All theories are simplifications of reality. All theories must, also, be potentially falsifiable. If not, they fail as theories: they become no more than acts of faith. Newtonian gravity was a simplification of reality that was ultimately falsified by Einstein (although, for the rest of us, Newton's theories still work very well most of the time). And Einstein's Theory of Relativity was, after the event, proven to be a good theory through careful measurement of developments in the cosmos. One day, no doubt, someone will come up with an even better theory than Einstein's, one that perhaps ties together Einstein's views of the very big with Heisenberg's views of the very small. We know, after all, that while the theories of Relativity and of Quantum Mechanics are excellent descriptions of some aspects of the universe, they are less-than-happy bedfellows.
In economics, theories are not so easy to falsify as in the natural sciences. Scientific method is used, of course, but the available evidence is often so limited that no one theory can easily be proved superior to others. This, in turn, leads to major problems for policymakers. Let's say that inflation is on the rise. How should this be interpreted? Is it demand-pull inflation, a sign that the economy is operating beyond full capacity and that a tightening of either monetary or fiscal policy is required? Is it cost-push inflation, an indication that the labour market is full of inefficiencies and might respond to either an incomes policy (as was tried in the UK in the 1970s) or deregulation (which, through legislation and a bit of brute force, worked well in the 1980s). Or is it the result of an external shock - higher energy prices, for example - that might threaten to raise inflationary expectations, thereby requiring higher interest rates as a form of anti-inflation "insurance"?
Throughout the 1980s and 1990s, inflation was typically thought to be driven by a mixture of cost-push and demand-pull factors. Standard economic models suggested that inflation was, directly, a function of wages, productivity, commodity prices and profit margins. What's more, these cost-plus-mark-up models typically assumed that the various components of the inflationary process were positively correlated. If commodity prices went up, for example, wages would go up. If profits went up, the economy was probably overheating, so wages would again go up. Underneath all of this was an assumption that any cost or demand shock had the ability to change our expectations about the future path for inflation and that we could, collectively, act on those changed expectations.
Sometimes, these models led to rather odd conclusions. Higher wages imply higher unit labour costs but equally so does lower productivity. Yet there are plenty of occasions when slowing productivity growth might say nothing at all about the likelihood of higher inflation. If an economy is too strong, and unemployment too low, it may be that additional hires suffer from diminishing marginal productivity. In this situation, slower productivity growth might be associated with higher demand-pull inflation. Equally, though, an economy that is slowing rapidly might also experience slower productivity growth: existing workers spend fewer hours at the factory, or produce fewer widgets per hour. In this situation, demand-pull pressures would seem to suggest that inflation would be falling, not rising.
A good example of the confusion created by differing interpretations of productivity data came last week. The fourth quarter US productivity data were a lot weaker than expected. According to the provisional estimates, output per hour worked fell 0.6 per cent at an annualised rate, well below the 1.5 per cent annualised gain expected on Wall Street. And because productivity was weaker, unit labour costs were a bit stronger, rising by 3.5 per cent on the quarter at an annualised rate. Add to this gain the persistent rises in oil prices and suddenly it's not so difficult to argue that we're returning to the economic conditions that led to runaway inflation in the 1970s. As Reuters put it: "Investors worried that higher labour costs would fuel inflation, leading the Fed to raise rates more aggressively than expected."
I think this simplistic view of inflation is completely wrong. The decisive change in recent years has been the breakdown in the relationships that once existed between the various ingredients that make up the standard inflationary process. Increases in commodity prices no longer imply increases in wages. Increases in wages no longer imply increases in profits. Increases in profits no longer imply higher inflation. And central banks have to be increasingly careful in gauging which of their hitherto reliable early warning indicators of higher inflation are still in any way useful.
Of the major shocks that have hit the global economy in recent years, the majority have disrupted the relationships between the various indicators of higher inflation. If companies outsource the production of, say, televisions to China, they're doing so because they believe that wage levels will come down. Those wage declines boost their profits, but also help Western consumers to buy televisions more cheaply than before. And because, in response, demand for televisions rises, demand for the raw materials that are required for television production must also rise. The outsourcing story therefore leads to lower wages but higher prices for raw materials. Which of these two - contradictory - trends should central banks worry about when assessing the outlook for inflation?
Or let's say that producers in the UK are open to much more international competition as a result of the technologies associated with globalisation. The sale of both goods and services over the internet means that barriers to foreign competition have, in many industries, been removed. Basic microeconomic theory suggests that companies increasingly become international "price-takers". The pricing power they once enjoyed fades away, so that any increase in domestic demand that forces through higher wages leads not to higher prices but, instead, to lower profits. America's experience in the late 1990s, before the stock market crash, fits this description perfectly.
Then think about a world of commodity price shocks where capital is mobile and central banks are credible. Higher oil prices squeeze profits, but companies know that central banks won't tolerate price increases. In response, rather than raising prices and triggering a wage-price spiral, companies deal with higher energy prices by demanding pay cuts and higher productivity from their workers: the workers can't easily object because capital mobility places companies in an unusually strong bargaining position.
And finally, as the Bank of England knows only too well, there's the impact of labour migration. Most economists believe that low unemployment threatens higher inflation: it is, after all, the key implication of the Phillips curve, particularly Friedman's expectations-augmented version. Yet the UK has enjoyed low unemployment for many years now, without any threat of higher wages or inflation. One reason is, of course, the Polish plumber: labour supply has become a lot more elastic than before, because labour supply is no longer limited to the constraints imposed by the indigenous population. Hence the Bank of England has been able to lower interest rates when energy prices have been rising, when unemployment has been low and when productivity has been poor.
All theories of inflation are falsifiable. To my mind, many of the theories of inflation that have dominated conventional economic wisdom for many years have now collapsed under the weight of accumulated evidence associated with globalisation and central bank independence. The problem, of course, lies in getting others to accept this view. The Bank of England has certainly been quick to recognise this new inflationary order. Other central banks, notably the one based in Frankfurt, seem a little slower off the mark.
Stephen King is managing director of economics at HSBCReuse content