use tbill3,
log using tbill3.log, replace

* set the data for time series
tsset time

* look for a random walk
* run a regression of tbill3
* on lagged tbill3


* lag of tbill3
gen tbill3_1=tbill3[_n-1]

* run y on ylag regression
reg tbill3 tbill3_1 
test tbill3_=1

* now construct the 1st difference
* the lag the 1st difference

gen dtbill3=tbill3-tbill3_1
gen dtbill3_1=dtbill3[_n-1]


* now run regression of
* difference on lag difference
reg dtbill3 dtbill3_1 
test (dtbill3_1=1)

log close