Optimization Of Fiscal Retrenchment Policies Using A Social Accounting Matrix Multiobjective Linear Programming Model: The Case Of New York State

Abstract

State finances continued to deteriorate over fiscal 2010, suffering from the cumulative impact of lower revenues, ballooning general fund spending and the general rise in the level of unemployment. The phasing out of the American Recovery and Reinvestment Act is likely to constrain even further the fiscal space of local governments throughout the country, potentially endangering a fragile economic recovery. In order to face rapidly rising budget gaps, forty states enacted mid-year budget cuts totaling $22 billion for FY 2010. The fiscal retrenchment approach to budget policy appears to gain ground amongst embattled states, with governors proposing drastic cuts in their fiscal 2011 executive budgets to meet balance-budget requirements. The legacy of the 1970s tax revolts, having made tax hikes difficult to enact, also contributed to corner states into spending reduction strategies. The study of cutback management has heavily focused on how and why reduction targets are adopted by state executives and legislators, as well as on their subsequent effects on local economies. The literature is however scarce on proposing a framework for efficient structuring of budget cuts at the state level. Stricken by uncertainty, many governors are driven toward across-the-board cuts, treating general fund expenditures as a fungible commodity. With states facing increasingly painful budgetary choices, weighting their implications and analyzing potential alternatives become critical to evaluate prospects for regional economic recoveries. In this dissertation, a Social Accounting Matrix (SAM) Multiobjective Linear Programming (MOLP) model is proposed and applied to the case of New York State. The SAM multipliers provide a powerful instrument to evaluate the short-term impact of austerity measures while linear programming (LP) offers an optimization framework to close efficiently the state's budget gap. Attention focuses on the existence of several conflicting objectives that the decision maker tries to optimize simultaneously. Four procedures are introduced to solve the model: the augmented weighted Tchebycheff method, an elistist genetic algorithm, the weighted sum method and constraint programming. The theoretical framework established in the following chapters as well as its application to the Deficit Reduction Plan proposed by Governor Paterson in fiscal 2009 show promising results. The model indeed converges to a set of Pareto optimal solutions that are by essence, more efficient with respect to growth, employment and labor income than the original plan. It constitutes one of the first practical applications of multiobjective optimization to policy design through a Walrasian general equilibrium framework.