Econometrics - Time Series Study Guide Exam With Advanced Answers.

Time Series - correct answer Observation on one person, country, or thing over time. Each observation is a a point in time. Time Series Assumptions - correct answer Still need the five core assumptions, but with modification. E(e, X) = 0 [means that future outcomes can't be correlated, which is too strong] Cov(es, et) = 0 [not likely true for time series data] Finite Distributed Lag Model - correct answer Current outcomes can depend on past variables. Impact propensity/Impact multiplier - correct answer What is the immediate impact of a change? Impulse response function - correct answer What is the impact (through time) of a one-time temporary increase by one unit? Long run propensity/Long run multiplier - correct answer What is the total impact of a permanent increase? Seasonality - correct answer - Many time series have seasonality, fluctuations through the year. - Include monthly, season, or quarter dummies as regressors to deal with this. Weak Dependence - correct answer A time series process is weakly dependent if independence between Xt and Xt+n increases as h approaches infinity. H represents the distance between observations in time. Data close together may be correlated, but data spread apart is not Dependance violates OLS assumption, but with weak dependence we can appeal to the Central Limit Theorem and Law of Large Numbers. Moving Average Process [MA(1)] - correct answer Xt = et + ae(t-1) x1 = e1 + a e0 x2 = e2 + a e1 x3 = e3 + a e2 - Weak dependence - Far apart observations have zero covariance Auto Regressive Process [AR(1)] - correct answer Yt = pY(t-1) + et - As long as the expected value of e is less than one then the impact of the y lag dies out over time and can be considered weakly dependent Autocorrelation - correct answer The error term at one date can be correlated with the error terms in the previous periods Autocorrelation Model - correct answer Yt = B1 + B2Xt + et et = pe(t-1) + vt Derivations - correct answer 1. Expected Value 2. Variance Properties of B - correct answer - unbiased - inefficient Testing for AR(1) Autocorrelated Errors - correct answer 1. Estimate the model using OLS 2. Save the residuals, so that you have a predicted value for the error term 3. Lag the residuals by 1 so that you can estimate the relationship b