Datasets · Knowledge Systems
Menyerap Tanpa MenciptaKapasitas produksi pengetahuan negara-negara ASEAN: dataset komparatif terdokumentasi dengan tolok ukur Korea Selatan, Tiongkok, dan India
Asia Tenggara telah menjadi salah satu penyerap teknologi paling efektif di dunia, tetapi tetap menjadi produsen pengetahuan yang tipis. Dataset ini menghimpun belanja riset, jumlah peneliti, artikel ilmiah, paten, pendidikan tinggi, dan peringkat inovasi untuk sebelas negara, lengkap dengan tahun, sumber, dan kamus data, lalu membaca celah antara menyerap dan mencipta sebagai persoalan arsitektur sistem pengetahuan nasional.
Daftar Isi

Abstrak
Kajian ini menyajikan dataset komparatif terdokumentasi tentang kapasitas produksi pengetahuan delapan negara ASEAN (Indonesia, Malaysia, Singapura, Thailand, Vietnam, Filipina, Brunei, Kamboja) dengan Korea Selatan, Tiongkok, dan India sebagai tolok ukur. Data dihimpun dari World Development Indicators Bank Dunia (yang meneruskan statistik UNESCO Institute for Statistics dan WIPO), profil Global Innovation Index 2026, serta laporan Indikator Iptek BRIN 2024, dengan tahun observasi terakhir dicatat untuk setiap nilai dan “n.a.” untuk nilai yang tidak terverifikasi. Kerangka analisisnya membedakan tiga lapis: input riset, output pengetahuan berupa publikasi, dan output teknologi berupa paten dan penerimaan kekayaan intelektual. Temuan pokoknya adalah celah sistematis antara kapasitas menyerap dan kapasitas mencipta. Malaysia, Filipina, dan Thailand mengekspor produk berteknologi tinggi dalam porsi besar, tetapi hanya 11 sampai 19 persen permohonan paten di negara mereka diajukan oleh penduduk, dan pembayaran royalti mereka sebelas sampai dua puluh kali lipat penerimaannya. Indonesia menambah artikel ilmiah dua puluh delapan kali lipat sejak 2010, tetapi permohonan paten residennya turun dari puncak 2019 dan belanja riset menurut ukuran BRIN hanya 0,10 persen PDB pada 2023. Kajian ini menutup dengan catatan metode, kamus data, skenario hingga 2030, dan rekomendasi bagi kementerian, BRIN, Bappenas, universitas, ASEAN COSTI, UNESCO, dan Bank Dunia.
Asia Tenggara tidak kekurangan akses terhadap pengetahuan dunia; ia kekurangan simpul domestik yang mengubah akses itu menjadi kemampuan mencipta, dan celah itu lebih ditentukan oleh arsitektur institusi serta koalisi politik daripada oleh jumlah publikasi.
- Input risetBelanja litbang terhadap PDB, peneliti per sejuta penduduk, dan partisipasi pendidikan tinggi.
- Output pengetahuanArtikel ilmiah dan teknis, absolut dan per sejuta penduduk, beserta lintasannya sejak 2010.
- Output teknologiPaten residen, rasio paten nonresiden, ekspor teknologi tinggi, dan neraca royalti kekayaan intelektual.
Kata kunci
- sistem inovasi nasional
- kapasitas absorpsi
- belanja riset dan pengembangan
- ASEAN
- jebakan pendapatan menengah
- paten residen
- BRIN
- dataset terdokumentasi
Pertanyaan, jawaban pokok, dan mengapa ini penting
Seberapa besar sesungguhnya kapasitas negara-negara Asia Tenggara untuk memproduksi pengetahuan baru, bukan sekadar mengonsumsi, mengadaptasi, atau merakit pengetahuan yang diciptakan di tempat lain? Pertanyaan itu sering dijawab dengan satu angka, biasanya peringkat dalam indeks inovasi global atau jumlah artikel terindeks, lalu diperdebatkan seolah-olah angka itu menggambarkan keseluruhan sistem. Dataset ini menolak jalan pintas tersebut. Ia menghimpun sebelas indikator dari sumber multilateral yang benar-benar diambil pada akhir Oktober 2026, mencatat tahun observasi setiap nilai, menandai kekosongan secara terbuka, dan menyusunnya dalam kerangka yang memisahkan input, output pengetahuan, dan output teknologi.
Jawaban pokoknya dapat dinyatakan dalam satu kalimat. Kawasan ini telah menjadi penyerap teknologi yang sangat efektif, tetapi sebagian besar negaranya belum membangun simpul domestik yang mengubah penyerapan menjadi penciptaan. Bukti paling tajam datang dari kontras antara dua kelompok angka. Di satu sisi, ekspor berteknologi tinggi mencapai 60,4 persen ekspor manufaktur Filipina, 59,4 persen Singapura, dan 58,6 persen Malaysia pada 2024 menurut World Development Indicators, angka yang lebih tinggi daripada Korea Selatan (36,3 persen). Di sisi lain, penduduk Malaysia hanya mengajukan 841 permohonan paten pada 2024, Filipina 873, dan Thailand 930, sementara Korea mengajukan 195.786 menurut data paten residen WIPO yang diteruskan Bank Dunia. Kawasan ini mengirimkan cip ke seluruh dunia, tetapi hampir tidak mematenkan apa pun yang terkandung di dalamnya (tingkat keyakinan: tinggi).
Indonesia menempati posisi yang khas dalam pola itu. Negeri ini tidak termasuk penyerap teknologi tinggi yang kuat, karena ekspor teknologi tingginya hanya 8,7 persen ekspor manufaktur pada 2024, dan juga belum menjadi produsen teknologi, karena belanja riset nasionalnya menurut Indikator Iptek, Riset, dan Inovasi Indonesia 2024 terbitan BRIN hanya sekitar Rp20,39 triliun atau 0,10 persen PDB pada 2023. Satu-satunya lapis yang tumbuh dramatis adalah publikasi: jumlah artikel ilmiah dan teknis naik dari 1.391 pada 2010 menjadi 38.992 pada 2023. Indonesia dengan demikian menghadapi celah ganda, yaitu belum cukup menyerap dan belum cukup mencipta, dan pertumbuhan publikasi belum tersambung ke kedua lapis lainnya (tingkat keyakinan: tinggi untuk arah, sedang untuk besaran karena perbedaan metode pengukuran yang diuraikan di Bagian V).
Persoalan ini penting secara global karena tiga alasan. Pertama, laporan World Development Report 2024 Bank Dunia menempatkan peralihan dari investasi ke penyerapan lalu ke inovasi sebagai syarat keluar dari jebakan pendapatan menengah, dan sebagian besar negara ASEAN kini berada tepat di persimpangan tahap kedua dan ketiga. Kedua, peta rantai pasok teknologi sedang digambar ulang oleh persaingan Amerika Serikat dan Tiongkok, sehingga negara yang hanya menjadi lokasi perakitan berisiko ditinggalkan ketika keputusan relokasi berikutnya dibuat. Ketiga, negara-negara dengan populasi terbesar di kawasan, yaitu Indonesia, Filipina, dan Vietnam, akan menentukan apakah Asia Tenggara secara keseluruhan menjadi kutub pengetahuan ketiga Asia atau tetap menjadi pinggiran yang produktif di antara Asia Timur Laut dan Asia Selatan.
Dataset ini disusun sebagai bahan ajar bagi School of Knowledge Systems & Knowledge Management, sehingga ia dirancang agar dapat diperiksa ulang, diperbarui, dan dibantah. Setiap angka dalam tabel menyertakan tahun dan sumber; setiap kode indikator dicatat dalam kamus data; dan setiap keputusan harmonisasi dijelaskan. Kajian ini sengaja tidak mengulang perdebatan tentang metrik publikasi atau tentang miskinnya teori dalam ilmu sosial, dua tema yang sudah dibahas di tempat lain. Fokusnya adalah kapasitas sistem nasional: siapa yang membiayai riset, siapa yang mengerjakannya, apa yang dihasilkan, dan apakah hasil itu kembali ke ekonomi dalam bentuk kemampuan teknologi yang dimiliki sendiri.
