1 BLGF LGU Revenue Forecasting System

BLGF LGU Revenue Forecasting System: 

BLGF LGU Revenue Forecasting System Prepared by Norman R. Ramos LGU Finance Advisor, ADB TA 7451 Support to Local Government Financing ADB TA 7451 Based on models and procedures developed under ADB TA 4556 and ADB TA 7451 11/15/2011 BLGF LGU Revenue Forecasting System 1

Topics: 

Topics Existing LGU Revenue Forecasting Efforts The BLGF Fiscal Capacity Model as a Component of the Integrated Framework for LGU Comparative Performance Assessment The BLGF LGU Revenue Forecasting System: An Overview BLGF Central Office Fiscal Capacity Model BLGF Fiscal Capacity Model: Revenue and Expenditure Elasticity Estimates The Concept of Elasticity Setting the Values of Explanatory Variables Major Findings in the Development of the BLGF Fiscal Capacity Mode Annual Forecast Error Correction and Periodic Review of Parameters BLGF Long-term LGU Fiscal Capacity Simulation Model 11/15/2011 BLGF LGU Revenue Forecasting System 2

Topics: 

Topics The LGU Revenue Forecasting Template Structure Input Sheets Quarterly Weights Sheet Initial Forecast Sheets Final Forecast Sheets Overall LGU Revenue Target Setting Process BLGF Region-based LGU Revenue Target Review and Reconciliation Process 11/15/2011 BLGF LGU Revenue Forecasting System 3

Existing LGU Revenue Forecasting Efforts: 

Existing LGU Revenue Forecasting Efforts 11/15/2011 BLGF LGU Revenue Forecasting System 4

Existing LGU Revenue Forecasting Efforts: 

Existing LGU Revenue Forecasting Efforts Revenue forecasting for Philippine LGUs has largely been “ ad-hoc ” exercises Using either compound growth rate techniques or just judgmental estimates based on the past year’s performance, LGU treasurers come up with revenue forecasts used in the preparation of the annual budget of the LGU. The Bureau of Local Government Finance (BLGF) in recent years has adopted an annual revenue forecasting exercise on the regional level. Regional revenue targets are set for four (4) key LGU revenue sources — real property tax ( RPTax ), business tax ( BTax ), fees and charges (F&C), and income from economic enterprises ( EcoEnt ). The revenue targets are based on revenue elasticities with respect to Gross Regional Domestic Product (GRDP) and on GRDP target growth rates set by the NEDA. 11/15/2011 BLGF LGU Revenue Forecasting System 5

Existing LGU Revenue Forecasting Efforts: 

Existing LGU Revenue Forecasting Efforts The targets for individual LGUs are “negotiated” at the regional level between the BLGF regional office and the individual LGU treasurers. The National Economic and Development Authority (NEDA) with technical assistance from the Asian Development Bank (ADB) under TA 4512, has come in 2007 with a 4-volume set of planning guidelines that include: 1) Development Planning and Physical framework Plan Guidelines; 2) Investment Programming and Revenue Generation Guidelines; 3) Budgeting and Public Expenditure Management (BPEM) Guidelines and 4) Project Evaluation Guidelines . The BPEM covers the preparation of revenue forecasts at the LGU level. 11/15/2011 BLGF LGU Revenue Forecasting System 6

Existing LGU Revenue Forecasting Efforts: 

Existing LGU Revenue Forecasting Efforts Under ADB TA 4556, a formal revenue and expenditure forecasting approach has been developed utilizing available BLGF budget operating statement (BOS) data for 1991 to 2000 and statement of income and expenditures (SIE) data from 2001 to 2005 as well as 2001 to 2006 data from the Commission on Audit (COA). Under the current ADB TA 7451, the central office model is being programmed and an accompanying EXCEL template is b eing prepared for use by LGUs. The developed model is intended to generate forecasts for the current operating revenue and expenditure items of BLGF’s improved Statement of Receipts and Expenditures (SRE) 11/15/2011 BLGF LGU Revenue Forecasting System 7

The BLGF Fiscal Capacity Model as a Component of the Integrated Framework for LGU Comparative Performance Assessment: 

The BLGF Fiscal Capacity Model as a Component of the Integrated Framework for LGU Comparative Performance Assessment 11/15/2011 BLGF LGU Revenue Forecasting System 8

Integrated Framework for LGU Comparative Performance Assessment: 

Integrated Framework for LGU Comparative Performance Assessment 11/15/2011 BLGF LGU Revenue Forecasting System 9 Overall LGU Performance – LGPMS of DILG Governance/Administration Constituency Welfare Financial Performance LGU Financial Performance LGFPMS of BLGF Revenue Mobilization Expenditure Management Debt & Investment Capacity Overall Financial Management Major Components of Framework LGU Income Classification LGU Financial Performance Assessment Overall LGU Performance Assessment

Integrated Framework for LGU Comparative Performance Assessment: 

Integrated Framework for LGU Comparative Performance Assessment The LGU income classification component “ pre-sorts ” LGUs by political level and by income class to make the application of the performance measures “fair and meaningful”. The LGFPMS “ statically ” assesses the fiscal performance of LGUs vis-à-vis benchmarks for each of the fiscal performance indicators appropriate for each political level and corresponding income classes within each political level. Parallel to the fiscal performance indicator assessment is the fiscal capacity assessment using the fiscal capacity model that develops a prognosis across time into the future of the potential fiscal performance of the LGU. 11/15/2011 BLGF LGU Revenue Forecasting System 10

Integrated Framework for LGU Comparative Performance Assessment: 

