Statistic

...

[pic 1]

[pic 2]

[pic 3]

As can be seen from the histogram and plots, the distribution of errors is close to normal distribution and standard deviation of ε is constant. The errors are also independent.

To determine whether independent variables and y are related, the null and alternative hypotheses are: H0: βi = 0, H1: βi ≠ 0 and then we compare the p-value under the 5% significance level. As can be seen from the table, following variables are related to volatility: E/P, b/m, LIQUIDITY, DFY and infl. Other variables, D/P, ntis, tbl, lty and IP are not so related to y. it means there is a linear relationship between volatility and one or more of these variables, but because of multicollinearity the t-test revealed no linear relationship.

D/P

E/P

b/m

ntis

LIQUIDITY

tbl

lty

DFY

infl

IP

D/P

1

E/P

0.64497

1

b/m

0.8865

0.7509

1

ntis

-0.0774

-0.015

-0.025

1

LIQUIDITY

0.00598

0.0579

-0.002

0.0463

1

tbl

0.66039

0.6567

0.7342

0.0529

-0.01725

1

lty

0.79583

0.6385

0.8138

0.1253

0.00656

0.9133

1

DFY

0.59198

0.1618

0.6511

-0.339

-0.06838

0.2816

0.3651

1

infl

0.25377

0.2958

0.3295

0.0773

-0.03944

0.396

0.3463

0.0349

1

IP

-0.1027

0.0523

-0.118

0.3027

0.09788

-0.013

0.0104

-0.3615

0.009

1

So the full model is

Volatility=0.1257 + 0.0348 (D/P) + 0.0513 (E/P) - 0.1690 (b/m) + 0.0509 (ntis)-0.3396 (LIQUIDITY) - 0.0071 (tbl) + 0.1833 (lty) + 7.5902 (DFY) - 2.2221 (infl) + 1.4314 (IP)

Intercept: the intercept, β0 =0.1257, provides the value of Y when all the variables are zero, however, as the data range of all the independent variables do not cover the value zero, we should not interpret the intercept seriously.

β1 =0.0348, for each additional 1 unit increase in D/P, volatility increases by 0.0348 unit, holding other variables constant.

Β2 =0.0513, for each additional 1 unit increase in E/P, volatility increases by 0.0513 unit, holding other variables constant.

Β3 = -0.1690, for each additional 1 unit decrease in b/m, volatility decreases by 0.1690 unit, holding other variables constant.

Β4 =0.0509, for each additional 1 unit increase in ntis, volatility increases by 0.0509 unit, holding other variables constant.

Β5 = -0.3396, for each additional 1 unit decrease in LIQUIDITY, volatility decreases by 0.3396 unit, holding other variables constant.

Β6 = - 0.0071, for each additional 1 unit increase in tbl, volatility decreases by 0.0071 unit, holding other variables constant.

Β7 = 0.1833, for each additional 1 unit decrease in lty, volatility increases by 0.1833 unit, holding other variables constant.

Β8 =7.5902, for each additional 1 unit increase in DFY, volatility increases by 7.5902 unit, holding other variables constant.

Β9 = -2.2221, for each additional 1 unit increase in infl, volatility decreases by 2.2221 unit, holding other variables constant.

β10 =1.4314, for each additional 1 unit increase in IP, volatility increases by 1.4314 unit, holding other variables constant.

As can be seen from the table, p-value of LIQUIDITY has largest

...