[Hemanshu Kumar]

But O.P. doesn't say that it's a linear probability model, nor that the 0.692 number is a marginal effect. Those are very strong assumptions. If this is a logistic regression model and the 0.692 is the actual regression coefficient, the meaning is altogether different and would correspond not to any fixed probability difference but rather to an odds ratio of almost exactly 2 associated with a 10-fold difference in GDP. And if the 0.692 is the coefficient in a probit regression, it's something different yet again.

Either O.P. has missed key points in reading the article, or the article is seriously deficient and the results uninterpretable.

As it turns out, OP had missed out on the authors' statement about it being a linear probability model. Aside from that, the authors actually provided the interpretation of their coefficient (OP said "So the interpretation in the paper says..."). From that interpretation as well, it was easy to infer that the linear probability model was being employed.