Who can explain, not just complain?

The economists so entertainingly exposed by Tim Harford have good reason for staying undercover. They include the game theorists who nearly bankrupted UK telecoms by auctioning mobile phone licences for £23 billion (after New Zealand counterparts gave them away for the price of a handset); supermarket strategists who cull cheaper brands from their more affluent locations; and development specialists who speak up for sweatshops on the grounds that, without them, employees would sweat more and earn even less. But when the Financial Times writer and former economics tutor meets his invisible handlers at Waterloo Station they are in the more familiar role of exposing other people's devious logic. A coffee bar is blocking travellers' path on to the platforms. Passengers overbooked on cheap seats are being bribed to take later trains. Only a tenth of the premium paid-for "fair trade" coffee is getting through to the producer. Anyone can complain; only the dismal-science detectives can explain.

Harford accomplishes his mission "to help you see the world like an economist" with subtlety and style - and an ambiguity that may explain his switch from academia to journalism. He sets out to show how an economist's perspective can make a confusing world more intelligible - helping us to shop more cheaply, avoid traffic jams, stop cold killing the old, and (if we run one) make a poor country rich. The Undercover Economist is, however, equally revealing - and substantially more entertaining - when it probes the frequently Persian-sized gulf between economists' perspectives and the real world. When their advice is to redistribute income only by "lump sum" taxes, pollute Africa because it is cheaper than recycling in Britain, or deny blind people heart treatment because of their already lower quality of life, their logic appears to have been economised too far.

No naive cheerleader for neoliberalism, Harford is most revealing when he follows market forces where they were not designed to go. He cogently explains why health insurance breaks down when not made compulsory (as those at the lowest risk opt out of subsidising the highest), why perfectly competitive markets would generate zero profit and chaotic pricing, and why the cleverest of fund managers cannot beat a dartboard at picking hot stocks.

Once the library is swapped for the newsroom, this relentless flow of ideas can be a barrier to taking stock. While Harford is billed as applying the latest economic theories, his strongest sections draw more on the "classic" than the contemporary. Thus his assessment of externalities (benefits that are not paid for and costs that are not compensated for) comes straight out of Ronald Coase's 1960 presentation; as does its controversial conclusion, that giving people property rights that allow charging for the benefits they give or the costs they absorb can restore efficient allocation without state intervention. The problem of "adverse selection" - only the vulnerable buying insurance, only bad cars being resold - looks no further than George Akerlof's 1970 "market for lemons" probe. And a 24-page diagnosis of why some economies do not grow appeals mainly to Mancur Olson's (2000) diagnosis of rule by "stationary bandits", whose only concern is that the economies don't actually shrink.

Harford's economics is almost exclusively "micro". Even his venture into globalisation stops at classical trade theory, ignoring the more intriguing impact of capital flows. Growth, fiscal deficits, inflation and other "macro" dimensions are left to professorial popularisers such as Paul Krugman and Harford's own FT boss Martin Wolf.

But while others - notably Steven Landsburg ( The Armchair Economist ) and John Kay ( Everlasting Light Bulbs ) have covered micro topics with similarly well-informed wit, Harford deserves his place in the ever-shrinking realm of accessible economics. Beginning students will find many textbook theorems neatly translated here, and their lecturers would do well to swipe the brighter anecdotes before the start of term.

Alan Shipman is a freelance economist, occasional lecturer and tutor at Cambridge University.