How did you become a forecaster?
After completing my undergraduate degree in physics, I found econometrics to be an interesting field to apply my model building, math, and computer skills toward solving real world problems. My experience with forecasting started with my dissertation back in 2003, which focused on the forecasting performance of international monetary models.
Soon after graduation, I joined Washington state government, and produced state medical expenditure forecasts for almost a decade. In 2015, I switched my concentration from medical expenditures to medical caseloads after joining the Washington State Caseload Forecast Council (CFC). The CFC produces caseload forecasts for state programs that comprise a large portion of the Washington State budget, in areas such as public education, criminal justice, public assistance, and medical services. The forecasts are based on data extracted from large state administrative databases, with a quick turnaround required to meet a variety of stakeholders’ needs. Beyond producing sophisticated forecasting models, we facilitate a collaborative workgroup process among budget analysts and state program experts across different fields to produce the best possible results. During this period, I also expanded my forecasting scope to many other social service programs.
What areas of forecasting interest you?
In my experience, two areas related to forecasting are important and interesting. The first is a data driven, comprehensive, module-based, and automatic forecasting infrastructure that connects administrative databases and various analytical software to timely produce data analyses and forecasts. Data analyses are important to improve the forecast performance and build stakeholder’s confidence in the forecast process. At the 2018 International Symposium on Forecasting, I presented the forecasting application I created and use at the CFC, which incorporates many functions including forecasting, reporting, and data analysis in an integrated system.
The second is to build sophisticated forecasting models and related forecast tracking processes. Currently the forecasting method utilized in many caseload forecasting processes is a system dynamics approach using multivariate time series models, and we have been continuously researching better methods by adding more forecasting modules to improve forecast performance.
How has the International Institute of Forecasters influenced you?
There exist different challenges in the practical world and academic world, and the IIF acts as a bridge to connect these two. CFC staff are encouraged to attend the IIF annual meetings and make presentations to share our forecasting practice and learn from other IIF members. Participation at these conferences expands our knowledge scope and benefits our work in the long run.
What do you do in your free time?
I love sports, especially watching and playing soccer. I routinely went to the gym before the COVID-19 pandemic, and sometimes I play table tennis and hike.