Say you're going to come into a sum of money. Would you rather get $3,000 next week, or $3,800 in a couple of months?

Well, that depends on whether you just ate a bag of sweets, according to a recent study.

Researchers gave a bag of sweets to two groups of people and asked them that question. Those who were allowed to eat the sweets first and then answer the question were quite happy to wait for the extra cash. The ones who had to answer the question before eating the sweets opted to take the smaller sum of money rather than wait.

Studies such as these, which focus on how human behaviour and emotions affect people's economic choices, are part of a rapidly growing field known as behavioural economics.

It challenges one of the fundamental principles of economic theory: That people are perfectly rational and always act in their best economic interests. And it has sweeping implications for how policy makers view the economy.

Naturally, none of this is news in fields such as advertising that play on emotions as a matter of course.

"Emotions have the capacity to turn us into virtually different people - spendthrifts or tightwads, among other things," said George Loewenstein, the professor at Carnegie Mellon University who worked on the study.

And that makes people's behaviour especially hard to predict for economists trying to divine pressing questions such as where consumer spending is headed, or whether Americans are more likely to spend or save the latest round of tax cuts.

For the first time, the US central bank, the Federal Reserve, has declared an interest in behavioural economics, which not that long ago was considered heretical by mainstream economists.

The Boston Federal Reserve Bank, one of 12 regional Fed banks, last week held its annual conference on the topic of "How Humans Behave - Implications for Economics and Policy".

For two days at a resort in Cape Cod, psychologists, anthropologists, sociologists, economists and Fed officials pored over MRI scans of the brain and discussed laboratory studies in which people make irrational economic choices.

"Behavioural economists have discovered the wild side of macroeconomic behaviour," said Boston Fed President Cathy Minehan.

The challenge now is making connections between the research, which draws heavily on psychology and the social sciences, and the economy at large.

"Behavioural economics provides a whole lot of missing links to understand things like why we have business cycles, why people's spending behaviour fluctuates over time," Carnegie Mellon's Loewenstein told Reuters.

"Macroeconomics doesn't always have good theories about those issues," said Loewenstein, who is a professor of both economics and psychology.

Traditional economics, for instance, doesn't spend a lot of time on individuals' tendencies to be over-confident or avoid risk, which can go a long way towards explaining recent stock market behaviour through the bubble years of the 1990s and the slow, painful bust of the past three years.

Some behavioural economists say policy makers should give more weight to consumer sentiment surveys, which ask people questions about current and future economic conditions and their spending plans.

The Federal Reserve has said it already pays close attention to confidence surveys, and they are some help in predicting near-term spending. But over a longer time frame of a few quarters, they do not provide a good guide to spending.

Prof. Loewenstein suggested that is because what consumers say they will do and what they actually end up doing, are two different things. Mood has much to do with it.

Federal Reserve officials who attended the conference were cautious on how the new theories might contribute to policy-making.

Fed Governor Donald Kohn said their application to key issues for monetary policy has been limited so far. But he added there was potential "at least around the margins" for behavioural theories to improve the conduct of policy.

"Much of the attention of the Fed is taken up with trying to understand the behaviour of households and businesses," Kohn said. "My question is whether behavioural economics can narrow our ignorance."