The Regional Poverty Gap Nobody Talks About

When someone says "Philippine poverty rate is X percent," what does that actually tell you? Not much, as it turns out. A national average smooths over the fact that some regions have poverty rates several times higher than others. The number you hear on the news is true in aggregate but misleading in practice — because nobody lives in "the Philippines on average." People live in specific regions with specific economies.

I pulled PSA data covering all 16 regions to see how wide the gap really is. The answer: wider than most people realize.

What I Built

A regional comparison of poverty indicators across the Philippines. The analysis covers poverty incidence, agricultural wages, farm household income, and a rough cost-of-living adjustment. I built choropleth-style visualizations ranking regions from least to most impoverished, and compared wage data against basic cost thresholds to see where workers are barely getting by.

Why This Project

I kept running into the same problem in my other Philippine data projects: regional differences were huge, and national averages kept hiding them. In the FIES analysis, NCR and ARMM looked like different countries. In the food prices project, regional price gaps were persistent. I wanted to tackle regional inequality head-on instead of treating it as a footnote in other analyses.

How I Put It Together

The PSA publishes poverty statistics by region, but the data is spread across multiple reports and tables. I pulled poverty incidence rates, agricultural wage surveys, and farm household income data into a single dataset. Joining them required matching region codes carefully — the PSA sometimes uses different naming conventions across reports.

The wage-versus-cost-of-living comparison was the trickiest part. There's no single "cost of living index" by region from the PSA, so I approximated it using regional food price data from my earlier project and minimum daily cost thresholds. It's rough, but it gives a directional sense of whether wages actually cover basic expenses in each region.

For the regional ranking visualizations, I used a color-gradient approach similar to a choropleth map but rendered as a bar chart since I didn't want to spend time on geographic mapping for this project. The ranking format made it easy to spot which regions consistently land at the top or bottom.

What the Data Showed

ARMM (now BARMM) had the highest poverty incidence by a significant margin — roughly 3 to 4 times the rate of NCR. That gap has been documented before, but seeing it in a side-by-side bar chart with every other region gives it a different weight. ARMM wasn't just the poorest. It was in a category of its own.

NCR had the lowest poverty incidence, which isn't surprising. But the regions just outside Metro Manila — CALABARZON and Central Luzon — also performed well, likely because of their economic proximity to the capital. The farther a region sits from major economic centers, the higher its poverty rate tended to be.

Agricultural wages told a sobering story. In several regions, average daily agricultural wages barely exceeded the cost of feeding a family for a day. These workers aren't saving. They aren't investing in education for their kids. They're surviving, day by day, with almost no margin for error.

• ARMM/BARMM poverty incidence was 3-4x higher than NCR
• Regions near Metro Manila (CALABARZON, Central Luzon) had lower poverty
• Agricultural wages in the poorest regions barely covered daily food costs
• The gap between the best and worst regions showed no sign of closing

Looking Back

This was the shortest project technically but one of the hardest to write about. The numbers are straightforward. The implications aren't. When agricultural workers earn barely enough to eat, and that pattern repeats across years of data without improvement, it raises questions that go well beyond what a data analysis can answer.

What I can say is this: the regional poverty gap is real, it's large, and it's persistent. Any policy conversation about poverty reduction that doesn't account for regional differences is starting from the wrong place.