+ - 0:00:00
Notes for current slide
Notes for next slide

Regional Disparities in Educational Outcome, Spatial Spill-over, and the Determinants

Evidence from 34 Provinces in Indonesia

Ragdad Cani Miranti

BPS- Central Bureau Statistics of Indonesia

Prepared for 2021 Asian Seminar in Regional Science
October 8th 2021
[slides available at: https://asrs-slides-2021.netlify.app/#1]

1 / 25

Motivation:

  • Regional disparities in educational outcomes still persist
  • Financial resources have been intensified
  • Government of Indonesia implemented the rule of 20 percent budgeting for educational funds.
  • There are relative few studies that evaluate the role of spatial effects on the educational process.
  • Lack of academic literature on educational development at the sub-national level in Indonesia.

Research Question:

  • Does role of neighborhood effect exist in altering the educational outcomes between proximate provinces?
  • What are the significant variables affecting the educational outcomes across provinces?
  • How is the magnitude of those variables and spill-over effects to the educational outcomes across provinces?

Methods:

  • Spatial autocorrelation (Moran's I and LISA)
  • Spatial Panel Models (SAR,SEM,SAC,SDM region-fixed effect)
2 / 25

Main Results:

  1. Significant Spatial Autocorrelation at the province level.

  2. Spatial model

    • Spatial Durbin Model is the best specification model
    • Number of educational institutions and pupil-teacher ratio are two key factors affecting educational outcomes across provinces in Indonesia
    • Number of educational institutions (elementary school and senior high school) have indirect effect to the educational outcome.

  3. Policy Implication

    • Increasing spatially better coordination and cooperation on educational development connectivity between local governments especially in the Eastern part and outer islands
    • Improving educational infrastructures
3 / 25

Outline of this presentation

  1. Introduction and Data

  2. Spatial autocorrelation

    • Moran's I Scatter Plot
    • Local Indicators of Spatial Autocorrelation

  3. Spatial Panel Model

    • Cross-sectional dependence test in the panel data
    • OLS Fixed Effect vs Some Spatial Panel Model with Region Fixed-Effect
    • Direct and Indirect Effect

  4. Concluding Remarks



[ Slides and paper available at: https://asrs-slides-2021.netlify.app/#1]

4 / 25

(1) Some stylized facts

Educational Outcome disparities across districts over time

Educational Outcome disparities across islands over time

5 / 25

Disparities in mean years of schooling across districts over time

Note: X-axis reflects mean years of schooling, Y-axis reflects expected years of schooling, human development index is reflected by the dot size and name of island is reflected by the dot color

6 / 25

Violin Chart of Mean Years Schooling Dispersion across Islands in Indonesia

7 / 25

(1) Data

Mean years of schooling 2010-2018 at province level

Expected years of schooling 2010-2018 at province level

Number of elementary schools, junior high schools, and senior high schools 2010-2018 at province level

Pupil-teacher ratio in elementary school, junior high school, and senior high school

All data are derived from Central Bureau Statistics of Indonesia (BPS-Statistics Indonesia)

8 / 25

(3) Spatial Autocorrelation

Global Spatial Autocorrelation : Moran's I Statistics

Local Indicators of Spatial Autocorrelation (LISA)

9 / 25

Global spatial Autocorrelation

Mean Years of Schooling

Global spatial autocorrelation is measured based on the Moran’s I, which is statistically significant at 5% level for all years.

10 / 25

Expected Years of Schooling

11 / 25

Local Indicators of Spatial Autocorrelation (LISA)

Mean Years of Schooling Initial period (2010)

Note : Local spatial dependence is significant for both spatial clusters and spatial outliers at 5% significance level

12 / 25

Local Indicators of Spatial Autocorrelation (LISA)

Mean Years of Schooling Initial period (2018)

**Significant cold-spot (spatial cluster with low mean years of schooling surrounded by low mean years of schooling-cluster) located in Papua Province.

13 / 25

Local Indicators of Spatial Autocorrelation (LISA)

Expected Years of Schooling Initial period (2010)

Note : Local spatial dependence is significant for both spatial clusters and spatial outliers at 5% significance level

14 / 25

Local Indicators of Spatial Autocorrelation (LISA)

Expected Years of Schooling Initial period (2018)

Spatial outlier high-low (spatial cluster with relatively high expected years of schooling surrounded by relatively low expected years of schooling cluster) is only significant in Maluku Province.

15 / 25

Some Spatial Panel Models

16 / 25

Cross-sectional dependence test in the panel data model

Pesaran Test for Model of Mean Years of Schooling

Pesaran Test for Model of Expected Years of Schooling

17 / 25

Spatial Panel Models for Mean Years of Schooling

Spatial Pane Model with Region Fixed-Effect

18 / 25

Spatial Autocorrelation Parameter

19 / 25

Spatial Panel Models for Expected Years of Schooling

Spatial Panel Model with Region Fixed-Effect

20 / 25

Spatial Autocorrelation Parameter

21 / 25

Direct and Indirect Effect

Mean Years Schooling

22 / 25

Direct and Indirect Effect

Expected Years of Schooling

23 / 25

(5) Concluding Remarks

  • Inequality still matters. Eastern part of Indonesia need to be the priority of educational development.

  • The availability of educational institutions in each level and the pupil-teacher ratio are vital factors to improve the educational outcomes in Indonesia.

  • Spill-over effect does exist. The indirect effect of availability of schools to the mean years of schooling and expected years of schooling is positive and significant.

-

24 / 25

Thank you very much for your attention

Slides and working paper available at: https://asrs-slides-2021.netlify.app/#1]

Quantitative Regional and Computational Science lab

https://quarcs-lab.rbind.io

25 / 25

Motivation:

  • Regional disparities in educational outcomes still persist
  • Financial resources have been intensified
  • Government of Indonesia implemented the rule of 20 percent budgeting for educational funds.
  • There are relative few studies that evaluate the role of spatial effects on the educational process.
  • Lack of academic literature on educational development at the sub-national level in Indonesia.

Research Question:

  • Does role of neighborhood effect exist in altering the educational outcomes between proximate provinces?
  • What are the significant variables affecting the educational outcomes across provinces?
  • How is the magnitude of those variables and spill-over effects to the educational outcomes across provinces?

Methods:

  • Spatial autocorrelation (Moran's I and LISA)
  • Spatial Panel Models (SAR,SEM,SAC,SDM region-fixed effect)
2 / 25
Paused

Help

Keyboard shortcuts

↑, ←, Pg Up, k Go to previous slide
↓, →, Pg Dn, Space, j Go to next slide
Home Go to first slide
End Go to last slide
Number + Return Go to specific slide
b / m / f Toggle blackout / mirrored / fullscreen mode
c Clone slideshow
p Toggle presenter mode
t Restart the presentation timer
?, h Toggle this help
Esc Back to slideshow