An Open Systems Model of Local Government Forecasting

• Published date: 01 July 2018
• Date: 01 July 2018
• DOI: 10.1177/0275074017692876
• Subject Matter: Articles

692876ARPXXX10.1177/0275074017692876American Review of Public AdministrationReitano
research-article2017
Article
American Review of Public Administration
2018, Vol. 48(5) 476 –489
An Open Systems Model of Local
© The Author(s) 2017
Reprints and permissions:
Government Forecasting
sagepub.com/journalsPermissions.nav
https://doi.org/10.1177/0275074017692876
DOI: 10.1177/0275074017692876
journals.sagepub.com/home/arp
Vincent Reitano1
Abstract
Local governments use single source forecasts to inform decision making, which can constrain their ability to prepare for
and respond to financial uncertainty. This unique context may have increased the challenges faced by governments at the
local level through economic downturns such as the Great Recession. Given this concern, which has yet to be addressed
in the literature, this article develops an open systems model of local government forecast accuracy, which can be tested
across any type of local government. This article tests the model with a panel of special purpose governments at the local
level, specifically school districts in Pennsylvania, from 2003 through 2013. Estimation of the model with longitudinal analysis
shows that government forecasters at the local level consider internal and external factors when forecasting own-source
and intergovernmental revenue streams. In particular, a mix of institutional, financial, and political factors are associated with
forecast accuracy. Forecasters at the local level also considered the role of economic shocks, as evidenced by decreased
expectations for own-source revenue through the Great Recession. Collectively, these findings demonstrate that they
consider a complex and multifaceted information set, which includes both internal and external determinant factors of
forecast accuracy at the local level. These factors can prove critical to increasing forecast accuracy in context of the financial
uncertainty experienced through the Great Recession.
Keywords
forecasting, financial management, strategic planning
Introduction
At the state level, forecasting is conducted by a variety of
agencies and even external organizations and researchers,
State and local governments face increased financial uncer-
leading to multiple forecasts which can be compared through
tainty in the aftermath of the Great Recession, which has
the budgeting process (Brogan, 2013). In contrast, local gov-
challenged the accuracy of revenue forecasting (Pew ernments often have less external involvement and transpar-
Charitable Trusts, 2015). At the state level, volatility in own-
ency in their forecasting process, with decision makers
source and intergovernmental revenue has led to decreased
typically accepting a forecast from a single source (Frank &
forecasting accuracy by more than 10% through the Great
Zhao, 2009; Wildavsky, 1986). If a single source is utilized,
Recession, which was the highest measured decline in 20
the inaccuracy of a forecast may increase if other forecasts
years (Pew Charitable Trusts, 2015). State government
are not considered, compounding the challenges of preparing
finances ultimately rebounded by 2011 as revenue sources
for uncertainty (Voorhees, 2002, 2006). This concern
became more stable. Comparatively, the effect of this period
becomes magnified when considering that local revenue
of fiscal stress on local government revenue has been more
sources tend to be less elastic than at the state level, which
drawn out, with historical declines in property tax collection
restricts flexibility in budgeting through a period such as the
not occurring until the collapse of the housing market in
Great Recession (Alm, 2013). As a result, the ability of local
2010 (Gordon, 2012; Harris & Shadunsky, 2013; Jonas,
governments to respond to a fiscal crisis may be constrained
2012). The prolonged decline in property tax revenues
if a single source annual forecast is used (Frank & Zhao,
reflects the “new normal” local governments face (Ammons,
2009; Schroeder, 1982).
Smith, & Stenberg, 2012; Martin, Levey, & Cawley, 2012),
The unique context of forecasting at the local level may
although there remains little systematic investigation of how
have contributed to the uncertainty local governments faced
this trend has affected government forecasters. Inquiry into
forecasting can prove critical to understanding how local
1North Carolina State University, Raleigh, NC, USA
governments prepare for uncertainty in revenues while meet-
ing the demands of constituents.
Corresponding Author:
As a critical function of the budget process, forecasting
Vincent Reitano, School of Public and International Affairs, North
Carolina State University, 212 Caldwell Hall, Campus Box 8102, Raleigh,
can provide estimates of revenues and expenditures, which
NC 27695, USA.
can be utilized in the planning process (Frank & Zhao, 2009).
Email: vcreitan@ncsu.edu