Kerangka analitis: sistem inovasi nasional dan tiga lapis kapasitas
Kerangka kajian ini bertumpu pada tradisi sistem inovasi nasional. Christopher Freeman, dalam artikel tahun 1995 berjudul “The ‘National System of Innovation’ in Historical Perspective” di Cambridge Journal of Economics, menelusuri gagasan itu hingga Friedrich List dan menunjukkan bahwa kinerja teknologi suatu bangsa ditentukan oleh jaringan lembaga, yakni perusahaan, universitas, lembaga pemerintah, dan sistem pendidikan, beserta cara jaringan itu belajar dan mengimpor teknologi. Bengt-Åke Lundvall, dalam buku National Systems of Innovation: Toward a Theory of Innovation and Interactive Learning (1992; terbit ulang Anthem Press), menambahkan bahwa inti sistem itu adalah pembelajaran interaktif antara pengguna dan produsen pengetahuan. Dari kedua sumber ini kajian ini mengambil satu prinsip metodologis: kapasitas produksi pengetahuan tidak dapat diukur dari satu simpul, karena yang menentukan adalah sambungan antarsimpul.
Prinsip kedua diambil dari Wesley Cohen dan Daniel Levinthal, yang dalam artikel tahun 1990 berjudul “Absorptive Capacity: A New Perspective on Learning and Innovation” di Administrative Science Quarterly berargumen bahwa kemampuan sebuah organisasi untuk mengenali, mengasimilasi, dan memanfaatkan pengetahuan dari luar sangat bergantung pada pengetahuan terkait yang sudah dimilikinya, dan bahwa riset internal adalah salah satu cara utama membangun kemampuan itu. Implikasinya bagi analisis kebijakan sangat besar. Penyerapan dan penciptaan bukan dua strategi yang bisa dipilih salah satunya; keduanya adalah dua wajah dari investasi yang sama. Negara yang berhenti berinvestasi dalam riset sendiri, dengan alasan cukup mengandalkan teknologi impor, pada akhirnya juga kehilangan kemampuan untuk menyerap teknologi impor itu secara mendalam.
Atas dasar itu kajian ini menyusun empat dimensi pengukuran. Dimensi input mencakup belanja riset dan pengembangan sebagai persen PDB, peneliti per sejuta penduduk, dan rasio partisipasi kasar pendidikan tinggi. Dimensi output pengetahuan mencakup artikel ilmiah dan teknis. Dimensi output teknologi mencakup permohonan paten oleh penduduk. Dimensi absorpsi mencakup ekspor teknologi tinggi, permohonan paten oleh nonpenduduk, dan pembayaran serta penerimaan biaya penggunaan kekayaan intelektual. Global Innovation Index 2026 dipakai sebagai indikator komposit pembanding, dengan kesadaran bahwa ia sendiri merupakan agregasi dari banyak indikator yang sebagian tumpang tindih dengan dataset ini.
Empat dimensi kapasitas produksi pengetahuan
- 1Inputsumber dayaBelanja litbang, peneliti setara penuh waktu, partisipasi pendidikan tinggi.
- 2Output pengetahuanpublikasiArtikel ilmiah dan teknis, absolut dan per sejuta penduduk.
- 3Output teknologikepemilikanPaten residen dan penerimaan royalti kekayaan intelektual.
- 4AbsorpsipenyerapanEkspor teknologi tinggi, paten nonresiden, pembayaran royalti.
Pembeda analitis yang paling penting dalam kerangka ini adalah rasio antara dimensi keempat dan dimensi ketiga, yang kajian ini sebut celah absorpsi-produksi. Celah itu diukur dengan dua cara yang saling menguatkan. Cara pertama adalah porsi penduduk dalam total permohonan paten di dalam negeri: semakin kecil porsinya, semakin besar ketergantungan sistem pada penemu asing yang melindungi teknologi mereka di pasar domestik. Cara kedua adalah rasio pembayaran terhadap penerimaan royalti: semakin besar rasionya, semakin besar aliran sewa pengetahuan ke luar negeri dibandingkan aliran masuk dari pengetahuan yang dimiliki sendiri. Kedua ukuran itu tidak sempurna, karena perusahaan multinasional dapat memindahkan pendapatan royalti antarnegara untuk alasan pajak, tetapi keduanya menangkap hal yang luput dari indikator input dan publikasi.
Kerangka ini juga meminjam tradisi indeks ekonomi pengetahuan. Metodologi Penilaian Pengetahuan Bank Dunia yang dijelaskan Derek Chen dan Carl Dahlman dalam makalah tahun 2005 berjudul The Knowledge Economy, the KAM Methodology and World Bank Operations membagi ekonomi pengetahuan ke dalam empat pilar, yaitu rezim insentif ekonomi dan kelembagaan, pendidikan, sistem inovasi, dan teknologi informasi. Kajian ini tidak membangun indeks komposit baru, karena indeks semacam itu cenderung menyembunyikan kontradiksi antardimensi di balik satu skor. Sebaliknya, kajian ini membiarkan indikator berdiri berdampingan agar kontradiksinya terlihat, karena justru kontradiksi itulah yang menjadi objek analisis.
Catatan metode ringkasnya sebagai berikut. Sumber utama adalah World Development Indicators Bank Dunia versi pembaruan 8 Oktober 2026, yang untuk belanja litbang dan peneliti meneruskan data UNESCO Institute for Statistics, untuk paten meneruskan data WIPO, dan untuk publikasi meneruskan Science and Engineering Indicators National Science Foundation Amerika Serikat yang menghitung artikel terindeks Scopus. Untuk setiap negara dan indikator diambil observasi terakhir yang tersedia, sehingga tahun observasi berbeda antarnegara dan dicatat di setiap sel. Peringkat inovasi diambil dari profil negara Global Innovation Index 2026 yang diterbitkan WIPO pada akhir September 2026. Data nasional Indonesia dibandingkan dengan laporan BRIN 2024. Periode lintasan yang dianalisis adalah 1996 sampai 2025, bergantung pada ketersediaan seri. Pendekatannya adalah perbandingan deskriptif terstruktur, bukan estimasi kausal, dan setiap penilaian utama diberi tingkat keyakinan.
Dataset inti: input, output, dan posisi komposit
Tabel 1 menyajikan dimensi input untuk sebelas negara. Angka belanja litbang dan peneliti berasal dari seri GB.XPD.RSDV.GD.ZS dan SP.POP.SCIE.RD.P6, sedangkan partisipasi pendidikan tinggi berasal dari seri SE.TER.ENRR. Setiap nilai adalah observasi terakhir yang tersedia dalam World Development Indicators per 8 Oktober 2026, dibulatkan untuk keterbacaan, dengan tahun observasi di dalam kurung.
| Negara | Belanja litbang (% PDB) | Peneliti per sejuta penduduk (FTE) | APK pendidikan tinggi (%) |
|---|---|---|---|
| Indonesia | 0,28 (2020) | 395 (2020) | 44,9 (2023) |
| Malaysia | 1,01 (2022) | 1.218 (2022) | 38,5 (2024) |
| Singapura | 1,81 (2022) | 8.782 (2022) | 97,3 (2023) |
| Thailand | 0,94 (2023) | 1.592 (2023) | 49,1 (2025) |
| Vietnam | 0,41 (2023) | 836 (2023) | 37,6 (2024) |
| Filipina | 0,28 (2020) | 320 (2020) | 47,4 (2024) |
| Brunei Darussalam | 0,28 (2018) | 506 (2022) | 36,4 (2023) |
| Kamboja | 0,09 (2015) | 30 (2015) | 21,4 (2024) |
| Korea Selatan | 4,94 (2023) | 9.472 (2023) | 111,9 (2024) |
| Tiongkok | 2,58 (2023) | 2.107 (2023) | 76,9 (2024) |
| India | 0,65 (2020) | 259 (2020) | 34,4 (2025) |
Tabel 1 sudah memperlihatkan tiga kelompok yang jelas. Kelompok pertama adalah Korea dan Singapura, dengan peneliti di atas 8.700 per sejuta penduduk dan belanja riset di atas 1,8 persen PDB. Kelompok kedua adalah Malaysia, Thailand, dan Tiongkok, dengan peneliti antara 1.200 dan 2.100 per sejuta, meskipun intensitas belanja Tiongkok jauh lebih tinggi daripada dua negara ASEAN itu. Kelompok ketiga adalah Indonesia, Vietnam, Filipina, Brunei, Kamboja, dan India, dengan peneliti di bawah 850 per sejuta. Vietnam berada di batas atas kelompok ketiga dengan 836 peneliti per sejuta pada 2023, lebih dari dua kali lipat angka Indonesia tahun 2020, padahal belanja riset kedua negara relatif dekat. Perbedaan itu menunjukkan bahwa jumlah peneliti dapat tumbuh lebih cepat daripada belanja ketika negara memprioritaskan pembentukan tenaga riset, tetapi juga mengingatkan bahwa peneliti yang kurang didanai mudah menjadi pengajar penuh waktu dengan gelar peneliti.
