To Give Is to Get

Abstract

Public managers and elected officials are generally restricted from supporting election campaigns with public resources. In the case of legislative referenda, the public stakeholders responsible for putting a policy question on the ballot must play a neutral role when acting in their official capacity. A system where private money supports public goals has emerged as regulatory provisions simultaneously restrict direct private giving to elected officials and public support for election campaigns. Using campaign finance disclosures, election results, and municipal bond issuance data, we find that post-election fees paid to firms making political contributions are significantly higher than for non-contributors. The finding improves the understanding of how private dollars support public policy outcomes, raises questions about the circumvention of laws restricting the use of public resources in election campaigns, and informs ongoing consideration of the need for additional regulatory action and disclosure requirements to address issue committee campaign contributions.

Keywords

campaign finance direct democracy municipal bonds pay-to-play

Introduction

State and local laws commonly require citizens to vote on referenda to increase tax rates or authorize bond issuances. Frequently, these referenda are to finance capital projects, such as school construction and improvements. Elected and appointed officials are put in a difficult position after acting to place these referenda on ballots. Officials are generally unable to use any public resources—including their time at work—to promote the passage of referenda due to state and local laws. In some cases, the market has adapted to these limits on public officials’ activities by transferring the role of promotion of these local referenda to independent issue committees supported by interested private parties.

The typical issue committee supporting a local bond referendum has a name resembling “Citizens for X” or “Friends of X,” where “X” may be “Safe Schools,” “Public Libraries,” or “Fire Safety.” While individual citizens often support these groups with their time and money, a substantial portion of the contributions to these issue committees comes from parties that may receive financial benefits from the referendum’s approval. With respect to issuing debt for capital projects, these parties often include financial intermediaries, bond underwriters, and construction firms. In the face of the widespread legal restrictions on governments’ ability to use public resources to advocate for ballot items, this article asks whether political contributions from bond underwriters influence governments’ subsequent debt issuance costs.

While some finance firms have official policies insisting that there is no quid pro quo regarding such campaign contributions, our findings suggest that issuance costs are in fact higher when contributions are made. Using campaign finance disclosures, election results, and municipal bond issuance data from California, we find that, on average, underwriters receive around US$1.70 for each US$1 contributed to support a bond campaign. From a management perspective, the behavior reflects lingering concerns over pay-to-play practices in the public finance field while raising serious questions about the circumvention of restrictions on the use of public funds for political campaign activities.

This article contributes to the existing literature in important ways. First, while the role of campaign contributions to political candidates and state ballot initiatives has received attention from researchers, the use of industry-supported issue committees in local referenda has been overlooked. This article provides an overview of this behavior and an analysis of whether it is problematic beyond just creating the appearance of impropriety. Second, the topic highlights the continuing tension between politics and administration for public managers, as well as the possible need for improved disclosure and increased regulation of financial activity beyond the traditional reporting framework for governments.

Background

Pay-to-Play Activity and Municipal Securities Regulation

Over the past decade, state and local governments issued, on average, US$386.7 billion in long-term debt each year (“Market Statistics Archive, Bond Referenda Results, 2007-2011,” 2013). The issuance of municipal bonds is a technical process in which local governments depend on financial intermediaries including financial advisors, underwriters, bond counsel, credit rating agencies, trustees, and credit enhancers. In this article, we focus on the underwriter, which is the investment banking firm who purchases the bonds from the issuing government and then resells them to investors. There have long been concerns about impropriety in municipal finance, where campaign contributions to elected officials were prominent in securing bond issuance business. This practice is commonly referred to as “pay-to-play” (Jordan, 1999). A retired manager of the Internal Revenue Service’s tax-exempt bond field operations recently stated, “It’s rare to sell a Senate seat, but it’s not rare to sell a bond deal . . . Pay-to-play in the municipal bond market is epidemic” (Walsh, 2009, p. A1).

Pay-to-play activity historically centered on direct contributions to elected officials who have influence over the private firms selected and paid to assist in the issuance of bonds and other financial services. The municipal securities industry was effectively unregulated until Congress created the Municipal Securities Rulemaking Board (MSRB) as part of the Securities Acts Amendments of 1975. The broad impetus for the MSRB’s creation was the growth in issuance volume and number of dealers (MSRB, 2013a).¹

The MSRB proposed Rule G-37 in August of 1993, following an industry-led voluntary effort, and adopted it in 1994 (Opp, 2005). This rule intended to sever the potential link between political contributions and municipal bond business by restricting the ability of public finance professionals to contribute to candidates’ political campaigns. If a public finance professional makes a contribution either to a candidate for whom they are not eligible to vote for, or in excess of US$250 for a candidate for whom they can vote, then their firm is banned from conducting business with that governmental entity for 2 years (Opp, 2005). In the wake of G-37, pay-to-play activity shifted from direct contributions by public finance firms to the use of consultants who were hired to secure business for the firms and did not fall under the purview of G-37. The MSRB responded with Rule G-38 in 1996 to address this loophole (Opp, 2005). A number of additional loopholes have been identified including political giving to candidates by family members, contributions made to political parties rather than candidates or office-holders, the use of payments to law firms to help secure municipal finance business, and the ability of large firms to direct political contributions through employees not classified as municipal finance professionals (Jordan, 1999; Opp, 2005). The latter includes the increasingly relevant professionals “who work with financial derivatives, like swaps and options” (Walsh, 2009). This history of regulatory oversight suggests that each time a new rule is implemented to minimize pay-to-play situations, new ones are created or emerge.

Numerous observers acknowledge that contributions to issue committees remain an area of potential influence in securing business, because G-37 and G-38 do not apply to such activity (Opp, 2005; Retnasaba, 2006; Robbins & Simonsen, 2007; Simonsen & Hill, 1998). This article explores the role of these increasingly visible issue committee campaign contributions in securing municipal bond business and their relationship to a government’s ultimate cost of financing. The MSRB, effective in early 2010, revised Rule G-37 to require ballot campaign contribution disclosure in response to “concerns that contributions by covered parties to bond ballot campaigns could assist dealers with obtaining municipal securities business” (O’Neill, 2013, p. 4). Since then, the MSRB has expanded the disclosure requirements for bond ballot contributions, which will be used to establish a basis for any future regulatory action deemed necessary “up to and including a corresponding ban on engaging in municipal securities business as a result of certain contributions” (78 FR 10657-8). This is the first large-scale empirical analysis using regulatory data to inform future MSRB decisions and goes beyond individual case studies or anecdotes. The following section describes bond referenda and existing rules limiting the use of public resources for election activities.

Bond Referenda and State and Local Restrictions on Public Sector Political Activity

While local governments have a number of alternatives to finance capital outlays, the most common method is the use of general obligation (GO) debt financing via the tax-exempt bond market. As of 1990, 39 states required voter approval to authorize a local government bond issuance (U.S. Advisory Commission on Intergovernmental Relations, 1993). Bond referenda, a common type of non-voter initiated referenda, are proposed and referred to the ballot by elected officials (Zimmerman, 2001). They tend to be competitive with only 63% passing nationally since 2007 (“Market Statistics Archive, Bond Referenda Results, 2007-2011,” 2013). California, on which we focus our analysis, experienced greater success during this time period with just more than 83% of bond elections approved by voters (“Market Statistics Archive, Bond Referenda Results, 2007-2011,” 2013).

