Lecture_4_Regression_Part_1

Lecture_4_Regression_Part_1 - 1 Lecture 4 (Part 1)...

Info iconThis preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon
Lecture 4 (Part 1) Regression and Forecasting Read: (WK Ch 4, 5; other books) EC 413/513 Economic Forecast and Analysis (Professor Lee) 1
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Topics: Regression for Forecasting Trend and Seasonality Regression for Demand Analysis More on Regression Analysis 1. Modeling & Forecasting Trend Using Regression Linear TREND Consider a regression with a time trend Y t = a + b × t + e t t=1,2,. ., T Or, in general, Y t = a + bX t + e t t=1,2,. ., T Y t = Dependent Variable X t or t = Independent Variable a = constant or intercept b = slope coefficient = marginal effect = How much Y will increase, on average, when X increases by one unit (= Y/ X) Get back to Y t = a + b × t + e t t=1,2,. ., T Then, = average increase of Y t per year. Why? 2
Background image of page 2
Note: If we use YEAR instead of t, the estimate of the slope coefficient (b) remains the same. Y t = a + b YEAR + e t t=1,2,. ., T Example ; Regression results t = 115.8 + 8.6 × t or t = -17041.2 + 8.6 × YEAR (note: YEAR = t + 1995) Thus, the average increase per year is 8.6. Forecasting Trend 6 = 115.8 + 8.6 × 6 = 167.40 or 2001 = -17041.2 + 8.6 × 2001 = 167.40 7 = 115.8 + 8.6 × 7 = ? Y t t time YEAR 123 1 1 1996 134 2 2 1997 143 3 3 1998 150 4 4 1999 158 5 5 2000 Y t t time YEAR t 123 1 1 1996 115.8 + 8.6 × 1 = 124.4 134 2 2 1997 115.8 + 8.6 × 2 = 133.0 143 3 3 1998 115.8 + 8.6 × 3 = 141.6 150 4 4 1999 115.8 + 8.6 × 4 = 150.2 158 5 5 2000 115.8 + 8.6 × 5 = 158.8 - 6 6 2001 ? - 7 7 2002 ? 3
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
TREND in LOG What if? log( Y t ) = a + b × t + e t t=1,2,. ., T What does the estimate of b imply? = ? log( t ) = 4.765 + 0.061 × t (or log( t ) = -117.65 + 0.061 × YEAR ) 0.061 = Forecasts From Logged model log( 6 ) = 4.765 + 0.061 × 6 = 5.131 6 = e 5.131 = 169.18 why? log(y) = x y = e x Ex) log( 7 ) = ? and 7 = ? Note: We often use log-linear model when Y increases exponentially. Y t t YEAR log(Y t ) 123 1 1996 134 2 1997 143 3 1998 150 4 1999 158 5 2000 - 6 2001 ? - 7 2002 ? 4
Background image of page 4
Monthly Data (quarterly data, daily data, weekly data…) t = 5.847 - 0.035 × t Thus, the average decrease per month is 0.035 log( t ) = 1.766 – 0.006 × t →Τ he average decrease in growth rates per month is 0.6%. Quadratic TREND Y t = a + b 1 × t + b 2 × t 2 + e t t=1,2,. ., T If b 2 > 0, U-shaped. If b 2 < 0, inverse U-shaped. Calculating Growth Rates r t = × 100 % = × 100 % This is the standard definition of the (arithmetic) growth rates. Y t Month t 5.81 Jan-05 1 5.80 Feb-05 2 5.73 Mar-05 3 5.67 Apr-05 4 5.70 May-05 5 5
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
But, consider the average growth rates. Is it right? Calculating Geometric Growth The basic equation for geometric growth is Y t = Y 0 (1+r) t Y 2 = Y 0 (1+r) 2 75 = 100(1+r) 2 (75/100) = (1+r) 2 (0.75) 1/2 = (1+r) r = - 1 = 0.866 – 1 = -0.134 or -13.4% From this calculation, we can obtain a CONSTANT growth rate r. In finance, people often use the following formula.
Background image of page 6
Image of page 7
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 02/29/2012 for the course EC 513 taught by Professor Staff during the Fall '08 term at Alabama.

Page1 / 26

Lecture_4_Regression_Part_1 - 1 Lecture 4 (Part 1)...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online