Integrated Framework for LGU Comparative Performance Assessment The results of both static fiscal indicator analysis and fiscal capacity projections are combined and serve as the bases of the LGU debt capacity certification and the LGU creditworthiness rating: The fiscal capacity projections provide an estimate of what the LGU can borrow , and this serves as a basis of what is traditionally certified by BLGF as the debt capacity; The creditworthiness rating system will assess the appropriate LGU creditworthiness rating - best, high, good, medium, below medium and speculative. This will then be translated into a set of recommended proportions of the maximum borrowing capacity as determined by the debt capacity projections; and Applying the appropriate proportion on the LGU maximum borrowing capacity will yield what the LGU should borrow . It will be stressed that this is just an advisory or guide figure and not a BLGF recommendation. 11/15/2011 BLGF LGU Revenue Forecasting System 11

Integrated Framework for LGU Comparative Performance Assessment: 

Integrated Framework for LGU Comparative Performance Assessment Basic Premise in integrating LGFPMS with LGPMS DILG’s LGPMS puts BLGF’s LGFPMS as part of the administration module. The basic premise in combining the financial performance indicators with the service delivery indicators is that improved LGU financial performance is not the goal per se but should be translated to improved constituency welfare via improved service delivery . Integration of LGFPMS with LGPMS Integration of the proposed LGFPMS “in toto ” (broken arrow from LGFPMS to Administration component of LGPMS) using the proposed benchmarks Integration of the creditworthiness ratings as computed by the LGU creditworthiness rating system. (Broken arrow from creditworthiness rating scheme to Administration component of LGPMS 11/15/2011 BLGF LGU Revenue Forecasting System 12

The BLGF LGU Revenue Forecasting System: An Overview: 

The BLGF LGU Revenue Forecasting System: An Overview 11/15/2011 BLGF LGU Revenue Forecasting System 13

Components of the BLGF LGU Revenue Forecasting System: 

Components of the BLGF LGU Revenue Forecasting System SRE-format Current Revenue and Current Operating Expenditure Target Setting Model – computer-programmed and made part of the BLGF Loan Approval and Credit-worthiness System LGU revenue target setting Debt certification and creditworthiness assessment Filling up of missing entries in LGU SRE submissions LGU EXCEL-based revenue forecasting template for generating annual and quarterly revenue forecasts for real property taxes, business taxes, fees and charges, and income form economic enterprises. Regional-based reconciliation process for the development of a single LGU set of revenue forecasts to be used for investment programming and budgeting purposes. Simple long-term simulation model for total LGU revenues and expenditures was estimated to explore long-term trends in LGU revenues and expenditures. Simulating probable long-term trends in LGU total revenues and expenditures. Model’s structure can in the future be applied on a more disaggregated analyses, e.g., RPT, general public services, etc. 11/15/2011 BLGF LGU Revenue Forecasting System 14

SRE-format Current Revenue and Current Operating Expenditure Target Setting Model: Overview: 

SRE-format Current Revenue and Current Operating Expenditure Target Setting Model: Overview 2 major components: the current operating revenue block, and the current operating expenditure (including debt service) block. The item-level formats for both revenue and expenditure follows that of the SRE. The revenue block combines econometric as well as simple elasticity approaches to develop annual revenue forecasts at the LGU level, by revenue item that is then subjected to “ negotiations ” (if necessary) at the BLGF regional level prior to adoption as formal annual LGU revenue targets. 11/15/2011 BLGF LGU Revenue Forecasting System 15

SRE-format Current Revenue and Current Operating Expenditure Target Setting Model: Overview: 

SRE-format Current Revenue and Current Operating Expenditure Target Setting Model: Overview The current operating expenditure block forecasts current operating expenditure (excluding debt service) at the LGU level based on elasticities of each expenditure item , by LGU type , and by specific LGU with respect to current operating revenues . Debt service is calculated using three linear econometric equations for each LGU type — province, city, and municipality— relating debt service (financial expenses) in Year t to the outstanding debt of the LGU in Year (t-1). The BLGF revenue targeting system/process seeks to reconcile the “ top-to-bottom ” BLGF central office forecasts with “ bottom-up ” local treasurer forecasts, and is expected to support the systematic generation of local revenue forecasts that will be owned and utilized by LGUs in the preparation of their annual budgets . 11/15/2011 BLGF LGU Revenue Forecasting System 16

BLGF Long-term LGU Fiscal Capacity Simulation Model: An Overview: 

BLGF Long-term LGU Fiscal Capacity Simulation Model: An Overview 11/15/2011 BLGF LGU Revenue Forecasting System 17 Based on a chaos-type structure Chaos-type equations are deterministic systems – exhibit regular and predictable behavior - that also exhibit seemingly irregular, random and even turbulent behavior and are highly sensitive to initial conditions. The election variable introduces a cyclical element into both the revenue and expenditure equations. The GDP variable acts as an overall proxy for the total tax base of LGUs. Introducing a quadratic GDP term introduces a “ hump ” — a limit to how much of the tax base can be captured by the LGU. Changes or variations in the value of variables like prices have been observed to be proportional to the square root of time. Similar to the expenditure forecast model, revenue at time t is fed back into the expenditure equation and determines expenditure at time t along with the election dummy.

