Saturday, December 4, 2010

Keep your shirt on, life could be worse

Oh! No! The economy was roaring along in June quarter, growing by 1.1 per cent, but now it has almost come to a halt, up just 0.2 per cent in the September quarter. What's more, take out a leap in rural production and we actually went backwards.

It made a great story this week - thrills and spills in econoland - but I wouldn't believe it. Why not? Because in real life economies don't soar and dive in the space of six months without there being a very big and obvious reason - the introduction of the goods and services tax, for instance, or the collapse of Lehman Brothers and its aftermath scaring the pants off businesses and consumers.

You need to know that both the financial markets and the media have a vested interest in statistical volatility.

It gives them something to bet on or write stories about and makes their lives more interesting. So it suits them to take economic statistics literally, ignoring their well-known limitations.

Sensible people, however, always take them with a grain of salt, knowing the economy is far more stable than the stats - especially quarter-to-quarter changes - show it to be.

The making of the ''national accounts'' - the bottom line of which is gross domestic product - is like the making of sausages: you're better off not knowing what goes into them. They're pulled together using bits and pieces from thousands of different sources.

Often, inferior sources are used because the more reliable information isn't yet available. Sometimes no information is available, so the statisticians take a guess. When the better information does come along, the figures are changed. Since the better data come along at different times, the figures for a particular quarter are constantly being changed, for at least the next two years.

The original figure for growth in the December quarter of 2008 - the quarter when Lehman Brothers collapsed - was minus 0.5 per cent. It was then revised down each quarter until it reached minus 0.9 per cent. Then it was revised up each quarter, reaching minus 0.7 per cent three months ago.

This week it was revised down to minus 1 per cent. So we're now being told the contraction was twice the size we were originally told. And there were people at the time imagining that figure had been written by God on tablets of stone.

Or, let's try another one. Three months ago we were told real GDP grew by 3.3 per cent over the year to June. Now we're told it grew by 2.7 per cent over the year to September.

Why the sudden slowdown? Well, not primarily because of the alleged virtual cessation of growth in the September quarter, but because revisions shifted 0.4 percentage points of growth out of the December quarter of 2009 and into the September quarter of 2009 (which dropped out of the annual calculation).

The Bureau of Statistics acknowledges the ropiness of its figures, which is why it tries to direct users to its ''trend estimates'', which simply average out the quarterly ups and downs. But for good reasons and bad, economists, the markets and the media invariably ignore the trend figures and focus on the more volatile unsmoothed ones.

The point is that much of the quarter-to-quarter volatility in the growth figures isn't real but just ''statistical noise''. You have to ignore the noise to hear the true ''signal'' underneath it.

Remember, too, it's easy to be bamboozled by quarterly changes. If some big transaction is accidentally put into the wrong quarter, this distorts the quarterly change for three successive quarters.

Because a big thing such as a national economy - or an ocean liner - is actually quite hard to speed up or slow down, when the figures show it rapidly speeding up in one quarter, the greatest likelihood is that the figures for the following quarter will show it rapidly slowing down.

And that's just what the past two quarters' figures show. Logical deduction: the economy didn't really grow that fast in the June quarter and didn't really slow that much in the September quarter.

There's an old trick Treasury used to reduce the statistical noise and get a clearer signal: add the last two quarters together and take an average. That says the economy has probably been growing at a quarterly rate of about 0.65 per cent over the past six months ([1.1 + 0.2] ÷ 2).

That makes more sense, but even it seems too low. How can I say that? Because we have an independent (and less volatile) set of stats to measure the national accounts against: the employment figures.

These show employment growing fairly steadily over the past year, growing particularly strongly in the September quarter and increasing by 3.2 per cent for the year. That's not an economy that's suddenly run out of juice.

So when we peer through the statistical haze, what do we see in the national accounts? First, we see that the pick-up in business investment spending - particularly in the mining sector - is occurring, in line with what the companies have long been telling us about their plans for huge spending over the coming year and longer.

Second, despite strong growth in household disposable income (fed by strong growth in employment and rising real wages), consumer spending isn't growing nearly as strongly, meaning households are saving a lot more. (The figures say the household saving rate was 10 per cent of disposable income - which is too high to believe, but undoubtedly saving is high.)

This is bad news for retailers but good news for the economy generally because it postpones the time when, with the economy nearing full employment, the economic managers are struggling to cope with a massive mining investment boom and a consumption boom.

The way they'll cope with a double boom is simple: they'll jack up interest rates (which will also add upward pressure to the exchange rate) to discourage consumer spending. So the longer households keep thinking now's a good time to get on top of their debts, the better off we'll be. The bad news in the accounts, however, is the continuing weakness in the building of new homes and also in commercial (as opposed to industrial) construction.

This suggests inadequate supply will soon be pushing up rents and thus increasing inflation pressure.

So a literal reading of this week's accounts sends us just the opposite message to the true position.