(Carol Clements, Nicholas Pohlman, Scott Seibel, 1997)
The Federal Reserve Bank of Dallas currently has no way to tell what characteristics of a bank make it conducive to forming a successful merger. Further, they are unable to predict what types of mergers will be successful. It has been our goal to develop a model that will determine which types of mergers will be most successful, and what characteristics of a bank promote that success.
The approach we used in developing our model entailed a five-step process. First, we created a merger tree to order and simplify the large data set. Next, we classified each of the mergers by the number of banks acquired. From these categories, we extracted a small sample on which we later performed in-depth analysis. In addition, we verified the data to determine the presence of geographical overlap. Finally, we used the software package, DEA, to produce efficiency ratings based on this data both before and after the merger occurred.
The technical section of our project included computing the improvement ratios of each bank over time. Using these ratios as the principal method of comparison for the banks, we performed extensive numerical and graphical analyses on the data. From these analyses we determined the most significant factors on each bank, the ideal number of banks involved, and the overall optimal model for a successful bank merger.