One emphasis is on randomised control trials — carrying out some program in a random selection of villages, for example — and more generally on empirical evaluation of the merits of possible interventions. This is contrasted by Banerjee and Duflo with more ideological claims about the merits of aid, represented by the extreme positions of Jeremy Sachs and William Easterly (roughly "the more the better" and "it's all bad").
Four chapters introduce some of the key topics. Firstly, food is not actually a major problem for most of the world's population; there is hardly anywhere where providing cheap grain is particularly useful. In contrast, there are potentially large returns to nutrition programs targeted to specific regions and groups such as pregnant women and children. When it comes to health care, Banerjee and Duflo emphasize the importance of information, and the role of infrastructure and public health services in avoiding individuals having to understand and evaluate complex health options and medical procedures.
With education there is a debate between "supply-wallahs" who argue there's not enough provision of schooling and "demand-wallahs" who argue there's no demand for it. Banerjee and Duflo present some of the evidence for when there is a demand for education and when it is effective. They also highlight a tendency to over-invest in one child, due to a mistaken belief that smaller amounts of education aren't so useful. And with demographics, they consider the extent to which poor people control their fertility, why they may choose to have large families, and differences in the interests of household members.
Part two turns to finance. Noting "the enormous potential gains that could be achieved if the poor did not need to be the hedge-fund managers of their own lives" — by, for example, allowing riskier agricultural strategies with much higher average returns — Banerjee and Duflo argue there is significant missing insurance provision. They explain the limitations of both mutual help and micro-insurance, where credibility on one side and moral hazard and adverse selection on the other are constraints, and argue for a role for governments, not in taking over insurers but in subsidising premiums where the returns from insurance are high.
Micro-finance has many hidden complexities. One concern is maintaining the reputation of the lender, as shown by examples of "runs" in which borrowers have stopped repaying loans en masse. Another limitation is that micro-finance fails to fund larger businesses. On the other side, why the poor don't save more is best understood through the psychology of saving and the logic of self-control. (The analysis here is just as applicable to people in wealthier countries.)
More people in poor countries than in rich ones run their own businesses. But the poor are not natural entrepreneurs: for most this is a fallback from what they actually want, which is a secure job. This can be seen in their aspirations for their children and in their failure to grow businesses when given the opportunity.
Many critics of aid focus on failings in what Banerjee and Duflo capitalise as INSTITUTIONS, such as democracy, the rule of law, property rights, and so forth. They counter that: "the political constraints are real, and they make it difficult to find big solutions to big problems. But there is considerable slack to improve institutions and policy at the margin." And they emphasize the possibilities for improving "lower case" institutions, such as local governments, community governance, and the lower echelons of the bureaucracy. To get anything done it is necessary to take into account "the three Is: ignorance, ideology, and inertia", as well as the motivations and constraints of the all the actors involved.
Poor Economics is pitched at a "popular science" level and has hardly anything in the way of theory — and it seems as much motivated by ideas from public health as by development economics. Even as simple a concept as a marginal return is carefully explained, for example, and the most complex idea is the possibility of a "poverty trap", where the returns to investment (or nutrition, or some other input) are S-shaped. (Such static analysis may not have much relevance if the dynamics of poverty involve large shocks, as argued by Krishna in One Illness Away. And in general the potential benefit of an intervention has no necessary connection to whether a "trap" is involved or not.)
In any event, Poor Economics should be accessible to almost anyone and would make a good introduction to poverty and approaches to its alleviation even for complete newcomers to the topic. It offers a lot of information on specific topics, with solidly grounded examples and arguments, as well as an overview of the area as a whole.