Forecasting welfare caseloads: The case of the Japanese public assistance program

Abstract

Forecasting welfare caseloads has grown in importance in Japan because of their recent rapid increase. Given that the forecasting literature on welfare caseloads only focuses on US cases and utilizes limited classes of forecasting models, this study employs multiple alternative methods in order to forecast Japanese welfare caseloads and compare forecasting performances. In pseudo real-time forecasting, VAR and forecast combinations tend to outperform the other methods investigated. In real-time forecasting, however, a simple version of forecast combinations seems to perform better than the remaining models, predicting that welfare caseloads in Japan will surpass 1.7 million by the beginning of 2016, an approximately 20% increase in five years from the beginning of 2011.

Comparison of Forecasting Performance with VAR vs. ARIMA Models Using Economic Variables of Bangladesh

“…Masayoshi Hayashi [33] emphasized the VAR model for forecasting welfare caseloads in Japan; found that VAR and forecast combinations tend to outperform the other methods of investigation. Forecasting women, infants, and children caseloads, the VAR model predicted fewer errors compared to ARIMA models [34]. Based on this research [34], the cost savings to using the VAR model were high.…”

“…Forecasting women, infants, and children caseloads, the VAR model predicted fewer errors compared to ARIMA models [34]. Based on this research [34], the cost savings to using the VAR model were high. If the costs of estimating these models were high, this may not represent a large cost saving.…”

Discontinuous or Kinked Response? Dynamics of Unemployment Insurance and Public Assistance under the COVID-19 Crisis

“…As in Figure 1, the left graphs show both the observed and forecasted PA caseloads, and the sharp intervention timing. Note that Suzuki and Zhou (2007) and Hayashi (2014) examine the monthly-level trends of PA caseloads in the 2000s and before using more elaborated time-series techniques, but our simple linear time trend model works well as is shown in Section 4.…”

Economic Downturns, Endogenous Government Policy and Welfare Caseloads

“…Zolotoy and Sherman (2009) implemented a two-step latent factor approach to model welfare caseloads. More recently, Hayashi (2014) proposed to use a variety of fore casting models including ARIMA, exponential smoothing, Markov forecasting, logistic smooth threshold autoregression models, VAR, and variations of forecast combinations. Ac cording to his results, pseudo real-time forecasting, VAR and forecast combinations tend to outperform the other methods.…”

“…An intensive literature has examined the relative importance of the different factors in explaining caseload changes (CEA, 1997;Figlio and Ziliak, 1999;Moffitt, 1999a;MaCurdy et al, 2000;Blank, 2001;Wallace and Blank, 1999;Ziliak et al, 2000;Grogger et al, 2003;Grogger, 2004;Page et al, 2004;Haider et al, 2004;Ayala and Pérez, 2005;Looney, 2005;Danielson and Klerman, 2008;Hayashi, 2014). Most of this research concludes that lower unemployment rates are important determinants of the caseload declines, but changes in the programs and other policies are also relevant.…”