Economics help!

1. (5 points) Assume availability of a sample (y;, I;, 2;), i = 1,. .., n. Discuss under which conditions the OLS estimates of B1 and 82 are unbiased.2. (3 points) Suppose we are particularly interested in estimating 81 but we only have information on a random sample for (yi, r;), i = 1, ..., n. Under which conditions would it be correct to estimate the model y = Bo+ Bir tu?The file VOTE1 . dta contains information on Congressional campaign expen- ditures for the U.S. House of Representatives in 1988. For each of the 173 two-party races we know the percentage of votes received by candidate A (voteA), the logarithm of campaign expenditures (in thousands of US dol- lars) for both candidate A (lexpendA) and candidate B (lexpendB) and, consequently, the share of total expenditure accounted for by candidate A (shareA). Consider the linear regression: voteA = Do + AlexpendA + ,lexpendB + u 1. (4 points) Use STATA to run the regression of voted on lexpendA and lexpendB and interpret the estimated coefficients. Are these sta- tistically significant? Answer: - rag voteA lexpendA lexpendB Bourca 83 Number of obs 173 F(2, 170) 310.69 Modal | 38048. 019 2 19024.0096 Prob > F 0.0000 10409 .2296 170 61. 230762 R-squared 0.7852 Adj R-squared 0.7827 Total | 48457.2486 172 281. 728189 Root MSE 7. 826 VOLGA Conf. At4. Err. t P>Iti [967: Conf. Interval] laxpanda 5. 24195 3726485 17.02 0.000 5.606335 7. 077564 laxpande -6. 756751 . 3798728 -17.79 0.00 0 -7.606625 -6. 00 5876 _cons 52. 03.393 2.750137 18.92 0.00 0 46. 61012 57.46775 The OLS estimate for 8, is roughly 6.3, meaning that a one percent increase in campaign expenditures for candidate A is associated with a 0.06 increase in the outcome, namely a 6% increase in his vote share, since shareA takes values between (0. 1). Similar figures and inter- pretations apply for 82, with an obvious negative sign. Both of these estimates are highly significant, with magnitudes of t-statistics around 17, way outside of the standard acceptance region (-1.96, 1.96) for a 5% statistical significance test, as also suggested by p-values being essentially zero. 2. (2 points) Do you believe these estimates should be interpreted as causal effects of campaign expenditures on voting outcome? If not