This seems to be a fairly balanced (and pretty long!) review of the evidence from Scott Alexander who is a psychiatrist (I believe) which comes down on the side that the balance of the evidence is that SSRIs do have an effect that can't be explained purely as active placebo: http://slatestarcodex.com/2014/07/07/ssris-much-more-than-yo...
Here's another 2014 article leaning the other way by Irving Kirsch, one of the more prominent researchers to question the evidence for anti depressant efficacy: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4172306/
In Anatomy of an Epidemic Robert Whitaker details many of the problems around the published studies that raise real concerns how bias free the remaining positive results of these studies might be. In addition, pretty much all of the discussion above is based around short term effects and perhaps the most worrying part of Whitaker's book for me is the suggestive evidence of devastating long term effects of many psychiatric drugs, SSRIs included.
Personally I think the bar should be rather higher for evidence of short term benefit for any drug than the 'optimistic' view of the evidence for SSRIs and especially in light of evidence of the drug making the condition it is supposed to treat significantly worse over the long term. Given that even proponents of SSRIs seem to accept that talk therapy is equally effective I know I'd rather take the drug and side effect free option but Scott Alexander gives a different take on that in his article, pointing out that drugs are much cheaper (and hence more profitable for both psychiatrists and drug companies as Whitaker points out in his book).
https://en.m.wikipedia.org/wiki/Selective_serotonin_reuptake...