Time Series_3_Generalisation
1. Linear and Log-Linear Trend Model
1.1 Assumptions: (1) E[ε] = 0;
(2) ε is normally distributed;
(3) E[ε2] = σ2 const; and violation of this is called 'heteroskedasticity' 异方差, when subsamples spread out more than rest of the sample (conditional/unconditional);
(4) ε is iid; and violation is called 'autocorrelation' or 'serial correlation' (+/-);
(5) independent variables is uncorrelated with residuals;
(6) no linear relation between any two or more independent variables; and violation is called 'multicollinearity';
1.2 Deal with Heteroskedasticity
1.2.1 Scatter plots
1.2.2 Breusch-Pagan chi-square test:
(1) BP statistic = n × R2 (n = # of observations, df = k, k = # of independent variables, R2 is from a 2nd regression of ε2 from the 1st regression on the independent variables)
(2) This is regression of ε2 on the independent variables, if conditional hete-ity exisits, the ind-variables will significantly contribute to explanation of ε2.
(3) One-tailed test. Hete-ity is only problem if R2 and BP test statistic are too large.
1.2.3 Correction
Calculate 'robust standard errors' (or 'white-corrected standard errors' or 'hete-ity consistent standard error') and use it recalculate the t-statistics using original regression coeff.
1.3 Deal with Autocorrelation
1.3.1 Residual plots
1.3.2 Durbin-Watson statistic
(1) DW ≈ 2(1 - r), where r is correlation coeff between residuals from one period and those from previous period;
(2) DW = 2, then ε terms are homoskedastic and not serially correlated, i.e. r = 0;
DW > 2, means r < 0, thus ε terms nega-seri-correlated;
(3) DB Decision Rule: H0 is 'There's no positive autocorrelation.'
Reject H0 if DW < dlower, and conclude positive autocorrelation;
Inconclusive between dlower and dupper;
Don't reject if DW > dupper;
1.3.3 Correction
(1) Adjust coeff standard errors using Hansen method;
(2) Improve specification of model;
1.4 Deal with Multicollinearity
1.4.1 Signal: When t-tests indicate 'none coeff significantly ≠ 0' while at same time F-test significant and R2 is high.
1.4.2 Correction: Omit one or more independent variables using stepwise regression.
2. Forecasting ModelsAutoregressive (AR) Models
2.1 First Differencing: transform data to covariance stationary series when it's random walk (i.e. hs unit root).
2.1.1 Stationarity(history is relevant for forecasting):
Distribution doesn't change over time;
E(Yt) and Var(Yt) is const;
Cov(Yt, Yt-j) doesn't depend on t;
FirstAR model is correctly specified when no autocorrelation in ε terms, otherwise this AR need correction.
2.2 Autoregressive Model AR(p)
2.1.1 Method one: Run AR model and examine autoc-ion.
2.1.2 Method two: Dickey Fuller test
2.1.3 Mean Reversion: if xt > b0 / (1 - b1), the AR(1) predicts xt+1 will be lower than xt.
2.3 Autoregressive Distributed Lag Model ADL(p, r)
2.4 Model Selection
Criterion BIC (Bayesian Information Criterion) or AIC to decide optimal length of lag(p)
Choose p that minimizes BIC.
3. Other Notations
3.1 Use Clustered HAC SE in panel data to correct for serial correlation, allowing errors to be correlated within a cluster but not across the clusters. While in time series data, use Newey-West HAC SE which estimates correlations between lagged values in time series.
Criterion:m (# of lags) = 0.75 T1/3, T is # of obsevation (or periods)
3.2 Seasonality
3.3 ARCH
ARCH exists if E[ε2] in one period is dependent on E[ε2] of previous period. Thus, standard errors of regression coeff in AR models and hypothesis tests of these coeff are in valid.
3.4 Cointegration: Two time series are economically linked (related to the same macro variables) or follow the same trend (relationship is not expected to change).
The ε term from regression of these two time series will be covariance stationary and t-tests will be reliable.
yt = b0 + b1xt + ε
The ε term are tested for unit root using Dickey Fuller test with critical t-values calculated by Engle and Granger (DF-EG test). If rejects NULL of unit root, then ε terms generated by these two time series are covariance stationary and they are cointegrated, thus we can use the regression to model their relationship.