It probably is a pseudo-science in the sense that the systems observed are very difficult to perform experiments on. I lump it with ecology and climatology as a "complex science." Obviously, psychology is in there too.
Keynsian approaches
AFAIK, are more or less the origin of modern economics; you next get Uncle Milty and the Chicago school; right now it is what jyl mentioned, "behavioral economics." This last one muddies the fine theoretical waters of the first two, b/c it assumes that psychology messes up rational choice.
In any field, it is always easier to model things when they are at equilibrium and in a steady-state. That always gets done first. Later, people start to explore transient behavior. Generally, simplicity in a model is at odds with reality, i.e. increased precision. It's hard to get both...