Scribd
Upload
35 views

An Attempt To Understand and Model The Relationship Between Wage Rate, Consumer Price Index, Unemployment and Minimum Wages

The document describes modeling the relationship between wage rate, consumer price index, unemployment, and minimum wages using a vector autoregressive (VAR) model. A VAR(1) model is fit to the data and the results are presented. The VAR(1) model shows good fit based on multiple R-squared values and significant parameter estimates. Residual diagnostics also indicate the VAR(1) model captures the dynamics in the data well based on residuals, autocorrelation, and partial autocorrelation plots.

Uploaded by

Rajesh Sharma
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
35 views

An Attempt To Understand and Model The Relationship Between Wage Rate, Consumer Price Index, Unemployment and Minimum Wages

The document describes modeling the relationship between wage rate, consumer price index, unemployment, and minimum wages using a vector autoregressive (VAR) model. A VAR(1) model is fit to the data and the results are presented. The VAR(1) model shows good fit based on multiple R-squared values and significant parameter estimates. Residual diagnostics also indicate the VAR(1) model captures the dynamics in the data well based on residuals, autocorrelation, and partial autocorrelation plots.

Uploaded by

Rajesh Sharma
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 11

An Attempt to understand and model the relationship between wage

rate, Consumer Price Index, Unemployment and Minimum wages

Introduction

The objective is to understand the relationship between Wage Rate,


CPI, Unemployment and Minimum Wages and model them using a
Vector Autoregressive model. We can as well develop a multiple
regression equation to describe the relationship between each of
these, but since they are time dependent we need to carry out
multiple time series analysis to understand and model their
relationship which in turn would help us in predicting their values in
the future.
The first step towards this objective is to plot the time series plots for
each of these variables, wage rate, CPI, unemployment and minimum
wages and see if they are stationary or not.
Note: All the 4 sets of values are first converted into Log values and
then plotted.

Page 2 of 11

Page 3 of 11

From the above plots, we find that wage rate and CPI have some trend
components associated with them as the plot is rising whereas in the
case of unemployment and minimum wages there is no significant
trend. Although the components wage rate and CPI may be stationary
which can be confirmed by ADF test, VARS may still be non-stationary.
The modeling process ends when the roots of the characteristic
equation or the eigen values of the coefficient matrix of the model are
all less than one.
Here we do a VAR (Vector Autoregressive model) for modeling the
relationship but the order is selected based on the following criteria.
AIC(n) HQ(n) SC(n) FPE(n)
3
3
3
4
$criteria
1
2 3 4 5
AIC(n) -1.335691e+01 -1.528034e+01 -Inf -Inf -Inf
HQ(n) -1.353556e+01 -1.560191e+01 -Inf -Inf -Inf
SC(n) -1.248776e+01 -1.371587e+01 -Inf -Inf -Inf
FPE(n) 1.868993e-06 8.316862e-07 NaN 0 0
For understanding the dynamics of the relationship one can select HQ
or SC criterion and for forecasting one can select AIC or FPE criterion.
Our aim is to understand the dynamics because of which we start with
VAR of order 3.
But since VAR(3) becomes computationally singular we dont get a
model at all.

Page 4 of 11

Next we try to fit a VAR(2) model but there again we encounter the
problem of some of the eigen values being greater than one. Hence
we finally fit a VAR(1) model whose parameters are as follows:
Note:
V1 Wage Rate
V2 CPI (Consumer Price Index)
V3 Unemployment
V4 Minimum Wages
VAR Estimation Results:
=========================
Endogenous variables: V1, V2, V3, V4
Deterministic variables: const
Sample size: 17
Log Likelihood: 29.71
Roots of the characteristic polynomial:
0.8035 0.8035 0.2903 0.2903
Call:
VAR(y = d, p = 1)
Estimation results for equation V1:
===================================
V1 = V1.l1 + V2.l1 + V3.l1 + V4.l1 + const
Estimate Std. Error t value Pr(>|t|)
V1.l1 0.742472 0.349668 2.123 0.0552 .
V2.l1 0.108812 0.208816 0.521 0.6118
V3.l1 -0.291610 0.153100 -1.905 0.0811 .
V4.l1 -0.008727 0.004109 -2.124 0.0552 .
const 0.325171 0.201007 1.618 0.1317
--Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
Residual standard error: 0.06524 on 12 degrees of freedom
Multiple R-Squared: 0.9057,
Adjusted R-squared: 0.8743
F-statistic: 28.83 on 4 and 12 DF, p-value: 4.514e-06
Estimation results for equation V2:
===================================
V2 = V1.l1 + V2.l1 + V3.l1 + V4.l1 + const
Estimate Std. Error t value Pr(>|t|)
V1.l1 1.510074 0.626845 2.409 0.033 *
Page 5 of 11

V2.l1 0.062697 0.374341 0.167


V3.l1 -0.450232 0.274461 -1.640
V4.l1 -0.008575 0.007366 -1.164
const -0.210026 0.360343 -0.583
--Signif. codes: 0 *** 0.001 ** 0.01

