Personally, I think that it depends on the economist's
general approach. Those that take a historical approach have a lot more weight with me: "this situation is relevantly similar to these situations in the past; people in these situations tried these solutions: here's how they worked out." Those that depend entirely on models always struck me as being relevantly similar to bands that built songs by formula; it might have the appearance of a good song, but usually wasn't... and when it was, it was usually a case of hitting the odds from time to time.
It's why I still look at Marx and Keynes with respect.