You must submit a py file that answers these questions in the console.

Use the Industry Portfolios.csv as your test assets, converting them to excess returns using RF

in FF$3.$csv$.$ Use the Mkt$-$RF$,$ SMB$,$ and HML factors in FF$3.$csv as your tradable factors. Use all

the data you can. The article Fama and French $(1993)$ is $$Common risk factors in the returns on

stocks and bonds$$ in Journal of Financial Economics, vol. $33,1993,$ pages $356.$

$1.$ Explain briefly how Fama and French $(1993)$ build SMB and HML from six underlying

portfolios. Include how these six underlying portfolios are constructed. One paragraph

should suffice.

$2.$ Estimate the tradable factor model. For the food, retail, books, telecommuniation, and

automobile industries, report the alpha and betas on the tradeable factors

$3.($Continuing from the previous question$)$ Using White$$s $(1980)$ robust standard errors, report which of these are significant at the $5\%$ level

$4.($Continuing from the previous question$)$ Interpret what the significant alpha and betas

mean

$5.$ Now consider all $30$ industries. Use the LR statistic and $5\%$ level to test each of the following null hypotheses

a$)$ Mkt$-$RF does not explain the covariation of industry returns

b$)$ SMB does not explain the covariation of industry returns

c$)$ HML does not explain the covariation of industry returns

$6.($Continuing from the previous question$)$ Use the LR statistic and $5\%$ level to test the APT

hypothesis.

$7.$ Of your $4$ conclusions from the previous two questions: which are changed by instead

using a finite$-$sample adjusted LR statistic?