Analysts typically mention that the observations should not be correlated. In bootstraping, you are violating this. However, is there any statistical test on autocorrelation of the observed data? Simply, does any statistical test provide an answer on the question in the title?
Also, imagine...