0.870
0.127
0.267
0.571
* 0.05 . 0.1 1

Residual standard error: 0.117 on 12 degrees of freedom


Multiple R-Squared: 0.8996,
Adjusted R-squared: 0.8661
F-statistic: 26.87 on 4 and 12 DF, p-value: 6.564e-06
Estimation results for equation V3:
===================================
V3 = V1.l1 + V2.l1 + V3.l1 + V4.l1 + const
Estimate Std. Error t value Pr(>|t|)
V1.l1 -0.196962 0.400717 -0.492 0.63192
V2.l1 0.250139 0.239301 1.045 0.31649
V3.l1 0.591796 0.175452 3.373 0.00554 **
V4.l1 0.001241 0.004709 0.264 0.79661
const 0.299362 0.230352 1.300 0.21816
--Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
Residual standard error: 0.07477 on 12 degrees of freedom
Multiple R-Squared: 0.6926,
Adjusted R-squared: 0.5902
F-statistic: 6.761 on 4 and 12 DF, p-value: 0.004347
Estimation results for equation V4:
===================================
V4 = V1.l1 + V2.l1 + V3.l1 + V4.l1 + const
Estimate Std. Error t value Pr(>|t|)
V1.l1 22.76450 23.83620 0.955 0.358
V2.l1 -13.24300 14.23455 -0.930 0.371
V3.l1 2.68582 10.43655 0.257 0.801
V4.l1 0.02309 0.28011 0.082 0.936
const -13.96566 13.70227 -1.019 0.328
Residual standard error: 4.447 on 12 degrees of freedom
Multiple R-Squared: 0.07177, Adjusted R-squared: -0.2376
F-statistic: 0.232 on 4 and 12 DF, p-value: 0.9151
Page 6 of 11

Covariance matrix of residuals:


V1
V2
V3
V4
V1 0.0042564 0.005190 -0.0009335 0.16277
V2 0.0051903 0.013679 0.0017712 0.25066
V3 -0.0009335 0.001771 0.0055899 0.08446
V4 0.1627740 0.250662 0.0844646 19.77894
Correlation matrix of residuals:
V1
V2
V3
V4
V1 1.0000 0.6802 -0.1914 0.5610
V2 0.6802 1.0000 0.2026 0.4819
V3 -0.1914 0.2026 1.0000 0.2540
V4 0.5610 0.4819 0.2540 1.0000
The next step is to remove or drop the insignificant terms in the model
and in the process we get the following model:
VAR Estimation Results:
=========================
Endogenous variables: V1, V2, V3, V4
Deterministic variables: const
Sample size: 17
Log Likelihood: 13.939
Roots of the characteristic polynomial:
0.9748 0.7689 0.1204
0
Call:
VAR(y = d, p = 1)
Estimation results for equation V1:
===================================
V1 = V1.l1 + V4.l1
Estimate Std. Error t value Pr(>|t|)
V1.l1 0.974850 0.030205 32.27 2.82e-15 ***
V4.l1 -0.010922 0.004152 -2.63 0.0189 *
--Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
Residual standard error: 0.06899 on 15 degrees of freedom
Multiple R-Squared: 0.9922,
Adjusted R-squared: 0.9911
F-statistic: 952.5 on 2 and 15 DF, p-value: < 2.2e-16

Page 7 of 11

Estimation results for equation V2:


===================================
V2 = V1.l1 + V3.l1
Estimate Std. Error t value Pr(>|t|)
V1.l1 1.5472
0.1416 10.930 1.53e-08 ***
V3.l1 -0.6838
0.1374 -4.977 0.000165 ***
--Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
Residual standard error: 0.1147 on 15 degrees of freedom
Multiple R-Squared: 0.972,
Adjusted R-squared: 0.9682
F-statistic: 259.9 on 2 and 15 DF, p-value: 2.287e-12
Estimation results for equation V3:
===================================
V3 = V2.l1 + V3.l1
Estimate Std. Error t value Pr(>|t|)
V2.l1 0.13534 0.05989 2.26 0.0391 *
V3.l1 0.88929 0.05014 17.74 1.78e-11 ***
--Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
Residual standard error: 0.07513 on 15 degrees of freedom
Multiple R-Squared: 0.9906,
Adjusted R-squared: 0.9894
F-statistic: 791.9 on 2 and 15 DF, p-value: 6.197e-16
Estimation results for equation V4:
===================================
V4 = const
Estimate Std. Error t value Pr(>|t|)
const -3.4233
0.9696 -3.531 0.00278 **
--Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
Residual standard error: 3.998 on 16 degrees of freedom
Multiple R-Squared: 0.4379,
Adjusted R-squared: 0.4028
F-statistic: 12.47 on 1 and 16 DF, p-value: 0.002777

Page 8 of 11

Covariance matrix of residuals:


V1
V2
V3
V4
V1 0.0059261 0.0035532 0.0005615 0.15197
V2 0.0035532 0.0164242 0.0004556 0.23990
V3 0.0005615 0.0004556 0.0070235 0.07512
V4 0.1519736 0.2399033 0.0751189 21.30826
Correlation matrix of residuals:
V1
V2
V3
V4
V1 1.00000 0.36016 0.08703 0.4277
V2 0.36016 1.00000 0.04242 0.4055
V3 0.08703 0.04242 1.00000 0.1942
V4 0.42767 0.40553 0.19418 1.0000
In almost all the cases, Multiple R-square value looks good (reasonably
high) and all the terms are also significant. The roots of this VAR(1)
model is also shown above which is once again indicated below:
Roots of the characteristic polynomial:
0.9748 0.7689 0.1204
0
Since all the Eigen values are less than 1, the coefficient matrix of the
model is said to be stable. Next, we plot the model to analyze the
goodness of fit and residuals, ACF and PACF plots to find how good the
model is.

Page 9 of 11

Page 10 of 11

Page 11 of 11

You might also like