As editor of an economics journal, I regularly hear authors complain about long lags for referee reports. I am very sympathetic with these complaints and frustrated myself. What is especially frustrating is that I know that these lags are much shorter in many other disciplines, especially the hard sciences. Physicists and others are simply shocked at the lags in economics refereeing and publishing, and this has led to much of the recently emerged econophysics literature being published in physics journals like Physica A and European Physical Journal B, which are not indexed in the Journal of Economic Literature, and hence not read and not even known about by most economists.
Ofer Azar has argued in a paper in Economic Inquiry this year ("The slowdown in first response times of economics journals: can it be beneficial?" 45(1), 179-187) that indeed the increasing lag of authors hearing back from journals could be a good thing, maybe even optimal. This is because people need to pay a cost for sending papers to journals that are too far above them and in which they have no chance of publishing. Hence, they will get the signal if they have to wait a long time and just get rejected and send their papers to appropriate outlets. Azar argues (in a paper forthcoming in JEBO, which I edit) that the shorter response times in physics reflect that their papers are shorter on average than those in economics. This latter point is correct, but I think that what has happened is that a bad social norm has evolved in economics where referees simply assume that they can sit on papers from economics journals for a long time and just put them aside, fearing that if they get their reports back quickly, they will simply be punished by having more papers sent to them to referee, given that the average response times are so long. I dislike this social norm, but it is very hard to overcome (and the Berkeley Electronic Press's effort to shorten refereeing times has so far resulted in published articles that barely get cited in other papers).