One pillar or two?

In a pair of new papers, CEPR Research Fellow Michael Woodford attacks two widely-held tenets of current thinking about monetary policy - the idea that financial integration leaves central banks powerless in the face of global developments; and the insistence of the European Central Bank that money-growth matters in setting interest rates.

In August 2007, when it might have seemed that central banks were at the mercy of the financial turmoil sweeping through markets across the world, Mervyn King, the governor of the Bank of England, insisted he still had firm control of domestic inflation. 'Zimbabwe has demonstrated very clearly that if you want to have a high inflation rate, boy can you have one. And I could produce that here too if anybody was foolish enough to ask us to do it,' he said.

Woodford provides a firm theoretical footing for King's claim in the first of his two papers, by taking aim at the popular idea that globalisation has eroded the power of central banks; and inflation is no longer simply 'made at home'.

It has become commonplace to talk about 'global liquidity' and 'global slack,' as if central banks must now approach monetary policy-making in a completely different way. However, Woodford carries out a painstaking analysis of these concepts, using a New Keynesian model of the economy, and shows that they reflect a misunderstanding about the linkages between global developments, and domestic inflation.

He finds that globalisation does mean key relationships in the economy - between a given level of interest rates and the level of aggregate demand, for example - can be altered by outside events. That does not mean, however, that raising interest rates will suddenly make no difference, leaving central banks impotent in the face of global integration.

In the paper, Woodford models separately the effect of integrating goods markets, financial markets and labour markets on inflation, and the impact of changes in interest rates. Even when he assumes a much more radically integrated global economy than we have today, however, he finds no evidence that central bankers' powers will suddenly ebb away.

In fact, in some cases, the logic is the reverse of what the 'global liquidity' argument would suggest. For example, loose monetary policy abroad might tend to lead to a cheaper currency, cutting the costs of foreign goods, and eating into the market for domestic goods, and hence into domestic demand. So lower overseas interest rates might actually slow the domestic economy, rather than creating excess 'global liquidity'.

Woodford argues that monetary policy abroad might affect the level of demand for a given level of interest rates at home; but it doesn't mean that shifting interest rates is suddenly a useless tool. As King put it, 'does anyone believe that if we changed interest rates to 50% per cent it would have no noticeable effect on the economy?'

Woodford rides into another, more technical dispute about the conduct of contemporary monetary policy in a second recent paper. He examines the European Central Bank's controversial 'two-pillar' system of analysing economic developments - and dismisses it as unnecessary.

The ECB's first pillar is the standard economic analysis conducted by all inflation-targeting central banks; but it supplements this with a controversial 'second pillar,' which involves monitoring movements in the money supply.

In recent years, the ECB's approach has received backing from economists, including Stefan Gerlach, who argue that while in the short run the empirical relationship between money-growth and inflation is variable, in the very long run it always holds.

This leads to the argument that a single model cannot explain both the short-term fluctuations of inflation - determined by short-run factors such as changes in prices and wages - and the longer-term correlation between money-growth and inflation.

Gerlach suggests that this problem means central banks should use a 'money-augmented,' or 'two-pillar' Phillips curve to analyse inflation, instead of a single, unified approach. Estimates of the output gap are used to predict short-run inflation, while money-growth is used to forecast the long-run.

In his paper, however, Woodford attacks the empirical argument that the relationship between the money-supply and inflation is so robust over the long-term - even across centuries - that it gives central banks a strong enough reason to keep it under control, even if it has little explanatory power in the short-run.

Woodford argues that using a standard New Keynesian model of the economy, he can predict the long-run, or low-frequency relationship between money-growth and inflation - without having to assume there is any causal link involved. There is a stable relationship between money and demand in the economy, built into his model; but it is not a causal one.

This approach can explain many of the empirical findings - such as comparisons between long-run moving averages of money-growth and inflation - that have been used to support demands for a second monetary pillar. Woodford carries out a detailed series of economic simulations, based on his model, and generates long-run correlations between monetary growth similar to those found in empirical studies of the euro-area - without having to assume any causal link.

He argues that these results suggest that a single model - and one monetary pillar - is enough; and the ECB should simply abandon the second pillar, which it only adopted at the outset because monetary analysis was at the heart of the respected German Bundesbank's approach, and the fledgling bank in Frankfurt wanted to stress its hawkish credentials.

DP 6447 Does a 'Two-Pillar Philips Curve' Justify a Two-Pillar Monetary Policy Strategy

DP 6448 Globalisation and Monetary Control

Michael Woodford