Estimating Labor Force and Fiscal Modules for Coastal Louisiana Economies: Extension of the COMPAS Modeling Framework Presented by: Arun Adhikari Dr. J.

Estimating Labor Force and Fiscal Modules for Coastal Louisiana Economies: Extension of the

COMPAS Modeling Framework

Presented by:Arun Adhikari

Dr. J Matthew Fannin

Projected Funded Through Cooperative Agreement with Minerals Management Service, US Dept of Interior

Outline

Introduction and Objective of the Study

Background and Overview of COMPAS Modeling

Data and Methodology

Empirical Specifications

Models Discussion

Results and Discussion

Concluding Remarks

Introduction Accuracy in any policy analysis is of a great value to public decision

makers

This study also aims to develop a model to forecast different expenditure demands in the fiscal module of the Louisiana Community Impact Model (LCIM) using alternative procedures capable of increasing the performance over traditional COMPAS estimators.

The specific objective includes modeling the fiscal module (four major categories of expenditure; public service, public works, general government and health and welfare) of LCIM for 60 parishes of Louisiana to compare the performance between spatial and non spatial estimators that takes into account heterogeneity.

Background and Overview of COMPAS Modeling

COMPAS models are regional economic models that combine input-output and econometric approaches to build a conjoined model of economic structure.

COMPAS models typically treat employment demand as an exogenous driver of changes in the labor market which ultimately impact the fiscal sector.

The fiscal module in this research is an extension to the module used by Fannin et al., (2008). The goal of this analysis is to adhere to the basic theme of regional science: spatial location matters.

Background and Overview of COMPAS Modeling

Data

Estimation is based on the COMPAS model for Louisiana that includes 60 parishes , where the variables for the fiscal module were selected on the basis of Fannin et al (2008) and were modified depending upon the requirements of our model and applied geographically to all Louisiana parishes.

Data sources:

a) Audited Financial Statements

b) Bureau of Economic Analysis

c) U.S. Census Bureau

d) Department of Education

Within the fiscal module, different expenditure equation data on public safety, public works, general government and health and welfare sectors were estimated.

Methodology

Four expenditure equations were estimated. Every equation is a function of several specific variables from all Louisiana parishes.

These equations were estimated by a cross-section Ordinary Least Square (OLS) model as a base control with quantile regression, and spatial autoregressive model regressions also estimated.

We applied OLS regression and quantile regression using STATA, and spatial regression using MATLAB.

Base year of estimation is 2007.

Empirical Specifications Four different expenditure equations that were estimated for

comparison are:

1) lnpcgg = f (lnpcasdv, lnpcretsl, lnpcin, lnpurb, lnarblndnsty)

2) lnpchw = f (lnpcasdv, lnpcretsl, lnpcin, lnpafam)

3) lnpcps = f (lnpcasdv, lnpcretsl, lnpcin, lnpafam, lnarblndnsty)

4) lnpcgg = f (lnpcasdv, lnpcretsl, lnpcin, lnpurb, lnarblndnsty, lnpclcrdml)

Performances Evaluation and Comparison

Forecast performances were evaluated based on the procedures outlined in Johnson, Otto and Deller (2006), and Kovalyova and Johnson (2006).

The performance of estimators is compared on the basis of quantitative evaluation methods. These methods include analysis of:

a) mean simulation error (ME),

b) mean percent error (MPE),

c) mean absolute error (MAE),

d) mean absolute percent error (MAPE),

e) mean square error (MSE),

f) root mean square error (RMSE),

g) root mean square percent error (RMSPE),

h) and Theil’s coefficient U1 and U2

Models Discussion

1) Ordinary Least Squares (OLS)

2) Quantile Regressions Divided into 3 quantiles (0.33, 0.66 and 0.99)

