Forecasting School Districts’ Revenues

and Expenditures

Nick Warner

November 5, 2015

Forecasting Districts’ Revenues and Expenditures2

Forecasting Overview

Forecasting Districts’ Revenues and Expenditures3

What is forecasting?

• Predicting the future as accurately as possible, given all of the information available, including historical data and knowledge of any future events that might impact the forecasts

• Many methods - Forecasters use the one that best fits the known data.

Forecasting Districts’ Revenues and Expenditures4

Why is forecasting challenging?

Forecasting Districts’ Revenues and Expenditures5

Why is forecasting challenging?

• A forecast must be based on what we know at the time, and things could change.

• Some data is predictably variable or volatile, so any forecast includes a range of possible amounts.

Forecasting Districts’ Revenues and Expenditures6

Why would a school district use forecasting?

Forecasting Districts’ Revenues and Expenditures7

Why would a school district use forecasting?

• Inform long- and short-term budget, capital, and staffing plans– Are expected revenues sufficient to

cover expected expenditures under current conditions?

– Are expected revenues sufficient to implement new improvement initiatives?

– Is the district positioned to handle changes in student population?

– And more...

Forecasting Districts’ Revenues and Expenditures8

How can we forecast data important to school districts?

Forecasting Districts’ Revenues and Expenditures9

How would a school district forecast?• Do it yourself, based on known data sources

and info about your district– Methods range from very basic to extremely

advanced, but all will provide very useful insights and help inform decisions and planning.

• Get help from experts

Warning: A forecast should be used to inform decisions, not dictate budget

planning or decisions. Budget planning and forecasting should remain separate.

Forecasting Districts’ Revenues and Expenditures10

Property Tax Revenue

• Importance: Primary local funding source for operations (with the exception of certain systems that have access to sales taxes for operations)

• Data Sources: Digest values, observed home sales, info from property tax assessor’s office

• Volatility: Low, relatively easy to forecast– Exception: The Great Recession created declining

statewide property tax digests.

• Method to Try: Linear trend forecast (ordinary least squares)

Forecasting Districts’ Revenues and Expenditures11

Property Tax Revenue

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 $500

$700

$900

$1,100

$1,300

$1,500

$1,700

$1,900

$2,100 District A

Tax Year

Ne

t M

&O

Dig

est

in $

MIl

lion

s

Linear trend line

Forecasting Districts’ Revenues and Expenditures12

Property Tax Revenue

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 $500

$700

$900

$1,100

$1,300

$1,500

$1,700

$1,900

$2,100

District A With Moving Average Forecast

Tax Year

Ne

t M

&O

Dig

est

in $

MIl

lion

s

Two period moving average

Which Line fits the data better? otherwise known as “goodness

of fit”

Forecasting Districts’ Revenues and Expenditures13

Sales Tax (ESPLOST)

• Importance: A major fund source for capital projects– Districts enter into five year plans based in part on

expected collections

• Data Sources: Sales tax collections (GDOR) and information on community business activity

• Volatility: High, forecast results in a wide range– Seasonality and other trends generally are

incorporated into forecast.

• Method to Try: An observed average over a long period of time.

Forecasting Districts’ Revenues and Expenditures14

Sales Tax (ESPLOST)

1-Ja

n-99

1-Ja

n-00

1-Ja

n-01

1-Ja

n-02

1-Ja

n-03

1-Ja

n-04

1-Ja

n-05

1-Ja

n-06

1-Ja

n-07

1-Ja

n-08

1-Ja

n-09

1-Ja

n-10

1-Ja

n-11

1-Ja

n-12

1-Ja

n-13

1-Ja

n-14

1-Ja

n-15

40

50

60

70

80

90

100

110

Chart Title

Month - Year

ES

PLO

ST

Re

ve

nu

e $

in

T

ho

usa

nd

s

Sample District, 1999-2015

1999-2015 monthly average

Goodness of fit? Not bad and its just the long

term monthly average

Forecasting Districts’ Revenues and Expenditures15

Sales Tax (ESPLOST)

