Input design is the process of designing informative inputs which maximizes the information about system dynamics using minimal cost and resources in a system identification experiment. In this work, we focus on designing inputs based on the economic framework for identification of constrained processes. The process is nominally constrained and feasibility during identification is ensured by ''backing off'' (or moving away) from the active constraints. The economic cost or penalty due to the ''back-off'' is minimized while ensuring feasibility and satisfying constraints on the quality of the identified model. The proposed economic based formulation results in a nonconvex optimization problem. This problem can be solved using the bisection algorithm for single input-single output (SISO) systems. We present a two-stage iterative solution algorithm to solve for multi-input-multi-output (MIMO) systems where a convex optimization problem is solved at each stage. The optimal input is realized as a white noise passing through an M-tap filter. The filter coefficients are obtained by a noniterative spectral factorization technique. The efficacy of the proposed approach has been successfully demonstrated in a mass spring damper system as well as in a Van der Vusse reactor system.