Only available on StudyMode
  • Topic: Bond, Municipal bond, Bonds
  • Pages : 94 (13115 words )
  • Download(s) : 121
  • Published : January 21, 2013
Open Document
Text Preview
Fractionalization and the municipal bond market
Daniel Bergstresser*
Randolph Cohen**
Siddharth Shenai***

(First version April 2010. Current draft June 2011. Comments welcome.)

We study the impact of ethnic and religious fractionalization on the U.S. municipal debt market, and find that issuers from more ethnically and religiously fractionalized counties pay higher yields on their municipal debt. A two standard deviation increase in religious fractionalization is associated with a six basis point increase in bond yields, and a two standard deviation increase in ethnic fractionalization is associated with a ten basis point increase. To provide a scale for these results, a four-notch rating change, from AAA to AA-, is associated with an eight basis point increase in yields. Additional analysis suggests that at least some of this effect is not driven by the risk of the bonds, but instead reflects inefficiency in the underwriting process. Keywords: Fractionalization, municipal bonds.

We are grateful for support from Harvard Business School. We are also grateful for comments from Alexander Whalley, Andrei Shleifer, and Alberto Alesina, and from seminar participants at the SEC, HBS, Boston College, USC, and UC-Merced. Finally, we are grateful for excellent research assistance from Mei Zuo. Sid Shenai is at Bracebridge Capital. Daniel Bergstresser and Randolph Cohen have primary academic appointments. Both also engage in consulting and other activities in the investment management industry. Details are available upon request. * Corresponding author. Harvard Business School. Tel.: 617-495-6169. E-mail:

** Massachusetts Institute of Technology.
*** Bracebridge Capital.


Electronic copy available at:

A variety of researchers have explored the impact of ethnolinguistic diversity on economic outcomes. Easterly and Levine (1997) propose that ethnolinguistic fractionalization can explain the poor growth observed in post-colonial Africa: borders left by the former colonial powers have resulted in ethnically divided countries that make for challenging governance.1 In the American context, Glaeser, Scheinkman, and Shleifer (1995) find no effect of racial fragmentation on the growth rates of cities, but Alesina and La Ferrara (2000) find that more ethnically fragmented communities have lower participation in civic associations such as church groups, fraternities, and service groups. These social institutions contribute to social capital and can have positive economic effects, as discussed by Putnam (1993, 1995). Similar research on religious diversity, however, is lacking, though McCleary and Barro (2006b) examine the relationship between religious observance and economic growth. In this paper we use American counties as a laboratory for studying the impact of fractionalization. This approach has at least two advantages. First, using American counties holds constant a variety of underlying factors that cannot be kept constant in cross-country analysis. Second, because local authorities borrow in credit markets, the yield that they pay on their debt is a natural index for comparing their performance. Higher yields on municipal debt must reflect either higher risk (if capital markets are working efficiently) or cross-locality differences in the efficiency with which the municipal underwriting process works. We construct fractionalization measures of ethnic diversity using data from the 2000 Census, and of religious diversity using the 2000 Religious Congregations and Membership                                                              1

From January 8, 2011 New York Times article on breakup of Sudan: ‘More than any other continent, Africa is wracked by separatists. There are rebels on the Atlantic and on the Red Sea. There are clearly defined liberation movements and rudderless, murderous groups known principally for their cruelty or greed. But...
tracking img