Kolom partisipasi pendidikan tinggi memperlihatkan paradoks yang tidak boleh dilewatkan. Indonesia (44,9 persen) dan Filipina (47,4 persen) memiliki partisipasi pendidikan tinggi yang lebih tinggi daripada Malaysia (38,5 persen) dan Vietnam (37,6 persen), tetapi intensitas peneliti mereka jauh lebih rendah. Dengan kata lain, perluasan akses pendidikan tinggi di kedua negara kepulauan itu belum dikonversi menjadi kapasitas riset. Sistem pendidikan tinggi mereka tumbuh terutama sebagai sistem pengajaran massal, bukan sebagai sistem produksi pengetahuan, dan ini adalah temuan struktural yang lebih penting daripada selisih peringkat mana pun (tingkat keyakinan: tinggi).
Tabel 2 menyajikan dimensi output dan posisi komposit. Artikel berasal dari seri IP.JRN.ARTC.SC untuk 2023, paten residen dari seri IP.PAT.RESD untuk 2024, dan nilai per sejuta penduduk dihitung kajian ini dengan membagi angka absolut dengan populasi 2023 dari seri SP.POP.TOTL. Peringkat Global Innovation Index 2026 berasal dari profil negara WIPO untuk Indonesia, Malaysia, Singapura, Thailand, Vietnam, Filipina, Korea, Tiongkok, dan India; peringkat Brunei dan Kamboja diambil dari rangkuman SeasiaStats karena profil WIPO keduanya tidak diambil dalam kajian ini.
| Negara | Artikel ilmiah dan teknis (2023) | Artikel per sejuta penduduk | Paten residen (2024) | Paten residen per sejuta penduduk | GII 2026: peringkat (input/output) |
|---|---|---|---|---|---|
| Indonesia | 38.992 | 139 | 1.931 | 6,9 | 55 (60/60) |
| Malaysia | 23.691 | 674 | 841 | 23,9 | 34 (29/40) |
| Singapura | 11.894 | 2.010 | 1.898 | 320,7 | 5 (1/9) |
| Thailand | 16.656 | 232 | 930 | 13,0 | 44 (46/41) |
| Vietnam | 10.644 | 106 | 1.226 | 12,2 | 43 (54/35) |
| Filipina | 4.752 | 41 | 873 | 7,6 | 52 (69/49) |
| Brunei Darussalam | 502 | 1.093 | 1 | 2,2 | 98 (n.a.) |
| Kamboja | 201 | 12 | 1 | 0,1 | 93 (n.a.) |
| Korea Selatan | 71.724 | 1.387 | 195.786 | 3.786 | 4 (4/6) |
| Tiongkok | 932.712 | 661 | 1.672.001 | 1.185 | 10 (22/5) |
| India | 228.174 | 159 | 63.217 | 44,0 | 38 (49/25) |
Kolom terakhir Tabel 2 menyimpan informasi yang jarang dibaca, yaitu selisih antara peringkat input dan peringkat output dalam Global Innovation Index. Negara yang peringkat outputnya lebih baik daripada peringkat inputnya mengubah sumber daya menjadi hasil secara relatif efisien; negara dengan pola sebaliknya memiliki sumber daya dan lingkungan kelembagaan yang relatif baik tetapi hasilnya tertinggal. Vietnam (input 54, output 35), India (input 49, output 25), dan Tiongkok (input 22, output 5) berada di kelompok pertama. Malaysia (input 29, output 40) dan Singapura (input 1, output 9) berada di kelompok kedua, sedangkan Indonesia berada tepat di tengah dengan peringkat 60 untuk keduanya. Profil WIPO juga mencatat bahwa Vietnam dan India adalah dua ekonomi yang melampaui ekspektasi tingkat pembangunannya selama enam belas tahun berturut-turut, sebagaimana dilaporkan VietnamPlus pada 3 Oktober 2026.
Perbandingan publikasi dengan paten di Tabel 2 menghasilkan pembacaan yang paling tajam. Indonesia menghasilkan artikel lebih banyak daripada Malaysia, Thailand, dan Vietnam secara absolut, tetapi per sejuta penduduk ia hanya menghasilkan seperlima dari Malaysia, dan paten residen per sejuta penduduknya (6,9) lebih rendah daripada Thailand, Vietnam, dan Malaysia, dan hanya sedikit di bawah Filipina. Singapura, dengan penduduk sekitar 5,9 juta, mengajukan hampir sama banyak paten residen dengan Indonesia yang berpenduduk lebih dari 281 juta. Korea mengajukan sekitar 2,7 paten residen untuk setiap artikel ilmiah, sedangkan Indonesia mengajukan sekitar lima paten untuk setiap seratus artikel. Rasio itu bukan ukuran mutu ilmu, tetapi ukuran sejauh mana kerja ilmiah tersambung dengan kerja teknologi di dalam sistem (tingkat keyakinan: tinggi).
Kamus data
Kamus data berikut mendefinisikan setiap variabel dalam dataset, satuannya, sumber aslinya, kode seri bila ada, dan cakupan negara serta tahunnya. Variabel turunan ditandai sebagai hasil hitung kajian ini, sehingga pembaca dapat membedakan angka yang dikutip langsung dari sumber dengan angka yang merupakan hasil operasi aritmetika atas angka sumber. Kamus ini dimaksudkan agar dataset dapat diperbarui setiap tahun oleh mahasiswa atau analis lain tanpa harus menebak keputusan yang diambil pada versi pertama.
