In 2002, the editors of the RFS, JF, and JFE simultaneously published an editorial that urged authors to make good use of the advice and input provided by referees.1 Recent informal communications have suggested to us that it is time to renew that advice. Many papers are submitted to our journals, and the scarcest resource we have as a profession is the supply of time donated by referees to read, consider, and comment on their colleagues' work. In general, the author does not know the identity of the referee, so referees can express honest opinions about the quality of the work without alienating the author. However, this system has the counterproductive consequence that authors can undervalue the services they receive.We are particularly troubled by two practices that we see too frequently. First, some authors submit papers to journals at a relatively early stage of production in the hope that “the referee will help me figure out how to revise it to make it publishable.” This strategy imposes substantial costs on both sides. It burdens the referees with responsibilities that are not theirs. For the submitter, it raises the probability that the referee and editor will reject the paper as being too distant from acceptability. By submitting a paper that is unpolished, an author can cut off a potentially valuable publication outlet.

This paper explores the implications of filtering and no-arbitrage for the maximum likelihood estimates of the entire conditional distribution of the risk factors and bond yields in Gaussian macro-finance term structure model (MTSM) when all yields are priced imperfectly. For typical yield curves and macro-variables studied in this literature, the estimated joint distribution within a canonical MTSM is nearly identical to the estimate from an economic-model-free factor vector-autoregression (factor-VAR), even when measurement errors are large. It follows that a canonical MTSM offers no new insights into economic questions regarding the historical distribution of the macro risk factors and yields, over and above what is learned from a factor-VAR. These results are rotation-invariant and, therefore, apply to many of the specifications in the literature.