I am an Assistant Professor in Econometrics at the University of Amsterdam. Before that I obtained my PhD at the Erasmus University Rotterdam.
My primary research area is applied time series econometrics. My research interests include financial econometrics, state space modeling and Bayesian econometrics. I have applied these methods to credit risk, asset allocation and forecasting macroeconomic data.
You can reach me at firstname.lastname@example.org.
Winner of research grant at Europlace Institute for Finance
Based on unique data we show that macro variables, the default rate and loss given default of bank loans share common cyclical components. The innovation in our model is the distinction between loans with either severe or mild losses. The variation in the proportion of these two types drives the cyclic behavior of the loss given default, and constitutes the links with the default rate and macro variables. These links vary according to loan and borrower characteristics. During downturns, the proportion of defaults with severe losses increases, but the distribution of losses conditional on their being mild or severe does not change. Though loans are monitored more closely than bonds and are more senior, the cyclical variation in their losses resembles those for bonds, albeit around a lower average level. This variation leads to an increase in the capital reserves required for loan portfolios.
We assess the predictive ability of 15 economic uncertainty measures in a real-time out-of-sample forecasting exercise for the quantiles of The Conference Board's coincident economic index and its components (industrial production, employment, personal income, and manufacturing and trade sales). The results show that the measures hold (real-time) predictive power for quantiles in the left tail. Because uncertainty measures are all proxies of an unobserved entity, we combine their information using principal component analysis. A large fraction of the variance of the uncertainty measures can be explained by two factors. First, a general economic uncertainty factor with a slight tilt toward financial conditions. Second, a consumer/media confidence index which remains elevated after recessions. Using a predictive regression model with the factors from the set of uncertainty measures yields more consistent gains compared to a model with an individual uncertainty measure. Further, although often better forecasts are obtained using the National Financial Conditions Index (NFCI), the uncertainty factor models are superior when forecasting employment and in general the uncertainty factors have predictive content that is complementary to the NFCI.
The relationship between excess returns and the dividend-price ratio is known to be unstable. However, there is no consensus on the type of instability. We investigate the consequences of different types of break processes for a long-term investor. Differences in parameter instability affect the long-term investor in particular, as misspecification errors are exacerbated as the investment horizon increases. The break process is inferred with a mixture innovation model which is estimated using Bayesian methodology. The advantages of this approach are that we can estimate the frequency and size of breaks, and allow for separate break processes for the regression coefficients and the error variances. The estimated parameters show quite some instability, though less than if we were to assume a model with continuous breaks. We show that assuming constant parameters can lead to enormous losses for the long-term investor, even if in reality the break probability is small. The costs of ignoring uncertainty regarding the instability are smaller, but non-negligible.
Keywords: Return predictability, parameter instability, mixture innovation model, long-term investing, Bayesian modeling
JEL classification: C11, C32, G11