Author:Dzigbede, Komla D.

    State governments issue debt to finance a variety of capital projects including transportation, highways, and natural resource preservation. When issuing debt, a state government typically assembles a debt management team comprising an underwriter, municipal advisor, and bond lawyer. An underwriter works with an issuer to structure a bond sale and acts as an intermediary between the issuer and the investing public, assuming the risk of buying the bonds and reselling them to investors (Simonsen and Hill 1998; Fruits, Booth, and Smith 2008). Municipal advisors give financial advice on debt issuance, investment of issuance proceeds, and use of derivatives and guaranteed investment contracts (Luby and Hildreth 2014). Bond lawyers ensure that the structure and sale of bonds conforms with securities and tax laws and give informed opinions on the legal status of securities to assure investors that the securities are binding legal obligations of the issuer (Johnson, Luby, and Moldogaziev 2014). A government issuer works with a debt management team to issue debt at the least interest cost. Borrowing at the least interest cost makes it possible for governments to undertake capital projects in a cheaper way and lessen the burden of interest payments during the tenure of the debt.

    Research shows that repeated use of the same underwriter can have either a positive or negative impact on borrowing costs (Miller and Justice 2012; Liu 2015; Burch, Nanda, and Warther 2005). On the one hand, an issuer may accumulate relevant information from using the same underwriter for different debt issues, which can lower information costs and reduce borrowing costs. On the other hand, use of the same underwriter for multiple issues can raise the risk of collusion and fraudulent practices and increase borrowing costs, or repeated use of the same underwriter can raise borrowing costs if organizational learning is limited and an issuer and underwriter together continue with previous errors in issuance strategy, leading to greater risk-taking and more inefficiency.

    I argue that the reason for this mixed stance on how repeated use of the same underwriter affects borrowing costs is because the relationship is more complex than has been theoretically proposed and empirically tested. The empirical evidence on repeated use of the same underwriter is limited because previous work (e.g., Liu 2015) ignores non-linear effects on municipal borrowing costs. This article expands knowledge on the topic by using an empirical framework that accommodates non-linear effects in measuring the impact of repeated use of the same underwriter on municipal borrowing costs. Analyzing non-linear impacts is essential to identify a critical threshold beyond which repeated use of the same underwriter would be more cost-effective, or less cost-effective, for the municipality.

    The argument for non-linear effects is grounded in theory. First, transaction costs theory, initiated in Coates (1937) and formalized in later works (e.g., Williamson 1989), supports the view that repeated use of the same underwriter would cause municipal borrowing costs to rise initially, reach a peak, and decrease thereafter, portraying a concave or inverted-u shape. This is because an underwriter incurs search and information costs on the issuer when structuring a debt sale, therefore as the issuer uses the same underwriter for multiple issues, fixed sunk costs are spread across more issues and the average cost of underwriting, as well as borrowing costs, would eventually decline.

    Second, organizational learning theories, starting from Wright (1936) and expanded in later studies (e.g., Dutton and Thomas 1984; Argotte 2012; Arthur and Huntey 2005), suggest that organizations learn from experience and use that knowledge to lower production costs and increase productivity and efficiency but with diminishing rates or returns. Consequently, if issuers and underwriters learn from each other, information gains should be greater at the initial stage when there are fewer repeated uses of the same underwriter, causing borrowing costs to decline in this stage of increasing efficiency. However, as learning and knowledge acquired in the past tend to have less effect on organizational performance than more recently acquired knowledge, gains in efficiency would peak, and decline thereafter, when the issuer repeatedly uses the same underwriter for many more issues, meaning that the decline in borrowing costs will reach a trough and start to rise, depicting a convex or u-shaped curvature.