Directly elected officials, such as school board members, place bond referenda on ballots. Once placed on the ballot, however, public resources cannot be used to advocate for the passage of referenda. State and local campaign finance laws differ, but the logic of such restrictions on political activity is that because “public tax monies belong to both proponents and opponents of a bond measure, it is unfair to force opponents of the measure to lend financial support to the proponents” (Orrick, Herrington & Sutcliffe, 2010, p. 1). Local government officials are allowed to provide information to voters about a bond referendum, but they may not “promote passage of the referendum” (Lawrence, 1988, p. 9). Given these restrictions, elected officials take a risk by placing a referendum on the ballot without being able to support it publicly in their official capacity or use public resources. The public interest in any given referendum is likely to be fragmented. For example, a school district attempting to gain voter approval to issue debt for the renovation of school buildings may appeal to parents of school-age children that will benefit from the new school, but such efforts may be received negatively by those opposed to increasing property taxes. School district officials and teachers may also have fairly strong incentives to support such spending.

Underwriter Campaign Contributions to Issue Committees

The use of issue committees to promote bond referenda in place of public efforts has received significant attention from the popular press, the trade press, state legislators, and regulatory agencies because it raises the appearance of impropriety (e.g., see Olinger & Gorski, 2012). A report in California found that only 5 out of 111 contributing underwriters did not end up getting the work for the associated bond issuances (Evans, 2012). This success rate is qualified by the fact that for “four of those, however, more than one underwriter made donations and the contract went to the firm that had contributed a larger amount to the campaign” (Evans, 2012). The Bond Buyer documented a similar relationship finding “a nearly perfect correlation between broker-dealer contributions to California school bond efforts in 2010 and their underwriting subsequent bond sales” (Jensen, 2012a, p. 2).

The notion that underwriting work is awarded based on campaign contributions raises persistent concerns in the public finance field over pay-to-play activity. Underwriters, at least in California, formally sidestep this concern by being selected as underwriter through a negotiated sale in advance of the actual election campaign and prior to making any contribution to the issue committee (Evans, 2012). This behavior points to the importance of knowing when an underwriter is awarded bond underwriting business, either formally or informally, relative to pledging or making a bond ballot campaign contribution. The recent MSRB changes to disclosure requirements for bond ballot contributions include a “reportable date of selection,” which reflects the date an underwriter is chosen “either in writing or orally” (78 FR 10658). It is common for the agreements for underwriting services to formally state that the underwriter is under no obligation to make contributions to the bond issue’s related issue committee.

In addition to knowing the time order of selecting an underwriter relative to subsequent contributions, it is also necessary to understand when the pricing of the underwriter services is set. Based on a limited review of publicly available underwriting contracts, the bulk of the compensation (aside from reimbursement for miscellaneous additional expenses) is set at the time the negotiated contract is signed. The fees are typically communicated as a percent of the amount of bonds planned for issuance pending the successful election. The reviewed contracts make it clear that no fees are due to the underwriter for their pre-election services if the associated referendum fails.

In the past decade, California and Colorado have seen proposals to end the practice of bond election contributions by underwriters fail within their state legislatures (Jensen, 2012a; Olinger & Gorski, 2012; Saskal, 2010). These efforts reflect a belief that contributions are passed through to governments in the form of higher fees. A former California State Assemblyman and school board member described the process as,

Basically you would have a financial firm come and say we will help with setting up the campaign, help find a consultant if you need one and then we will do the issuance when it is all said and done . . . Obviously, all of those upfront costs and expenses are rolled into the cost of the deal—it is not rocket science to figure it out. (Jensen, 2012a, p. 2)

An underwriter counters such assertions with a policy statement that they

will not make, or indicate a willingness to make, any financial contribution as a condition to being retained as an underwriter. These financial contributions are assumed . . . as a general overhead cost, are provided on a voluntary basis only, and are not passed through to any client nor does Piper Jaffray seek reimbursement for such costs. (Piper Jaffray, n.d.)

In the summer of 2013, actions initiated outside the legislative and regulatory systems began to address the practice. A dozen large finance firms voluntarily promised to withhold from making election contributions “to bond measures they seek to underwrite” and, separately, the Los Angeles County treasurer established a policy restricting its use of any firm that gives “political contributions to school bond campaigns” (Weikel, 2013).

To better understand the prevalence of this behavior nationally, we gathered the Rule G-37 disclosures for the top 50 negotiated sale municipal bond underwriters since the MSRB ballot campaign contribution disclosure requirement has been in place (see Table 1). Contributions have been reported in 18 states, but activity is concentrated in only a few, with California, Colorado, and Ohio representing more than 85% of contribution dollars nationally. California, by itself, is home to nearly two thirds of the activity. Roughly 80% of the more than 600 contributions are directed to school district election campaigns. The total dollar amount of reported contributions is just less than US$2 million for the period starting in 2010 and ending in the second quarter of 2012. The average contribution was US$2,970, ranging from US$25 to US$50,000, while some were designated as in-kind (rather than cash) contributions. The aggregated dollar amount of contributions may seem immaterial, but they are associated with large authorized debt amounts. In California, less than US$2 million in election contributions given over 5 years were associated with more than US$15 billion in authorized debt issuance (Evans, 2012).

Table 1.

Ballot Contributions for Top 50 Negotiated Sale Municipal Bond Underwriters, 2010-2012.

State	Number of contributions	Share of contributions (%)	Contributions (US$ value)	Share of contributions (% of US$ value)
Arkansas	10	1.6	39,750	2.1
California	123	19.6	1,222,533	65.4
Colorado	44	7.0	237,038	12.7
Georgia	244	38.8	36,199	1.9
Illinois	13	2.1	23,667	1.3
Kansas	10	1.6	3,250	0.2
Louisiana	1	0.2	900	0.0
Michigan	32	5.1	29,485	1.6
Missouri	4	0.6	32,600	1.7
Nebraska	30	4.8	23,744	1.3
New Mexico	6	1.0	17,000	0.9
New York	1	0.2	700	0.0
Ohio	82	13.0	146,908	7.9
Oklahoma	4	0.6	5,798	0.3
Texas	16	2.5	23,143	1.2
Utah	2	0.3	7,000	0.4
Wisconsin	1	0.2	250	0.0
Wyoming	6	1.0	18,212	1.0
Total	629		1,868,176

Source. Authors’ calculations; data aggregated from MSRB G-37 disclosures.

Note. MSRB = Municipal Securities Rulemaking Board.

It is natural to question why election contributions are so heavily focused on school district bond campaigns as opposed to other types of government. The most obvious explanations are that school districts depend on GO debt (Mikesell, 2007) and, despite making up fewer than 15% of governments nationally (U.S. Census Bureau, 2012), over the past decade, school districts represented two of every three bond issues (64%) presented to voters for approval. In addition to the need to hold a referendum, the selected method of sale is likely to determine whether underwriters are willing to contribute to bond election campaigns. Ballot campaign contributions are not expected when the competitive sale method is being used, because the issuer has little control over the selection of the underwriter. The California data suggest that election contributions are rarely, if ever, made in elections that ultimately result in bonds sold competitively. In California, the prevalence of underwriter contributions in school district and community college bond elections is likely the result of a state law that only allowed those governments to engage in negotiated sales for GO bonds (Senate Local Government Committee, 2010).² Finally, California has a supermajority voting requirement to pass school district referenda, which, with some exceptions, was reduced from two thirds to 55% with Proposition 39’s approval in 2000.