BLGF Central Office Fiscal Capacity Model: 

BLGF Central Office Fiscal Capacity Model 11/15/2011 BLGF LGU Revenue Forecasting System 18

The BLGF Fiscal Capacity Model: 

The BLGF Fiscal Capacity Model 11/15/2011 BLGF LGU Revenue Forecasting System 19 The BLGF model basically forecasts at the LGU level the current operating revenue and expenditure items contained in the SRE. Missing actual SRE data will be temporarily replaced by the appropriate model-estimated values as against the former “reactivation” system in the old SIE. From such forecasts, the net operating income can be calculated and inputted into BLGF’s Loan Approval and Debt Monitoring Module

The BLGF Fiscal Capacity Model: 

The BLGF Fiscal Capacity Model 11/15/2011 BLGF LGU Revenue Forecasting System 20 Based on econometrically estimated elasticities utilizing time series and panel BOS and SIE data. Elasticities distinguish between LGU types – province, city, municipality. Elasticities are estimated for each major revenue category – RPT, business taxes, etc. based on SRE categories. Elasticities of each LGU category are estimated with respect to the following variables: Revenue base, i.e., property values, GDP Lagged values Occurrence of elections Interest rate National government budgetary deficits Periodic Cycles Time trend Individual LGU growth rates, by revenue and expenditure item is derived from that of the growth rates of each LGU category using elasticities of each revenue and expenditure item, by LGU with respect to that of the LGU category to which they belong. Annual forecast error–correcting mechanism .

The BLGF Fiscal Capacity Model: 

The BLGF Fiscal Capacity Model 11/15/2011 BLGF LGU Revenue Forecasting System 21

The BLGF Fiscal Capacity Model: Revenue Forecasting Module: 

The BLGF Fiscal Capacity Model: Revenue Forecasting Module Step 1 : Forecast annual growth rates for each revenue category in the Statement of Receipts and Expenditures (SRE), e.g., real property tax, business tax, other taxes, fees and charges, etc., and for each LGU type — province, city, and municipality, excluding Inter-Local Transfers based on estimated elasticities econometrically estimated from available BOS and SIE data from 1991 to 2005. Step 2 : Forecast the annual growth rates for each revenue category by individual LGU . The calculation utilizes individual LGU revenue elasticities , by revenue category, with respect to the LGU type to which they belong calculated from SIE 2001 to 2005 data. Step 3 : Apply the annual growth rates for each revenue category, by individual LGU, on the actual base year (time = t) LGU revenue estimates as stored in the SRE to come up with the forecast revenue in year t+1. The forecast in year t+1 becomes the base year for forecasting t+2, and so on, for multi-year forecasts . For Inter-Local Transfers , the forecasting process is as follows: Determine the LGU type , e.g., province, city, municipality, and its income class within the type to which it belongs, e.g., 1st class province, 2nd class city, 3rd class municipality. [1] Calculate the expected value of the inter-local transfer that the LGU will probably receive in forecast year t. This is done by multiplying the probability of the LGU receiving inter-local transfers based on its type and income class by average inter-local transfer received by an LGU for the LGU type to which it belongs. The forecasting process iterates across time , e.g., t+2, t+3, etc., to arrive at a set of multi-year revenue targets. [1] Either the existing income classification system based on total revenues or the proposed real per capita locally-sourced revenue classification system may be used . 11/15/2011 BLGF LGU Revenue Forecasting System 22

The BLGF Fiscal Capacity Model: Expenditure Forecasting Module: 

The BLGF Fiscal Capacity Model: Expenditure Forecasting Module Except for debt service (financial expenses), all other current operating expenditure items are determined by total current operating revenue. Step 1 : Forecast annual growth rates for each expenditure category in the Statement of Receipts and Expenditures (SRE), e.g., General Public Service, Health ,Nutrition and Population Control, labor and Employment etc., and for each LGU type — province, city, and municipality, excluding Debt Service (Financial Expenses). The calculation is based on estimated elasticities econometrically estimated from available BOS and SIE data from 1991 to 2005. Step 2 : Forecast the annual growth rates for each expenditure category by individual LGU . The calculation utilizes individual LGU expenditure elasticities , by revenue category, with respect to the LGU type to which they belong calculated from SIE 2001 to 2005 data. Step 3 : Apply the annual growth rates for each expenditure category, by individual LGU, on the actual base year (time = t) LGU expenditure estimates as stored in the SRE to come up with the forecast expenditure in year t+1. The forecast in year t+1 becomes the base year for forecasting t+2, and so on, for multi-year forecasts. For Debt Service (Financial Expenses) , the model utilizes three (3) econometric equations estimated from Year 2004 and Year 2005 COA data relating debt service in Year t of LGU i to outstanding debt of LGU i in Year t-1 . 11/15/2011 BLGF LGU Revenue Forecasting System 23

BLGF Fiscal Capacity Model: Revenue and Expenditure Elasticity Estimates: 

BLGF Fiscal Capacity Model: Revenue and Expenditure Elasticity Estimates 11/15/2011 BLGF LGU Revenue Forecasting System 24

All Provinces: Revenue Elasticities: 

All Provinces: Revenue Elasticities 11/15/2011 BLGF LGU Revenue Forecasting System 25

All Cities: Revenue Elasticities: 

All C ities: Revenue Elasticities 11/15/2011 BLGF LGU Revenue Forecasting System 26

All Municipalities: Revenue Elasticities: 

All Municipalities: Revenue Elasticities 11/15/2011 BLGF LGU Revenue Forecasting System 27

Projection Parameters: Inter-local Transfers: 

Projection Parameters: Inter-local Transfers 11/15/2011 BLGF LGU Revenue Forecasting System 28

Projection Parameters: Current Operating Expenditures: 

Projection Parameters: Current Operating Expenditures 11/15/2011 BLGF LGU Revenue Forecasting System 29

The Concept of Elasticity: 

The Concept of Elasticity 11/15/2011 BLGF LGU Revenue Forecasting System 30

Concept of Elasticity: 

