Notes on Shapiro and Stiglitz

First, I believe the link I gave you previously for the paper is now broken.  You can get it from JSTOR.  Note that you have to be on the Campus network for the link to give you full access and be able to download the PDF.  From home use VPN for that purpose.

Next, let me try to sketch a little picture of "circular flow" that is going on in the background of this paper.  Workers can be in one of two states, employed or unemployed. The equilibrium of the model is stationary - the level of employment stays fixed throughout as does the level of unemployment.  But at each moment in time some workers separate from their jobs, which creates vacancies, and some unemployed workers find those jobs, filling those vacancies.  That unemployed workers want good paying jobs is not surprising, so the filling the vacancies part makes sense intuitively.  That workers with good jobs nonetheless separate makes less sense.  In reality, some workers do retire or quit to move elsewhere and find other work.  But here there is no retirement (everybody lives till infinity, an assumption to keep the modeling simple) and when the quit happens the worker becomes unemployed.   So that doesn't seem to correspond with reality very much.  The reason they need the quit rate, b, to be positive, is so that job vacancies do happen in equilibrium.

With that, the intuitions of the paper are not hard, once you get the drift of the asset equations.  I've written up some notes for that, so you can see how those equations are derived.  From the worker's point of view there are three asset equations to derive - the value of an employed worker that shirks, the value of an employed worker that doesn't shirk, and the value of  of an unemployed worker. It seems intuitive enough that if the firm catches the worker shirking then the worker gets fired. (In reality there are steps of "progressive discipline" with firing the worker the last step.  The aim of the intermediate steps is initially to turn the situation around and make the worker productive, then later to document the cause of dismissal.)  However, in the model with all workers homogeneous, the firm is really indifferent to firing the worker or not.  Again, this is a little fudge to keep the modeling simple.

The heart of the paper is the no-shirk condition and understanding the implications of that.  A good job produces a surplus for the worker over what an unemployed worker gets.  Its the desire to maintain that surplus in the future that motivates the worker to put in effort at present.