The huge gap between the world's richest and poorest countries remains one of the great moral dilemmas for the West. It also presents one of the greatest challenges for development economics. Do we really know how to help countries overcome poverty?

In his eloquently written and deeply researched new book, The Great Escape: Health, Wealth, and the Origins of Inequality, Princeton University's Angus Deaton urges caution. For those worried about world poverty, it is unquestionably the most important book on development assistance to appear in a long time.

Deaton suggests that far too often, Western aid serves to assuage donors' guilt rather than improve recipients' plight. This is particularly the case when naive assistance serves to reinforce a dysfunctional status quo. Although Deaton supports select initiatives, particularly for delivering medical and technological knowledge, he questions whether the vast majority of aid passes the basic Hippocratic litmus test of "first do no harm".

For starters, assessing and implementing aid policy requires developing tools to gauge accurately where need is greatest. Economists have developed some useful indicators, but they are vastly less precise than politicians and the media seem to understand.

Most experts agree, and Deaton concurs, that at least 1 billion people on the planet live in desperate circumstances resembling conditions that prevailed hundreds of years ago. Our failure to alleviate their plight is morally reprehensible. But where, exactly, are the greatest concentrations of poor people? Data are hard to come by and even harder to interpret.

Attempts to convert national incomes into a common denominator are fraught with complications. To take one prominent example, there is a 25 percent margin of error on purchasing-power-parity comparisons between GDP in the United States and China. In other words, we cannot say whether China's output today equals 55 percent of US GDP or 92 percent. So much for precise forecasts of the date when China will overtake the US as the world's largest economy; we won't even know for sure when it happens!

This problem is hardly unique to comparisons of China and the US; it applies with perhaps even greater force when comparing incomes of the poor in Mumbai, India, with those of the poor in Freetown, Sierra Leone. Another major problem is measuring progress in a given country over time. How can one compare cost-of-living indexes in different periods when new goods are constantly upending traditional consumption models? Consider the impact of cell phones in Africa, for example, or the Internet in India.

Deaton goes on to offer a revealing critique of some of the most hyped and fashionable approaches to improving aid. For example, the "hydraulic model" of aid - the idea that if we simply pumped in more aid, better results would gush out - ignores the fact that funds are often fungible. Even if aid is narrowly targeted at say, food or health, a government can simply economize on expenditures that it might have made anyway and redirect them elsewhere - for example, to the military.

Direct delivery of medical help is one of the best options, but it still can be a huge drain on already-scarce local resources - hospitals, doctors and nurses. An influx of Western NGOs often bids talent away from nascent businesses that could help the country long after the NGOs reset their priorities and move on.

Indeed, there is a striking parallel between the problems caused by aid inflows and the "natural resource curse" (or "Dutch disease" as it is termed in Western countries), whereby inflows into one economic sector - typically oil or minerals - drive up economy-wide prices (including the exchange rate), rendering other sectors uncompetitive. Moreover, a great deal of this aid is delivered in kind and for strategic reasons, often supporting ineffective and kleptocratic governments.

Deaton observes that, in general, Western countries developed without receiving any aid. China and India, too, have succeeded in lifting hundreds of millions of people out of poverty with relatively little Western aid (particularly China). Deaton argues that aid providers must be extremely careful not to interfere with political and social forces that, over time, can generate organic - and therefore more lasting - internal change.

Despite these caveats, Deaton's message is fundamentally positive. For most of humankind, now is a better time than ever before to be alive. The path to development remains for others to follow. Highly targeted Western aid and advice can help, but donors must take more care not to stand in the way of the beneficiaries of their assistance.

The author, a former chief economist of the IMF, is a professor of economics and public policy at Harvard University. Project Syndicate