To illustrate the workings of a California issue committee and the types of activities funded by campaign contributions, we consider “Yes on Z for Safe Schools,” which was formed to support Beaumont Unified School District’s Measure Z in the election of November 11, 2008.³ This typical example shows that underwriters are only one of many contributors to issue committees. Measure Z presented the following question to the voters:

To retain/attract excellent teachers, keep students from gangs by upgrading safety, technology security systems, energy efficiency; acquiring land; repairing, constructing, equipping schools/classrooms; adding vocational education/after-school space; fixing leaky roofs; replacing heating/air-conditioning; and qualifying for State matching grants shall Beaumont Unified School District issue $125,000,000 in bonds at legal interest rates, with annual financial audits, Independent Citizens’ Oversights, and no money for State or administrators’ salaries?

The measure ultimately passed with 9,029 votes in favor and 5,523 opposed. The committee raised just less than US$94,000 in contributions. Of the 31 itemized contributions, only 6 came from individual donors, including the superintendent of the school district, 4 assistant superintendents, and the treasurer of the issue committee; combined, these individual contributions totaled less than 1% of all contributions. The primary contributors represented the construction (with 46% of the total contributions) and public finance (with just less than 40%) industries. A construction firm and investment bank topped the list of cash donors by each contributing US$25,000.

The first bond issuance based on the successful referendum, for US$20 million, confirms that many of the public finance contributors played prominent roles (Beaumont Unified School District, 2009). The selected underwriter contributed US$25,000, the financial consultant gave US$5,000, the bond counsel donated US$2,000, and the paying agent added US$1,000 to the referendum campaign. There is substantially less visibility into construction activities, but a district press release confirms that the largest contributor received the construction contract (Sturman, 2009).

The total amount raised to support passage of the bond referendum is less than one tenth of 1% of the US$125 million of future debt issuance authorized by the vote. The committee’s campaign finance disclosure forms also provide a glimpse into the types of activities these groups perform (Yes on Z for Safe Schools, 2008-2009). The largest share of expenditures was for campaign literature and mailings, representing just over half of spending, while 40% of the costs went to a political consulting firm. The committee ran a multifaceted campaign including hosting an informational booth at an Oktoberfest event, creating a campaign logo, producing and installing signage, paying for prospective voter phone lists, making phone calls, mailing postcards, newspaper advertising, and providing targeted voter guides to different audiences including Latinos, Republicans, seniors, and women, among others.

Literature Review

We base our empirical expectations on two distinct literatures—the influence of money in direct democracy and the costs of municipal debt issuance. We use these literatures to hypothesize about the role of underwriter issue committee contributions in bond referenda and the costs of issuing bonds following voter acceptance of referenda.

Money and Direct Democracy

Scholarly attention to the role of campaign contributions in direct democracy is typically focused on the influence of money on the success or failure of state-level referenda and initiatives, as well as on inducing increased voter turnout. Our research offers a unique glimpse into direct democracy and campaign contributions at the local level. It is also a research setting where we can more closely link the campaign support to tangible financial interests of donors, namely, underwriters, and subsequent fees for services.

Garrett and Gerber (2001) raise concerns regarding how “money disproportionately influences direct democracy” (p. 73). Most notable for our study is the belief that “quid pro quo corruption and the appearance of such corruption” may also be relevant in direct democracy just as it is in elections for political office (Garrett & Gerber, 2001, p. 74). Garrett and Gerber (2001) summarize the literature, suggesting that campaign spending does not dictate election outcomes, spending in favor of a ballot item is less effective than spending against, and the source of contributions matters. Spending funded by citizen groups is more effective than that of industries.

If such contributions and spending are ineffective in securing more successful voting outcomes, why do we see contribution activity at the local level? This notion of ineffectiveness runs counter to anecdotes from experienced public officials (Jensen, 2012a). Recent research by de Figueiredo, Ji, and Kousser (2011) notes that contributions and campaign spending naturally increase for ballot items that are thought to have lower odds of passage, which results in a bias against finding that such spending is successful. Using nearly three decades of California ballot items, they show that spending in support of a ballot item is in fact effective. Overall, the existing literature is not yet settled on how money influences such election outcomes.

Rather than examining election outcomes, some have focused on whether money influences voting behavior. Smith and Tolbert (2004) have noted that direct democracy has an “educative effect” in that it increases citizen knowledge and participation in state elections. Therefore, money in direct democracy may allow the information required for this citizen engagement to be distributed more widely (e.g., to groups traditionally less apt to vote). Election contributions of underwriters appear self-serving, but it is easy to argue that the supported issue committees work to better inform the voting public. Such commentary is often heard in response to complaints about the practice. A bond underwriter explicitly states that such contributions “are more akin to charitable contributions because they benefit a legitimate public purpose rather than an individual” (Fairman & Lawrence, 2009, p. 5). Despite this informational role, Inman (1978) finds that differences in voter turnout have an extremely small effect on the actual outcome of school district budget referenda. Therefore, even if turnout increases, the probability of the election’s success may be left effectively unchanged.

Debt Issuance Methods and Pay-to-Play Activity

After broadly considering what we know about the role of money in direct democracy, we turn to the connection between political contributions and fees paid by governments. Two related topics have dominated scholarship on municipal bond issuance costs—the optimal method of sale and “pay-to-play” activity. The present study is connected to both topics.

The method of sale debate centers on whether competitive bidding by underwriters for business generates cost savings relative to a negotiated sale. In the negotiated sale, the government issuer selects the underwriter not solely based on cost but typically using a request-for-proposal and more subjective criteria. Related to the current study, a government could not require or request election services from underwriters as part of a competitive bidding process, because the bids would take place just prior to the actual sale of the bonds and after any associated referendum. A negotiated method of sale allows for such election services to be formally requested, as seen in Colorado (Steers, 1998), or informally provided as an underwriter service.

The debate over the two most common sale methods comes down to whether underwriter competition reduces issuance costs or certain types of issues—those being more complex and difficult to market—benefit instead from the negotiated sale (Leonard, 1996). Efforts to control for selection bias in method of sale research (Kriz, 2003; Peng & Brucato, 2003) suggest that “[t]he more severe the information asymmetry, the more likely it is that an issuer will choose a negotiated sale” (Peng, Kriz, & Neish, 2008, p. 62). Other research continues to challenge the widespread use of negotiated sales, especially for typically straightforward, or “plain vanilla,” GO debt (Robbins & Simonsen, 2007). Our study adds a very different reason for an issuer to prefer a negotiated sale: The uncertainty of a required bond election’s outcome may influence the decision to choose a negotiated sale.