Concept of Elasticity What is an elasticity: Measurement of the percentage change in one variable (dependent) that results from a 1% change in another variable (explanatory). Applying the concept to LGU revenue sources Measurement of the percentage change in a particular revenue source , e.g., RPT with respect to a variable affecting its movement e.g., value of assessed taxable property in the previous year, elections, etc.,. The same goes for expenditure item. For example, an RPT elasticity with respect to property value during the previous year of 0.57 indicate that for every percentage increase in assessed property value during the previous year, RPT revenues will increase by 0.57% Size of Elasticities Elasticity > 1, elastic: Growth rate of revenue > growth rate of revenue base Elasticity < 1, inelastic: Growth rate of revenue < growth rate of revenue base Elasticity = 1, unitary: Growth rate of revenue = growth rate of revenue base 11/15/2011 BLGF LGU Revenue Forecasting System 31

Estimation of Elasticity: 

Estimation of Elasticity Elasticity Estimates for Each Revenue and Expenditure Item, by LGU Type – Province, City, Municipality – with respect to Explanatory Variables The relevant elasticities are estimated by fitting a multivariate logarithmic function Y = α + Β 1 * ln X 1 + Β 2 * ln X 2 +, …,+ B n ln X n using multiple regression analysis on the paired time series and cross-section data for each revenue item, by LGU type and where the variables are expressed in terms of natural logarithms ( ln ). The partial slope coefficients ( Bs ) of the estimated multiple regression equations for each revenue item and for each LGU category measures the elasticity (% change) of the revenue item for each LGU type with respect to a % change in each of the explanatory variables, e.g., gross value added in real estate, gross domestic product, etc . 11/15/2011 BLGF LGU Revenue Forecasting System 32

Estimation of Elasticity: 

Estimation of Elasticity Elasticity Estimates for Each Revenue and Expenditure Item, by Individual LGU with respect to the LGU Type to which it belongs Calculate the % annual growth rates of the total revenue from each revenue item k, e.g. RPT, Business Tax, etc., (AGRTR) of LGU type j, e.g. province, city, municipality, in the appropriate period 2001-2005 in this case. For 2001-2005, AGRTRkj = EXP (Ln(TCORkj05/TCORkj01)-1)/100 Using the same formula, compute the annual growth rate of revenue item k for each LGU i ( AGRTRki ) for the same year. For each LGU i , calculate the ratio ( AGRTRki )/ AGRTRkj . This ratio represents the elasticities . The LGU elasticities for each province with respect to all provinces as a group for each revenue item is shown in the following slide. A elasticity > 1, indicates that for that particular revenue item, the province is growing faster than all provinces as a group. A elasticity < 1, indicates that for that particular revenue item, the province is growing at a slower rate than all provinces as a group. A elasticity = 1, indicates that for that particular revenue item, the province is growing at a rate equal to that of all provinces as a group . 11/15/2011 BLGF LGU Revenue Forecasting System 33

Estimation of Elasticity: 

Estimation of Elasticity 11/15/2011 BLGF LGU Revenue Forecasting System 34

Data Categorization Types: 

Data Categorization Types 11/15/2011 BLGF LGU Revenue Forecasting System 35 Time Series Data : also sometimes referred to as longitudinal data. Presents the data for a particular variable, e.g., RPT, Business Tax, etc across time periods, e.g. years, months, etc. For example, the revenue and expenditure elasticities for each LGU category – province, city, municipality - were estimated from BOS and time series data from 1991 to 2004-5 Cross-section Data : Presents the data for a particular variable across categories, e.g., provinces, or income class, etc. at a particular point in time, e.g., 1992. For example, the expenditure elasticities , by expenditure item and by LGU type was estimated from 2005 SIE cross-section data. Panel Data : Presents cross-section data for multiple time periods. For example, the SIE data for 2001 to 2005 or the FIES for 1994, 1998 and 2000 can be considered as panel data. Note that intervals between cross-sections need not be regular. For example, the estimation of the LGU-level revenue elasticities with respect to that of the LGU category to which they belong made use of SIE panel data from 2001 to 2004.

Setting the Value of Explanatory Variables: 

Setting the Value of Explanatory Variables 11/15/2011 BLGF LGU Revenue Forecasting System 36

Values of Explanatory Variables: 

Values of Explanatory Variables 11/15/2011 BLGF LGU Revenue Forecasting System 37

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 38 38 Explanatory Variable Setting Tips Growth Rate of GDP at Current Prices Get from the National Statistical Coordination Board (NSCB) web page the relevant actual values. For the required projections get from the National Economic and Development Authority (NEDA) web page or from the BSP web page. If not yet available, use the following regression equation to generate preliminary forecasts : Where Ln GDP is the natural log of GDP, TIME is a time trend variable with an initial value of 0 in 1991, 1 in 1992, 3 in 1993 and so on, and the subscript t-1 represents the past year. Outputs of the equation are GDP estimates in mill. Php derived by taking the antilog ( Exp ) of LN GDP. For example, the GDP est. in 2007 of 6,635,341 = Exp (15.71) Growth rates have to be computed as (GDP t / GDP t-1 )-1. For example, the growth rate in 2007= (6,635,341/6,013,241)-1

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 39 39 Explanatory Variable Setting Tips Growth Rate of REGVA at current Prices Get from the National Statistical Coordination Board (NSCB) web page the relevant actual values. For the required projections get from the National Economic and Development Authority (NEDA) web page or from the BSP web page. If not yet available, use the following equation to generate preliminary forecasts: Where REGVA G. Rate is the annual % growth rate of REGVA and GDP G. Rate is the annual % growth rate of GDP. For 2007, REGVA A. G. Rate = 1.01146* 10.3%= 10.5%

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 40 40 Explanatory Variable Setting Tips Growth Rate of Average Treasury Bill Rate Get from the DOF their forecasts regarding the trend in the weighted average treasury bill rate. If not yet available, use the following equation to generate preliminary forecasts: Where Ln ATBR is the natural log of the weighted average treasury bill rate, Ln GDP is the natural log of GDP and TIME is a time trend variable with an initial value of 0 in 1992, 1 in 1993, 3 in 1994 and so on. Outputs of the equation are weighted average t-bill rates in % derived by taking the antilog (Exp) of Ln ATBR. For example, the ATBR est. in 2007 of 6.64% = Exp (1.89) The Ave. T-Bill G. Rate = ATBR t – ATBR t-1. For example, in 2007 the annual g. rate of the T-bill rate = T-bill rate in 2007- T-bill rate in 2006 = 6.64% - 6.99% = 0.35%

