Philippines Central Bank Uses AI and Job Boards to Track Unemployment in Real Time

Real-Time Labour Market Signals

The Bangko Sentral ng Pilipinas has deployed artificial intelligence and online job-posting data to estimate unemployment conditions in near real time, bypassing the two-month lag in official labour statistics.

The central bank's enhanced Labour Market Intelligence System draws on postings from JobStreet and the Department of Labour and Employment's PhilJobNet, according to a study in the BSP's August Monetary Policy Report. The system compares vacancy volumes against job-search activity to approximate the balance between employer demand and worker supply, a metric that matters for inflation forecasting and wage-pressure analysis.

Official employment figures from the Philippine Statistics Authority's Labour Force Survey arrive roughly two months after the reference period. For a central bank calibrating monetary policy in real time, that delay can obscure turning points in labour demand or the emergence of wage inflation. The BSP's approach trains machine-learning models on historical data, then tests them on later periods to simulate operational use for current monitoring.

How the System Works

The intelligence system calculates a ratio of online job vacancies to search activity. A rising ratio signals tighter labour markets, where employers compete for fewer available workers. A falling ratio points to slack, with more job seekers relative to openings. That directional signal matters more than the absolute level, because online postings capture only a slice of the total labour market.

To organise the raw data, researchers used generative AI to match job descriptions with official Philippine industry and occupational classifications. Machine learning then identifies which indicators carry the most predictive power for unemployment estimates. The system sorts postings by sector and occupation, revealing granular imbalances that national aggregates can miss.

"Some sectors may face worker shortages even when overall labour market conditions appear stable, while others may be weakening," the study noted.

Sectoral Visibility

The sector-by-sector view lets policymakers distinguish broad-based tightness from concentrated pressure. If vacancies surge in information technology but stagnate in manufacturing, the inflation implications differ. Wage growth in a small, high-skill sector may not spill over into broader price pressures, whereas widespread hiring competition across industries typically does.

The BSP framed the tool as a complement to, not a replacement for, official surveys. Labour Force Survey data remain the benchmark for headcount employment and participation rates. The online system offers speed and sectoral detail, but it skews toward formal-sector jobs and may under-represent informal or rural employment.

Regional Context

Central banks across Asia have experimented with alternative data to fill gaps in traditional statistics. The Monetary Authority of Singapore uses payment-card transactions and shipping data for GDP nowcasting. Bank Indonesia tracks mobility data from ride-hailing apps. The Reserve Bank of India mines corporate earnings calls for sentiment. The BSP's job-posting approach fits that pattern, leveraging digital exhaust to sharpen near-term forecasts.

The Philippines labour market has long been shaped by overseas employment flows, domestic underemployment, and a large informal sector. Official unemployment stood at 4.5 per cent in the second quarter of 2026, but underemployment, which captures workers seeking more hours, ran closer to 12 per cent. Online vacancies may track formal hiring demand more reliably than they capture the full employment picture, particularly in agriculture and informal services.

Policy Implications

Timely labour-market intelligence matters for inflation targeting. The BSP's mandate centres on price stability, with a two to four per cent inflation band. Wage growth, driven by tight labour markets, can feed into services inflation and broader price pressures. Conversely, rising unemployment signals economic slack that may justify easier monetary policy.

The central bank has held its policy rate steady through most of 2026, balancing inflation risks against growth concerns. The new intelligence system gives the Monetary Board an earlier read on whether labour demand is accelerating or cooling, potentially informing rate decisions before official employment data arrive.

The study did not disclose the system's accuracy relative to official figures, nor whether it has been integrated into formal policy briefings. The BSP described the work as research, published in the Monetary Policy Report alongside other analytical projects.

Broader Adoption

The use of AI to classify and interpret unstructured job-posting text represents a practical application of natural language processing in economic surveillance. Generative models can parse job titles, required skills, and industry keywords, then map them to standardised taxonomies. That capability scales across languages and jurisdictions, making similar systems feasible for other Southeast Asian central banks.

JobStreet, owned by SEEK Asia, operates across the Philippines, Indonesia, Malaysia, Singapore, and other markets. PhilJobNet is the Labour Department's public employment portal. Both platforms aggregate postings from employers of varying sizes, though large corporations and formal-sector roles dominate the listings.

The BSP's move reflects a broader shift in central banking toward nowcasting tools that fuse traditional statistics with real-time digital signals. As labour markets digitalise and more hiring moves online, job boards become a proxy for economic activity. The challenge remains ensuring the proxy reflects the full economy, not just its most visible segments.