Pay-to-play behavior is expected to be the most prominent among underwriting firms and in negotiated sales, where past relationships are influential in the underwriter selection process (Lemov, 1990; Simonsen & Hill, 1998) and the largest dollar amounts in the bond issuance process are at stake (Peng et al., 2008). Two potential outcomes of that influence are outright corruption or higher costs for governments issuing debt (Peng et al., 2008). There is little incentive for an underwriting firm to contribute to a campaign if they have to deal with the uncertainty of the business being awarded using competitive bids. Therefore, we only expect to see such contributions being made where the negotiated method of sale is used. This is confirmed by our California data, in which not a single underwriter election contribution was made for school district debt ultimately issued using the competitive sale method. Once an underwriter is selected, they have a strong incentive to support a referendum campaign because they are “motivated to ensure that the sale happens” (Simonsen & Hill, 1998, p. 76) or risk not getting paid.

Although there is a broad literature on the topic of corruption that is often associated with pay-to-play in municipal finance, we focus more narrowly on empirical work examining the role of campaign contributions in the municipal bond industry. The notion that government officials may make sub-optimal financial decisions with respect to debt management due to political or campaign finance considerations is not novel. Vijayakumar (1995) asserts that, given government discretion, “the incentive to use [a] call decision for patronage purposes [to generate multiple bond issuances] is likely to be significant” (p. 211). Retnasaba (2006) finds a large and abrupt response to the elimination of underwriter campaign contributions to candidates following the implementation of Rule G-37—estimating that roughly one third of government issuers were using the negotiated method of sale for the purpose of receiving candidate contributions. Butler, Fauver, and Mortal (2009) find that the gross spread received by the underwriter was more than 10% higher for negotiated sales when pay-to-play was active, although their data do not provide a link between specific campaign contributions to candidates and the resulting issuance costs. The results suggest that “campaign contributions might generate a quid pro quo in the form of higher fees for underwriting services” (Butler et al., 2009, p. 2890).

Overall, the existing literature leads us to believe that the negotiated method of sale is more prone to the influence of political contributions. The effects of such influence reach beyond simply the underwriter selection process and have implications for bond pricing and underwriter fees. This article builds on these existing studies by looking at non-candidate contributions that can be linked directly to future fees paid to the underwriter. We test the hypothesis that, contrary to underwriter statements, all else equal, underwriter contributions to issue committees increase the subsequent costs to local governments for issuing debt approved by bond referenda.

Data and Method

Our primary analysis uses campaign finance disclosures, election results, and municipal bond issuance costs to test our hypothesis. Due to data limitations, we provide a limited discussion of the relationship between underwriter contributions and referenda outcomes in our results section while focusing on the effects of election contributions on underwriter issuance costs.⁴ The model used to test our hypothesis regarding the effect of election contributions on underwriter issuance costs, represented as the underwriter’s discount, takes the following form:

Underwriter ’ s Discount = β_{0} + β_{1} (Underwriter Contributions) + V + ε

where β₀ is a constant, V is a vector of additional control variables (in addition to the Underwriter Contributions independent variable), and ε is the error term with standard properties. We estimate the model using ordinary least squares (OLS) regression. Because many governments issue multiple bonds during this time period, we use clustered robust standard errors at the jurisdiction level.

We drew data from three main sources. First, municipal bond issuance data were requested from the California Debt and Investment Advisory Commission (CDIAC) for the period beginning January 2007 through July 2012. Bond referenda results were collected from the California Elections Data Archive (CEDA) for elections held in 2007 through 2011. The elections were then tied to specific underwriter contributions disclosed through the MSRB repository of G-37 disclosure forms, political contribution activity for 2007 through 2009 as reported by Evans (2012), and the California Secretary of State’s Cal-Access system.

Governments can be frequent issuers of municipal debt and issuances, especially refundings and short-term borrowings, may be unrelated to recent referenda. To ensure that we attribute bond election campaign contributions to the appropriate bond issues, we examined the official statement for each bond issuance in our sample, which details whether the bonds are associated with a specific referendum. In an effort to establish the most credible comparison group, we limit the bond sample to the more than 1,300 GO bond issues of school districts and community colleges from 2007 to 2012. School districts represent nearly 90% of the issues. Occasional missing data reduce the sample size to just more than 1,100 bond issues.

We do not include, or review, campaign contributions made prior to 2007. Our intended counterfactual is bonds that did not receive associated ballot election campaign contributions, but the actual counterfactual includes some bond issuances that may have received contributions prior to 2007. The potential omission of contributions for the comparison group will understate any effect size of contribution activity; furthermore, this group of bond issues and referenda are likely small. We address this limitation by also using a sample restricted to only government issues that followed referenda for which we reviewed election contribution information. The occasional practice of underwriters not reporting an underwriter’s discount for one series within a bundled issuance (often accompanying Federally taxable bonds issued at the same time as tax-exempt bonds) means that we cannot link the specific bond characteristics to the related underwriting fees in those cases. We remove such bundled issuances from our sample. California is a suitable research setting based on its centrality in the phenomenon under study, its prominent usage of direct democracy, its history as a dominant issuer of municipal debt, its 2013 revival of legislation to ban the practice of underwriter financing of bond referenda, and the public availability of its municipal bond issuance cost data.

Our research design assumes that school districts whose associated issue committees receive election contributions are not fundamentally different from school districts whose associated issue committees do not receive contributions. To test whether differences between these two groups exist, we calculated difference of means t-tests on an extensive series of school district and referenda characteristics. In short, underwriters were less likely to give to school district referenda facing a higher supermajority requirement (the two-thirds threshold), more likely to contribute to referenda with lower median home values in the district, and less likely to donate to committees supporting rural referenda. Interestingly, the enrollment growth rate is also significantly different—referenda for districts with flat enrollment were more likely to receive contributions than referenda for districts with growth. One explanation for this is that the need for capital investment is more obvious in a growing district when compared with a stable or flat district. The credit quality differences related to such differential enrollment growth, geographic location, and median home values are incorporated into our model, as described later, through the included credit ratings. These limited differences suggest that the two groups are generally comparable.

Dependent Variables

For the issuance cost hypothesis, our dependent variable is the underwriter’s discount, a term used to describe the spread between the purchase “price the underwriter pays the issuer for the bonds and the price at which the bonds are reoffered to the investing public” (Temel, 2001, p. 86). The primary model uses the underwriter’s discount in logged dollars. Alternately, we convert the underwriter’s discount to a percentage of the bond issue’s par amount, mirroring the common usage of basis points to measure and discuss costs. A secondary dependent variable uses the total debt issuance costs, which include the costs for financial advisors, credit ratings, credit enhancement, bond counsel, and trustee services, in addition to the underwriter’s discount.

Independent Variables

Our primary explanatory variables of interest measure underwriter campaign contributions to an issue committee supporting a bond referendum. The first is a simple dichotomous measure indicating that an associated underwriter contribution was made. We use the aggregated dollar amount of the underwriter contributions associated with that bond issuance as a second measure. If underwriters pass along the election contributions they make to issuers as an expense, it is unclear whether contributions would be recouped in the first debt issuance associated with the successful referendum or spread over all of the future election-related issues. An additional specification only codes the first referendum-related issuance with associated contributions to better understand the relationship between contributions and underwriter fees.