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 41 41 Dummy and Lagged SRE Variables Election Year : Assumes a value of 1 every election year and 0 for all other years. Budget Deficit : Assumes a value of 1 every time a huge budgetary deficit is expected. Periodicity : Other Receipts (for provinces), assumes a value of 1 every 10 th year beginning 2004. No such dummy variables for cities and municipalities. Extraordinary receipts and aids (for the 3 LGU types), assumes a value of 1 every 6 th year beginning 2002. LGC DUM : Extraordinary receipts and aids (for provinces) value of 1 starting 1993 to 2004 and thereafter. All the lagged explanatory variables , e.g. past year’s growth rates of RPT, business tax, and extraordinary receipts, the average inter-local transfer per LGU, and past year’s outstanding debt per LGU are available from the SRE database or from the annual COA report.

Major Findings in the Development of the BLGF Fiscal Capacity Model: 

Major Findings in the Development of the BLGF Fiscal Capacity Model 11/15/2011 BLGF LGU Revenue Forecasting System 42

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 43 43 Major Findings of the Modeling Work For real property taxes (RPT), both provinces and municipalities have revenue to tax base elasticities that are less than 1 — 0.20 for provinces and 0.33 for municipalities — indicating their failure to fully capture the taxes due from the growing real property base proxied by the gross value added in real estate (REGVA). The cities are more aggressive in capturing the taxes due from their rapidly growing real property bases with an elasticity of 1.70 . For business-related taxes , all LGU types are lagging behind in capturing the taxes due from their growing business tax base with elasticities with respect to the proxy tax base — Gross Domestic Product (GDP) — all less than 1 — 0.30 for provinces , 0.56 for cities , and 0.23 for municipalities . For fees and charges , the performance of the three LGU types are relatively close to each other with respect to the proxy tax base — Gross Domestic Product (GDP) of 0.34 for provinces , 0.44 for cities , and 0.46 for municipalities . For revenues from economic enterprises , the performance of the three LGU types are relatively close to each other with respect to the proxy tax base — Gross Domestic Product (GDP) of 0.40 for provinces , 0.46 for cities , and 0.48 for municipalities .

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 44 44 Major Findings of the Modeling Work Current year collection for real property and business taxes are positively affected by the previous year’s collection . This could be partly attributable to the target-setting practice of Treasurers of just applying what they consider as achievable growth factors on the previous year’s collection. There is a three-year revenue cycle associated with election years. Election years negatively affect major local revenue sources of LGUs: Provinces — real property tax, business tax, fees and charges and economic enterprises; Cities — economic enterprises, other receipts; Municipalities — real property tax, business tax, fees and charges and economic enterprises, other receipts. Other receipts react positively to the average Treasury bill rate , which could be indicative of the usual practice of LGUs investing excess cash in interest-earning deposits as well as treasury bills.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 45 45 Major Findings of the Modeling Work Shares from the National Tax Collection including the IRA are negatively affected by the budgetary deficit of the national government. The 1992 LGC had a positive and permanent effect on extraordinary receipts and aids for provinces. There is a positive upward trend across time in the collection of three (3) major LGU revenue sources — economic enterprises , fees and charges , and s hares from the national tax collection . Across LGU classes, the probability of an LGU receiving inter-local transfers goes up as its income class goes down .

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 46 46 Major Findings of the Modeling Work LGU current operating expenditures are primarily determined by their revenue levels. General public services are growing at almost the same rate as revenues for all three (3) LGU categories. Other Purposes closely follows probably indicative of the relatively loose public expenditure management practices among LGUs. This is followed by economic services . The elasticities for health , nutrition and population control for provinces and cities is higher than that of municipalities probably because both LGU types operate hospitals while municipalities do not. The elasticities for social services are about the same for the three LGU categories. Relatively poorer LGUs pay higher financial expenses for their borrowings. Municipalities pay an average of 10.7% per annum, provinces , 9.6% and cities – the most affluent – 9.1%.

Annual Forecast Error Correction and Periodic Review of Parameters: 

Annual Forecast Error Correction and Periodic Review of Parameters 11/15/2011 BLGF LGU Revenue Forecasting System 47

