MEASURING ECONOMIES OF SCALE AND SCOPE IN AGRICULTURAL BANKING

The efficient size of agricultural banks is an issue that will remain important for the rest of the 1990s. Consolidation that is occurring i n the financial services industry has spilled over into agricultural b anking. This has raised concerns as to whether banks that have traditi onally supplied credit to production agriculture will continue to do s o after consolidation. Another major concern is whether consolidation is motivated by greater efficiency, or whether consolidation is result ing in greater market power without achieving cost savings. Study of t he production technology of financial institutions can determine wheth er and to what degree economies of size exist and how agricultural len ding will fit into the overall business plans of consolidated banks. G enerally, empirical studies have used either duality theory with the e stimation of cost functions or nonparametric estimation methods to ass ess efficiency in the financial services industry. The purpose of the present study is to estimate economies of scope and scale in agricultu ral banking. This will be accomplished through estimation of an indire ct multi-product cost function. The uniqueness of this study, when com pared to previous studies of efficiency of the financial services indu stry, i's the disaggregation of the outputs so that agricultural lendi ng can be evaluated. In addition, a normalized quadratic cost function is estimated with curvature properties imposed. Clark reviewed thirte en studies that measured economies of scope and scale for commercial b anks, credit unions, and savings and loan associations. He found that they offered four broad conclusions: (a) overall economies of scale ex ist at low levels of input, (b) no consistent evidence of economies of scope exists, (c) some evidence of cost complementarities exists, and (d) the results seem to be robust among financial institutions. Humph ery also reviewed studies that examine the issue of economies of scale for banks and found that little cost savings exist for increases in s ize alone. He concluded that significant benefits accrue from loan div ersification. Humphery also reported that the differences in cost stru cture within the same firm size category are larger than measured cost economies. Berger, Hunter, and Timme reviewed the efficiency of finan cial institutions. They discovered that many of the previous studies s howed the average cost curve to be relatively flat. They also suggeste d that differences in technologies may occur between small and large b anks or that some factor that varies with bank size may be excluded fr om previous studies. McAllister and McManus found that the translog co st function specification gives a poor approximation when applied to a ll bank sizes, suggesting that nonparametric estimation procedures sho uld be examined. In addition, they included risk as a cost factor in t he estimation. Berger, Hunter, and Timme suggested that, although fron tier estimation methods are theoretically correct, studies that have c ompared results of frontier methods to those of more traditional estim ation methods have found only small differences for scale measures.