¹ The following are the most important studies on this topic. Davis, Lance E., “The Investment Market, 1870–1914: The Evolution of a National Market,” this Journal, 25 (09 1965), 355–99.Google ScholarSylla, Richard, “Federal Policy, Banking Market Structure, and Capital Mobilization in the United States, 1863–1913,” this Journal, 29 (12 1969), 657–86.Google ScholarSmiley, Gene, “Interest Rate Movement in the United States, 1888–1913,” this Journal, 35 (09 1975), 591–620.Google ScholarJames, John A., “Banking Market Structure, Risk, and the Pattern of Local Interest Rates in the United States, 1893–1911,” Review of Economics and Statistics, 58 (11 1976), 453–62;Google Scholar“The Development of the National Money Market, 1893–1911,” this Journal, 36 (12 1976), 898–907; andGoogle ScholarMoney and Capital Markets in Posibellum America (Princeton, 1978).Google Scholar

² For these rates see Davis, “The Investment Market,” Smiley, “Interest Rate Movement,” and James, “Banking Market Structure,” “The Development,” and Money and Capital Markets.Google Scholar

³ Smiley, Gene, “State and City Loan Rate Estimates, 1916–1940,” Working Paper, Marquette University Department of Economics, Milwaukee, Wisconsin, 1981. This paper is available from the author. Country and city rates could not be estimated separately in 1938, 1939, and 1940. In those three years each state rate is an average for all country and city national banks of each state.Google Scholar

⁴ In the 1930–1933 period numerous banks disappeared due to merger, suspension or failure. To match the number of banks reporting earnings at the end of a six-month period and the number of banks reporting in the balance sheet at call dates, the loan base was adjusted. With loans lagged six months some banks reporting loans (and other assets and liabilities) did not exist when the earnings were reported. Estimates of loans per bank at call dates were constructed and total loans (for a state or city) were decreased by the banks which did not report earnings. The correction of the loan base by subtracting loans of banks not reporting earnings was adjusted for the size of the nonreporting banks. With the aid of data on capital and surplus, the loans subtracted off were decreased or increased depending on whether the capital and surplus of the failed banks were smaller or larger than the average capital and surplus per bank in the state or city. In cases where it was obvious that the decreased number of banks was due to bank merger (mainly in the reserve cities) no loan adjustments were made. These adjustments were not necessary after 1933–2 (with a few isolated exceptions) since nearly all of the bank failures occurred from 1930 through 1933.Google Scholar

⁵ Board of Governors of the Federal Reserve System, Banking and Monetary Statistics (Washington, 1943), p. 426.Google Scholar

⁶ The cities and states in the regions used in Tables 1 and 2 are as follows. Note that rates for the certral reserve city banks of New York City (NYCY) and Chicago (CHIC) are reported separately in Table 1. NEW ENGLAND (NED). States: Maine, Vermont, New Hampshire, Massachusetts, Rhode Island, and Connecticut. Cities: Boston. MIDDLE ATLANTIC (MAC). States: New York, New Jersey, Pennsylvania, Delaware, Maryland, and the District of Columbia. Cities: Brooklyn and the Bronx, Philadelphia, Pittsburgh, and Baltimore. UPPER SOUTHEAST (USE). States: Virginia, West Virginia, North Carolina, Tennessee, and Kentucky. Cities: Nashville. LOWER SOUTHEAST (LSE). States: South Carolina, Georgia, Florida, Alabama, Mississippi, Louisiana, and Arkansas. Cities: Jacksonville. EAST NORTH CENTRAL (ENC). States: Ohio, Indiana, Illinois, Michigan, and Wisconsin. Cities: Cincinnati, Columbus, Indianapolis, Chicago Reserve City banks, and Peoria. WEST NORTH CENTRAL (WNC). States: Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas. Cities: Minneapolis, St. Paul, Kansas City (Missouri), St. Joseph, St. Louis, Lincoln, Omaha, Topeka, and Wichita. SOUTHWEST (SWT). States: Oklahoma, Texas, New Mexico, and Arizona. Cities: Oklahoma City, Tulsa, Dallas, Fort Worth, Galveston, Houston, San Antonio, and Waco. MOUNTAIN (MOU). States: Montana, Wyoming, Colorado, Idaho, Utah, and Nevada. Cities: Denver and Salt Lake City. PACIFIC COAST (PCT). States: Washington, Oregon, and California. Cities: Seattle, San Francisco, and Los Angeles. For greater clarity in the graphical presentation of the regional average rates in Figures 4 and 5, some regions were aggregated. In the city rates in Figure 4, the Northeastis composed of the NED, MAC, and ENC regions. The South is the USE, LSE, and SWT regions. The Midwest is the WNC region. The West is the MOU and PCT regions. In the state rates in Figure 5 the Northeast is the NED, MAC, and ENC regions. The Southeast is the USE and LSE regions. The Midwest is the WNC, the Southwest the SWT, the Mountain the MOU, and the Pacific Coast the PCT region.Google Scholar

⁷ The standard deviation is an absolute dispersion measure. This is appropriate for relative prices (such as interest rates). If absolute prices were used (such as dollar wage rates) a relative dispersion measure such as the coefficient of variation or the variance of the logarithms of the individual prices would be appropriate. For the postbellum period John James has used the mean deviation of rates from New York City rates to measure rate dispersion; he argues that New York City was the financial center of the United States and therefore provided the standard to which all other rates should be compared. The two absolute dispersion measures tell a very similar story. For a more extensive discussion of relative versus absolute dispersion measures of interest rate variation, see Smiley, Gene, “Regional Differences in Interest Rates in the United States, 1888–1913,” Working Paper, Marquette University Economics Department, January 1976 (available from the author).Google Scholar

⁸ Beginning in 1926, when average annual interest rates could be calculated for six-month periods, the seasonal variation in the Midwestern and Western country bank rates can be easily seen. This causes the country bank rate dispersion in Figure 6 to fluctuate sharply beginning in 1926 when the dispersion is calculated semiannually rather than annually.Google Scholar

⁹ In the 1891 set the only southern cities were New Orleans and Louisville, and the only city west of Kansas City, St. Joseph, and Omaha was San Francisco. Rate dispersions for this set of cities could not be calculated past 1926 since rates for cities with less than three national banks could not be separately estimated after that, and rates for eight of the 23 cities in this set could not be constructed. The set of 37 cities whose rates could be estimated for the entire 1916–1937 period included 15 of the 23 1891 cities plus two southeastern cities, eight Texas and Oklahoma cities, three from Kansas and Nebraska, and four more from the Western U.S. Rate dispersions for this more comprehensive set of cities could not be calculated for years before 1909.Google Scholar

¹⁰ The dispersion on a semiannual basis for 1926 on makes this more difficult to see. For state rates averaged over the calendar year for 1926 through 1929, the country bank dispersions for 1923 through 1929 are: 1923, 0.817; 1924, 0.884; 1925, 0.850; 1926, 0.794; 1927, 0.861; 1928, 0.798; and, 1929, 0.881.

¹¹ For example, see Temin, Peter, Did Monetary Forces Cause the Great Depression? (New York, 1976), Figures 4 and 5, pp. 124–25.Google Scholar

¹² See Temin, Did Monetary Forces, andGoogle ScholarFriedman, Milton and Schwartz, Anna, A Monetary History of the United States, 1867–1960 (Princeton, 1963), chap. 7.Google Scholar

¹³ Temin, Did Monetary Forces, chap. 4, andGoogle ScholarFriedman, Milton and Schwartz, Anna, “Money and Business Cycles,” Review of Economics and Statistics, 45 (02 1963), 32–78.Google Scholar