Control Variables

We include an array of standard bond issue and issuer characteristics that are expected to influence the costs of bringing the debt to market. Again, we note that our sample is limited to school district and community college GO bond issues to construct the most appropriate counterfactual for the issues with election contributions.⁵ We follow Robbins and Simonsen (2013), who model total bond issuance costs. The control variables can be broken into a series of categories. These variables and their data sources are included in Table 2.

Table 2.

Variable Sources and Descriptive Statistics for Issuance Cost Models, 2007-2012.

		All observations (N = 1,127)		Underwriter contributions (n = 192)		No underwriter contributions (n = 935)
Variables	Source	M	SD	M	SD	M	SD
Dependent variables
Underwriter’s discount (US$000)	CDIAC	217.8	498.57	209.34	251.28	219.54	535.5
Underwriter’s discount (b.p./100)^a	CDIAC	0.96	0.62	1.07	0.59	0.94	0.62
Independent variables
Underwriter contribution (US$000)^a	MSRB, Evans (2012), and Cal-Access	2.70	7.41	15.86	10.67	0.00	0.00
Principal amount (US$000)^a	CDIAC	33,600	88,600	25,000	35,000	35,400	95,900
Years to maturity^a	CDIAC	21.22	9.34	23.17	10.57	20.82	9.02
Refunding (1,0)^a	CDIAC	0.34	0.48	0.13	0.34	0.39	0.49
Callable (1,0)^a	CDIAC	0.73	0.45	0.64	0.48	0.75	0.44
Taxable (1,0)^a	CDIAC	0.07	0.25	0.11	0.32	0.06	0.23
Competitive sale (1,0)^a	CDIAC	0.15	0.36	0.00	0.00	0.18	0.39
Combination structure (1,0)^a	CDIAC	0.52	0.50	0.70	0.46	0.48	0.50
Serial structure (1,0)^a	CDIAC	0.40	0.49	0.20	0.40	0.44	0.50
Term structure (1,0)	CDIAC	0.08	0.27	0.09	0.29	0.07	0.26
Insured (1,0)^a	CDIAC	0.53	0.50	0.44	0.50	0.55	0.50
Underwriter 1 (1,0)^a	CDIAC	0.25	0.44	0.41	0.49	0.22	0.42
Underwriter 2 (1,0)	CDIAC	0.21	0.40	0.26	0.44	0.19	0.40
Underwriter 3 (1,0)^a	CDIAC	0.08	0.28	0.20	0.40	0.06	0.24
Underwriter 4 (1,0)	CDIAC	0.06	0.23	0.06	0.24	0.06	0.23
Underwriter 5 (1,0)	CDIAC	0.09	0.28	0.06	0.23	0.10	0.29
S&P not rated (1,0)	CDIAC	0.00	0.07	0.02	0.12	0.00	0.05
S&P natural AAA (1,0)^a	CDIAC	0.02	0.14	0.00	0.00	0.03	0.16
S&P AA (1,0)^a	CDIAC	0.36	0.48	0.44	0.50	0.35	0.48
S&P A (1,0)	CDIAC	0.06	0.23	0.08	0.28	0.05	0.22
S&P BBB (1,0)	CDIAC	0.00	0.03	0.00	0.00	0.00	0.03
Bond anticipation note (1,0)^a	CDIAC	0.07	0.25	0.13	0.33	0.05	0.23

Note. CDIAC = California Debt and Investment Advisory Commission; MSRB = Municipal Securities Rulemaking Board.

Difference of means t-test between bond issues with associated issue committees receiving and not receiving underwriter contributions are significant at the 5% or better level.

First, general bond characteristics include the par amount, years to maturity, refunding bond status, callable bond status, taxable bond status, and method of sale. The structure of the bond and whether the issue is a bond anticipation note are included because they influence the underwriter workload.⁶ We use month-year fixed effects as a granular time trend to capture market characteristics at the time of issue.⁷ As GO bonds, credit ratings are included as a series of dichotomous variables to represent the credit quality of the issuing government. The omitted credit rating category is insured-AAA, which is the most common in the sample. Local and regional underwriting firms are more likely to make election contributions and past research finds that local firm underwriting costs are lower than those of national firms (Butler, 2008). To ensure that the localness of contributing firms is not confounded with the effects of contributions, underwriter fixed effects are included for the contributing firms. This eliminates concerns that the contributing firms are simply high-cost underwriters.⁸

Results

Election contributions were made to successful campaigns by underwriters in 118 instances where the debt was subsequently issued during our study period. In some cases, multiple contributions were made to the same issue committee by the same underwriting firm. For the purposes of the analysis, such contributions are aggregated. The average contribution amount was just more than US$15,000. Underwriter contributions averaged US$3.44 per affirmative vote cast in elections. The mean underwriter’s discount is US$216,286, or just less than 1% of the mean bond principal amount. The bond referenda campaigns collectively supported by underwriters sought approval for more than US$15 billion in debt issuance.

While the previously discussed data limitations keep us from modeling the relationship between issue committee campaign contributions and election outcomes, we use descriptive statistics, presented in Table 3, to demonstrate changes in average outcomes by underwriter contribution status. Table 3 presents the pass rate and average vote share of the school district referenda during the sample period with the columns ordered from the highest to the lowest observed pass rate and vote share for each group. Elections accompanied by large underwriter contributions, meaning greater than the mean contribution in total dollar terms, have statistically significantly higher passage rates and average vote shares than those without large contributions as reported in column 1. This relationship can also be seen in Figure 1, which plots the referendum vote share by underwriter contribution amount. The relationship between these contributions and election outcomes can take two main forms. First, large dollar amounts are given primarily to issue committees by underwriters when easy passage of the election is expected and the degree of certainty about receiving the associated fees is high. Alternately, the larger absolute dollar amounts may be sufficient to raise the capacity of the issue committee to an effective level regardless of the size of the voter population or proposed bond amount.

Table 3.

California School District Referenda Outcomes and Issue Committee Underwriter Contribution Status.

	Issue committee underwriter contribution status
California school district referenda outcomes	Large contribution (total US$)	Received contribution	All referenda	No contribution	Large contribution (US$ per voter)
Percent passed	90***	81	79	76	69**
Average vote share (%)	64*	63	63	63	60*
Number of observations	60	135	279	144	42

Note. A “large contribution” is defined as those greater than the mean contribution amount (either on a total or per voter basis, respectively). Statistical significance is based on difference of means t-tests.

p < .10. **p < .05. ***p < .01.

Figure 1.

Underwriter contributions and referendum vote share.

The pass rate and average vote share decline continuously, relative to the large contribution group, as the sample is defined as all referenda receiving contributions, all referenda, and referenda not receiving underwriter contributions. In column 5 of Table 3, the pass rate and average vote share are presented for referenda issue committees that were recipients of large contributions measured on a per voter basis. Contrary to the previous relationship with large total contributions, outcomes are significantly worse for these recipients than for other referenda. We interpret this relationship as preliminary evidence that underwriters will give more resources, in per voter terms, when passage at the ballot box is uncertain. This is consistent with de Figueiredo et al.’s (2011) finding that contributions and campaign spending increase with lower odds of passage. Supporting this perspective that underwriters differ contribution levels based on the probability of electoral success is the fact that contributions were made to issue committees for half of the referenda facing the 55% supermajority passage threshold (of which 82% ultimately passed) and only to a quarter of those referenda faced with the higher two-thirds passage threshold (of which only 35% ultimately passed). Furthermore, Figure 2 displays a positive relationship between the size of the proposed bond offering and the amount of giving, suggesting that contribution amounts are influenced by potential compensation.