| Variabel | Definisi | Satuan | Sumber | Kode | Cakupan |
|---|---|---|---|---|---|
| Belanja litbang | Belanja domestik bruto untuk riset dan pengembangan, modal dan rutin, di sektor bisnis, pemerintah, pendidikan tinggi, dan nirlaba swasta | % PDB | UNESCO UIS melalui WDI (rilis UIS Februari 2025) | GB.XPD.RSDV.GD.ZS | 11 negara; 2015–2023; seri 1996–2023 untuk enam negara |
| Peneliti | Profesional yang menghasilkan pengetahuan, produk, proses, atau metode baru, dihitung setara penuh waktu | per sejuta penduduk | UNESCO UIS melalui WDI | SP.POP.SCIE.RD.P6 | 11 negara; 2015–2023 |
| APK pendidikan tinggi | Jumlah mahasiswa terhadap populasi kelompok usia pendidikan tinggi resmi; dapat melebihi 100 | % | UNESCO UIS melalui WDI | SE.TER.ENRR | 11 negara; 2023–2025 |
| Artikel ilmiah dan teknis | Artikel jurnal dan prosiding terindeks Scopus di bidang sains dan rekayasa | jumlah | NSF Science and Engineering Indicators melalui WDI | IP.JRN.ARTC.SC | 11 negara; 2010–2023 |
| Artikel per sejuta penduduk | Artikel 2023 dibagi populasi 2023 | per sejuta | Hitungan kajian ini | turunan | 11 negara; 2023 |
| Paten residen | Permohonan paten yang diajukan penduduk di kantor paten nasional | jumlah | WIPO melalui WDI | IP.PAT.RESD | 11 negara; 2010–2024 |
| Paten nonresiden | Permohonan paten yang diajukan pihak asing di kantor paten nasional | jumlah | WIPO melalui WDI | IP.PAT.NRES | 9 negara; 2024 |
| Porsi paten residen | Paten residen dibagi jumlah paten residen dan nonresiden | % | Hitungan kajian ini | turunan | 9 negara; 2024 |
| Ekspor teknologi tinggi | Ekspor produk berintensitas litbang tinggi terhadap ekspor manufaktur | % | WDI | TX.VAL.TECH.MF.ZS | 9 negara; 2023–2024 |
| Pembayaran royalti KI | Biaya penggunaan kekayaan intelektual yang dibayarkan ke luar negeri, neraca pembayaran | USD harga berlaku | WDI | BM.GSR.ROYL.CD | 8 negara; 2025 |
| Penerimaan royalti KI | Biaya penggunaan kekayaan intelektual yang diterima dari luar negeri | USD harga berlaku | WDI | BX.GSR.ROYL.CD | 8 negara; 2025 |
| Rasio royalti | Pembayaran dibagi penerimaan royalti KI | kali | Hitungan kajian ini | turunan | 8 negara; 2025 |
| Peringkat GII | Peringkat Global Innovation Index, beserta subindeks input dan output | peringkat dari 139 | WIPO, profil negara GII 2026 | tidak ada | 11 negara; edisi 2026 |
| Belanja riset nasional (BRIN) | Jumlah belanja riset pemerintah, perguruan tinggi, dan badan usaha menurut pedoman Frascati 2015 | Rp dan % PDB | BRIN, Indikator Iptek 2024 | tidak ada | Indonesia; 2023 |
| Populasi | Jumlah penduduk pertengahan tahun | jiwa | WDI | SP.POP.TOTL | 11 negara; 2023 |
Dua keputusan dalam kamus ini perlu dijelaskan. Pertama, artikel ilmiah dan teknis tidak mencakup humaniora dan sebagian besar ilmu sosial, karena taksonomi National Center for Science and Engineering Statistics berfokus pada sains dan rekayasa, sehingga indikator ini mengukur kapasitas sains-rekayasa, bukan seluruh produksi akademik. Kedua, paten residen menghitung permohonan, bukan paten yang diberikan, dan menghitungnya di kantor paten nasional masing-masing negara. Ukuran ini peka terhadap kebijakan insentif: lonjakan permohonan dapat terjadi karena syarat administratif bagi dosen atau peneliti, bukan karena lonjakan invensi. Karena itu kajian ini membaca angka paten bersama lintasannya, bukan sebagai titik tunggal.
Catatan metode dan mutu data: seleksi, harmonisasi, vintase, dan batas
Seleksi negara mengikuti tiga kriteria. Enam negara inti ASEAN dipilih karena merupakan ekonomi terbesar di kawasan dan memiliki seri data paling lengkap. Brunei dan Kamboja dimasukkan karena datanya tersedia untuk sebagian besar indikator, meskipun dengan tahun yang tua, sehingga keberadaannya memperlihatkan rentang kapasitas di dalam ASEAN. Laos dan Myanmar, yang menurut rangkuman SeasiaStats masing-masing berada di peringkat 108 dan 125 dalam Global Innovation Index 2026, tidak dimasukkan karena kajian ini tidak mengambil seri belanja litbang dan peneliti untuk keduanya, sehingga memasukkan mereka hanya dengan peringkat komposit akan menghasilkan perbandingan yang timpang. Korea Selatan dipilih sebagai contoh paling berhasil dari peralihan penyerap menjadi pencipta, Tiongkok sebagai contoh peralihan berskala benua, dan India sebagai pembanding negara berpendapatan menengah bawah yang besar dengan kekuatan output yang tidak sebanding dengan inputnya.
Harmonisasi dilakukan secara minimal dan eksplisit. Kajian ini tidak mengimputasi nilai yang hilang, tidak menginterpolasi antartahun, dan tidak menyesuaikan nilai ke tahun dasar bersama. Akibatnya, sel-sel dalam satu kolom dapat berasal dari tahun yang berbeda hingga delapan tahun, seperti belanja litbang Kamboja (2015) dibandingkan dengan Korea (2023). Pilihan ini sengaja diambil karena imputasi akan memberi kesan presisi yang tidak dimiliki data. Variabel per sejuta penduduk dihitung dengan populasi 2023 untuk semua negara; untuk paten tahun 2024 hal ini menghasilkan sedikit penaksiran berlebih, tetapi pertumbuhan penduduk satu tahun di negara-negara ini terlalu kecil untuk mengubah urutan atau besaran secara berarti.
Masalah vintase jauh lebih serius daripada yang biasanya diakui. Metadata World Development Indicators menyatakan bahwa seri belanja litbang dan peneliti diambil dari layanan unduhan massal UNESCO Institute for Statistics yang terbit Februari 2025, meskipun basis data Bank Dunia itu sendiri diperbarui pada Oktober 2026. Profil Global Innovation Index 2026 untuk Korea mencantumkan belanja litbang 5,13 persen PDB untuk 2024, sementara World Development Indicators masih mencantumkan 4,94 persen untuk 2023. Dua angka resmi itu tidak saling bertentangan; keduanya berasal dari vintase yang berbeda. Namun pembaca yang mengambil satu angka tanpa tahun dan sumbernya akan menghasilkan perbandingan yang keliru. Kajian ini memilih angka World Development Indicators untuk konsistensi antarnegara dan mencatat angka WIPO yang lebih baru di dalam teks.
Kelambatan pelaporan UIS memengaruhi negara-negara yang paling penting bagi analisis ini. Observasi terakhir belanja litbang dan peneliti untuk Indonesia, Filipina, dan India berasal dari 2020, tahun pandemi ketika PDB turun dan rasio belanja terhadap PDB dapat terdistorsi. Profil WIPO untuk Indonesia secara eksplisit menandai enam belas indikator sebagai usang, termasuk belanja litbang, peneliti, dan belanja litbang sektor bisnis, sedangkan profil India mencatat bahwa angka belanja litbang 0,65 persen dan 259,3 peneliti per sejuta adalah data 2020. Ini berarti bahwa posisi tiga negara berpenduduk besar dalam setiap indeks inovasi global sebagian ditentukan oleh data yang berumur enam tahun pada saat indeks diterbitkan (tingkat keyakinan: tinggi).
Kasus Indonesia memerlukan catatan tersendiri karena reorganisasi kelembagaan riset sejak pembentukan BRIN mengubah cara data dikumpulkan dan dilaporkan. Laporan BRIN tahun 2024 menghitung belanja riset nasional 2023 sebesar Rp20.394.979.384.006, atau sekitar 0,10 persen dari PDB harga berlaku Rp20.892,4 triliun, dan menyatakan rasio itu tidak berubah dari tahun sebelumnya. Angka itu kurang dari separuh angka 0,28 persen untuk 2020 yang diteruskan UIS dan dipakai Global Innovation Index. Laporan BRIN sendiri menjelaskan bahwa datanya berasal dari beberapa instansi, yaitu BRIN untuk pemerintah, pusat data kementerian pendidikan untuk perguruan tinggi, sistem informasi industri Kementerian Perindustrian untuk badan usaha, dan Kementerian Dalam Negeri untuk pemerintah daerah, bahwa beberapa angka masih sementara, dan bahwa pengumpulan data sektor bisnis adalah tantangan utamanya. Kajian ini tidak dapat menentukan apakah selisih antara 0,28 dan 0,10 persen mencerminkan penurunan nyata, perbedaan cakupan survei, atau keduanya, sehingga kedua angka dilaporkan berdampingan (tingkat keyakinan bahwa belanja riset Indonesia berada di bawah 0,3 persen PDB: tinggi; tingkat keyakinan tentang besaran pastinya: rendah).
