Portfolios of the Poor (by Daryl Collins and three co-authors) is a masterly assessment of the financial needs of people on very low incomes based on the collection of detailed two-weekly or weekly diaries kept by researchers who tracked individuals and families in India, Bangladesh and South Africa. It is stuffed full of interesting and surprising insights, and should be read by anyone concerned with economic development and poverty reduction. I can't praise it highly enough. This is a model of the careful collection of evidence with important practical consequences.

The first surprise is the extent to which poor people – living on or about the $2 a day threshold, or about 40% of the world's population – use a wide variety of financial services, although many of them are informal. The reason is that such low incomes are typically extremely variable and so there is a more intense need than in the case of better-off people for financial management. Poor people have a greater demand for financial intermediation. Less surprising, perhaps, are the facts that poor people face much greater risks including the risk of theft or unanticipated life emergencies, and that their typical transaction is very small indeed. The latter is the key insight of providers in other industries with 'bottom of the pyramid' business models, but is far less widely adopted in financial services.

One of the consequences of these characteristics is that the management of cash flow is the over-riding need for people on such low incomes. The typical researchers' focus on balance sheets will show that the kind of people who kept diaries for this research will see little change in their (tiny) level of assets from one year end to the next, but there will have been, relatively speaking, huge flows of cash in and out in between. Another insight is that a year is far too long a time horizon for financial decision making – and this includes assessing loan rates using APRs. Most of the time poor borrowers will want a loan for a very short period and may regard the (high) interest rate as a fee for a service. There is an interesting chapter looking at the actual versus notional interest rates charged by different types of lenders. Many of the diary participants would have several different loans at rates ranging from zero from a family member or employer to a high rate from a microfinance institution, to a seemingly usurious rate from a moneylender – but one which might never be imposed in full.

The book notes that Grameen Bank has begun to adapt its policies and products to suit better the needs of poor customers, offering savings products for example, which are more urgently needed than loans in many cases, and more flexibility in loan terms and purposes. However, it argues that microfinance needs to modernise more – the group meetings are increasingly resisted by Grameen members, for example. Other microfinance institutions will have more need to reform, as Grameen has already responded to customer needs to some degree.

It only mentions the exciting growth of mobile banking en passant. I edited a report on m-transactions published by Vodafone in 2007; the MPesa example in Kenya cited there has grown by leaps and bounds, as have other mobile-based schemes elsewhere in Africa, the Philippines and elsewhere. The potential of mobile transactions for poor customers is that it offers a low-cost service using existing retail outlets (in contrast to branch banking for example) and so offers great promise for making small transactions affordable. This gap in Portfolios of the Poor is a minor quibble, however. Its key message is forceful: that the absence of reliable financial tools reinforces the other vulnerabilities faced by people on low incomes; they have a tremendous need for reliable and low-cost financial services as a pre-requisite for being able to cope with the many other vulnerabilities they experience.