Business Obstacles and Firm Employment Growth in ASEAN: Does Country Heterogeneity Absorb the Investment Climate Effect?
Last updated:
At a glance
RPS
—
Current chapter focus
—
Chapter completion (tracker)
—
Share of chapters marked Completed (7 items).
Total Raw Firms
6,631
Final Analytical Sample
3,741
Countries
6
Survey Period
2023–2025
Unique Cells (IV)
278
Mean Obs per Cell
17.2
| Country | N | % |
|---|---|---|
| Indonesia (IDN) | 1,185 | 31.7% |
| Malaysia (MYS) | 769 | 20.6% |
| Thailand (THA) | 615 | 16.4% |
| Philippines (PHL) | 527 | 14.1% |
| Cambodia (KHM) | 405 | 10.8% |
| Lao PDR (LAO) | 240 | 6.4% |
| Obstacle | M1 OLS | M4 Full FE | Verdict |
|---|---|---|---|
| Tax rates (j30a) | +0.0214** | +0.0184* | Robust |
| Tax administration (j30b) | n.s. | n.s. | Not robust |
| Business licensing (j30c) | -0.0101 | -0.0217** | Robust |
| Political instability (j30e) | n.s. | n.s. | Not robust |
| Corruption (j30f) | n.s. | n.s. | Not robust |
| Access to finance (k30) | +0.0057 | +0.0138* | Marginal |
| Labor regulations (l30a) | n.s. | n.s. | Not robust |
| Inadequate workforce (l30b) | n.s. | n.s. | Not robust |
| Informal competition (e30) | -0.0160** | -0.0228*** | Robust |
| Crime & theft (i30) | -0.0132** | +0.0005 | FE Paradox |
Model fit
R² M1 = 5.6% · R² M4 = 14.3%
*** p < 0.01, ** p < 0.05, * p < 0.10. Standard errors clustered at country × industry level.
After adding Country Fixed Effects, most obstacle ratings lose statistical significance.
Note: Majority lose significance, but informal competition (e30) and business licensing (j30c) survive — ASEAN differs from Eastern Europe.
k30 remains significantly negative after Country FE.
Note: k30 is significant in M4 but positive, not negative. May reflect reverse causality.
Finance constraints hit small firms harder; labor regulations hit large firms harder.
Note: k30 effect for small firms: +0.0254*** (p < 0.001). Medium and large firms: negative interaction (p < 0.001).
Significance of M4-style coefficients: baseline vs R1 (drop IDN), R2 (DH growth), R3 manufacturing / services splits. Size-interaction check (R4) summarized in Verdict where relevant.
| Obstacle | Baseline | R1 No-IDN | R2 DH | R3 Mfg | R3 Svc | Verdict |
|---|---|---|---|---|---|---|
| j30a | ✓ | ✓ | ✓ | ✓ | ✗ | Robust |
| j30b | ✗ | ✗ | ✗ | ✗ | ✗ | Not robust |
| j30c | ✓ | ✓ | ✓ | ✓ | ✓ | Robust |
| j30e | ✗ | ✗ | ✗ | ✗ | ✗ | Not robust |
| j30f | ✗ | ✗ | ✗ | ✗ | ✗ | Not robust |
| k30 | ✓ | ✓ | ✗ | ✗ | ✓ | Mixed |
| l30a | ✗ | ✗ | ✗ | ✗ | ✗ | R4 interaction |
| l30b | ✗ | ✗ | ✗ | ✗ | ✗ | R4 interaction |
| e30 | ✓ | ✓ | ✓ | ✓ | ✓ | Robust |
| i30 | ✗ | ✗ | ✓ | ✗ | ✓ | Unstable |
Cell-level LOO instruments are collinear with two-way FE. First-stage F < 10 for all 10 obstacles. OLS M4 used as main specification.
No multicollinearity problem. j30a × j30b correlation = 0.733, below 0.8 threshold.
Robustness check R1 (drop IDN) confirms results hold.
Saved automatically in this browser (localStorage).
Semua berkas skripsi: RPS, referensi jurnal, kuesioner WBES, dan data mentah Stata.
Rencana Penelitian Skripsi — Titus Herdiawan (120210170070)
📄 Buka RPS (PDF)Jurnal referensi
PDF dibuka di modal; berkas Excel diunduh langsung.
.dta adalah format Stata binary. Tidak bisa dibuka di browser —
gunakan tombol download lalu buka dengan Stata 14+ atau Python (pyreadstat).