Michael Magura

Using interindustry input-output relations as a Bayesian prior in employment forecasting models

Abstract This paper presents the results of using input-output tables as a source of Bayesian prior information in a national employment forecasting model. A Bayesian vector autoregressive (BVAR) estimation technique is used to incorporate the interindustry input-output table relationships into the labor market forecasting model. This technique requires that a simple translation of the direct use coefficients from the input-output table be used as prior weighting elements to depict the interindustry relations. The Bayesian model provides out-of-sample forecasts superior to those from unconstrained vector autoregressive, univariate autoregressive, a block recursive bvar model and a naive BVAR model based on the Minnesota random walk prior. This suggests that interindustry input-output table linkages provide useful information that can be effectively incorporated into labor market forecasting models.

A Mixture-Model Approach to Combining Forecasts

A multiprocess mixture-model approach to combining forecasts from alternative sources is proposed. This approach extends the Granger–Ramanathan method by allowing the weights used in producing the combination forecast to vary over time. In addition, the procedure discounts outlying data points that arise during time periods when all of the competing forecasts miss the mark. An empirical comparison with traditional and more recently proposed combination methods demonstrates that the proposed methodology outperforms these.

Using Bayesian Techniques for Data Pooling in Regional Payroll Forecasting

This article adapts to the regional level a multicountry technique recently used by Garcia-Ferrer, Highfield, Palm, and Zellner (1987) and extended by Zellner and Hong (1987) to forecast the growth rates in gross national product across nine countries. This forecasting methodology is applied to the regional level by modeling payroll formation in seven Ohio metropolitan areas. We compare the forecasting performance of these procedures with that of a ridge estimator and find that the ridge estimator produces forecasts equal to or better than those from the newly proposed estimators. We conclude that the ridge estimator, which does not reference the pooled data information introduced by the newly proposed techniques, may serve as a benchmark against which to judge the relative importance of this kind of information in improving forecasts.

The black dropout rate and the black youth unemployment rate: A Granger-causal analysis

The high unemployment rate of black high school graduates can create the perception that a diploma is of little value and encourage dropping-out of school. Black youth who do drop out are less employable and further push up the black youth unemployment rate. This raises a question: Is the high dropout rate of black youth due to their high unemployment rate or is their high unemployment rate due to their high dropout rate? A study of this question using the definition of Granger-causality finds that it is the high unemployment rate which causes the high dropout rate rather than the opposite.