3) Spatial Autoregressive Regression (SAR) Derivation of spatial weight matrix

Results

Table 1: Variable description and summary statistics, Louisiana, 2007 

Variable Name Mean Standard Deviation Min Max

General Government Expenditure 9,176,819 29,103,832 593,955 210,722,026

Health and Welfare Expenditure 2,125,658 2,924,059 5,664 13,602,439

Public Safety Expenditure 7,008,965 25,111,070 232,882 189,130,903

Public Works Expenditure 9,549,507 11,714,106 847,070 65,739,927

Assessed Value 418,151,563 553,860,439 36,056,864 3,466,560,930

Retail Sales 901,353,145 1,355,501,809 29,883,946 7,612,001,075

Arable Land Density 760 381 192 1,909Local Road Miles 1,534 726 284 3,635

Population 60,191 74,123 5,828 429,914

Per Capita Income 2,8739 5,218 20,060 43,206

Percent African American 30 14 4 67

Percent Urban 45 27 0 96

ResultsTable 4: Performance Estimation of 66th quantile for Public Safety,

Louisiana, 2007b1 b2 b3 b4 b5 b6

-33.01 0.97 -0.02 2.86 -0.06 -0.04Areaname y Yhat yhat-y Abs (Yhat-y) (yhat-y)/y Abs (Yhat-y)/y (yhat-y)2 {(yhat-y)/y}2 y2 yhat2

Catahoula 28.32 27.13 -1.19 1.19 -0.04 0.04 1.42 0.00 802.10 736.12Red River 28.85 27.35 -1.51 1.51 -0.05 0.05 2.27 0.00 832.57 747.92Livingston 30.61 43.92 13.30 13.30 0.43 0.43 177.00 0.19 937.18 1928.74Iberia 31.08 87.97 56.89 56.89 1.83 1.83 3236.34 3.35 965.90 7738.33Vernon 37.74 54.10 16.36 16.36 0.43 0.43 267.57 0.19 1424.56 2926.89Jackson 41.09 80.47 39.38 39.38 0.96 0.96 1550.55 0.92 1688.46 6475.07Bossier 44.40 90.30 45.90 45.90 1.03 1.03 2106.96 1.07 1971.47 8154.60Jefferson Davis 46.42 46.42 0.00 0.00 0.00 0.00 0.00 0.00 2155.00 2154.99Bienville 47.01 101.06 54.05 54.05 1.15 1.15 2921.30 1.32 2209.67 10212.35Beauregard 47.18 42.56 -4.62 4.62 -0.10 0.10 21.37 0.01 2225.95 1811.08Ascension 48.31 145.88 97.57 97.57 2.02 2.02 9520.23 4.08 2333.39 21280.06Tensas 48.41 65.39 16.97 16.97 0.35 0.35 288.11 0.12 2343.99 4275.66Caldwell 48.58 29.66 -18.92 18.92 -0.39 0.39 357.93 0.15 2360.39 880.01Webster 49.22 54.32 5.10 5.10 0.10 0.10 26.03 0.01 2422.79 2951.04DeSoto 51.12 83.04 31.92 31.92 0.62 0.62 1018.75 0.39 2613.08 6895.01Vermillion 51.80 51.48 -0.33 0.33 -0.01 0.01 0.11 0.00 2683.49 2649.79West Feliciana 52.95 118.60 65.65 65.65 1.24 1.24 4309.44 1.54 2804.22 14066.26Natchitoches 54.10 46.29 -7.81 7.81 -0.14 0.14 60.99 0.02 2927.03 2142.99Tangipahoa 54.31 40.93 -13.39 13.39 -0.25 0.25 179.18 0.06 2950.07 1675.16East Carroll 55.84 24.53 -31.31 31.31 -0.56 0.56 980.05 0.31 3118.18 601.96

SUM 364.02 522.16 8.64 11.72 27025.59 13.74 41769.49 100304.033.71 204.38 316.71

Avrg 18.20 26.11 0.43 0.59 1351.28 0.69 2088.47 5015.20MSE 1,351.28RMSE 36.76U1 0.07U2 0.52