Figure 2.

Underwriter contributions and referendum bonding authority requested.

In sharp contrast to the general similarity, we observe between school districts that receive and do not receive election contributions, the subsequent bond structures of the two groups differ tremendously (see descriptive statistics in Table 2). We consider this an important finding, because there is no simple reason why the election strategy related to a referendum should be correlated with the structure of the resulting bonds. It suggests that underwriters in these negotiated sales are structuring deals in fundamentally different ways after making a contribution. The bond issues associated with election contributions have longer maturities, are more likely to have serial and term bond elements, are more often taxable, and, as previously noted, are exclusively found coupled with the negotiated method of sale.

We focus our results discussion on the underwriter election contribution variables, but the control variables, generally, perform as expected. Table 4 includes the abridged results; the full model results are included in the appendix (Tables A1 and A2). The size of the bond issue is directly related to the underwriter’s discount in dollars but inversely related to the discount in basis points reflecting the economies of scale in debt issuance. Longer maturities, lower credit quality, and bond anticipation notes are associated with higher underwriting costs. Refundings, competitive sales, and the term bond structure are related to lower underwriting costs. In accordance with the existing literature (Butler, 2008), the local and regional underwriting firms responsible for the election contributions provide significantly cheaper underwriting services than non-contributing firms. This allows us to now consider the effects of election contributions on costs controlling for existing differences in underwriter pricing.

Table 4.

Results of Underwriter Contributions on Underwriter’s Discount, 2007-2012.

	(1)	(2)	(3)	(4)	(5)	(6)
	Underwriter contribution (1,0)	Underwriter contribution (US$000)	Underwriter contribution—First issue only (1,0)	Underwriter contribution—First issue only (US$000)	Underwriter contribution (1,0)	Underwriter contribution (US$000)
DV = Underwriter’s discount (ln US$)	0.120** (0.050)	0.007*** (0.003)	0.104* (0.055)	0.009*** (0.003)	0.239*** (0.074)	0.010** (0.004)
DV = Underwriter’s discount (basis points/100)	0.120* (0.062)	0.006** (0.003)	0.148** (0.071)	0.011*** (0.003)	0.238** (0.094)	0.009** (0.005)
Bond issues included	All school district and community college GO bond issues				Restricted sample—Verified non-contribution issuers
N	1,127	1,127	1,127	1,127	454	454

Note. First issues are considered to be the first referendum-related bond issuance following an election where the underwriter was also a contributor to the referendum issue committee. Full regression details for primary specifications are included in Tables A1 and A2. DV = dependent variable.

p < .10. **p < .05. ***p < .01.

As seen in Table 4, the dichotomous and dollar measure of election contributions are statistically significant and positively associated with the underwriter’s discount regardless of the form of the dependent variable. The average increase in the underwriter’s discount when a contribution was made is 12.75%, or US$27,576 based on the mean underwriter’s discount in the sample and the estimated coefficient on the election contribution indicator variable. The average underwriter contribution for those governments that issued bonds in the sample is just more than US$16,128, suggesting that every dollar contributed to a successful campaign results in a higher underwriting fee of around US$1.70. This return of contributions is taken as preliminary evidence that issue committee contributions are passed through to issuing governments and taxpayers despite state laws limiting such use of public funds. Related election contributions are associated with costs that are 12 basis points higher in the alternate basis point model.

Interpreting the dollar value form of the campaign contribution variable suggests that a US$1,000 increase in contribution amount is associated with a 0.66% increase in the average underwriter’s discount. The effect size is understated in this specification because the full underwriter contribution amount is used as the independent variable for each related issuance regardless of the share of bonding authority previously used. While we do not have a basis for believing that underwriters allocate their contribution costs across issues using any uniform method, the operationalization of contribution amounts overstates the size of the contributions by counting them multiple times. An unreported specification uses an alternate underwriter contribution independent variable comprised of the total contribution amount multiplied by the percent of the referendum-authorized amount being issued with the current bonds (rather than just using the full contribution amount for each). The estimated coefficient for this contribution variable increases fourfold relative to the reported contribution amount specification, suggesting that the presented association between contributions and underwriter costs is quite conservative.

The notion that underwriting firms recoup higher percentages of their election contributions in the first issuance associated with a referendum is generally supported by the specifications reported in columns 3 and 4 of Table 4, but the higher costs persist over the repeated issuances. The concern that our counterfactual led to an understatement of the effect size is borne out in the specification that restricts the sample to issues of governments that passed referenda since 2007 where we could verify election contributions (columns 5 and 6 of Table 4). The coefficient on the dichotomous election contribution indicator nearly doubles and the contribution amount coefficient increases by more than 45%. While the primary policy concern is that election contribution costs are passed on to issuers, it is interesting to note that the lower average costs of using a contributing underwriter (based on the average coefficient of −.224) are erased by the higher costs associated with receiving a contribution (.239).

There are a number of potential threats to the validity of our findings. The higher underwriter’s discount found in the presence of contributions may simply compensate for lower total issuance costs due to extra services being provided by the underwriter. These might even include legitimate bond election services such as polling (Chen, 2012). For example, the negotiated sale may reduce the need to incur some financial advisory fees. When we substitute the total issuance costs for our underwriter’s discount dependent variable, we find that the effect size of underwriter contributions remains statistically significant and actually grows in magnitude to an increase in costs of 17% based on the contribution indicator coefficient.⁹ This suggests that increased underwriter fees are not offset by reductions in other issuance costs and that unobserved contributions by other financial intermediaries may also affect the full range of issuance costs. Although they are now exempt from MSRB ballot contribution disclosure rules (Jensen, 2012b), recent MSRB and Securities and Exchange Commission actions suggest that greater scrutiny may be paid to the future contribution activity of financial advisors (O’Neill, 2013).

Although we have visibility into the underwriter’s discount, we do not have access to the other component of the underwriter’s total compensation, namely, “the gain (less the loss) from trading” of the bonds (Raineri, Robbins, Simonsen, & Weaver, 2012). There is no a priori reason to believe that aftermarket activity differs for issuances associated with contributions, but we lack visibility into this possibility and the all-in costs of the bond issues. Although the controls for bond characteristics are extensive, they do not include a number of factors that may influence the underwriter’s discount such as the use of capital appreciation bonds, the use of a syndicate (multiple underwriters), and underwriter prestige. As a relationship-intensive industry, the role of prior experiences between the issuer and contributing underwriter may lead to a scenario where the underwriter is selected based on favoritism rather than cost considerations. While it is possible that political contributions are more likely under such circumstances, the data suggest that the firms making contributions are largely low-cost firms in the absence of contributions.

