1

i.

ii. The proportionate effect on ̂

(

)

. To obtain the percentage

effect, we multiply this by 100: 1.2%. Therefore, a 50 point ceteris paribus increase in
ros is predicted to increase salary by only 1.2%. Practically speaking, this is a very
small effect for such a large change in ros.
iii. The 10% critical value for a one-tailed test, using
as 1.282. The t statistic on ros is

, is obtained from Table G.2

, which is well below the critical value.

Therefore, we fail to reject H0 at the 10% significance level.
iv. Based on the sample, the estimated ros coefficient appears to be different from zero
only because of sampling variation. On the other hand, including ros may not be
causing harm; it depends on how correlated it is with the other dependent variables
(although these are very significant even with ros in the equation).
2.

(

i.

) or about 0.228 to 0.596

ii. No, because the value 0.4 is well outside the 95% CI.
iii. Yes, because 1 is well outside the 95% CI.
3.

i. While the standard error on hrsemp has not changed, the magnitude of the coefficient
has increased by half. The t statistic on hrsemp has gone from approximately -1.47 to
-2.21, so now the coefficient is statistically less than zero at the 5% level.
(

ii. If we add and subtract

) from the right-hand side and collect terms we

have:
(

)

[

(

)

(

)

(
(

[

)]
)

(

(

)

(

)]

)

1

05. 4. Because ̂ ̂ (̂ ̂ ) . In columns (2) and (3). a ceteris paribus 10% increase in market value is predicted to increase salary by 1%. although its t statistic is only about -1. ( . (We can use the standard normal critical value. with tceoten 3. (̂ ̂ ) (̂ ) (̂ ) (̂ ̂ ). The t statistic is ) .96). i. we conclude that the size of the firm. Plugging this into the population model gives: ( ) ( ) The preceding equation is what we would estimate by regressing y on x1. but it is not negligible. Because the coefficient is an elasticity. Therefore. 1. does not matter. profit margin has no effect on CEO salary. We use column (3). It appears that.2. 5. we are not interested in the coefficient on log(employ). ii. 2 . This is very small. once firm sales and market value have been controlled for. ii. The t statistic on log(mktval) is about 2. The standard error is the square root of the variance. iv. as measured by employees. we can write . which is very small. This is not a huge effect. which controls for the most factors affecting salary.11 and tcomten -2. the coefficient on profmarg is actually negative. once we control for training and sales per employee (in a logarithmic functional form). which has a t statistic of 0.iii. The test statistic is iii. 3x1+x2 and x3. Therefore. From part ii. log(mktval) is statistically significant. which is just significant at the 5% level against a two-sided alternative. No.79. The coefficient and standard error on x1 are what we want. and we fail to reject whether we specify a one-tailed or a two-sided alternative. either. iii. These variables are individually significant at low significance levels. i.

with the standard errors in parentheses. but not as CEO. but not at the 1%.92%. ̂ ( ) . We already know that ̂ se( ̂ ) .018. From part (ii). 7.031) ) (0.3% to 13.087 The p-value for testing H0: against the two-sided alternative is about 0. ii. we run the regression log(price) on (sqrft-150bdrms) and bdrms and obtain the standard error on bdrms.1 and that is about 2. i. Other factors fixed.86 (two-sided p-value = 0.133) n = 401. are: (̂ ( ) (0. ( ) ( ) ( ) iii. i.29) (0. lowers salary by approximately 0. The results from the OLS regression. indicating a strong degree of multicollinearity. another year with the company.iv.1390 (or about 3. Yet each coefficient is very statistically significant: t statistic for ̂ for ̂ ( ) is about 5. On the other hand.004). 3 .027) (0. R2=0. More non-CEOs years with a company makes it less likely the person was hired as an outside superstar.6%.9%).84. which means that an additional 150 square foot bedroom increases the predicted price by about 8. The 95% confidence interval is about 0. 6.0326 to 0. and salaries of these people are bid up. another year as CEO with the company increases salary by about 1. The correlation is -0. This second finding at first seems surprising. so that we reject H0 at the 5%level. ii. now we also get . but could be related to the “superstar” effect: firms that hire CEOs from outside the company often go after a small pool of highly regarded candidates.71%.

they are jointly significant at the 5% level because the outcome of the F2. The coefficient on log(hseval) is an elasticity: a one percent increase in housing value. 4 .134) ) (0. psoda is estimated to increase by 1%.12%. iv. prppov. All of the control variables. It holds fixed three measures of income and affluence.018) 2 n = 401.029) (0.038) ) ( (0. holding the other variables fixed. Therefore. so I will go thru’ it in discussion on Monday). Because the regression in (iii) contains the most controls.184. other factors held fixed. (iii) seems the most reliable. R = 0. a reasonable estimate is that if the proportion of blacks increases by 0.396 statistic is about 3.10. so it is not surprising that some are individually significant. The OLS regression results when log(hseval) is added are: ( ̂( ) (0.52 with pvalue= -0.29) (0.03 (I made a mistake with the hint. v. Nevertheless. Adding log(hseval) makes log(income) and prppov individually insignificant (at even the 15% significant level against a two-sided alternative for log(income) and prppov is does not have a t statistic even close to one in absolute). and log(hseval) are highly correlated.iii. increases the predicted price by about 0. log(income). log(hseval) is individually significant and log(income) and prppov are jointly significant. The twosided p-value is zero to three decimal places.