    Third, the 'bank hold up' theory, outlined in Rajan (1992) and discussed extensively in subsequent studies (e.g., Stein 2015), suggests that over the course of a bank-firm relationship, banks acquire an information advantage (compared to outside investors that compete with banks to extend working capital to firms) which enables banks to hold-up their borrowers and enforce higher lending rates. Applying this theory to municipal debt issuance contexts, it may be that over the course of an issuer-underwriter relationship, and with more repeated use of the same underwriter by the issuer, the underwriting firm would be able to obtain an information advantage (compared to other underwriting firms) which might cause the underwriter to charge continually higher rents and fees that altogether raise issuance costs exponentially.

    This article investigates the convex, concave, and exponential scenarios described above (see Figure 1). It focuses on state government debt issuance and asks: how does repeated use of the same underwriter affect state government borrowing costs? Data are from a cross-section of 1,063 general obligation bonds, with maturities greater than 3 years, which the State of California issued between 2005 and 2014. The analysis uses a general model of state borrowing costs that includes bond-specific characteristics, issuer-related variables, municipal market-wide factors, and macroeconomic trends, and accommodates non-linear patterns and year effects. I estimate specific forms of the general model using statistical procedures that adjust standard errors to be robust to heteroskedasticity and enhance normality of the residuals. The analysis also uses Box-Cox transformations to consider functional forms that best fit the data and deepen understanding of the non-linear impacts of repeated use of the same underwriter on municipal borrowing costs.

    Results provide new insights on how repeated use of the same underwriter affects state government borrowing costs. First, I find significant evidence that as the state government repeatedly uses the same underwriter for new debt issuance, borrowing costs tend to increase initially, but with more repeated use of the same underwriter, borrowing costs tend to reach a peak, and decrease eventually, reflecting a concave or inverted-u relationship. Second, under conditions where information capture and rent-seeking by underwriting firms persist, and when the functional forms of model variables are optimized to best fit the data, I find significant evidence that repeated use of the same underwriter may increase state government borrowing costs exponentially. Both findings show that the relationship between repeated use of the same underwriter and municipal borrowing costs is more complex than the linear effects prior studies present, therefore examining non-linear effects expands under-standing and should guide municipal debt management.

    Figure 1. Impact of Repeated Use of the Same Underwriter on Municipal Borrowing Costs

    (a) Transaction Costs Theory and Municipal Borrowing Costs

    (b) Organizational Learning Theory and Municipal Borrowing Costs


    (c) Bank Hold-Up (Information Capture) Theory and Municipal Borrowing Costs

    In the next sections, the article discusses the academic literature on repeated use of the same underwriter and municipal borrowing costs and presents research hypotheses. The study methodology, data and variable formulations, and empirical results are also presented. Finally, I discuss debt management implications of the findings.


    Only a limited amount of work exists that explores the implications of issuers' repeated use of the same intermediaries in municipal capital markets. This section draws from three streams of literature to shed more light on the subject and present research hypotheses. The first stream of literature focuses on transaction costs and examines how information and search costs affect municipal borrowing costs as an issuer repeatedly uses the same underwriter. A second stream of literature examines organizational learning and how knowledge and information gains from an issuer's repeated use of an underwriter affects efficiency of debt management as reflected in municipal borrowing costs. The third stream of literature focuses on relationships in financial markets, specifically how the relationship between issuers and financial intermediaries affects municipal borrowing costs. Much work already exist that discuss each of these streams of literature in detail, therefore my goal in this review of the literature is to highlight key aspects of each stream that shed light on the interest cost implications of repeated use of the same underwriter in municipal debt markets.


    Coase (1937), Arrow (1969), and Williamson (1979, 1989) are among the scholars that expanded knowledge on transaction costs in financial markets. Coase (1937) explained transaction costs as the exchange costs that characterize organization and operation of the market system and noted that these costs can be reduced but not eliminated because of the enduring nature of costs associated with negotiating and monitoring contracts and trading information. Arrow (1969) showed that information transacting parties need to enter and participate in markets can be costly, unequal, and closely related to communication and...

To continue reading