Forecasting Output

1) Raw data, not seasonalized

2) Seasonal Adjustment used: Census II X-12 multiplicative (MASA): Used because of the presence of seasonal variations that are increasing with the level of my series. Increasing degree of variability overtime…

TX non seasonalized and seasonalized

3) Combined seasonally adjusted with non-seasonally adjusted

De-seasonalizing the data helped with the removal of seasonal component that creates higher volatility in model. Now, variations more stable

4) No cyclical component. No evidence of unfixed wavelike rises and falls around the trend

5) Test for stationarity employs a correlogram. This correlogram exhibits a pattern that shows values of rks that decline rapidly for larger values of k

Correlogram for stationarity Texas

Correlogram for stationarity US

ADF Test for stability: If unit root present or not. In levels

Null Hypothesis: TX_SA has a unit root

Exogenous: Constant, Linear Trend

Lag Length: 1 (Automatic - based on SIC, maxlag=15)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic
-1.479488
0.8346
Test critical values:
1% level

-3.990935

5% level

-3.425841

10% level

-3.136094

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(TX_SA)

Method: Least Squares

Date: 12/05/13 Time: 23:58

Sample (adjusted): 1990M03 2013M08

Included observations: 282 after adjustments

Variable
Coefficient
Std. Error t-Statistic Prob.

TX_SA(-1)
-0.016026
0.010832
-1.479488
0.1401
D(TX_SA(-1))
-0.176957
0.058894
-3.004657
0.0029
C
2.964990
1.932313
1.534426
0.1261
@TREND(1990M01)
-7.33E-05
0.000884
-0.082887
0.9340

R-squared
0.041782
Mean dependent var
0.064045
Adjusted R-squared
0.031441