---

Reitano
477
through the Great Recession. McDonald (2015) provides the
Gorr, Grizzle, & Klay, 1989). Within the open system, unique
case of Saint Joseph County, Indiana, which forecasted rev-
factors are associated with the forecast accuracy of both
enues higher than actually received through the Great
own-source revenue forecasts and intergovernmental reve-
Recession, severely constraining their ability to pay for
nue forecasts. Fourth, the article suggests that forecasters
expenditures on legally binding contracts. Special purpose
considered the uncertainty from the 2007-2009 financial cri-
governments such as school districts also faced a similar
sis in their forecasts, which provides further evidence that
concern, given the challenges they experienced in respond-
forecasting is a best practice for short-term government
ing to lower than anticipated revenues through the recession
financial management (Government Finance Officers
(Bruck & Miltenberger, 2013). Despite these cases, which
Association [GFOA], 1998) and long-term strategic plan-
reflect the concerns associated with using a single source
ning (International City/County Management Association,
forecast in local government, research has largely focused on
2015; Kavanagh, 2007).
state government forecasting accuracy and the broader ques-
Practitioners can use the model developed in this article to
tion of forecasting methodology (McDonald, 2015; Sun,
understand the internal and external factors surrounding
2005). This leaves the concern of which factors are associ-
forecasting, which can provide a critical link between plan-
ated with local government forecasting. Addressing this
ning for unpredictable events and also reducing uncertainty
research question requires development and testing of a
at the local level (Makridakis & Taleb, 2009). In addition, the
model of local government forecasting.
article presents a discussion of the limitations associated
This article develops an open systems model of local gov-
with a common measure of forecast accuracy, in addition to
ernment forecasting accuracy. The model suggests that sin-
alternative measurers emerging in the literature (Tofallis,
gle source government forecasts consider internal financial
2015). Collectively, these contributions can advance the
and institutional factors in addition to external political
understanding of forecasting practice at the local level, which
demands and economic trends across two separate revenue
is deemed as being “methodologically backward relative to
streams: own-source and intergovernmental revenue. To test
private sector or state and federal benchmarks” and subject
the model, a panel dataset of 499 school districts in
to limited research inquiry (Frank, Gianakis, & McCue,
Pennsylvania from 2003 through 2013 is used. As a special
2005, p. 559).
purpose government at the local level, school districts are
What follows in the second section is the development of
susceptible to the effects of financial uncertainty and also
an open systems model of local government forecasting and
play an integral role in local government service delivery;
presentation of testable hypotheses. Discussion of the data,
therefore, school districts provide a relevant context to test
variables, and methodology is presented in the “Data and
the open systems model of local government forecasting,
Method” section. The fourth section provides an empirical
although the model can be tested with any type of local gov-
test of the model and discusses results. Implications for
ernment (see Duncombe & Hou, 2014; Reschovsky, 2004).
research on local government forecasting and policy recom-
Estimation of the model provides evidence that special pur-
mendations are presented in the “Conclusion” section.
pose government forecasters at the local level are responsive
to internal and external factors through the years prior to,
during, and after the Great Recession. In addition, the model
Theoretical Framework
shows that forecasters consider external events such as the
Streams of Government Forecasting Literature
Great Recession and stabilization funding from the American
Recovery and Reinvestment Act (ARRA).
The literature on state and local government forecasting has
This article advances the literature on local government
followed two streams (Bretschneider & Schroeder, 1985;
forecasting on four fronts. First, the article develops the
Mikesell & Ross, 2014; Sun, 2005). The first stream has
understanding of the...