In the course of adding demographics we’ll also be able to incorporate immigration into the model. Immigration will be treated as exogenous. The model will be set up so that users can input the number and age structure of immigrations in current and future years. In setting up the model in this manner, we’ll be able to consider the potential for immigration policy to improve or exacerbate Social Security’s long-term financial condition comments.
Permitting the Model to Start from an Arbitrary Set of Economic Conditions In addition to incorporating demographics, a key improvement to the model now underway is permitting it to begin a simulation from an arbitrary set of economic circumstances. At the moment the model solves for an initial steady state and starts policy-reform simulations from that steady state. This is unsatisfactory because there is no reason to think the U.S. economy is currently in a steady state, either with respect to its demographics or its economic fundamentals.
The reason the A-К Model was originally designed to start in a steady state is that the assumptions that the economy is initially in a steady state permit one to calculate the distribution of wealth across and within cohorts. The distribution of wealth provides the critical state variables needed to start any simulation. Abandoning the assumption that the economy is initially in a steady state means that one needs to collect data on the intra- and intercohort wealth distribution to provide the starting values of these state variables. We plan to collect this wealth distribution data from the Survey of Consumer Finances (SCF) of the early 1990s. The challenge here will be to find correlates with lifetime labor income (such as education and current labor earnings) since the SCF doesn’t permit us to classify wealth by lifetime labor income.
Incorporating Dynamic Programming and Liquidity Constraints
A third improvement on which we’re working is to introduce a dynamic programming method of solving for agents’ annual levels of consumption, labor supply, and bequests. As we’ve added more detail to the model to emulate more closely U.S. fiscal institutions, we’ve stretched our ability to use first-order conditions and shadow prices to solve agents utility maximization problems. The reason is that U.S. fiscal institutions produce highly kinked budget constraints.