Much of the book relies on a simple modelling idea. The assumption (which is often relaxed to go into greater detail) is that the greatest determinant of future income is present income. This can be represented by graphs such as the following:
The graph on the left, the "inverted L-shape graph" represents a conventional picture of development in which, left to themselves, economies will tend to grow. One view of development is that this is how things tend to be, and the existence of poverty is largely due to interfering and corrupt governments preventing growth from happening.
An alternative view, represented by the "S-shape graph" on the right, suggests that there are "poverty traps" where, if a developing country were that bit richer, it would embark upon a path of growth, but for the moment growth is prevented by obstacles which could be overcome through outside investment. This is much of the idea behind programs like Jeffrey Sachs' Millenium Villages, which hope to provide sufficient wealth so as to push developing countries on to the growth path.
The foreword, aside from roughly describing the structure of the book, talks briefly about the authors' experience of writing the book, and gives a brief picture of what poverty looks like. When they speak of the global poor, they speak of people living on the equivalent (PPP) of 99 cents per day. Most of these people struggle to read, if they can read at all, which makes it difficult to buy certain products - how is one to judge health insurance when one has little or no way of obtaining further information about a disease with a long and unpronounceable name?
Chapter 1 gives a brief characterisation of two contrasting views on foreign aid - that typified by Jeffrey Sach, as outlined above, which believes that foreign aid can kick-start the local economy, and an opposing view held by figures such as William Easterly and Dambisa Moyo. They argue that aid in fact harms the recipients by corrupting local institutions and creating a lobby of aid agencies, and that the best thing that could happen would the emergence of genuine free markets. If these sound like left-wing and right-wing positions, that's because in general they are. They also seek to be universal proclamations about what aid will always or nearly always do, this book proclaims its intention to look at individual instances of aid and whether they have helped.
After this, it outlines a basic moral case for wanting to help the third world. In addition to referencing Peter Singer's Drowning Child Analogy, they mention Amartya Sen's view of poverty as a waste of talent. They then observe that the more controversial debate is what we actually can do. They don't say much, but my way of putting it would be that (1) ought entails can; hence (2) if we cannot relieve global poverty, we are under no obligation to do so. For this reason, the moral case for foreign aid is dependent upon us being able to actually help the people in receipt of aid.
Next, it suggests that there are complex issues which can only be answered by careful gathering and analysis of data. They give the example of insecticide-treated bed nets, which are tremendously effective as a way of reducing malaria. Is the best way to distribute these to give them away, or to sell them? If we sell them, should we do so at cost or subsidise them? In case it seems obviously better to give them away, remember that this costs money, and (at least in theory) it is always better, rather than just giving bed nets away, to give away enough money to buy bed nets and to sell them. This allows aid recipients to buy other things which they may consider a higher priority than a bed net.
There are other arguments against just giving the nets away; some think that, if the nets are given away for free, they will not be valued and so will not be used or will be misused. A (in my opinion more plausibly true) argument is that, if people become used to receiving bed nets for free, then in the future they will refuse to pay for nets and so well-functioning markets could be ruined.
These are claims to be empirically investigated, and so Banerjee and Duflo then briefly go into the methodology used. To answer this particular question, randomised controlled trials (RCTs) were carried out in which individuals were randomly selected to receive different levels of subsidy to buy bed nets, and their behaviour in response recorded.