Perbedaan definisi juga memengaruhi angka peneliti. UIS melaporkan 395 peneliti setara penuh waktu per sejuta penduduk untuk 2020. Laporan BRIN 2024 melaporkan 439.349 periset dalam hitungan kepala untuk 2023, yang jika dibagi populasi 279.179.042 jiwa yang dipakai laporan itu menghasilkan sekitar 1.574 periset per sejuta penduduk. Laporan yang sama tidak menyajikan angka setara penuh waktu dan mencatat bahwa 96,62 persen periset berada di perguruan tinggi, yang mencakup dosen dan mahasiswa doktoral terdaftar. Selisih empat kali lipat antara kedua angka itu bukan kesalahan; ia adalah ukuran berapa besar waktu dosen Indonesia yang secara efektif tersedia untuk riset. Hitungan kepala memotret jumlah orang yang secara administratif dapat meneliti, sedangkan hitungan setara penuh waktu memotret kerja riset yang benar-benar dilakukan.
Batas terakhir menyangkut indikator absorpsi. Data pembayaran dan penerimaan royalti untuk Vietnam, Brunei, dan Kamboja tidak tersedia dalam penarikan yang dilakukan, sehingga ditulis “n.a.”. Ekspor teknologi tinggi Vietnam tersedia hingga 2023, bukan 2024. Pembayaran royalti juga dapat mencerminkan praktik penetapan harga transfer perusahaan multinasional, bukan semata aliran pengetahuan. Untuk alasan itu kajian ini memakai indikator absorpsi untuk membaca pola besar, bukan untuk membandingkan selisih kecil antarnegara.
Celah absorpsi dan produksi: membaca kontradiksi kawasan
Tabel 4 menyandingkan indikator absorpsi dengan indikator kepemilikan teknologi. Ekspor teknologi tinggi berasal dari seri TX.VAL.TECH.MF.ZS, paten nonresiden dari seri IP.PAT.NRES, serta pembayaran dan penerimaan royalti dari seri BM.GSR.ROYL.CD dan BX.GSR.ROYL.CD. Porsi paten residen dan rasio royalti adalah hitungan kajian ini.
| Negara | Ekspor teknologi tinggi (% ekspor manufaktur, 2024) | Paten nonresiden (2024) | Porsi paten residen (%) | Pembayaran royalti KI (miliar USD, 2025) | Penerimaan royalti KI (miliar USD, 2025) | Rasio bayar/terima |
|---|---|---|---|---|---|---|
| Indonesia | 8,7 | 8.971 | 17,7 | 2,49 | 0,22 | 11,4 |
| Malaysia | 58,6 | 6.356 | 11,7 | 2,66 | 0,24 | 11,2 |
| Singapura | 59,4 | 11.523 | 14,1 | 20,69 | 15,68 | 1,3 |
| Thailand | 28,3 | 7.797 | 10,7 | 6,97 | 0,35 | 19,9 |
| Vietnam | 44,3 (2023) | 8.678 | 12,4 | n.a. | n.a. | n.a. |
| Filipina | 60,4 | 3.698 | 19,1 | 0,38 | 0,02 | 16,1 |
| Korea Selatan | 36,3 | 50.459 | 79,5 | 14,14 | 9,98 | 1,4 |
| Tiongkok | 26,3 | 156.053 | 91,5 | 46,63 | 13,07 | 3,6 |
| India | 18,6 | 41.940 | 60,1 | 18,76 | 1,74 | 10,8 |
Pola dalam Tabel 4 konsisten dan mencolok. Di semua negara ASEAN dalam tabel, porsi penduduk dalam total permohonan paten berada di antara 10,7 dan 19,1 persen. Di Korea angkanya 79,5 persen, di Tiongkok 91,5 persen, dan bahkan di India, dengan belanja riset yang tidak lebih tinggi daripada Thailand, angkanya 60,1 persen. Artinya, kantor paten di Asia Tenggara terutama berfungsi sebagai tempat perusahaan asing melindungi teknologi mereka di pasar kawasan, bukan sebagai tempat penemu lokal mendaftarkan invensi. Kajian ini menilai bahwa porsi paten residen di bawah dua puluh persen di seluruh ASEAN adalah indikator paling ringkas dari celah absorpsi-produksi kawasan (tingkat keyakinan: tinggi).
Neraca royalti mengonfirmasi pembacaan itu dari sisi aliran uang. Thailand membayar hampir dua puluh kali lipat lebih banyak royalti kekayaan intelektual daripada yang diterimanya pada 2025, Filipina enam belas kali, dan Indonesia serta Malaysia sekitar sebelas kali. Korea, sebaliknya, membayar hanya 1,4 kali penerimaannya, dan Singapura 1,3 kali. Bagi Singapura, rasio rendah itu sebagian mencerminkan perannya sebagai pusat regional kepemilikan kekayaan intelektual perusahaan multinasional, sehingga angka itu tidak dapat dibaca langsung sebagai keunggulan penemu lokal. Namun bagi Korea, rasio itu merupakan hasil dari empat dekade investasi yang membuat perusahaan domestik menjadi pemilik teknologi, bukan sekadar penggunanya.
Kasus Malaysia dan Filipina adalah ilustrasi paling jelas dari absorpsi tanpa produksi. Kedua negara mengekspor produk teknologi tinggi dengan porsi yang lebih besar daripada Korea dan Tiongkok, tetapi porsi paten residen mereka termasuk yang terendah. Profil WIPO menempatkan Malaysia di peringkat pertama dunia untuk ekspor dan impor teknologi tinggi sebagai persen total perdagangan, dan Filipina di peringkat keempat untuk ekspor teknologi tinggi. Struktur seperti ini lazim terjadi ketika ekspor teknologi tinggi terutama berupa perakitan, pengujian, dan pengemasan komponen elektronik oleh anak perusahaan asing, dengan riset dan desain dilakukan di negara asal perusahaan induk. Profil WIPO untuk Filipina juga mencatat bahwa pilar modal manusia dan riset negara itu berada di peringkat 100, sebuah kontras tajam dengan peringkat keempat untuk ekspor teknologi tinggi.
Kawasan ini mengirimkan cip ke seluruh dunia, tetapi hampir tidak mematenkan apa pun yang terkandung di dalamnya.
Vietnam menunjukkan varian yang berbeda dan lebih dinamis. Porsi paten residennya (12,4 persen) masih rendah, tetapi angka mutlak paten residennya naik dari 306 pada 2010 menjadi 1.226 pada 2024, peningkatan empat kali lipat yang berlangsung hampir tanpa jeda, dan peringkat outputnya dalam Global Innovation Index (35) jauh di atas peringkat inputnya (54). Vietnam juga satu-satunya negara inti ASEAN yang memiliki target kuantitatif eksplisit untuk menaikkan intensitas riset. Opini tiga peneliti BRIN di The Jakarta Post pada Maret 2025 mencatat bahwa Resolusi Politbiro Nomor 57 menargetkan kenaikan belanja litbang dari 0,4 persen menjadi 2 persen PDB dan 12 peneliti per 10.000 penduduk pada 2030. Target itu sangat ambisius dan kemungkinan besar tidak akan tercapai penuh, tetapi ia mengubah percakapan kebijakan dari penyerapan investasi asing menjadi pembangunan kemampuan domestik.
Pelajaran analitis dari bagian ini adalah bahwa indikator absorpsi yang tinggi tidak otomatis menandakan sistem pengetahuan yang kuat, dan bahwa indikator input yang sedang tidak otomatis menandakan sistem yang lemah. Negara yang tampak maju dalam ekspor teknologi tinggi dapat sangat rapuh bila keputusan lokasi produksi dibuat di tempat lain, sedangkan negara dengan input sedang dapat menunjukkan kemampuan konversi yang kuat bila simpul-simpulnya tersambung. Dalam bahasa Lundvall, yang menentukan adalah intensitas pembelajaran interaktif antara produsen dan pengguna pengetahuan di dalam sistem, dan indikator absorpsi semata tidak memberi tahu apakah pembelajaran itu terjadi di dalam negeri atau hanya melintas.