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 48 Annual Forecast Error Correction Annual forecast errors at the LGU level for years t+1, t+2,..,t+n will be corrected when actual SRE data corresponding to the forecast values in year t+1 becomes available via a three-step process. 1) The original LGU-level forecast for Year t+1 for each revenue and expenditure item k for each LGU will be replaced by the actual values to become the base for the error-corrected forecast for Year t+2 to n. 2) Each original forecast growth rate for each revenue and expenditure item k for each LGU for years t+2 to n will be multiplied by the following error correction factor to come up with an error-corrected forecast growth rates for years t+2 to n. 3) For each LGU and for each revenue and expenditure item k, the error-corrected forecast values for Year t+2 = Actual t+1 value * (1+Error-corrected forecast growth rate) for Year t+2. The error corrected values for Year t+3 will then be computed using the error-corrected forecast values for Year t+2 as base and error-corrected growth rates for Year t+3. The error correction process will then go on for every year up to Year t+n.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 49 49 Annual Forecast Error Correction and Periodic Review of Parameters Step (1) corrects the base of the forecasts. Step (2) corrects the slope of the forecasts. Step (3) combines the base and slope corrections to derive a corrected set of forecast values for each LGU and for each revenue and expenditure item k. This forecast error correction mechanism was built into the computer-based version of the model. The parameters of the model should be subjected to periodic review and re-estimation to take into account structural and permanent changes (such as amendments to the 1991 LGC or even the political structure) affecting LGU revenues and expenditures. Re-estimation should probably be done soon since the parameters of the model were estimated from 1991 to 2004-05-06 data and every 6 years thereafter.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 50 Re-estimation Process for Model Parameters All Revenue Items Except Extraordinary Receipts and Inter-Local Transfers Run the multiple regression equation for each LGU type j and revenue item k using the data from 1991 to 2010. Revenue item k is the dependent variable, while the others are the explanatory variables. All variables, except the dummy and the time trend variables are expressed in natural logarithms. When used in the model, the elasticity estimates for the dummy variables should be divided by 100 as they represent % responses. The estimated regression parameters represent the updated elasticities of each revenue item k for each LGU type j. Extraordinary Receipts Run a simple regression equation using cross-section SRE data for each of the LGU type j for the relevant updating year, e.g. 2010. Extraordinary receipts of LGU i is the dependent variable while ( RPTi + BusTi ) is the explanatory variable. Both variables are expressed in natural logarithms. The estimated regression parameters represent the updated elasticities of extraordinary receipts for each LGU type j.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 51 51 Re-estimation Process for Model Parameters Inter-Local Transfers Run a regression equation using cross-section (across LGU for each LGU category) SRE data for each of the LGU type j for the relevant updating year, e.g. 2010. The dummy variable for Inter-Local Borrowings (1 if LGU i received Inter-Local transfers and 0 if not) is the explanatory variable. The explanatory variable is the income classification of LGU i , e.g. 1st, 2nd , etc. Run a regression equation using time series SRE data for each of the LGU type j from 2001 to the relevant updating year, e.g. 2010. The average inter-local transfer for LGU type j in historical Year t is the explanatory variable. The explanatory variable is the average inter-local transfer for LGU type j in historical Year t-1. The estimated regression parameters are used to update the parameters used to forecast inter-local transfers.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 52 Re-estimation Process for Model Parameters The elasticity estimates used to estimate the individual LGU growth rates need to be re-estimated. Calculate the % annual growth rates of the total revenue from each revenue item k, e.g. RPT, Business Tax, etc., (AGRTR) of LGU type j, e.g. province, city, municipality, in the appropriate year, e.g. 2010. For 2010, AGRTRkj = ((TCORkj10/TCORkj09)-1)/100 Using the same formula, compute the annual growth rate of revenue item k for each LGU i ( AGRTRki ) for the same year. For each LGU i , calculate the ratio ( AGRTRki )/ AGRTRkj . This ratio represents the individual LGU elasticities with respect to the LGU type to which they belong. The expenditure elasticity estimates and debt service parameter be reviewed and if deemed needed, re-estimated.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 53 Doing Regression Analysis using EXCEL Enter in Row 1 under each relevant column each of the variable name. In the case of time series data, the 1 st column usually contains the year. In the case of cross-section data, the name of the category, e.g., LGU name like Bulacan for a province or Manila for a city or Sibulan for a municipality. In this case, there are 3 explanatory variables – REGVA(t-1), RPT(t-1) and the Elec DUM. Enter the respective values of each variable in each of the appropriate rows and columns. Since the RPT equation is of a double-log form, convert all the variables into its natural log form – LN RPT, LN REGVA (t-1), and LN RPT (t-1) – except Elec DUM using the LN function in EXCEL For example, if you want to convert Col. B, RPT, put the cursor in the appropriate cell, e.g., C2, and enter the command =LN(enterB2). Then copy the formula the rows to the last row, B15 Repeat the same process for the other variables to be converted to log form.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 54 Doing Regression Analysis using EXCEL Click Data Analysis on the Tools Menu Choose Regression . Follow the instructions: Input Y-range – the dependent variable, e.g. LN RPT – by highlighting cells D1 to D15. Note that dependent variable must only be 1 column. Input X-range – the explanatory variables, e.g. LN RPT, LN REGVA (t-1), LN RPT (t-1) and ELEC DUM – by highlighting cells E1 to G15. Select Labels if Row 1 contains variable labels just like in this example. Select Constant is Zero to force the regression line to pass through the origin. In this example, doing so, assumes that if the explanatory variables all take zero values then there will be zero growth rate for RPT. Select New Worksheet Ply The Coefficients corresponds to the elasticities used in the projections, e.g. RPT for Provinces as a group Predictive Power of Equation = % of movement in RPT between 1992 and 2005 explained by movements in the explanatory variables = 99.99% Get ratio of Standard Error to Mean value of dependent variable during the period = 0.046/7.44 = 0.62% Elasticity estimates

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 55 Doing Regression Analysis using EXCEL Use the Chart Module of EXCEL to generate a plot of actual to estimated data to show tracking power of the equation. Note that this particular equation is able to closely follow the movement of provincial RPT between 1992 and 2005 The standard error of estimate is < 1% of the mean value of RPT during the period. Key point to look at is how well does the equation forecast the turning points as the business tax equation for municipalities indicate. The periodicity is caused by the 3-year election cycle.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 56 Doing Regression Analysis using EXCEL Add an additional column (Col. H) to the spreadsheet that you used for regression analysis and label it RPT (est1). Generate the values for Col. H by multiplying the regression coefficients by the actual values for the year of LN REGV (t-1), LN (RPT(t-1) and Elec. DUM and adding the results for the year. For Cell H2, (0.2025*E2) + (0.6888*F2) + (-0.0248*G2) Fill in all the cells up to H15 in an analogous manner. Use the Chart module of EXCEL as follows: Highlight B1 to B15 and H1 to H15 Got to INSERT , choose CHART and then choose LINE For Category X-axis label , highlight A2 to A15. Fill in the appropriate X and Y axes labels Choose As new sheet to create the graph on a new sheer

