Analysis of seasonal and other short-term variations with applications to Finnish economic time series
Kukkonen, Pertti (01.04.1968)
Numero
28Julkaisija
Suomen Pankki
1968
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:bof-201708151535Sisällysluettelo
1. INTRODUCTION 9
2. THE METHOD OF ITERATED MOVING AVERAGES
2.1. General Remarks 12
2.2. Weighted Moving Averages 13
2.3. Computational Steps 15
3. THE METHOD OF ITERATED MOVING AVERAGES AS A METHOD OF ESTIMATION
3.1. Constant Seasonal Patterns 18
3.1.1. The Method of Moving Averages in Operator Notation 19
3.1.2. Some Concepts of Spectral Analysis 21
3.1.3. The Efficiency of Iteration 22
3.1.4. HANNAN's Estimator for Constant Seasonal Patterns 24
3.2. Changing Seasonal Patterns 25
3.2.1. The Frequency Response Function of the Filters C 26
3.2.2. Expectation of the Estimate of the Seasonal Component 31
3.2.3. Iteration and the Slutzky-Yule Effect 36
3.2.4. On the Treatment of the First and Last Observations in a Time Series 42
3.3. Examples of the Application of the Method of Iterated Moving Averages 45
3.4. Alternative Methods 48
4. REGRESSION ANALYSIS AS A METHOD OF ESTIMATING SHORT-TERM VARIATIONS
4.1. Effect of the Residual Component on the Forecast Error in the Case of Optimal Linear Filters 51
4.2. On Defining the Seasonal and Calendar Variations in Regression Analysis 53
4.3. Various Types of Models Employed in the Regression Analysis of Seasonal and Calendar Variations 59
4.4. The Least Squares of Differences Method 61
4.4.1. The Criterion Suggested by RUIST 62
4.4.2. The Least-squares Criterion for Differences and the Classical Linear Regression Model 63
4.4.3. Restrictions on the Parameters 67
4.4.4. The Method of Restricted Least Squares 69
4.4.5. A Computational Procedure for the Method of Restricted Least Squares 70
4.5. Elimination of Seasonal and Calendar Variations from the Time-series Data for Econometric Models through the Regression Method 75
4.5.1. LOVELL's Theorem 76
4.5.2. The Effect of Linear Restrictions 79
4.5.3. The Trend-cycle Component, Calendar Variations and Short-term Changes in the Seasonal Pattern 82
4.6. Applications of the Regression Method 84
4.6.1. Specification of the Trend-cycle Component in Applications of the Least Squares of Differences Method 85
4.6.2. Abrupt Changes in the Seasonal Pattern 97
4.6.3. Calendar Variations100
4.6.4. The Special Seasonal Variations due to Unseasonal Weather Conditions103
4.6.5. The Impact of the Choice of a Seasonal Adjustment Method on the Estimation of a Demand for Labour Model109
4.6.6. A Summary of the Application of the Least Squares of Differences Method114
REFERENCES 115
LIST OF SYMBOLS 117
APPENDIX 1. The Bank of Finland Method of Iterated Moving Averages for the Analysis of Seasonal Variations 119
APPENDIX 2. The Weights of Alternative Smoothing Formulae in the Method of Iterated Moving Averages 123
APPENDIX 3. Finnish Economic Time Series Data Used in Applications 128
2. THE METHOD OF ITERATED MOVING AVERAGES
2.1. General Remarks 12
2.2. Weighted Moving Averages 13
2.3. Computational Steps 15
3. THE METHOD OF ITERATED MOVING AVERAGES AS A METHOD OF ESTIMATION
3.1. Constant Seasonal Patterns 18
3.1.1. The Method of Moving Averages in Operator Notation 19
3.1.2. Some Concepts of Spectral Analysis 21
3.1.3. The Efficiency of Iteration 22
3.1.4. HANNAN's Estimator for Constant Seasonal Patterns 24
3.2. Changing Seasonal Patterns 25
3.2.1. The Frequency Response Function of the Filters C 26
3.2.2. Expectation of the Estimate of the Seasonal Component 31
3.2.3. Iteration and the Slutzky-Yule Effect 36
3.2.4. On the Treatment of the First and Last Observations in a Time Series 42
3.3. Examples of the Application of the Method of Iterated Moving Averages 45
3.4. Alternative Methods 48
4. REGRESSION ANALYSIS AS A METHOD OF ESTIMATING SHORT-TERM VARIATIONS
4.1. Effect of the Residual Component on the Forecast Error in the Case of Optimal Linear Filters 51
4.2. On Defining the Seasonal and Calendar Variations in Regression Analysis 53
4.3. Various Types of Models Employed in the Regression Analysis of Seasonal and Calendar Variations 59
4.4. The Least Squares of Differences Method 61
4.4.1. The Criterion Suggested by RUIST 62
4.4.2. The Least-squares Criterion for Differences and the Classical Linear Regression Model 63
4.4.3. Restrictions on the Parameters 67
4.4.4. The Method of Restricted Least Squares 69
4.4.5. A Computational Procedure for the Method of Restricted Least Squares 70
4.5. Elimination of Seasonal and Calendar Variations from the Time-series Data for Econometric Models through the Regression Method 75
4.5.1. LOVELL's Theorem 76
4.5.2. The Effect of Linear Restrictions 79
4.5.3. The Trend-cycle Component, Calendar Variations and Short-term Changes in the Seasonal Pattern 82
4.6. Applications of the Regression Method 84
4.6.1. Specification of the Trend-cycle Component in Applications of the Least Squares of Differences Method 85
4.6.2. Abrupt Changes in the Seasonal Pattern 97
4.6.3. Calendar Variations100
4.6.4. The Special Seasonal Variations due to Unseasonal Weather Conditions103
4.6.5. The Impact of the Choice of a Seasonal Adjustment Method on the Estimation of a Demand for Labour Model109
4.6.6. A Summary of the Application of the Least Squares of Differences Method114
REFERENCES 115
LIST OF SYMBOLS 117
APPENDIX 1. The Bank of Finland Method of Iterated Moving Averages for the Analysis of Seasonal Variations 119
APPENDIX 2. The Weights of Alternative Smoothing Formulae in the Method of Iterated Moving Averages 123
APPENDIX 3. Finnish Economic Time Series Data Used in Applications 128