You should believe in religion because it makes you feel good – because it satisfies your cravings for certainty, for spirituality, for a moral code. You should believe in it because it facilitates social relationships with fellow believers, and because it provides opportunities for companionship. These reasons are enough to believe by themselves, as long as your religion does not prompt bad decisions that a purely secular approach would have avoided, such as joining the Taliban. Those of course must be weighed for their costs and benefits.
You should not believe in religion because it is a risk-free proposition. Depending on how you count, there are more than 10,000 various religious beliefs and practices concerning these matters. Simply choosing one and hoping that it is the correct course, without any other knowledge, is a very low probability play – particularly when some religions state that a wrong choice will be worse than none at all. You should also consider the fact that nobody consciously goes back and forth to the afterlife on a regular basis, which would allow us to reliably narrow down the possibilities. Therefore, the most simple technique of “nothing to lose” belief does not allow us to reliably predict an improved future.
You should not believe in religion to try and get reliable answers about the future, or about the afterlife. Most religions do not state predictions about the far future (years away) that bring in additional details that go beyond what purely secular prediction techniques provide. I would love to see any new or old religious tradition bring forth true predictions in this life that are not so vague as to allow many different possible future events to match the prophecy; however, I am not aware of any such religions.
While miracles such as the healings of Jesus are strong predictors of future events when they can be repeated, unfortunately no one is able to currently repeat these, or other miraculous events, at a rate greater than using purely secular means. Verifying historical accounts and making them consistent with current reality, without any mention of miracles or other phenomena we do not observe in our modern lives, is hard enough, and unreliable/inconsistent enough, as it is. Oftentimes, these secular histories that have come down to us are useless or worse than useless as predictors – such as the history of ancient Japan. Most of these stories of miracles or prophecies have also proved to be useless predictors; indeed, many religions claim that other stories of miracles are false. Therefore, for us to assume that miracles of this sort occurred in the past adds additional uncertainty, to historical accounts that are already unreliable in predicting non-miraculous events. As a result, we cannot even predict with any reliability that the miracles of Jesus, or other such events, ever happened, and so then to use these miracles as predictors of the nature of the afterlife and how to get there is even less reliable.
You should believe in religion to gain solace and strength in times of hardship, but you should not expect your belief to reliably change the situation or circumstances beyond the secular aspects previously noted. Religion did not save the early Christians from a massive amount of persecution at the hands of the Romans. Throughout the twentieth century, practice of most or all religions was greatly punished throughout the tyrannies of Asia, whereas a secular life was more likely to avoid the wrath of the secret police and other militias. To this day, countries with religion-based elements, such as Iran and Israel, continue to favor certain practices over others, despite numerous prayers to the contrary. In all of these cases and more, the persecution lasted for decades, so the religious beliefs have not provided prompt improvement during the lives of the practioners, even for the practice of religion and not for any other worthy cause with which the religion may be allied. World peace has not come to fruition, nor any other end-times magnitude of change. Therefore, using religious belief and practice (or anything!), we cannot reliably predict these types of events either.
(As an aside, we may strive towards these big goals for our various reasons, but that doesn’t mean that we can say with any greater confidence than “what the hey, take a shot” that we can get there. Our quest for such end-states may justify the pursuit in and of itself, but we have no reliable expectation of actually getting there, given the human history to date.)
As for lesser improvements, such as the percentage of persons starving, dying, and massively suffering relative to straight-line future predictions: secular events, such as the expansion of education and improvements of technology and economics, are also correlated with these changes. Consequently, there are other possible explanations; and, the rate of these Enlightenment/Rationalism/technology-type secular changes is far more highly correlated with the rate of these lesser improvements, than are the growth and spread of major religious beliefs. Religion has been around for millennia, whereas the rate of secular changes has accelerated greatly only in the last few hundred years, when the lesser improvements also ramped up. As such, it is far more likely that the secular explanations are the major factor and that any influence of pure religious belief (vs. actions taken inspired by religious belief) is a secondary factor.