BLGF Long-term LGU Fiscal Capacity Simulation Model: 

BLGF Long-term LGU Fiscal Capacity Simulation Model 11/15/2011 BLGF LGU Revenue Forecasting System 57

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 58 Historical Trend in Total LGU Revenues and Expenditures From the advent of the LGC, LGU revenues have increased in absolute value. Expenditures kept pace with revenues at a lower level.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 59 Historical Trend in the Changes in Total LGU Revenues and Expenditures Annual absolute changes have exhibited increased fluctuations (Left-hand chart) and a declining trend in terms of % changes (Right-hand chart)

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 60 % annual changes in both LGU revenues (left) and expenditures (right) from 1992 to 2004 seem to be chaotic and displays no visually discernable patterns Historical Trend in the Changes in Total LGU Revenues and Expenditures

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 61 Application of a chaos-type model structure into LGU revenues (left) and expenditures (right) data revealed that the seeming random changes can be tracked. Historical Trend in the Changes in Total LGU Revenues and Expenditures

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 62 Long-term Simulation Model for LGU Revenues and Expenditures Seemingly chaotic movements of the annual percentage changes of total LGU revenues and expenditures were statistically modeled using the following dynamic equations: [1] Revenue Equation Δ LN Rev t = -0.13470 Elec DUM + 9.88624 Δ LN GDP (t-1) – 52.04886 Δ LN GDP (t-1) 2 – 0.10152 R 2 = 0.90515 Where Rev t = total LGU revenues at time t; Elec DUM = dummy variable for election years taking a value of 1 during election years and 0 in non-election years; GDP = Gross Domestic Product; and TIME = time trend variable. All values with LN are in terms of natural logarithms. The revenue and GDP figures are at current prices. The use of the change in the logarithms of the variables in effect converts the variables to % changes [1] They are considered dynamic because of the time lag elements.

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 63 Long-term Simulation Model for LGU Revenues and Expenditures Expenditure Equation Δ LN Exp t = 0.92775 Δ LN Rev t + 0.03679 Elec DUM R 2 = 0.93293 Note that elections encourage exuberant spending as indicated by the positive (+) sign of the election variable in the expenditure equation and at the same time dampens revenue as indicated by the negative (–) sign of the election variable in the revenue equation. The 2 equations can be used to do long-term simulations of overall LGU revenues and expenditures

The LGU Revenue Forecasting Template: 

The LGU Revenue Forecasting Template 11/15/2011 BLGF LGU Revenue Forecasting System 64

LGU Revenue Forecasting Template: 

LGU Revenue Forecasting Template 11/15/2011 BLGF LGU Revenue Forecasting System 65 Based on historical moving average annual growth rates to be used by LGU treasurers to generate revenue forecasts four key local revenue sources: i ) real property taxes; ii) business taxes; iii) fees and charges; and iv) economic enterprises. Uses as inputs quarterly SRE historical data for the current year + past 3 years data and current year assessed value of taxable properties and the current tax rates. Only ii), iii) and iv) will make use of historical average growth rates. Real property taxes will be forecast using the existing tax rates and the assessed value of real property. The initial forecast of the treasurer will be discussed and agreed with the other members of the Local Finance Committee (LFC). If, necessary, the forecasts will be subjected to a reconciliation with the BLGF central office forecasts at the regional level . The template also breaks down the annual forecasts into quarterly forecasts using seasonality weights calculated from the quarterly input data.

LGU Revenue Forecasting Template: 

LGU Revenue Forecasting Template 11/15/2011 BLGF LGU Revenue Forecasting System 66 EXCEL-based Made up of 9 linked spreadsheets 2 i nput sheets – i ) Base input sheet and ii) final annual forecast sheet The sheet tabs are colored yellow . 3 outpu t sheets – i ) initial annual forecast and initial quarterly forecast sheets the tabs of which are colored gray ; and ii) final quarterly forecast sheet the tabs of which is colored light blue . 4 quarterly weight calculation sheets colored brown for RPT, peach for business tax, red for fees and charges, and green for economic enterprises.

LGU Revenue Forecasting Template: Base Input Sheet: 

LGU Revenue Forecasting Template: B ase Input Sheet 11/15/2011 BLGF LGU Revenue Forecasting System 67 In most instances this is kept at zero since the RPT forecasts is better based on the Assessed Property values * RPT rate. In certain instances the user might want to compare RPT forecast based on historical trends with that based on assessed values

LGU Revenue Forecasting Template: Quarterly Weights: 

LGU Revenue Forecasting Template: Quarterly Weights 11/15/2011 BLGF LGU Revenue Forecasting System 68 Real Property Tax ( RPTax ) Tax on Business ( BTax )

LGU Revenue Forecasting Template: Quarterly Weights: 

LGU Revenue Forecasting Template: Quarterly Weights 11/15/2011 BLGF LGU Revenue Forecasting System 69 Fees and Charges (F & C ) Economic Enterprises ( EcoEnt )

LGU Revenue Forecasting Template: Initial Annual Forecast Sheet: 

LGU Revenue Forecasting Template: Initial Annual Forecast Sheet 11/15/2011 BLGF LGU Revenue Forecasting System 70 If there are negative values in any of the past 3 years growth rate, enter a “1” on the appropriate box so that 3 year average will be based on the median which will be less affected by the decline.

