Zero Lower Bound - Is it a problem in the Euro Area?

Diplomarbeit von Lars Protze

Abstract:

The case of Japan showed that the zero bound is a problem for the conduct of monetary policy that even nowadays has to be considered. For several years Japan experienced deflation and a short rate very close to zero leaving monetary policy almost helpless to boost economic activity. The same fears came up in America and Europe as economic performance deteriorated and nominal interest rates were lowered rapidly to stimulate the economy. However, lowering the interest rate to stimulate the economy is only possible when interest rates are above zero.

In this paper it shall be explored how optimal monetary policy is conducted with the constraint that interest rates cannot fall below zero and how large the risk to hit the bound is in the euro area. The first part is done in a New Keynesian model with sticky prices but flexible wages the second in an estimated model of the euro area.

The outline of the paper is as follows. In the next chapter an overview of the work on the zero bound and monetary policy is presented. Thereafter the New Keynesian model as it was presented by Eggertson and Woodford will be used to determine optimal policy. It will be shown that quantitative easing, as it was done by the Bank of Japan, is not an appropriate tool in the model surrounding to escape a deflation spiral and what should be done instead. It will be shown that credible commitment is able to overcome most of the distortions induced by the zero bound.

The central bank should commit itself to a target for the price level instead of a target for the rate of inflation. The optimal solution involves credible commitment to cause subsequent inflation when deflation vanishes. This management of expectation will help to escape a deflation spiral faster and causes lower welfare losses.

After treating the phenomenon in a model surrounding it shall be explored what the chances are to slide into that vicious circle if monetary policy follows a Taylor rule and how likely the zero bound is under different wage contracting specifications. This will be done in a small estimated euro area economy model.

It shall also be considered how the announcement of a positive inflation target well above zero may help to avoid the zero bound. This was done by the European Central Bank that changed its target from an inflation rate between zero and two to a rate below, but close to, two percent.

Finally the results will be discussed focussing on the assumptions that were made to derive them and what would change if these assumptions are not appropriate.

Inhaltsverzeichnis:

Text Sample:

Chapter 4.1.1.5, Results from other studies:

The graph shows the probability that the zero bound binds when the central bank follows a Taylor rule for different target values for the rate of inflation.

Hunton and Laxton used the Japan block of the MULTIMOD model of the IMF for evaluating the frequency of bind under a Taylor rule and different inflation targets. This model is a multi-economy macroeconomic model that the IMF uses for the World Economic Outlook. Black et al. used the Quarterly Projection Model of the Bank of Canada to derive their results (Graph 6: Authors graph).

Orphanides and Wieland used a small structural model of the U.S. economy to calculate the possibility that the zero bound binds. Finally Riefschneider and Williams employed the model of the U.S. economy used by the FRB to analyze how severe the zero bound is and how often it will be a constraint for the conduct of monetary policy.

They all find that setting a higher target will lower the probability that the zero bound binds substantially. All curves are downward sloped and approach zero quickly. At a target for the rate of inflation of two per cent the chances that the bound binds are nearly zero in the studies of Orphanides and Wieland as well as in the study of Black et al.

All these studies assumed that monetary policy follows a Taylor rule. This may not simulate the actual behaviour of a central bank when the zero bound becomes apparent. So the results displayed above may over or understate the actual probability that the bound binds when the central bank uses other instruments than just the short term rate or if the central bank does not follow a Taylor rule (for example when the bound becomes apparent and the central bank may decide to use a pre-emptive strike).

The analysis above showed results for the U.S. economy, the Canadian and the Japanese economy. The next subchapters show the distortions the zero bound causes in the euro area.

Distortions with a Taylor rule: Since the results of the studies that were presented above imply that the zero bound is a probable constraint for the conduct of monetary policy in a low inflation environment it has to be determined how large these distortions are. To do this a model of the European economy is used to derive the distortions when monetary authority follows a Taylor rule.

Inflation distortions: The zero bound leads to somewhat tighter monetary policy compared to what would occur in the absence of the bound. This is the case for inflation targets set close to zero because then the probability that the bound binds is larger. This tighter monetary policy leads to a downward bias in the rate of inflation as the graph below shows. The squares represent the outcome for Taylor style contracts and the diamonds the Fuhrer-Moore contracting (Graph 7).

The downward bias is stronger for the Fuhrer-Moore contracting. But as the probability to hit the bound decreases with larger inflation targets the bias decreases too at a larger inflation target. With Taylor contracting the bias is reduced to zero when the target is set to two per cent. However, with Fuhrer-Moore contracting is not even at a target rate of four per cent reduced to zero.

The standard deviation of the rate of inflation is higher for Taylor contracting than in the case when the zero bound is not a constraint while the standard deviation with Fuhrer-Moore contracting is reduced. And again at an inflation target of two per cent the change in the standard deviation with Taylor contracting vanishes.

Output distortions: As with inflation there is also a downward bias in the mean of output. With the zero bound output falls short of potential output by 0.1 percentage points with Taylor contracting at an inflation target set to zero. When the inflation target is set to two per cent it is almost reduced to zero. The distortions are somewhat stronger for the Fuhrer-Moore contracting specification. With 0.14 percentage points at an inflation target of zero it is not very large. However, it does not return to zero as quickly as under Taylor contracting when the inflation target is lifted. The downward bias occurs due to the tighter monetary policy when policy is constrained by the zero interest rate bound (Graph 8).

The standard deviation of output is higher for both wage contracting specifications. But the rise is not significant for Taylor contracting.

The graphs above show that the distortions can be lowered by raising the inflation target. With Taylor contracting a two per cent inflation target reduces the distortions almost to zero. With Fuhrer-Moore contracting the inflation target should be raised up to four per cent so that only minor disturbances are left. However, this recommendation is not very helpful for policy makers because inflation itself is costly and causes distortions in the economy which will lower welfare. The full benefits of price stability are achieved at zero inflation when the zero bound is not a constraint. The ECB decided to raise its target for the rate of inflation and the graphs above showed that this will lower the distortions of the bound significantly. But this is done by accepting the cost of higher inflation which have not been considered in the simulations. The main costs are price dispersion which leads to ineffective production. This lowers the level of output and lowers welfare through this.