Central banks the world over face a strange conundrum. High oil prices have pushed headline inflation up and yet non-oil (core) inflation remains tame. As the IMF's World Economic outlook released in April puts it: "[T]he impact of oil prices on core inflation to date has been surprisingly mild relative to previous experience."

If one takes the G-7 (broadly the big industrialised economies) average, consumer price inflation has risen steadily from 1.5 per cent in early 2004 to well over 3 per cent in 2006. Yet, core inflation has remained virtually flat in a range between 1.75 per cent and 2 per cent.

The corollary, of course, is that manufacturers are stuck with low pricing power and the rise in their energy costs has translated into a direct hit on the margins.

Indian wholesale price inflation data throw up a similar puzzle. Let me take a recent example. Oil prices were hiked in September 2005. In the six preceding months, manufactured product inflation, a rough proxy for core inflation, was at 3.8 per cent.

In the six that followed, average manufactured product inflation declined to 2.8 per cent. In fact, our attempt to model price transmission in the period post 2001 gives us a somewhat bizarre inverse relationship between fuel price inflation and manufacturing inflation.

In short, manufactured product inflation seems to decline when fuel prices are revised upwards and vice versa. Thus, there is certainly no evidence of a "pass through" of the hike to other products.

This is of more than just academic interest. If this pattern indeed persists and second-round non-oil inflation remains in check, it could constitute a case against contractionary monetary policy. The price of oil will follow its own narrow supply-demand dynamic and remain high if concerns on supply disruptions remain.

All that central banks will end up doing if they keep hiking rates relentlessly is to squeeze company margins and the cost of credit for individuals wanting to borrow. The result could be a slowdown in growth without moderation in headline inflation.

There is another implication. If indeed two types of inflation, oil and non-oil, have de-coupled, the phase of high headline inflation (or potentially high inflation where retail prices of oil products have been suppressed) will come to an end only when oil prices abate. Central banks should remain passive and keep praying that oil prices cool off.

How central banks respond to this conundrum depends on how they explain these stable levels of core non-oil inflation. A popular hypothesis attributes the apparent stability of core inflation to increasing globalisation.

It goes thus-increased "trade openness" has increased price competition. This has led to moderation in both consumer and input prices. The resulting decline in pricing power has spurred manufacturers to innovate and enhance productivity, setting off a virtuous cycle.

The extreme view, following this line of thought, is that as integration continues we find ourselves in the middle of a phase of secular deflation. This could again be interpreted as a case against monetary tightening.

The Bank of England in its Quarterly Inflation Report (February 2006) presents a radically different perspective. It suggests that there could be a negative relationship between energy prices and core inflation.

In fact, high energy prices could actually be causing low core inflation. Thus, as energy prices abate, prices of non-energy products will begin to rise and keep headline inflation strong.

Thus, there is no reason for central banks to go easy on their vigil-in fact, their concerns about inflation should start compounding as oil prices show signs of weakness.

The British Central Bank's reasoning is simple. High oil prices reduce household real incomes and rein in their demand for non-oil products. This softness in demand restricts the increase in the prices of non-oil products.

My colleague, ABN-AMRO Bank's chief economist for Europe, puts it rather succinctly: "If energy prices keep rising, core inflation has to make room for it" ("This is hardcore," April 27, 2006).

Following the same logic, a drop in energy price inflation would raise real incomes and demand for non-oil inflation. Producers would rush to hike prices and make up for margin losses, particularly if the economy is in the middle of a cyclical upswing.

If the BoE model holds, a sustained decline in oil prices will not signal the end of high inflation. Declining oil prices will see the resurgence of "core" inflation. If the rise in non-oil inflation were strong enough, it could even mean a pick-up in headline inflation.

Central banks should be positioned for a hike in interest rates as soon as the oil markets start cooling off. Core inflation is incidentally known to be more sensitive to monetary signals than energy and commodity price signals. Monetary authorities who want to err on the side of caution may even want to pre-empt inflationary pressures by hiking
rates.

Central banks will have to choose a strategy that lies somewhere between the extreme alternatives that these characterisations of inflation present. There is enough evidence on the ground to show that productivity has improved substantially and has enabled producers to live with lower pricing power.

However, the sharp rise in commodity prices (both oil and metals) in the last six months has dented margins badly and it is unlikely that productivity gains alone can compensate.

Thus, downstream producers are likely to use the first opportunity that they get to hike prices and rescue their margins. There are some signs of incipient global core inflation pressures. Asian exporters, for instance, have stopped cutting prices. The bottom line is that the peril of rising non-oil inflation persists.

Finally, while supply concerns have played a critical role in determining oil prices in the short run, robust global demand conditions have done their bit in pushing up the average price-line. Tight monetary policy might just be able to put a lid on oil prices if it dampens the growth momentum.

The author is chief economist, ABN Amro. The views here are personal