India's foreign exchange reserves are now nearly $150 billion. China's are almost $900 billion. There are other developing countries as well. Between them, the developing countries hold more than $5 trillion, which is a lot of money.

Back in 2002, when India 's reserves had begun to climb, and were around $40 billion, some economists argued vociferously that this was a costly mistake. By spending that money, they said, India could grow faster.

I tried telling them they should shut up because the insurance was well worth having. But when did you last meet an economist who listens to common sense?

They said insurance was fine but how many dollars does a country need for it? Eventually, thankfully, the debate petered out in the face of concerted developing country policy to accumulate reserves.

Now, four years later comes the suggestion that - hold on, folks - it makes sense after all to hold so many dollars! "Perhaps the insurance premium pays for itself," says Dani Rodrik in a recent paper*.

He says although for developing countries as a whole, "reserves have climbed to almost 30 per cent of developing countries' GDP� and the income loss amounts to close to 1 per cent of GDP� this does not represent too steep a price as an insurance premium against financial crises."

Well, well, well, what do you know! The pot is calling the kettle white. Is it any wonder that real decision-makers don't take economists seriously?

Rodrik will not say uncle, though. He says this policy is "highly unbalanced and far from optimal" because this represents an over-investment in the "costly strategy of reserve accumulation and under-invested in capital-account management policies."

What should they do instead? Reduce short term borrowing, he says. But India and China have done just that, as have several others, albeit not in the same measure, perhaps.

"The reason for this sub-optimal response is unclear," says Roderik. Could it be, he asks, because "unlike reserve accumulation, controls on short-term borrowing hurt powerful financial interests, at home and abroad?"

How come, he asks, it is all right to accumulate obscene levels of reserves but not all right to tax short-term flows?

The answer is so blindingly obvious that I am surprised it doesn't occur to a Harvard professor. It is control: reserves are controlled by central banks in a way that short-term inflows are not.

But he is right about one thing. Controls on short-term flows do upset very powerful people, especially politicians who bring back their ill-gotten gains via instruments like the participatory note that the Reserve Bank of India [Get Quote] has been unsuccessfully trying to abolish for the past two years. The preference for accumulating reserves is, thus, a response to this, not a substitute.

In the end, though, the issue boils down to whether the safety premium is too high. But this is a question of political, not economic judgement.

If governments listen to economists and run their reserves down, they may accelerate growth somewhat, or earn more in some other way. But if, God forbid, there is a crisis, neither the higher growth nor the lower financial costs would be of much use.

Well, says Rodrik, there is the Guidotti-Greenspan-IMF rule for holding reserves. This is that reserves should be equal to at least the short-term debt. Well, my driver could have told you that.

Be that as it may, it seems, this rule "reduces the (annual) probability of experiencing a sharp reversal in capital flows by 10 percentage points on average".

On average? What good is that? How does a central bank governor tell what his country's probability of experiencing a sharp reversal in capital flows is? What if Uncle Sam takes a sudden dislike and puts the word out that you are bad boy? It has been known to happen, you know.

*The Social Cost of Foreign Exchange Reserves, NBER Working Paper No. 11952 January 2006