. Form what I understand about Sen’s capability approach most of the things that he suggests require microeconomic policies not macro ones as he focuses on things like ignorance, oppressive state policies, lack of financial resources, ill health, lack of proper education etc. These are not really macroeconomic domains.
There are definitely more better and appropriate ways how to evaluate policies on microeconomic level, so what would be a point of using DSGE if it’s inferior tool for given task?
To answer the question if there are better ways how to evaluate policies the answer is yes.
For, example randomized control trails or if that is not possible there are methods such as regression discontinuity, differences in differences, synthetic control, instrumental variables. These are actually used all the time to evaluate policies that improve education, healthcare, lack of access to financial resources of poor. For example, I used to work in the Netherlands at CPB which was analytical unit that evaluates economic policies. I was in macro department but we had department for healthcare economics, labor economics and economics of education where my past colleagues worked on evaluating various programs aimed at improving education sector and hence also human capital accumulation and they mostly used the kind of models I mentioned.
Once you would determine the optimum policies on micro level you could use DSGE for evaluating what would be the impact of increasing taxes to pay for all those policies or spending on those policies since that’s one of the downsides of micro approach that you don’t see the general equilibrium effects of all that spending and taxing.
