Ensuring Decile Data Accuracy for Targeted Social Assistance

Public responses and criticisms regarding decile data, which categorises society into ten levels of welfare, serve as a reminder of the vital importance of verified data. These critiques, in turn, assist the government in formulating more measurable policies regarding state expenditure for social assistance (Bansos) programmes and various other subsidies, ensuring distribution is precisely targeted.

As is widely known, the public has been monitoring deciles through the National Socio-Economic Single Data (DTSEN). The latest DTSEN basis, as of 3 July 2026, records a population of 290.13 million people and 95.98 million families in Indonesia. DTSEN is an accumulation and synchronisation of social and economic data integrated from various ministries and agencies, such as the Integrated Social Welfare Data (DTKS), the P3KE data (Targeting the Acceleration of Extreme Poverty Eradification), and Regsosek (Socio-Economic Registration), including civil administration data (Dukcapil).

Using this DTSEN, deciles are created to group the population into ten welfare levels based on 39 variables covering social, economic, health, education, and housing aspects. The determination of these deciles is not performed by village, sub-district, regency, or provincial officials, but rather by the Central Bureau of Statistics (BPS).

Decile 1 represents families with the lowest welfare level or the very poor; decile 2 represents the poor category; decile 3 represents the near-poor group; decile 4 represents the vulnerable-to-poverty group, and so on, up to decile 10, which represents families with the highest welfare level or the very wealthy.

Deciles and all associated data serve as a reference for the government to design and establish policies for social assistance spending and various other subsidies, such as education subsidies or affirmative pathways for school and university admissions. Through deciles, the government also intends to limit access to energy or fuel subsidies for the affluent in deculations 9 and 10.

Citizens in various regions began monitoring the DTSEN and decile details in early August 2026 via the Ministry of Social Affairs (Kemensos) website. Shortly thereafter, responses and criticisms began to emerge in the public sphere via social media platforms. Many residents complained that the decile data, or the profile and economic condition descriptions, did not align with the current facts of their lives.

The Ombudsman of the Republic of Indonesia (ORI) has even received numerous complaints from the public regarding discrepancies in decile data. Ombudsman representatives in various regions, such as Gorontalo, Surabaya, and Bangka Belitung, immediately followed up on these complaints with suggestions for the necessity of field verification. The Regional House of Representatives (DPRD) of the Special Region of Yogyakarta (DIY) also received thousands of public complaints from citizens worried that assistance through the Smart Indonesia Card (KIP) might be revoked following the decile changes. The DIY DPRD has requested that the local Social Services (Dinsos) at both the regional and regency levels collect direct data from citizens experiencing decile changes.

Examples of data discrepancies in the deciles are numerous and varied. For instance, Timbul (49), a rickshaw driver from Ngentakrejo Village, Lendah, Kulon Progo, is anxious after seeing his profile moved to decile 9 after his child was accepted into university, whereas a village head is placed in decile 8. Previously, Timbul was in decile 1. Timbul’s position in decile 9 will impact his access to assistance programmes, including educational costs for his child.

Amad’s experience is equally unique. A resident of Mantrijeron, Yogyakarta City, who works irregular manual jobs, was surprised to find his decile position fluctuating rapidly. Initially, Amad was in decile 4. After the recent economic census, he moved to decile 6. A few days later, he was placed in decile 9. Similarly, Anif Ashar (47), a street food vendor and Quran teacher from Malang Regency, East Java, who earns less than Rp 1 million, was confused to see his profile placed in decile 9.

Diandra Fristi also had a notable experience. A third-semester student of Primary School Teacher Education at the University of Negeri Surabaya (Unesa) was once threatened with being unable to continue her studies. Coming from an underprivileged family, Diandra applied for the ‘Be Smart’ scholarship through an application opened by the Surabaya City Government. Her application was rejected by the system because she was identified as not belonging to a poor or underprivileged family.

On her scholarship website, it was stated that Diandra’s family was in decile 7 and owned three cars, whereas on the Ministry of Social Affairs website, her family was in decile 2. Her father works as a water gallon delivery man, and her mother is unemployed; furthermore, the family does not even own a motorcycle. Diandra’s experience went viral on social media, with many citizens admitting to similar anxieties regarding decile data that does not match their economic reality.

Fortunately, the government has been responsive to these criticisms and objections. The government has provided opportunities for every citizen to submit corrections to their decile profiles. The decile methodology itself is not inherently flawed. However, when the results of applying that methodology do not align with field facts, the immediate question is the quality of the data.

Social, economic, and health aspects of the population, including education and housing, are always changing alongside life’s dynamics. Consequently, the various data points used as references for deciles are dynamic and have the potential to change. A family that originally lived in a rented house always has the opportunity to live in their own home when they are able to pay a down payment and instalments months later. Conversely, there are families who are forced to delay a child’s education due to a lack of funds.

Thus, administrative data does not always reflect the reality…