Posisi Indonesia: publikasi melonjak, sistem tertinggal
Lintasan Indonesia sejak 2010 adalah lintasan yang tidak seimbang. Jumlah artikel ilmiah dan teknis naik dari 1.391 (2010) menjadi 2.652 (2013), 5.850 (2016), 17.156 (2019), 23.654 (2021), dan 38.992 (2023). Peningkatan dua puluh delapan kali lipat dalam tiga belas tahun itu jauh melampaui Vietnam (hampir sepuluh kali lipat, dari 1.096 menjadi 10.644) dan Malaysia (sekitar dua kali lipat, dari 10.816 menjadi 23.691). Pada periode yang sama, permohonan paten residen Indonesia naik dari 508 (2010) menjadi 1.058 (2015) dan mencapai puncak 3.093 pada 2019, lalu turun menjadi 1.549 pada 2022 sebelum pulih sebagian menjadi 1.931 pada 2024. Lapis publikasi tumbuh eksponensial, sedangkan lapis teknologi berfluktuasi tanpa tren yang meyakinkan.
Kajian ini tidak membahas mutu atau insentif di balik pertumbuhan publikasi itu, yang merupakan perdebatan tersendiri. Pertanyaan yang relevan bagi kapasitas sistem adalah mengapa pertumbuhan publikasi tidak tersambung ke lapis teknologi. Jawaban yang paling didukung data adalah struktur pelaku riset. Laporan BRIN 2024, sebagaimana juga dirangkum Good News From Indonesia pada September 2024, mencatat bahwa dari total belanja riset 2023, perguruan tinggi menyumbang 46,02 persen, pemerintah 32,42 persen, dan badan usaha atau industri hanya 21,56 persen. Dari sisi sumber daya manusia, 95,28 persen tenaga iptek berada di perguruan tinggi dan hanya 1,40 persen di badan usaha. Di Korea, profil WIPO mencatat belanja litbang yang dilakukan sektor bisnis sebesar 4,2 persen PDB, peringkat pertama dunia, sedangkan di Thailand belanja bisnis sekitar 0,6 persen PDB. Sistem Indonesia adalah sistem riset yang dijalankan hampir sepenuhnya oleh universitas, dalam ekonomi yang perusahaannya hampir tidak meneliti.
Angka porsi industri itu sendiri diperdebatkan. Dalam acara Business Gathering 2025, sebagaimana diberitakan Bisnis.com pada 28 April 2025, Menteri Pendidikan Tinggi, Sains, dan Teknologi Brian Yuliarto menyebut belanja riset Indonesia 0,28 persen dibandingkan rata-rata ASEAN 0,70 persen, dan berita yang sama mengutip porsi pendanaan industri sebesar 7,3 persen, jauh di bawah angka BRIN. Perbedaan antara 7,3 dan 21,56 persen kemungkinan berasal dari perbedaan antara siapa yang membiayai dan siapa yang melaksanakan riset, atau dari perbedaan tahun dan sumber, tetapi berita itu tidak menyebut sumbernya. Fakta bahwa menteri dan badan riset nasional memakai angka yang berbeda untuk ukuran dasar sistem adalah temuan tersendiri tentang kapasitas statistik iptek Indonesia (tingkat keyakinan: tinggi).
Reorganisasi kelembagaan menambah lapisan ketidakpastian. Peleburan lembaga-lembaga riset pemerintah ke dalam BRIN mengonsolidasikan anggaran dan peneliti, tetapi juga menempatkan sebagian besar riset pemerintah dalam satu jalur anggaran yang rentan terhadap kebijakan fiskal. Opini di The Jakarta Post yang sudah disebut mencatat bahwa dalam kebijakan efisiensi anggaran 2025, BRIN awalnya mengusulkan pemotongan sekitar Rp2 triliun dari usulan total sekitar Rp5,8 triliun, dan DPR akhirnya menyetujui pemotongan sekitar Rp1,4 triliun. Bagi sistem yang belanja riset nasionalnya hanya sekitar Rp20 triliun, pemotongan sebesar itu pada satu simpul adalah guncangan berarti. Di sisi lain, pembentukan Kementerian Pendidikan Tinggi, Sains, dan Teknologi membuka peluang untuk menyambungkan kembali kebijakan pendidikan tinggi dan riset yang sebelumnya terpisah.
Masalah Indonesia bukan hal baru, dan sejumlah dokumen dalam koleksi Arsip & Pustaka memperlihatkan bahwa diagnosisnya telah dirumuskan lebih dari satu dekade lalu. Pada tahun 2014, Hendra Gunawan dalam harian Kompas yang berjudul “Mengintip Masa Depan” menyebut universitas dan lembaga riset sebagai lumbung benih bangsa yang investasinya sangat kecil, dan membandingkan anggaran pendidikan tinggi Indonesia 2014 sekitar Rp3,2 triliun dengan sekitar Rp18 triliun untuk National University of Singapore pada 2013. Pada hari yang sama, Budi Santosa dalam harian Kompas yang berjudul “Dari ‘Brain Drain’ ke ‘Brain Gain’” berargumen bahwa kekhawatiran tentang hilangnya talenta penerima beasiswa berlebihan, dengan menunjuk Tiongkok dan Korea yang menarik kembali ilmuwannya ketika ekonomi mereka membaik, dan mengusulkan pendanaan profesor dan peneliti pascadoktoral asing di universitas Indonesia. Pidato pengukuhan guru besar Al Makin tahun 2018 berjudul Bisakah Menjadi Ilmuwan di Indonesia?: Keilmuan, Birokrasi dan Globalisasi menambahkan diagnosis mikro bahwa beban birokrasi dan administrasi menggerus waktu riset dosen, sebuah diagnosis yang kini terbaca secara kuantitatif dalam selisih empat kali lipat antara hitungan kepala dan hitungan setara penuh waktu peneliti.
Diagnosis itu juga berlaku bagi perguruan tinggi keagamaan, yang menampung sebagian besar dosen di luar sains dan rekayasa. Paparan tahun 2026 berjudul Rencana Strategis UIN Ar-Raniry menempatkan ekosistem riset sebagai salah satu strategi transformasi, berdampingan dengan internasionalisasi, kapasitas dosen, dan transformasi digital, sementara paparan tahun 2026 berjudul PKDP 2026: Memahami Paradigma Keilmuan di PTKI memetakan empat paradigma keilmuan yang berkembang di perguruan tinggi keagamaan Islam. Kedua dokumen itu menunjukkan bahwa di tingkat institusi, pertanyaan tentang kapasitas riset sudah menjadi bagian dari perencanaan strategis. Namun dataset ini memperlihatkan bahwa rencana institusional tidak akan menggeser angka nasional selama riset tetap diperlakukan sebagai fungsi tambahan bagi pengajar, bukan sebagai pekerjaan dengan waktu, dana, dan jalur karier sendiri.
