Capital constraints, bank risk-taking and monetary policy

In this thesis, the New Keynesian model is tested, extended and estimated using Chinese data. In particular, the New Keynesian Phillips Curve is tested for the weak identification problem. The empirical results reveal that both forward- and backward-looking behaviour exist in the pattern of Chinese inflation. The forward behaviour is quantitatively larger than the backward looking behaviour, but the latter is not negligible in the quantitative sense. We empirically test whether China was able to achieve monetary policy autonomy under capital controls by estimating a Vector Error Correction Model (VECM), and theoretically examine the possible consequences of capital controls by constructing an open economy New Keynesian model with capital controls. It is found that China has indeed achieved monetary policy autonomy of under a partially opened capital account and an exchange rates regime, in general, pegged to the US dollar. In the modelling section, the appropriateness and consequences of capital controls are examined using an open economy New Keynesian model. Capital controls transform dramatic and immediate changes in deposit portfolios into a prolonged and gradual adjusting process. The appropriateness of capital controls depends largely on the sources of shocks. Capital controls can reduce output fluctuations under a foreign demand shock, but can only increase the volatility of the economy under a foreign interest rate and an inflation shock. We build an open economy model with a banking sector and estimate it using Chinese data in order to analyse the effects of the monetary and macroprudential policies. Banks in the model can increase their leverage ratios quickly through greater reliance on short-term financing sources. Four different policy sets, including different combinations of a standard Taylor rule, a loan-to-value ratio extended Taylor rule, and a macroprudential tool, are assessed based on a central bank loss function. It is found that the extended Taylor rule has no dominant role among the four combinations of policies. The extended Taylor rule with macroprudential policy has lower losses compared with other policy sets.