There are no shortages of estimates of the economic cost of Japan's multiple catastrophes - every economist in the region has been asked to produce them, and they have little choice but to meet demand. Now at this stage, you'd expect me to decry the spurious precision of these estimates, and lament the stupidity of those who ask for them. After all, one of the real eye-openers of the last 10 days was the admission by the guy in charge of Three Mile Island when it melted down that he didn't have an accurate knowledge of what had happened inside that reactor until five years after the accident. It took that long for the reactor to cool down sufficiently for anyone to have a good look at the damage.
But despite this, the guesses economists are making today are valuable, because they at least erect a framework around which our understanding of the consequences can grow, and from which it can evolve. They're not accurate, they're in no way reliable, but they are helpful in tracking how our understanding changes.
And in the meantime, here's my advice on how to count the cost.
First, there are estimates of the damage - both the cost of reconstruction, and the cost of lost production. No-one really knows yet, but let's call it 'x'. Mainly, this is presented as a percentage of GDP, and, if you're lucky, there's a time-frame with it (over the next year at least).
These are more or less educated and informed guesses. But even if they were right, they'd not tell you what you need to know.
Because, second, there are the cashflow implications of the cost - ie, the extent of corporate reaction to the disasters. We can track this directly using the Ministry of Finance's quarterly survey of private sector balance sheets and p&ls. More than ever, over the next couple of years, these surveys will be far more important than the quarterly GDP numbers. Judging by the what happened after the Kobe earthquake of 1995, this number will be a significant multiple of 'x' - probably 8x to 10x. This cashflow impact really matters, because it will show up in the liquidity of the financial system, the appetite for government bonds, and the extent of repatriation of capital. In other words, on interest rates, bond yields, and the direction of the yen.
So the third thing you need to know, or forecast, is policy reaction to the cashflow crimp. Does the cashflow crimp potentially limit JGB issuance? And if it does, will we find Bank of Japan simply printing money to fit? Or, as I suspect the form will be, 'accepting for discount the bonds of the newly-formed Reconstruction Agency within 12 hours of issuance'