Guiding the Public Sector: Assessing the Use of Recommended Practices in the Budgeting Process

ABSTRACT

Public administration is a field defined by principles, best practices, and recommended guidelines. According to advocates, these guidelines have the potential to improve the effectiveness and efficiency of public organizations. However, this potential is predicated on the degree to which they are actually followed in practice. Using an original survey of local budgeting officials across the United States, this study examines the correlation between current budgeting practices and established, recommended practices. While a significant number of current practices closely match those that have been recommended, notable variations were present. The trends and their implications conclude the assessment.

KEYWORDS:

• Budgeting
• guidelines
• management
• practices
• recommendations

Notes

^1. The GFOA was the focus of the examination of e-government and its impact on fiscal accountability (Justice et al., 2006), but no attention was given to other recommendations associated with the GFOA or the NACSLB.

^2. These activities represent only a small fraction of the efforts made by the GFOA to improve the budgeting and financial capabilities of state and local governments. A detailed history of the organization, its activities, and accomplishments can be found in the April 2006 issue of Government Finance Review, which celebrates the 100th anniversary of the GFOA. Particularly informative articles address the GFOA and its role in public budgeting (Hildreth, 2006), accounting (Freeman & Allison, 2006) and pension management (Miller, 2006).

^3. Every effort was made to ensure the recommendations selected for the survey adequately described their larger practices. The reader is invited to examine the statements in subsequent tables and compare them with the full collection of recommendations contained in the framework at http://www.gfoa.org/services/nacslb/.

^4. In order to obtain the most accurate information from respondents, the initial welcome message and instructions highlighted the confidential nature of some of the responses. Because respondents were asked to provide the name of their municipality and their title, the message was designed to ensure respondents that their name, government, and contact information would not be revealed or connected to their individual responses.

^5. Derived from U.S. Census data, all local governments in the United States with a population greater than 10,000 were placed into STATA statistical software and the program’s random sample generator was utilized to obtain the sample.

^6. Of the 1000 local governments included in the sample, the principal contact method was a direct e-mail to the primary budgeting or finance official. Approximately 82% of respondents were contacted in this manner. When this contact information was not available, other methods were utilized as needed. As a consequence, 3.4% of governments received the survey link through a departmental e-mail address, 5.1% through an imbedded electronic form specifically for the budget or finance department, other officials such as administrators, clerks, and treasurers received the link in approximately 3% of the cases and, when no electronic methods were available, paper surveys were sent to 4.2% of the local governments in the sample. Of those that returned completed surveys, approximately 89% received the survey link in a direct e-mail, nearly 3% received a departmental e-mail, an additional 3% were contacted through a departmental electronic form, less than 2% were other government officials, and less than 2% received a paper survey. To ensure the contact method did not impact responses, difference of proportion tests were used to examine those who received the survey via direct e-mail and those who received it through one of the other methods. Using the aggregate categories of agree/somewhat agree and disagree/somewhat disagree, only four of the 38 practices revealed statistically significant differences. The mean difference among the four practices was 0.159. Given the small number of instances and generally low differences, the impact on overall trends and conclusions appears limited.

^7. The margin of error, therefore, is 5.9% for the final sample used in the analysis.