12:00-12:45

Zhaokun Zhang (VU University Amsterdam)
Weighted Maximum Likelihood Estimation with Optimal Forecasting Accuracy

This paper derives a novel weighted maximum likelihood (WML) estimator for autoregressive models that provides optimal forecasting accuracy. The WML estimator reduces to the classical ML estimator when the weights are uniformly set to one. We show that the improvement in forecast accuracy occurs because the model estimated by the WML method provides a better Kulback-Leibler approximation to the true data generating process. We devise a testing procedure that ensures that can guide the practitioner in the weight selection. Furthermore, we derive the asymptotic distribution of the test statistic and analyze its power and size. A Monte Carlo study reveals how the WML estimator can significantly improve the forecasting accuracy of autoregressive models. Applications to real data document the usefulness of this new methodology in empirically relevant settings.

Field: Econometrics

******

13:15-14:00 (Different time!) 

Simas Kučinskas (VU University Amsterdam) will present:
Capital Requirements and Liquidity Creation

Do high capital requirements destroy liquidity creation? I shed some light on this question by exploring welfare implications of an extreme form of capital regulation: 100% equity-financed narrow banks. Narrow banking is constrained inefficient because narrow banks create liquidity by, in effect, taxing investment. In a numerical example, the welfare cost of narrow banking can reach 1% of consumption. However, narrow banks also enhance financial stability. Narrow banks welfare-dominate standard fractional reserve banks whenever crises are sufficiently frequent or when consumers can engage in hidden trades. Overall, distortionary effects of high capital requirements may be offset by greater financial stability.

Field: Macroeconomics, Finance