SOME IMPROVEMENTS FOR BOOTSTRAPPING REGRESSION-ESTIMATORS UNDER 1ST-ORDER SERIAL-CORRELATION

Standard t-tests can be severely biased in the presence of trended reg ressors and serially correlated errors. Veall (Economics Letters, 1986 , 21, 41-44) presented simulation evidence that traditional bootstrapp ing does not substantially reduce this bias. Subsequent research has d eveloped a number of refinements to the bootstrap including accelerate d bias correction, prepivoting and iterated bootstrapping. The models considered by Veall were re-examined employing these refinements. It w as found that substantial improvements are possible, particularly from prepivoting and iterated bootstrapping.