Tenured professor of psychiatry here. I'll give you the summary of research on depression:
1. Emotional state is like blood pressure. It's a continuously varying thing. There is no discrete depressive disease state, just an extreme end of a continuum.
2. Just like blood pressure, though, very high blood pressure (or low) represents an illness state, a potentially very serious one. The etiology is a different matter. It may be due to sociocultural processes, or existential threat, or genetics, or neurological insults, or all of the above, each in some measure.
3. When we talk about "depression," it's usually a shorthand for a broader set of problems, related to lack of well-being, malaise, despair, sadness, anxiety, fear, and so forth.
4. The particular cocktail of emotions for each person is different though.
5. The best meta-analytic evidence suggests that antidepressants work, but not nearly as well for people as we were led to believe, and there's huge publication biases. Some drugs might not work at all. Some drugs work for some people; those same drugs might not work for other people at all.
6. Serotonin per se probably has nothing to do with the effect of antidepressants.
7. Psychotherapy works as well or as poorly as drugs on average, with some of the same publication biases. This is considered against many different types of controls, including placebo controls and various types of behavioral controls. Being wait-listed is the worst intervention of all, though.
8. Different psychotherapies do not differ in efficacy on average. The superiority of cognitive-behavioral therapy, etc. is a myth. Different therapies might differ for particular individuals, though, and there are some basic ingredients that seem to be necessary for efficacious therapy. The effect of therapist dwarfs the type of therapy (that is, some therapists are really really good, and others are less good, and that is much more important to outcome than type of therapy per se).
9. The best outcomes are generally therapy and drugs in combination, and if one thing doesn't work, trying different things will often work eventually.
The recent history ("recent" being the last 30-40 years) of psychiatry and clinical psychology is dominated by economically and politically-driven reductionism and oversimplification, within the realms of therapy as well as pharmacology. A lot of these dynamics are driven by a desire to present as rigorously scientific, wrapping ones' selves in banners of other disciplines. The problem with this is behavioral science is its own thing; it resembles other disciplines in some ways, but not in others (much like every other domain of science).