Lintasan pembanding: Korea, Tiongkok, Malaysia, Thailand, dan Vietnam
Titik tunggal dapat menyesatkan, sehingga Tabel 5 menyajikan lintasan belanja litbang terhadap PDB untuk enam negara sejak 1996, berdasarkan seri GB.XPD.RSDV.GD.ZS. Sel bertanda n.a. menandakan bahwa World Development Indicators tidak memuat observasi untuk tahun tersebut.
| Negara | 1996 | 2000 | 2005 | 2010 | 2015 | 2018 | 2020 | 2022 | 2023 |
|---|---|---|---|---|---|---|---|---|---|
| Korea Selatan | 2,14 | 2,05 | 2,43 | 3,18 | 3,79 | 4,27 | 4,52 | 4,85 | 4,94 |
| Tiongkok | 0,56 | 0,88 | 1,29 | 1,68 | 2,02 | 2,10 | 2,36 | 2,49 | 2,58 |
| Malaysia | 0,22 | 0,47 | n.a. | 1,04 | 1,28 | 1,04 | 0,95 | 1,01 | n.a. |
| Thailand | 0,12 | 0,24 | 0,22 | n.a. | 0,62 | 1,11 | 1,33 | 1,16 | 0,94 |
| Vietnam | n.a. | n.a. | n.a. | n.a. | 0,36 | n.a. | n.a. | n.a. | 0,41 |
| Indonesia | n.a. | 0,07 | n.a. | n.a. | n.a. | 0,23 | 0,28 | n.a. | n.a. |
Korea adalah kasus rujukan karena intensitas risetnya sudah di atas dua persen PDB pada 1996 dan terus naik hingga hampir lima persen pada 2023, bahkan 5,13 persen untuk 2024 menurut profil WIPO. Linsu Kim, dalam buku tahun 1997 berjudul Imitation to Innovation: The Dynamics of Korea’s Technological Learning (Harvard Business School Press), menggambarkan bagaimana perusahaan Korea bergerak dari peniruan teknologi matang menuju inovasi melalui pembelajaran yang disengaja dan sering dipaksakan oleh krisis. Keun Lee, dalam buku tahun 2013 berjudul Schumpeterian Analysis of Economic Catch-up (Cambridge University Press), berargumen bahwa negara pengejar yang berhasil cenderung masuk ke bidang teknologi dengan siklus pendek, di mana ketergantungan pada pengetahuan lama lebih kecil dan peluang lompatan lebih besar. Bagi dataset ini, pelajaran Korea bukan terletak pada besaran belanja, tetapi pada komposisinya: sektor bisnis menjadi pelaku riset utama, dan porsi paten residen hampir delapan puluh persen adalah konsekuensinya.
Tiongkok memperlihatkan bahwa peralihan itu dapat dilakukan dalam skala yang jauh lebih besar dan dalam waktu yang lebih singkat. Intensitas risetnya naik dari 0,56 persen pada 1996 menjadi 2,58 persen pada 2023, dan porsi paten residennya mencapai 91,5 persen pada 2024. Profil WIPO menempatkan Tiongkok di peringkat 10 dalam Global Innovation Index 2026, dengan peringkat output 5 dan input 22, dan Yicai melaporkan bahwa klaster Shenzhen-Hong Kong-Guangzhou mempertahankan peringkat pertama dunia sebagai klaster inovasi. Bagi Asia Tenggara, kasus Tiongkok penting bukan sebagai model yang dapat ditiru, karena skala dan kapasitas negaranya tidak tersedia di kawasan, tetapi sebagai bukti bahwa porsi paten residen dapat bergeser drastis dalam satu generasi bila belanja bisnis dan target kebijakan bergerak bersama.
Malaysia adalah peringatan. Intensitas risetnya naik cepat dari 0,22 persen pada 1996 menjadi 1,28 persen pada 2015, lalu turun dan bertahan di sekitar satu persen. Permohonan paten residennya justru menurun dari 1.231 pada 2010 menjadi 841 pada 2024. Malaysia memiliki salah satu sektor ekspor elektronik paling terintegrasi di dunia, tetapi lintasan ini menunjukkan bahwa integrasi itu tidak menghasilkan pendalaman kemampuan teknologi domestik yang sebanding. Richard Doner dan Ben Ross Schneider, dalam artikel tahun 2016 berjudul “The Middle-Income Trap: More Politics than Economics” di World Politics 68(4), berargumen bahwa peningkatan produktivitas melalui keterampilan dan inovasi menuntut investasi besar dalam kapasitas kelembagaan, sementara kapasitas politik untuk membangun lembaga itu dilemahkan oleh koalisi pendukung yang terfragmentasi, terutama di kalangan bisnis dan tenaga kerja, serta oleh ketimpangan. Dataset ini tidak dapat menguji argumen itu secara kausal, tetapi plateau Malaysia sejak 2015 sesuai dengan prediksinya: negara dapat mencapai tingkat belanja sedang melalui program pemerintah, tetapi melampauinya memerlukan koalisi dengan perusahaan domestik yang mau meneliti (tingkat keyakinan: sedang).
Thailand memperlihatkan lintasan yang lebih bergelombang. Intensitas risetnya naik dari 0,12 persen pada 1996 menjadi 1,33 persen pada 2020, salah satu kenaikan tercepat di kawasan, lalu turun menjadi 0,94 persen pada 2023. Profil WIPO mencatat belanja litbang sektor bisnis Thailand sekitar 0,6 persen PDB dengan data 2023, yang menunjukkan bahwa sebagian besar kenaikan sebelumnya memang digerakkan oleh perusahaan. Namun permohonan paten residen Thailand bergerak dari 1.214 pada 2010 menjadi 930 pada 2024, sehingga kenaikan belanja bisnis belum tercermin dalam kepemilikan teknologi. Penjelasan yang paling mungkin adalah bahwa sebagian riset bisnis di Thailand berupa pengembangan proses dan adaptasi produk yang tidak dipatenkan, atau dilakukan oleh anak perusahaan asing yang mematenkannya atas nama induk, tetapi kajian ini tidak memiliki data untuk memastikannya.
Vietnam adalah kasus terakhir dan paling relevan bagi Indonesia. Intensitas risetnya hampir tidak berubah, dari 0,36 persen pada 2015 menjadi 0,41 persen pada 2023, tetapi peneliti per sejuta penduduk mencapai 836, naik 8,88 persen dari tahun sebelumnya menurut profil WIPO, paten residennya naik empat kali lipat sejak 2010, dan peringkat Global Innovation Index-nya naik ke 43 pada 2026. Vietnam juga diklasifikasikan sebagai negara berpendapatan menengah bawah dalam profil itu dan menempati peringkat kedua dari 36 negara dalam kelompoknya, sedangkan Indonesia menempati peringkat ketujuh dari 35 negara berpendapatan menengah atas. Dengan input yang mirip dan pendapatan per kapita yang lebih rendah, Vietnam menghasilkan output inovasi yang lebih tinggi. Kajian ini menilai bahwa perbedaan itu terutama berasal dari integrasi industri yang lebih dalam ke rantai pasok elektronik dan dari sinyal politik yang konsisten di tingkat tertinggi, bukan dari keunggulan sistem riset yang sudah terbukti (tingkat keyakinan: sedang).
Argumen tandingan: apakah negara pengejar memang perlu mencipta?
Argumen tandingan terkuat terhadap kajian ini patut disusun dalam versi terbaiknya. Bagi negara pengejar, menurut argumen itu, menyerap teknologi yang sudah ada jauh lebih murah dan lebih cepat daripada menciptakan teknologi baru, sehingga strategi rasional adalah memaksimalkan investasi asing, perdagangan, dan pendidikan massal, lalu membiarkan kemampuan inovasi tumbuh sendiri seiring kenaikan pendapatan. Singapura sering dijadikan bukti: negara-kota itu menjadi kaya terutama melalui perusahaan multinasional. Bagi Indonesia, argumen itu diperkuat oleh profil WIPO yang menempatkan skala pasar domestik di peringkat 7 dunia, perkembangan klaster di peringkat 7, kebijakan dan budaya kewirausahaan di peringkat 9, serta kolaborasi riset universitas-industri di peringkat 10. Ditambah lagi, indikator belanja litbang dan paten terkenal tidak menangkap inovasi informal, inovasi model bisnis, dan inovasi digital yang justru tumbuh pesat di ekonomi besar seperti Indonesia. Dari sudut ini, keprihatinan atas angka belanja riset yang rendah adalah fetisisme statistik.
Argumen ini mengandung kebenaran penting, dan kajian ini menerima tiga bagiannya. Penyerapan memang merupakan tahap yang perlu, bukan kegagalan; indikator formal memang tidak lengkap; dan skala pasar domestik Indonesia memang aset strategis yang tidak dimiliki Malaysia atau Thailand. Namun argumen itu gagal pada tiga titik yang dapat diperiksa dengan data. Pertama, Cohen dan Levinthal menunjukkan bahwa kemampuan menyerap bergantung pada pengetahuan terkait yang sudah dimiliki dan dibangun antara lain melalui riset sendiri. Negara yang tidak meneliti tidak sekadar gagal mencipta; ia juga menyerap secara dangkal, sebatas mengoperasikan teknologi yang tidak dipahaminya. Lintasan paten Malaysia dan Thailand sejak 2010 adalah ilustrasi empiris dari penyerapan yang tidak mendalam.
