Stephen King: Fed fiddles with the thermostat but needs to see bigger picture

Ben Bernanke has chosen to reject the risk management approach of Alan Greenspan

Ben Bernanke, the chairman of the Federal Reserve, recently argued that US monetary policy decisions will from now on be "data-dependent". In Congressional testimony to the Joint Economic Committee at the end of April (to quote just one example), Mr Bernanke said: "Future policy actions will be increasingly dependent on the evolution of the economic outlook, as reflected in the incoming data. Specifically, policy will respond to arriving information that affects the [Federal Open Markets] Committee's assessment of the medium-term risks to its objectives of price stability and maximum sustainable employment".

Ever since, markets have been a bit more jittery than usual. That's hardly surprising. Economic data are often very volatile, up one minute and down the next. One month, there's a stronger than expected inflation number: in response, bond yields rise and equities fall in anticipation of higher policy rates. The next month, the inflation rate is a little lower than expected and markets then head off in exactly the opposite direction.

The problem, of course, lies in knowing what the Fed thinks about each data release. How much weight will the Fed place on a piece of data which might either be pointing to a new medium-term risk or, instead, be nothing more than a bit of rogue information that carries absolutely no implications for the future?

In a trivial sense, policymakers are always data-dependent. They would be fools to ignore economic data altogether. It seems to me, though, that central banks should be able to stand back from the daily market noise and make a dispassionate assessment of underlying trends. This can then be formulated into a reasonably clear message for financial markets to absorb. Data-dependency may be a transparent approach, but the absence of an underlying strategy creates the impression of a policy that's all at sea.

I don't believe for a minute that the Fed really is all at sea, but the emphasis on data-dependency is certainly odd. Knowing that someone is data-dependent is not enough to know whether he or she will systematically make the right decision or, at least, avoid making the wrong decision.

Let me give you two, weather-related, examples.

When I was at university, I knew a student who was entirely data-dependent in his choice of clothing. During the summer, he would insist on wearing a heavy overcoat. Through the winter, he would stroll around in a T-shirt and jeans. There's no doubt that his choices were data-dependent. If I recall correctly, he wanted to demonstrate his alternative lifestyle. At the time, I was quietly impressed with his supposed acts of rebellion but now realise that his data-dependent choices, betraying an attention-seeking weakness, were just idiotic.

I now want you to imagine someone newly arrived in Britain who, unable to read or speak English, cannot understand the weather forecasts. To stretch this example a little - perhaps beyond breaking point - this person lives in an air-conditioned, windowless room and, on getting up each morning, has no idea of what the weather is like outside. The only information our intrepid visitor has is yesterday's weather. Sometimes, yesterday's weather is a good guide to today's weather. But this is Britain, so there will be frequent mistakes. On some occasions, our visitor returns to his windowless room shivering uncontrollably, soaked to the skin in T-shirt, shorts and flip-flops; on other occasions, he comes back sweating profusely, having worn a three-piece woollen suit and overcoat in 35 degree sweltering heat.

In both of my examples, people are behaving in a data-dependent fashion. But they get things badly wrong, either through gross stupidity or through insufficient information. Data-dependency, on its own, is no guarantee of success. It's no surprise, therefore, that investors are feeling a little uneasy.

The Federal Reserve, like other central banks, is a bit like my visitor living in the windowless room. The data it sees refers to the past, not to the present or to the future. Each piece of data provides only a partial view of what's happening in a world that has already passed us by. And the Federal Reserve has to look at each of these pieces of data and assemble them into a coherent view of the world. Put this way, our central bankers face the same difficulties as children when trying to complete a jigsaw puzzle.

What's more, unlike my visitor to Britain, who simply has to accept the weather, come rain or shine, the Federal Reserve has the ability to change the economic climate. It has control of the thermostat that determines whether the economy is too hot or too cold.

Having control in a data-dependent world is not, though, a guarantee of success. The Fed alters the settings on the thermostat based on conditions a couple of months ago but the changed settings will only have an impact on the economy 12 months down the road. Living with a data-dependent central bank is a bit like being in one of those hotel showers where you simply can't get the hang of the thermostat: you turn the temperature up because it's a bit too cold and a few moments later you're being scalded with hot water.

What does a data-dependent central bank need to make itself credible? It helps, first of all, to have a clearly stated objective. It also helps if there's a single objective because, after all, there's only a single instrument - the rate of interest - and, with only one instrument, two objectives become a little tricky, even if you are good at multi-tasking.

The Federal Reserve, though, has two objectives - price stability and maximum sustainable employment - so faces potentially awkward choices: if both inflation and unemployment rise at the same time, life becomes a little complicated.

Other central banks do have single objectives - typically in the form of an inflation target - and seem to have had fewer problems recently in getting their views across to the markets. The data may be misleading but at least the markets know, roughly, what the central banks are trying to achieve over the medium term (although I note that the interpretation of inflation targets is becoming more and more stretched).

In his defence, Mr Bernanke would doubtless argue that the Federal Reserve has an informal inflation objective, the so-called 1-2 per cent "comfort zone" for annual gain in the personal consumers' expenditure deflator, excluding food and energy prices (see chart).

This "zone", though, is a very informal affair and, in any case, has probably not been the best gauge of inflationary pressures in recent times. By excluding persistent increases in energy prices, the Federal Reserve has, in effect, assumed away the inflation that, in reality, is burning a hole in people's pockets (both the European Central Bank and Bank of England include energy prices in their inflation objectives).

Mr Bernanke has chosen to reject the risk management approach that was favoured by Alan Greenspan. No longer does the Chairman of the Federal Reserve want to be a monetary mystic. His bid for transparency, however, is in danger of foundering on the rocks of data-dependency.

It's not good enough for a central bank to be data-dependent: it also needs a clear objective and a slowly evolving view - a testable hypothesis, no less - of how to get there. Without these, markets will continue to move randomly, increasingly concerned that the global economy is looking like a ship without a rudder.

Stephen King is managing director of economics at HSBC

stephen.king@hsbcib.com