While reduced mortality is good I'm much more cautious about interpreting poverty statistics, which draw elaborate inferences from very limited datapoints. For example, this definition of poverty as living on $2/day, while useful, is predicated on the idea of private property that flows from one person to another. We don't have a way to measure things like the quality of someone's social relations (people that would share their food with you, for example) so that never even shows up in the data, and we end up trying to describe the contents of the room by shining a flashlight through a keyhole.
One way I often think about it - not because it's more correct, but just in order to shift my perspective, is to consider that in 1820 the world population was just over 1 billion, so you had, say, 950m people worldwide living in extreme poverty. Today world population is 7.5 billion, so you have about 750 million people in extreme poverty.
Now, the percentage fall is great...but that assumes that the incidence of poverty would normally stay constant. But another way to think about it is that we've only reduced the incidence of extreme poverty by about 20% even though we've managed to grow overall population by a factor of ~7. Could it be that our economic system depends on the maintenance of a desperately poor underclass from whom wealth can be indirectly extracted and then more efficiently multiplied?
This is an unconventional approach, but bear with me for a moment. It's certainly true that if you're born right now, your probability of being born into extreme poverty is far lower than 2 centuries ago - great. But suppose you're in extreme poverty anyway - are you that much better off today than you were then? It's not as if being in extreme poverty now sucks ~10x less than it did 2 centuries ago - while it is happening to a somewhat smaller number of people in absolute terms, qualitatively you're just as badly off on the individual level - perhaps even worse off, because as part of a smaller minority people have less and less sympathy for your impoverished condition due to simple lack of common experience.
What if, instead of setting a goal to reduce extreme poverty as a proportion of the overall population, we had a goal to reduce the absolute number of people in extreme poverty to almost zero? We would still have people who were poorer than their peers but we might be able to reduce the incidence of poverty as a threat to survival. Currently we take the approach of growing the whole economy, and thus shrinking the percentage of people who are in poverty. but I argue that this is equivalent to growing the part of the economy that's not in extreme poverty and leaving the part of the economy that is essentially unchanged. Suppose, for example, that the lower a person's wealth/income, the shorter the time horizon on which returns were channeled to them? Thus, the benefits of an increase in GDP would be felt (albeit modestly) by the extremely poor first, and that the well-off received their rewards last?
It strikes me that an under-appreciated feature of capitalism is that size of income is strongly correlated with seniority, ie getting paid first, and that this is a Bad Thing. The most basic example of this - so 'normal' that it's rarely questioned - is that rent is generally payable in advance but wages are paid in arrears. Even if the worker's earned output exactly matches the cost of living, the worker is condemned to carry a debt of one calendar unit + interest - a small difference, but little different from a casino that offers 'generous' odds, which just means they'll take your money away more slowly. The worker doesn't enjoy an economic surplus until this additional overhead is paid off (not to mention the additional cost of thing slike rental deposits that are refundable in theory but rarely in practice).
But why should the size of a debt be correlated with seniority? Arguably, the more you can lend out, the greater the level of surplusage you enjoy. A millionaire may lend a friend $1000 without anxiety, whereas someone worth only $1000 would need to think cautiously before lending $100 - the millionaire is risking 0.1% of her wealth, while the poorer person would be risking 10%, 100x as much in relative terms. IF you had borrowed $1000 from a millionaire, and $100 from your poor neighbor, and your investment had paid off, should not the poor neighbor be paid first, to reflect the considerably greater risk undertaken on your behalf?
Of course, economics tells us that one way to handle this is through interest - the lender who is taking on a high risk should demand a high rate of interest for the loan. But having limited ability to supply capital in a competitive market, such lenders would have to be price takers. So if the going rate is 1%/day, the millionaire makes $10 while the poor lender makes $1 - his risk is 100x greater, but his payoff is 10x smaller.
Naturally, one lesson of this is to nor lend more than you can afford, but if you ever work for wages you're basically lending the employer the value of your labor for 2 weeks or a month or whatever the standard pay period is where you live. You're showing up and putting in the time, but you get paid after the work is performed while you are expected to pay in advance for all the things you consume. Anyone who isn't able to accumulate a capital surplus of a calendar period's living expenses is thus doomed to penury. I suggest that if smaller debts had default seniority over larger ones, to reflect the relatively higher risks taken on by smaller creditors, we would reduce the incidence of poverty significantly faster than we are doing now, at minimal damage to overall growth.