We compare the finite-sample performance of impulse response confidence intervals based on local projections (LPs) and vector autoregressive (VAR) models in linear stationary settings. We find that in small samples, the asymptotic LP interval often is less accurate than the bias-adjusted bootstrap VAR interval, notwithstanding its excessive average length. Although the asymptotic LP interval has adequate coverage in sufficiently large samples, its average length still far exceeds that of bias-adjusted bootstrap VAR intervals with comparable accuracy. Bootstrap LP intervals (with or without bias correction) and asymptotic VAR intervals are shorter on average, but they often lack coverage accuracy in finite samples.

We propose a formal test of the hypothesis that energy prices are predetermined with respect to U.S. macroeconomic aggregates. The test is based on regressing changes in daily energy prices on daily news from U.S. macroeconomic data releases. Using a wide range of macroeconomic news, we find no compelling evidence of feedback at daily or monthly horizons, contradicting the view that energy prices respond instantaneously to macroeconomic news and consistent with the commonly used identifying assumption that there is no feedback from U.S. macroeconomic aggregates to monthly innovations in energy prices.

The paper proposes a new measure of exogenous oil supply shocks. The timing, the magnitude, and the sign of this measure may differ greatly from current state-of-the-art estimates. It is shown that only a small fraction of the observed oil price increases during oil crisis periods can be attributed to exogenous oil production disruptions. Exogenous oil supply shocks cause a sharp drop of U.S. real GDP growth after five quarters rather than an immediate and sustained reduction in economic growth and a spike in CPI inflation after three quarters. Overall, exogenous oil supply shocks made remarkably little difference for the evolution of the U.S. economy since the 1970s, although they did matter for some historical episodes.

A Monte Carlo analysis of the coverage accuracy and average length of alternative bootstrap confidence intervals for impulse-response estimators shows that the accuracy of equal-tailed and symmetric percentile- t intervals can be poor and erratic in small samples (both in models with large roots and in models without roots near the unit circle). In contrast, some percentile bootstrap intervals may be both shorter and more accurate. The accuracy of percentile- t intervals improves with sample size, but the sample size required for reliable inference can be very large. Moreover, for such large sample sizes, virtually all bootstrap intervals tend to have excellent coverage accuracy.

Bias-corrected bootstrap confidence intervals explicitly account for the bias and skewness of the small-sample distribution of the impulse response estimator, while retaining asymptotic validity in stationary autoregressions. Monte Carlo simulations for a wide range of bivariate models show that in small samples bias-corrected bootstrap intervals tend to be more accurate than delta method intervals, standard bootstrap intervals, and Monte Carlo integration intervals. This conclusion holds for VAR models estimated in levels, as deviations from a linear time trend, and in first differences. It also holds for random walk processes and cointegrated processes estimated in levels. An empirical example shows that bias-corrected bootstrap intervals may imply economic interpretations of the data that are substantively different from standard methods.