THE PROBLEM

Dr. Rasmus Lentz in the Department of Economics at the University of Wisconsin does research in the fields of macroeconomics, microeconomics and labor economics, and teaches graduate and undergraduate labor courses.

To help with his research as a graduate student, Dr. Lentz decided to transition some of his economics projects and applications from high-level MATLAB and Gauss programming environments to Fortran, despite not being an expert in Fortran at the time. “I wanted to be able to write applications that executed faster than ones written in higher-level languages and Fortran is an excellent language for that,” said Dr. Lentz. At the same time, Dr. Lentz did not want to sacrifice the ability the higher-level languages gave him to quickly create prototypes and application code.

THE SOLUTION

As a graduate student at Northwestern University, Dr. Lentz believed the IMSL Fortran Library could help with his transition to Fortran while still enabling him to quickly develop prototypes and application code. “The IMSL Library is a great transition tool and helped me quickly move my projects and applications to Fortran; and that was without me having experience with Fortran at the time,” noted Dr. Lentz. “It was much faster for me to use procedures from the IMSL Fortran Library than create those same procedures on my own,” he added.

Today, several years later at the University of Wisconsin, the IMSL Fortran Library is still a useful tool, especially in the early stages of Dr. Lentz’s economics projects. “When I’m in the early development stages of models or applications, I can quickly make a call to the IMSL Library and put a prototype together to test out ideas without getting bogged down with minute Fortran programming details.”

For many of his research projects, especially in the labor economics area, Dr. Lentz requires solutions to problems that must be solved numerically. One example is work Dr. Lentz did on the impact of personal savings and earnings potential on the search intensity of an unemployed worker searching for new employment. The model developed for this work includes IMSL Fortran Library gradient-based routines and interpolation routines as well as procedures he developed on his own, and relates a worker’s observed savings, unemployment benefits, and earnings when employed to observed unemployment durations. The projects showed many results including the fact that wealthier individuals are observed to experience longer unemployment durations, shown in the model through a negative relationship between the choice of search intensity and savings. Sample data from the study, modeling wealth versus wage is shown below.

The results from studies like this example can be useful to both research and policy communities. The research community wants to understand data and explain what is happening in economics and the world so that economic theory is correct. The policy community can then use this data to determine the optimal design for programs such as government sponsored unemployment insurance.

RETURN ON INVESTMENT

For many economic problems that are solved mathematically, using computer programs for numerical solutions is important. ”Tools like the IMSL Fortran Library have a useful role to play,” said Dr. Lentz.