TIME SERIES REGRESSION NOTES
Time series regression analysis is tricky! Be careful and dont generate spurious results.
1. Transform all variables (both dependent variable and explanatory variables) to
stationary form. Then your equation will be balanced.

Credit Crunches, Deleveraging and the
Zero Lower Bound
Standard NK model
Euler equation with discount factor shock qt
Et ct+1 ct =
1
(it Et t+1 + qt )
Calvo pricing frictions:
t = Et (t+1 ) + ct
Discount factor shocks follow autoregressive process
qt =

Miscellaneous Classification Topics
Oversampling
When classes are present in very different proportions as in 2% success (1) and 98% failure (0)
and in the case that the benefits of identifying a success are much greater than the loss due to missspecifyin

Real Rigidities
NK models vs. Data
Data: long-lived eects from monetary shocks
half-life = 10 quarters
Models: half-life 1/2 - 2/3 average price durations
E.g. Calvo with = prob. of not changing prices
yt = yt1 + t
Half-life with = 1 1/12 is 8 mos.
Ne

3
Segmented Markets
Standard monetary models (cash-in-advance, money in utility function) imply, counterfactually,
that a persistent increase in the money supply increases nominal interest rates since expected
ination increases. In contrast, in the data n

7
Innovation, Technology Adoption and Firm Dynamics
7.1
Parente (JET, 1994)
Studies economy in which rms choose the timing of a technology adoption. Adopting a better
technology entails a) a xed cost and b) loss of expertise. There is also learning-by-doi

4
Misallocation
We study models in which wedges in the rms Euler equations for hiring capital/labor generate aggregate TFP losses. We study two mechanisms that generate such wedges: i) heterogeneity in markups across producers and ii) nance frictions. We

1
Economies with Fixed Price Adjustment Costs
A widely held view in macroeconomics is that changes in monetary policy (or more broadly
nominal, nancial and demand shocks) have eects on real activity because prices are sticky.
Price stickiness is usually m

1
Firm Dynamics
We next extend the tools we have learned to an environment in which heterogeneity is on
the rm side. In particular, we extend the equilibrium concept we used above to allow
for rm entry and exit. The model we study below is a variant of th