ResultsTable 3: Average performance estimation measures for different

categories of expenditure, Louisiana, 2007

Expenditure Category Spatial Autoregressive Linear (OLS) Quantile Regression

0.33 0.66 0.99

General Government

yhat-y 479.38 -105.359 12.006 21.296 988.329

(yhat-y)/y 7.73 0.196 0.344 0.271 6.558

{(yhat-y)/y}2 97.97 0.606 0.447 0.286 144.876

Theil’s Coeff (U1) 0.58 0.866 0.046 0.052 0.106

Health and Welfare

yhat-y -9.104 -10.115 4.438 16.451 34.844

(yhat-y)/y 1.31 1.305 3.233 0.737 0.701

{(yhat-y)/y}2 42.08 42.884 85.423 1.558 0.978

Theil’s Coeff (U1) 0.79 0.826 0.073 0.086 0.054

Expenditure Category

Spatial Autoregressive

Linear (OLS) Quantile Regression0.33 0.66 0.99

Coeff. t-stat Coeff. t-stat Coeff. t-stat Coeff. t-stat Coeff. t-stat

General GovernmentConstant -20.24*** -3.44 -15.11* -1.74 -10.69 -1.29 -14.86** -2.56 -56.35** -2.19percapasdval 0.54** 2.21 0.59*** 3.54 0.50* 1.68 0.65* 1.81 0.59 0.91percapretsls 0.27 0.79 0.08 0.31 0.06 0.22 -0.05 -0.10 -0.75 -0.74percapinc 1.95*** 2.90 1.28* 1.83 1.24 1.47 1.49** 2.33 5.68** 2.16percenturban -0.15 -1.73 -0.06 -0.95 -0.16* -1.87 -0.09 -0.94 0.006 0.04arblndensity -0.04 -0.24 0.07 0.28 -0.38* -1.92 -0.13 -0.70 0.94* 1.78rho -0.49*** -4.16

Health and WelfareConstant -28.57*** -3.37 -26.01** -2.40 -20.21** -2.41 -25.15*** -2.79 3.85 0.18percapasdval 0.44 1.37 0.41 1.36 0.63 1.36 0.50 1.23 0.19 0.32percapretsls -0.13 -0.33 -0.16 -0.49 -0.06 -0.09 -0.29 -0.56 0.55 0.85percapinc 2.73*** 2.79 2.50** 2.14 1.67 1.17 2.39* 1.96 -0.61 -0.26percentafam 0.34 1.40 0.40** 2.28 0.26 0.72 0.76*** 1.82 0.04 0.07rho -0.01 -0.11

ResultsTable 2: Parameter estimates for OLS and Quantile regressions,

Louisiana, 2007

Major Results An increase in assessed value leads to increase in the expenditure of the

all categories, as seen from all three models.

An increase in per capita income leads to increase in expenditure in the public safety for all the categories of expenditure, as seen from all three models. The magnitude keeps increasing for higher quantiles.

Lower and median quantiles are found to be performing better as compared to spatial and OLS regressions.

Some parishes like Ascension, Bienville, Iberia and West Feliciana are not performing as good on average. On the contrary, parishes like Catahoula, Jefferson parish, Red river and Vermillion are performing better than the average error measures.

Concluding Remarks This research sets out a strategy for comparing the forecasting

performances between spatial (SAR) and non-spatial models (OLS and Quantile Regressions) in a fiscal sector of LCIM.

Evaluation of alternative methodologies are expected to give regional economic modelers better information from which to choose when seeking to construct models projecting different modules.

These results will be helpful to those community modelers desiring to estimate cross-section fiscal modules for forecasting in states that have much greater heterogeneity among local government units.

Other spatial models like the spatial error model and panel data models could also be evaluated while comparing the performances between spatial and non-spatial estimators, which would be a concept for future research in this paper.

Thank You

Questions/Comments