LGU Revenue Forecasting Template: Initial Quarterly Forecast Sheet: 

LGU Revenue Forecasting Template: Initial Quarterly Forecast Sheet 11/15/2011 BLGF LGU Revenue Forecasting System 71 The sum of the quarterly forecast are checked against the initial annual forecast in the initial forecast sheet. If there is no discrepancy, then an “OK” mark is marked on the appropriate box in the “CHK “ column. If there is, then a “ CHk calculation” warning is entered.

LGU Revenue Forecasting Template: Final Annual Forecast Sheet: 

LGU Revenue Forecasting Template: Final Annual Forecast Sheet 11/15/2011 BLGF LGU Revenue Forecasting System 72 The values arising form the regional-based review and negotiation process are manually entered. The template automatically converts the final annual forecasts into final quarterly forecasts

Overall LGU Revenue Target Setting Process: 

Overall LGU Revenue Target Setting Process 11/15/2011 BLGF LGU Revenue Forecasting System 73

Overall LGU Revenue Target Setting Process: 

Overall LGU Revenue Target Setting Process The BLGF Central office target-setting process will only cover four (4) key LGU local revenue sources that is presently the subject of the existing BLGF revenue target-setting system: Real property tax Business tax Fees and charges Economic enterprises In addition, forecasts of two key central government transfers to LGUs are expected to be provided by the Department of Budget and Management (DBM) Internal Revenue Allotment (IRA) Shares in national wealth and other national taxes The other revenue items will be left to the local treasurers to forecast on their own since they are highly dependent on local practices and cash management patterns 11/15/2011 BLGF LGU Revenue Forecasting System 74

Overall LGU Revenue Target Setting Process: 

Overall LGU Revenue Target Setting Process 11/15/2011 BLGF LGU Revenue Forecasting System 75 Seeks to systematically develop a single set of LGU-level revenue targets to serve as the revenue basis of the annual LGU budget process. For the results of the process to be useful to the LGU budget process, the final revenue targets should be ready by the time of the budget call — 1st week of July The process starts with the generation of a financial and economic model-based set of initial LGU-level forecasts by the BLGF central office (CO) sent out to the LGUs and to the regional offices no later than 15 May .

Overall LGU Revenue Target Setting Process: 

Overall LGU Revenue Target Setting Process 11/15/2011 BLGF LGU Revenue Forecasting System 76 At the same time, LGUs will be developing their own revenue forecasts for the 4 key revenue items using the LGU Revenue Forecasting Template . Annual Quarterly Using input historical data for the current year + past 3 years , the template generate the annual and quarterly revenue forecasts for the LGU. The results serve as initial revenue forecasts for the LGU.

Overall LGU Revenue Target Setting Process: 

Overall LGU Revenue Target Setting Process 11/15/2011 BLGF LGU Revenue Forecasting System 77 The 2 sets of projections are then subjected to a regional review and reconciliation process where the LGU treasurers , the BLGF regional and central office staff participate during the month of June . The review and reconciliation will be guided by the BLGF Forecast Reconciliation Guide . The results of the reconciliation process will form the final and single set of LGU revenue targets to serve as the revenue basis of the LGU annual budget that should be ready no later than 30 June in time for the budget call by the 1st week of July .

BLGF Region-based LGU Revenue Target Review and Reconciliation Process: 

BLGF Region-based LGU Revenue Target Review and Reconciliation Process 11/15/2011 BLGF LGU Revenue Forecasting System 78

The Central/Regional Office-Local Treasurer Revenue Target Reconciliation Process: 

The Central/Regional Office-Local Treasurer Revenue Target Reconciliation Process Necessary if Treasurer in consultation with other LFC members want to contest the BLGF central office-set targets. The reconciliation process will only cover the following revenue targets: Real property tax Business tax Fees and charges Economic enterprises Otherwise, the central office-set targets are the ones that get submitted by the LFC. The Treasurer may opt to prepare an alternative set of revenue targets using the LGU annual growth rate based revenue forecasting procedure . Reconciliation is done at the regional level before a set of revenue targets mutually agreed upon between the central office, the regional office and the particular Treasurer get submitted by the LFC as basis for financial programming and annual budgeting . 11/15/2011 BLGF LGU Revenue Forecasting System 79

The Central/Regional Office-Local Treasurer Revenue Target Reconciliation Process: 

The Central/Regional Office-Local Treasurer Revenue Target Reconciliation Process The reconciliation process is necessary to reconcile BLGF central office ( CO) -generated financial and economic model-based revenue targets with any locally-generated revenue forecasts, to refine the initially set targets based on local inputs to be provided by the local treasurers, and to promote local ownership of revenue targets. This refinement process will be important in cases where major tax base as well as local policy and implementatio n changes occur, e.g., property revaluations, business closure or openings, use of improved billing and collection systems, etc. Only one set of revenue targets get to be considered by the LFC and serve as the basis of the annual LGU budget – the one submitted by the Treasurer. 11/15/2011 BLGF LGU Revenue Forecasting System 80

PowerPoint Presentation: 

11/15/2011 BLGF LGU Revenue Forecasting System 81 The Central/Regional Office-Local Treasurer Revenue Target Reconciliation Process

Revenue Target Reconciliation Guidepoints: 

Revenue Target Reconciliation Guidepoints 11/15/2011 BLGF LGU Revenue Forecasting System 82 For a uniform revenue target reconciliation process to evolve, it is necessary that discussions be based on points replicable across regions and across years. The accompanying table lists down indicative guide-points for the reconciliation process: What support data/documents LGU treasurers should bring to the regional reconciliation meeting. Indicative benchmarks that could serve as “boundary points” for the negotiation process. Justification points that the LGU treasurers should establish to support their request for a lowered set of revenue targets vis-à-vis the central office set targets .

Revenue Target Reconciliation Guidepoints: 

Revenue Target Reconciliation Guidepoints 11/15/2011 BLGF LGU Revenue Forecasting System 83