# Prelim 2018

**Office hours: **(in SS** **6435 )

7/6 Friday: 1pm-4pm

7/13 Friday: 1pm-4pm

7/20 Friday: 1pm-4pm

7/27

~~Friday~~Thursday: 1pm-3:45pm7/31 Tuesday: 1pm-4pm

8/1 Wednesday: 1pm-4pm

In addition, I have plenty of time outside of the scheduled office hours, so feel free to contact me to set up a meeting. If you have specific questions/problems in mind, it's always a good idea to let me know a day or two in advance so I can take a closer look at it before we meet.

**(unofficial) Old Prelim solutions:**

2013 second attempt Complete

2015 first attempt Complete*

2015 second attempt Incomplete (No answer to the ridiculous question 1B)

2016 first attempt Complete*

2017 First attempt (Some parts still missing of question 3)

**List of prelim solutions corrections (*)**

2016 first attempt

The answer to 5g) should include: "Assuming the same discrete state space for A' and shocks as in the lecture notes". Briefly: If the Markov process can increase/decrease productivity by +/ 2,1 or 0 grid points. Take two firms at the lower threshold. Give one the -2 shock and the other the -1 shock. We then observe 2 different net investment rates for firms on the lower threshold. Thanks to Marco!

The answer to Ananth's guess and verify has a minor typo. In one of the last lines there it should be beta*rho*F*log(A)

2015 first attempt

Some of you have reacted to the way the phase diagram is calculated. What is done is the solution is to solve for c(t+1)-c(t)=0 while being careful about that the timing. In the lectures and in the discussion section we have treated k(t+1)=k(t) when solving for c(t+1)-c(t). I think both approaches are fine, and there is no distinction in continuous time. For more, see these lecture notes by Eric Sims (as always!).