We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 19 October 2009
The most conspicuous deficiency of my study of underwriting compensation is my failure to examine the determinants of underwriting spreads on tax-exempt bond issues. That deficiency has now been remedied by Richard West. Unfortunately, my study was not published very long before West's and there was little opportunity for him to compare his results and speculation with my own. Although a few of the comments below are critical of West, and I point out some alternative interpretations of his statistical findings, this note should be considered a supplement to, rather than a critique of, West's paper.
1 See Chapter 7 in LongstreetIrwin Friend, James R. Irwin Friend, James R., Mendelson, Morris, Miller, Ervin, and Hess, Arleigh P. Jr., Investment Banking and the New Issues Market (Cleveland: The World Publishing Company, 1967)Google Scholar.
2 See this Journal (September 1967). It might be observed parenthetically that Mr. West's remark that my study was “‘long’ on correlations and regressions, but ‘short’ on the theory underlying the relationships studied” (p. 243) is based upon the short summary of my study that appeared in Friend, Irwin, Investment Banking and the New Issues Market, Summary Volume (Philadelphia: University of Pennsylvania, 1965)Google Scholar. Had West complained that I had failed to supply a mathematical model of the determinants of spreads, he would have been justified. My study was not, however, simply a brute force attempt to find such determinants. My statistical model is rationalized in very much the same way as is West's.
3 Cohan, Avery B., Cost of Flotation of Long-Term Corporate Debt Since 1935, Research Paper 6 (Chapel Hill: School of Business Administration, University of North Carolina, 1961)Google Scholar. It may be noted that this study of Cohan, which was the pioneering study in this area, was also no more devoid of theory than West's.
4 Mendelson, op. cit.
5 Significant between the .05 and .01 levels. (West, op. cit., p. 255).
6 West makes a slight error in his discussion of his Table 2. He notes that “the relationships between spreads and issue quality and single bids are quite stable.” (Op. cit., p. 259.) In Table 2, however, he suggests that the hypothesis that all years have the same coefficient be rejected with respect to the coefficient for single bids.
7 Cohan, op. cit., p. 33. The average size of the coefficient of maturity in Cohan's study was -.0013 and the corresponding value of t was -.55.
8 West, op. cit., p. 263.
9 Friend, op. cit., p. 71. In fairness to West, I quote Friend's summary of my finding rather than a statement of my own. West's observations are based on Friend's summary and in fairness to Friend it is important to establish that the misinterpretation of my finding is not his. I might also point out that in my own chapter I concluded that between 1955 and 1963 the preponderance of evidence suggested that there was no trend in tax-exempt spreads. Again, this is not necessarily inconsistent with West's finding.
10 Cohan, op. cit., p. 33.
11 In terms of gold. New York Times, December 13, 1967, p. 69.
12 My observation that there is a difference between the spread on five-year notes and straight and convertible bonds on the international markets is a tentative one and is based upon an investigation I am currently making of the international markets. Underwriting spreads on the international markets are highly conventional. In the underwriting of American offerings, five-year notes always carry a 2 percent spread. Other maturities almost without exception carry a 2.5 percent spread. This is true regardless of issuer, yield, or maturity. These static spreads are, however, a little deceptive. True spread may differ from the stated spread because of expense allowances. All issues do not have expense allowances, and those that do have such allowances don't always have the same amount. However, after allowing for expense allowances, not a single five year note has an adjusted spread as high as 2.5 percent. I have every intention of subjecting these international issues to the same kind of analysis that Cohan, West, and I have already utilized in our investigations of underwriting compensation. I would be singularly concerned with the appropriateness of our statistical techniques if the regression analysis suggested that there was no significant difference between the spreads on notes and bonds. However, some distortion of the results can be anticipated since the notes were offered in tighter markets than the bonds. The yields reflect this as well as differences in risk.
13 op. cit., pp. 425–426.
14 See James R. Longstreet and Morris Mendelson, “Characteristics of New Issues Transactions,” in Friend, et al., Investment Banking and the New Issues Market, op. cit. There is evidence in that study that individuals absorb a significantly smaller fraction of general obligation issues than they do of revenue bond issues and that commercial banks have significantly larger average size transactions in general obligations than in revenue bonds.
15 Longstreet and Mendelson, op. cit., p. 287.
