Eric Kaufmann, writing in The American, journal of the American Enterprise Institute, makes some interesting points about secularism, religion, and population sizes. His point: the religious have bigger families than the non-religious and so, over time, begin to outnumber them.
Maybe so, maybe not; this looks like someone making linear projections of existing trends without remembering that the important thing in forecasting is locating the inflection points. I think Kaufmann overlooks the trend of people to become more secular as they become more "first world and developed."
I expect the same forces that have already turned religious Europeans into secular Europeans and are turning religious Americans into secular Americans will likewise turn religious third world individuals secular as they become more developed. To anticipate otherwise is to believe something has changed.