The trend in interest cost studies is to attempt to control for the endogeneity of the method of sale using a selection model. Our sample is already limited to school district and community college GO bonds where 85% of issues use the negotiated method of sale. Excluding the competitive issues actually increases the magnitude of the election contribution coefficients. For these reasons and our focus on underwriter discounts and election contributions, we do not use a method of sale selection model.

Previously, we considered the similarities and limited differences between school districts with referenda issue committees receiving and not receiving election contributions. The small number of significant differences bolsters confidence in our results, but we also use instrumental variables (IVs) regression to validate that underwriters are not simply contributing to issue committees on the basis of unobservable school district characteristics that are correlated with higher-cost underwriting demands. We limit our observations to those bond issues associated with a referendum that took place during the period for which we observe political contributions. With this limited sample, we then instrument for the receipt of contributions using (a) the failure of the district’s prior referendum, (b) the district’s median home value, (c) rural locale, and (d) the number of school districts within the county. Election date fixed effects and additional election-related variables that are not exogenous are also included in both stages of the estimation. The predicted values of the contribution indicator from the first stage are used in place of the actual contribution indicator variable in the second-stage OLS estimation.

The estimated coefficient on the instrumented contribution variable remains positive and statistically significant for this sample. The coefficient roughly doubles in magnitude (.41 with p = .036) from the most comparable analysis (Model 5 in Table 4). This finding further reduces concerns over bias in our estimates, but we prefer our existing specifications because (a) they are more conservative than the IV results, (b) they are more generalizable due to the reduced sample size of the IV sample, and (c) post-estimation tests suggest that the receipt of an underwriter contribution is not endogenous, meaning that OLS is more efficient.

Conclusion

At all levels of government, public managers and elected officials are generally restricted from supporting election campaigns with public resources. In the case of legislative referenda, the public stakeholders responsible for putting a policy question on the ballot must play a neutral role when acting in their official capacity. We have documented a system where private money supports public goals as regulatory provisions simultaneously restrict direct private giving to elected officials and public support for election campaigns. We find that post-election fees paid to firms that contribute to campaigns are significantly higher than for non-contributors. From a management perspective, the behavior reflects lingering concerns over pay-to-play practices in the public finance field, while raising serious policy questions about circumvention of state and local regulations restricting the use of public resources in election campaigns. The findings add to the growing body of research that considers how policy outcomes are shaped, especially in the context of direct democracy by money and extends this work to the local level.

The average rate of return on the election contributions is a conservative US$1.70 per dollar given. However, the increase in fees appears to be largely offset by the typically lower costs of doing business with one of the donating underwriters. Despite the increases in the underwriter’s discount, we believe that it is the indirect payment of campaign contributions with public funds that is most troubling, along with the commonly held notion that these election contributions and services are tied directly to securing the underwriting work. The empirical results presented here demonstrate quid pro quo behavior in a direct democracy setting. However, if the legal restrictions on using public funds for campaign spending are overlooked, then the higher issuance costs may be seen as a small price to pay for the political advertising needed to inform ballot election voters. Again, it is important to note that this advertising is not wholly educational because it advocates for a single position (the passage of bond referenda).

The importance of these findings extends beyond the municipal bond market. Garrett and Gerber (2001) note that, “The Supreme Court has generally maintained that direct democracy is inherently free of quid pro quo corruption and has therefore been hostile to laws regulating contributions to and expenditures in issue campaigns” (p. 74). Our findings may not generalize to all forms of direct democracy, but they do suggest that the behavior deserves additional scrutiny. Whether the public value of providing private support to pass referenda trumps the integrity of existing public spending restrictions is a serious question to be decided by elected officials and voters alike. Some states, such as Missouri, have already banned the practice (Evans, 2012). Other states, such as California, continue to attempt to address such election contributions legislatively (Jensen, 2013) and through other actions (Weikel, 2013).

Regulation of the municipal securities industry has proven challenging over time. The appropriate level and complexity of disclosure remains uncertain. For example, are state or national disclosure requirements necessary or preferred? Should annual financial statements provide stakeholders with sufficient information to analyze potential pay-to-play relationships? With regard to MSRB regulations on bond election campaign disclosure, such requirements to publicly report bond election contributions do not currently apply beyond underwriters to other public finance professionals such as financial advisors and bond counsel, which makes transparency more challenging to the public (Jensen, 2012b). Expanded ballot election contribution disclosure requirements were recently adopted by the MSRB (MSRB, 2013b). The changes are expected to help gather the necessary information to gauge the need for additional regulation of bond election contributions. Finally, we reiterate that this behavior is not limited to the finance industry and other sectors that provide contributions, such as the construction industry, are less formally regulated—if at all. Our results here capture the influence of only a small fraction of the contributions made to issue committees in local elections. If political favoritism, based on election contributions, guides the selection of private sector services, then, in aggregate, the higher costs for taxpayers may be quite large. As public resources remain scarce and voter sentiment toward increased government spending ebbs and flows, the use of private funds for such campaign support is only likely to grow in the absence of stakeholder action.

Footnotes

Appendix

Acknowledgements

The authors acknowledge the organizations that provide open access to the municipal bond, election, and campaign finance data used to complete this study including the California Debt and Investment Advisory Commission (CDIAC), the California Elections Data Archive (CEDA), the California Secretary of State’s Office, and the Municipal Securities Rulemaking Board (MSRB).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Todd L. Ely is an Assistant Professor in the School of Public Affairs at the University of Colorado Denver. His research and teaching address the financing of state and local public services, education finance and policy, and public and non-profit financial management.

Thad D. Calabrese is an Assistant Professor in the Robert F. Wagner Graduate School of Public Service at New York University. His research and teaching focus on public financial management, with a focus on the interconnectedness of the government and non-profit sectors in providing goods and services.

References

Beaumont Unified School District. (2009, December 10). Election of 2008 General Obligation Bonds Series A and Series B Official Statement. Retrieved from http://emma.msrb.org/

Butler

A. W.

(2008). Distance still matters: Evidence from municipal bond underwriting. Review of Financial Studies, 21, 763-784.

Butler

A. W.

Fauver

Mortal

(2009). Corruption, political connections, and municipal finance. Review of Financial Studies, 22, 2873-2905.

Chen

(2012). California Local Agency General Obligation Bond Cost of Issuance 2009-2011 (CDIAC No. 12-08, CDIAC Research Unit). Sacramento: California Debt and Investment Advisory Commission. Retrieved from http://www.treasurer.ca.gov/cdiac/publications/issuance.pdf

de Figueiredo

J. M.

C. H.

Kousser

(2011). Financing direct democracy: Revisiting the research on campaign spending and citizen initiatives. Journal of Law, Economics, and Organization, 27, 485-514.

Evans

(2012, May 14). Critics struggle to end “pay to play” in school bonds. California Watch. Retrieved from http://californiawatch.org/dailyreport/critics-struggle-end-pay-play-school-bonds-16160

Fairman

Lawrence

(2009, August 7). Re: MSRB’s proposed changes to MSRB Rule G-37 and request for comments. Piper Jaffray. Retrieved from http://msrb.org/Rules-and-Interpretations/RegulatoryNotices/2009/~/media/Files/RFC/ 2009/2009-35/PiperJaffrayCo.ashx

Gabaldon

T. A.

(2009). Financial federalism and the short, happy life of municipal securities regulation. Journal of Corporation Law, 34, 739-769.

