This paper examines the relationship between structural equation modeling and the balanced scorecard in a banking institution. Using financial and non-financial data taken from the banking establishment, the paper propose a rational construction of Balanced Scorecard by choosing the right indicators for the right axes. This choice is made by implementing the Partial Least Squares (PLS) in our model. In addition, the scheme gives us the cause-and-effect chain, the one described by Kaplan and Norton as: measures of organizational learning and growth; measures of internal business processes; measures of the customer perspective; financial measures. We will observe that the Kaplan and Norton's model of BSC is nothing but a particular case of our findings. Following this we propose a general approach to build step-by-step a rational Balanced Scorecard and we will apply this modeling to a special case in the banking sector.