C5 - Econometric Modeling
-
-
Generalized Autoregressive Gamma Processes
We introduce generalized autoregressive gamma (GARG) processes, a class of autoregressive and moving-average processes in which each conditional moment dynamic is driven by a different and identifiable moving average of the variable of interest. We show that using GARG processes reduces pricing errors by substantially more than using existing autoregressive gamma processes does. -
Turning Words into Numbers: Measuring News Media Coverage of Shortages
We develop high-frequency, news-based indicators using natural language processing methods to analyze news media texts. Our indicators track both supply (raw, intermediate and final goods) and labour shortages over time. They also provide weekly time-varying topic narratives about various types of shortages. -
Climate Variability and International Trade
This paper quantifies the impact of hurricanes on seaborne international trade to the United States. Matching the timing of hurricane–trade route intersections with monthly U.S. port-level trade data, we isolate the unanticipated effects of a hurricane hitting a trade route using two separate identification schemes: an event study and a local projection. -
Risk Amplification Macro Model (RAMM)
The Risk Amplification Macro Model (RAMM) is a new nonlinear two-country dynamic model that captures rare but severe adverse shocks. The RAMM can be used to assess the financial stability implications of both domestic and foreign-originated risk scenarios. -
Forecasting Banks’ Corporate Loan Losses Under Stress: A New Corporate Default Model
We present a new corporate default model, one of the building blocks of the Bank of Canada’s bank stress-testing infrastructure. The model is used to forecast corporate loan losses of the Canadian banking sector under stress. -
Calculating Effective Degrees of Freedom for Forecast Combinations and Ensemble Models
This paper derives a calculation for the effective degrees of freedom of a forecast combination under a set of general conditions for linear models. Computing effective degrees of freedom shows that the complexity cost of a forecast combination is driven by the parameters in the weighting scheme and the weighted average of parameters in the auxiliary models. -
Sectoral Uncertainty
We propose a new empirical framework that jointly decomposes the conditional variance of economic time series into a common and a sector-specific uncertainty component. We apply our framework to a disaggregated industrial production series for the US economy. We identify unexpected changes in durable goods uncertainty as drivers of downturns, while unexpected hikes in non-durable goods uncertainty are expansionary. -
Comparison of Bayesian and Sample Theory Parametric and Semiparametric Binary Response Models
We use graphic processing unit computing to compare Bayesian and sample theory semiparametric binary response models. Our findings show that optimal bandwidth does not outperform regular bandwidth in binary semiparametric models. -
Nonparametric Identification of Incomplete Information Discrete Games with Non-equilibrium Behaviors
This paper jointly relaxes two assumptions in the literature that estimates games. These two assumptions are the parametric restriction on the model primitives and the restriction of equilibrium behaviors. Without imposing the above two assumptions, this paper identifies the primitives of the game.