Garrett

Gerber

(2001). Money in the initiative and referendum process: Evidence of its effects and prospects for reform. In Waters

M. D.

(Ed.), The battle over citizen lawmaking (pp. 73-95). Durham, NC: Carolina Academic Press.

10.

Inman

R. P.

(1978). Testing political economy’s “as if” proposition: Is the median income voter really decisive? Public Choice, 33(4), 45-65.

11.

Jensen

(2012a, January 17). Brokers’ gifts that keep giving. The Bond Buyer, p. 2.

12.

Jensen

(2012b, May 29). Some California FAs use pay-to-play tactics, critics say. The Bond Buyer, p. 1.

13.

Jensen

(2013, May 9). California school bond election bill passes committee. The Bond Buyer. Retrieved from http://www.bondbuyer.com/issues/122_89/california-passes-legislation-to-ban-financial-firms-from-school-election-1051440-1.html

14.

Jordan

J. B.

(1999). Regulation of pay-to-play and the influence of political contributions in the municipal securities industry. Columbia Business Law Review, 1999, 489.

15.

Kriz

K. A.

(2003). Comparative costs of negotiated versus competitive bond sales: New evidence from state general obligation bonds. The Quarterly Review of Economics and Finance, 43, 191-211.

16.

Lawrence

D. M.

(1988). Use of public funds in bond referendum campaigns. School Law Bulletin, 19(3), 9-10.

17.

Lemov

(1990, June). For municipal bonds, it’s not a plain vanilla world anymore. Governing, pp. 52-58.

18.

Leonard

(1996). An empirical analysis of competitive bid and negotiated offerings of municipal bonds. Municipal Finance Journal, 17(1), 37-67.

19.

Market statistics archive, bond referenda results, 2007-2011. (2013). The Bond Buyer. Retrieved from http://www.bondbuyer.com/marketstatistics/msa_main.html

20.

Mikesell

J. L.

(2007). Fiscal administration: Analysis and applications for the public sector (7th ed.). Belmont, CA: Thomson Wadsworth.

21.

Municipal Securities Rulemaking Board. (2013a). Creation of the MSRB. Retrieved from http://www.msrb.org/pdf.aspx?url=http://www.msrb.org/About-MSRB/About-the-MSRB/Creation-of-the-MSRB.aspx

22.

Municipal Securities Rulemaking Board. (2013b). MSRB notice 2013-09: SEC approves amendments to require the public disclosure of additional information related to dealer contributions to Bond Ballot Campaigns under MSRB Rules G-37 and G-8, April 1.

23.

Olinger

Gorski

(2012, March 4). Uneasy bonds to build on. Denver Post, p. 1A.

24.

O’Neill

K. M.

(2013, February 14). Notice of filing of a proposed rule change relating to amendments to MSRB Rules G-37 and G-8 and Form G-37. Federal Register, 78(31), 10656.

25.

Opp

(2005). Ending pay-to-play in the municipal securities business: MSRB Rule G-37 ten years later. University of Colorado Law Review, 76, 243.

26.

Orrick

Herrington

Sutcliffe

L. L. P.

(2010). Campaign finance and political conduct rules for school district bond elections. Retrieved from http://www.orrick.com/Events-and-Publications/Pages/campaign-finance-and-political-conduct-rules-for-school-district-bond-elections-2566.aspx

27.

Peng

Brucato

P. F.

(2003). Another look at the effect of method of sale on the interest cost in the municipal bond market—A certification model. Public Budgeting & Finance, 23, 73-95.

28.

Peng

Kriz

K. A.

Neish

(2008). Method of sale in the municipal bond market. In Feldstein

S. G.

Fabozzi

F. J.

(Eds.), The handbook of municipal bonds (pp. 51-67). Hoboken, NJ: John Wiley & Sons.

29.

Petersen

J. E.

Doty

R. W.

Forbes

R. W.

Bourque

D. D.

(1977). Searching for standards: Disclosure in the municipal securities market. Duke Law Journal, 1976, 1177-1204.

30.

Piper Jaffray. (n.d.). Contributions to bond elections. Retrieved from http://www.piperjaffray.com/2col_largeright.aspx?id=1293

31.

Raineri

Robbins

Simonsen

Weaver

(2012, August). Underwriting, brokerage, and risk in municipal bond sales. Paper presented at the 2012 Municipal Finance Conference at Brandeis University, Boston, MA.

32.

Retnasaba

(2006). Do campaign contributions and lobbying corrupt? Evidence from public finance. Journal of Law, Economics & Policy, 2, Article 145.

33.

Robbins

M. D.

Simonsen

(2007). Competition and selection in municipal bond sales: Evidence from Missouri. Public Budgeting & Finance, 27(2), 88-103.

34.

Robbins

M. D.

Simonsen

(2013). Municipal bond new issue transaction costs. Public Budgeting & Finance, 33(1), 1-24.

35.

Saskal

(2010, August 23). California Panel Nixes Underwriter Ban Bill. The Bond Buyer, p. 1.

36.

Senate Local Government Committee. (2010, August 9). Bill analysis: SB 623. Retrieved from http://totalcapitol.com/?bill_id=200920100SB623

37.

Simonsen

Hill

(1998). Municipal bond issuance: Is there evidence of a principal-agent problem? Public Budgeting & Finance, 18(4), 71-100.

38.

Smith

D. A.

Tolbert

(2004). Educated by initiative: The effects of direct democracy on citizens and political organizations in the American states. Ann Arbor: University of Michigan Press.

39.

Steers

(1998, June 4). A helping handout. Denver Westword. Retrieved from http://www.westword.com/1998-06-04/news/a-helping-handout/

40.

Sturman

(2009). Beaumont USD selects Bogh for Measure Z projects. Beaumont Unified School District. Retrieved from http://www.beaumont-ca.schoolloop.com/file/1244873688119/1229223312936/8468370641150501912.pdf

41.

Temel

J. W.

(2001). The fundamentals of municipal bonds (5th ed.). New York, NY: John Wiley & Sons.

42.

U.S. Advisory Commission on Intergovernmental Relations. (1993). State laws governing local government structure and administration. Washington, DC: U.S. Government Printing Office.

43.

U.S. Census Bureau. (2012). 2012 Census of governments: Organization component preliminary estimates. Retrieved from http://www.census.gov/govs/cog2012/

44.

Vijayakumar

(1995). An empirical analysis of the factors influencing call decisions of local government bonds. Journal of Accounting and Public Policy, 14, 203-231.

45.

Walsh

M. W.

(2009, January 9). Nationwide inquiry on bids for municipal bonds. The New York Times, Section A, p. 1.

46.

Weikel

(2013, August 7). County bars bond sales to donors. Los Angeles Times, p. AA5.

47.

Yes on Z for Safe Schools. California Form 460: Recipient committee campaign statement, multiple dates (between September 8, 2008, and June 30, 2009). Retrieved from http://nf4.netfile.com/pub2/Default.aspx?aid=CTRIV

48.

Zimmerman

J. F.

(2001). The referendum: The people decide public policy. Westport, CT: Praeger Publishers.