1-Ja

n-99

1-Ju

l-99

1-Ja

n-00

1-Ju

l-00

1-Ja

n-01

1-Ju

l-01

1-Ja

n-02

1-Ju

l-02

1-Ja

n-03

1-Ju

l-03

1-Ja

n-04

1-Ju

l-04

1-Ja

n-05

1-Ju

l-05

1-Ja

n-06

1-Ju

l-06

1-Ja

n-07

1-Ju

l-07

1-Ja

n-08

1-Ju

l-08

1-Ja

n-09

1-Ju

l-09

1-Ja

n-10

1-Ju

l-10

1-Ja

n-11

1-Ju

l-11

1-Ja

n-12

1-Ju

l-12

1-Ja

n-13

1-Ju

l-13

1-Ja

n-14

1-Ju

l-14

1-Ja

n-15

1-Ju

l-15

40

50

60

70

80

90

100

110

Sample District 1999-2015

Month Year

ES

PLO

ST

Re

ve

nu

e in

$T

ho

usa

nd

s

One Year Moving Average

Forecasting Districts’ Revenues and Expenditures16

Sales Tax (ESPLOST) – Helpful Data• Counties governments might be producing

sales tax forecasts, even within their budget documents, which can help districts predict their portion of the sales tax revenues.

• Speak to local business owners and parents.

• Georgia budget documents forecasts state sales tax revenues into the future. Check and see how well state sales tax collections track with your counties.

Forecasting Districts’ Revenues and Expenditures17

Number of Students

• Importance: Drives state funding (QBE), expenditures, staffing decisions...

• Data Sources: Previous student counts, data on <5 y.o. children in your area, and historic drop out and migration info

• Volatility: Low– Based on steady drop outs, in and out migration

• Method to Try: Linear trend (OLS) to start – move to age specific calculations improve forecast

Forecasting Districts’ Revenues and Expenditures18

Number of Students

fall 2006 fall 2007 fall 2008 fall 2009 fall 2010 fall 2011 fall 2012 fall 2013 fall 2014 fall 20158,600

8,800

9,000

9,200

9,400

9,600

9,800

10,000

District A

School Year

Fa

ll F

TE

Co

un

t

Looks like a good fit and would forecast steadily

declining student counts

Maybe we could dig a little deeper?

Forecasting Districts’ Revenues and Expenditures19

Number of Students

fall 2006 fall 2007 fall 2008 fall 2009 fall 2010 fall 2011 fall 2012 fall 2013 fall 2014 fall 2015600

650

700

750

800

District A

First Grade Second Grade Third Grade Forth Grade

Increasing early grades

Followed by Increasing older

Grades

Should we expect steadily declining student counts?

Forecasting Districts’ Revenues and Expenditures20

Number of Students – Helpful Data• Georgia Office of Planning and Budget

provides population by age group forecasts for all of the counties in Georgia

• Other publicly available datasets do the same by race.

Forecasting Districts’ Revenues and Expenditures21

Forecasting State Funding – QBE

• Predicating future student population, knowledge of proposed funding formula changes, and awareness of state budget shortfalls are the ways to forecast future QBE/State funding.

Forecasting Districts’ Revenues and Expenditures22

State Funding – QBE

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014$6.75

$7.25

$7.75

$8.25

$8.75

$9.25

$9.75

$10.25

$10.75

$11.25

Schoolyear

QB

E F

orm

ula

Ea

rnin

gs $

in

Mil

lio

ns

Forecasting Districts’ Revenues and Expenditures23

What about expenditures?

• Forecasting expenditures – Based on revenue and student population forecasts

• Other changes would be based on district decisions on curriculum and improvement efforts.

Forecasting Districts’ Revenues and Expenditures24

Questions?

Contact InformationNick Warnernwarner2@gsu.edu