Kedua, kasus Singapura justru membantah argumen itu bila dibaca dengan data. Singapura memiliki 8.782 peneliti per sejuta penduduk, intensitas riset 1,81 persen PDB, dan peringkat pertama dunia untuk subindeks input Global Innovation Index 2026. Negara-kota itu tidak hanya menunggu perusahaan multinasional; ia membangun universitas riset dan lembaga riset publik dalam skala yang membuatnya menjadi salah satu produsen pengetahuan paling intensif di dunia per kapita, dengan 2.010 artikel per sejuta penduduk. Ketiga, laporan World Development Report 2024 tidak menyatakan bahwa penyerapan cukup; laporan itu menyatakan bahwa negara berpendapatan menengah harus berpindah dari investasi ke penyerapan teknologi lalu ke inovasi, dan bahwa peralihan itu tidak terjadi dengan sendirinya. Bukti bahwa kenaikan pendapatan otomatis melahirkan inovasi tidak ditemukan dalam dataset ini: Malaysia dan Thailand telah menjadi negara berpendapatan menengah atas selama bertahun-tahun, tetapi porsi paten residen mereka tetap di sekitar sebelas persen.
Adapun kekuatan Indonesia dalam profil WIPO perlu dibaca dengan cermat. Skala pasar, klaster, kewirausahaan, dan kolaborasi universitas-industri menggambarkan potensi dan kondisi kelembagaan, bukan output teknologi. Profil yang sama menempatkan belanja pendidikan sebagai persen PDB di peringkat 131, artikel ilmiah per miliar dolar PPP PDB di peringkat 125, dan investasi asing langsung di bidang litbang dan teknologi tinggi di peringkat 112. Potensi yang tidak dikonversi tidak akan bertahan lama sebagai keunggulan, karena perusahaan digital yang tumbuh di pasar domestik besar akan membeli kemampuan teknologi dari luar bila kemampuan itu tidak tersedia di dalam negeri. Dengan demikian, argumen tandingan benar tentang urutan, tetapi keliru tentang otomatisme. Kesimpulan kajian ini tetap: tanpa investasi yang disengaja pada simpul produksi pengetahuan, terutama di sektor bisnis, celah absorpsi-produksi akan bertahan (tingkat keyakinan: tinggi).
Skenario hingga 2030 dan penanda yang perlu diamati
Skenario berikut berfokus pada Indonesia dalam konteks kawasan, karena dataset ini menunjukkan bahwa Indonesia adalah negara dengan selisih terbesar antara skala dan kapasitas. Horizon 2030 dipilih karena bertepatan dengan target Resolusi 57 Vietnam dan dengan paruh pertama pelaksanaan Rencana Aksi ASEAN untuk Sains, Teknologi, dan Inovasi 2026–2035. Probabilitas yang diberikan adalah penilaian kualitatif berdasarkan lintasan data dalam kajian ini, sinyal kebijakan yang terverifikasi, dan kendala fiskal yang diketahui.
Skenario hingga 2030
- AStagnasi absorptif45%
- BKonsolidasi bertahap30%
- CKemunduran fiskal15%
- DLompatan terarah10%
Probabilitas adalah penilaian penulis per 1 November 2026, bukan prediksi statistik.
Skenario A, stagnasi absorptif, adalah kelanjutan pola saat ini dan karena itu paling mungkin. Dalam skenario ini, belanja riset nasional menurut ukuran BRIN tetap di bawah 0,2 persen PDB, publikasi terus tumbuh tetapi melambat, paten residen berfluktuasi di sekitar dua ribu permohonan per tahun, dan peringkat Global Innovation Index bergerak di rentang 50-an. Indonesia tetap menarik investasi pengolahan mineral dan manufaktur, tetapi kemampuan teknologi di dalam rantai nilai itu tetap dimiliki mitra asing. Skenario B, konsolidasi bertahap, terjadi bila pemerintah menyelaraskan statistik iptek, menetapkan target belanja bisnis, dan memberikan insentif fiskal yang dapat dimanfaatkan perusahaan menengah dan besar. Belanja riset naik mendekati 0,3 persen menurut ukuran BRIN, porsi belanja industri bergerak menuju sepertiga, dan paten residen melampaui puncak 2019.
Skenario C, kemunduran fiskal, terjadi bila tekanan anggaran mendorong pemotongan lanjutan pada BRIN dan pendidikan tinggi tanpa kompensasi dari sektor swasta, sehingga peneliti senior pindah ke luar negeri atau ke sektor nonriset dan lapis publikasi pun mulai melemah. Pemotongan anggaran BRIN 2025 menunjukkan bahwa skenario ini bukan sekadar hipotetis. Skenario D, lompatan terarah, adalah varian Vietnam: komitmen politik tingkat tertinggi dengan target kuantitatif, kenaikan belanja publik yang cepat, dan kemitraan wajib antara investor asing dan lembaga riset domestik. Probabilitasnya rendah karena belum ada sinyal kebijakan sebesar itu, tetapi tidak nol karena pemerintah Indonesia memiliki kapasitas fiskal dan pasar untuk mengubah arah bila keputusan politik diambil.
Penanda yang perlu diamati dapat diurutkan menurut kecepatan munculnya. Dalam jangka pendek, penanda terpenting adalah edisi berikutnya laporan Indikator Iptek BRIN: apakah rasio belanja riset terhadap PDB bergerak dari 0,10 persen, apakah porsi industri dihitung dengan metode yang stabil, dan apakah angka peneliti setara penuh waktu mulai disajikan. Penanda kedua adalah apakah UIS menerima observasi Indonesia, Filipina, dan India setelah 2020, sehingga indeks global berhenti mengandalkan data pandemi. Dalam jangka menengah, penanda yang paling informatif adalah porsi paten residen: bila angka itu di Indonesia atau Vietnam melampaui seperempat pada 2030, celah absorpsi-produksi mulai menyempit. Penanda kawasan adalah apakah pelaksanaan rencana aksi ASEAN untuk sains dan teknologi menghasilkan sistem statistik bersama yang memperkecil kelambatan data. Penanda terakhir adalah pencapaian Vietnam atas target 2030-nya, karena keberhasilan atau kegagalannya akan menjadi argumen terkuat dalam perdebatan kebijakan riset di seluruh kawasan.
Rekomendasi per aktor dengan tenggat
Bagi Kementerian Pendidikan Tinggi, Sains, dan Teknologi, rekomendasi pertama adalah mengubah perlakuan riset dalam beban kerja dosen. Sebelum tahun akademik 2027/2028, kementerian perlu menetapkan pedoman alokasi waktu riset yang dapat diverifikasi dan mengadakan survei peneliti setara penuh waktu yang sejalan dengan pedoman Frascati, sehingga selisih empat kali lipat antara hitungan kepala dan hitungan setara penuh waktu dapat diukur, lalu diperkecil. Pada 2028, kementerian perlu mengaitkan sebagian pendanaan riset perguruan tinggi dengan kemitraan industri yang nyata, misalnya program doktor industri dan proyek bersama dengan pembiayaan bersama, alih-alih dengan jumlah publikasi. Kementerian juga perlu memastikan bahwa angka yang dipakai pejabatnya di forum publik bersumber dari statistik resmi yang sama dengan yang dilaporkan ke lembaga internasional.
Bagi BRIN, prioritas paling mendesak adalah statistik. Pada pertengahan 2027, BRIN perlu menerbitkan catatan rekonsiliasi yang menjelaskan selisih antara angka 0,28 persen PDB untuk 2020 yang diteruskan UIS dan angka 0,10 persen PDB untuk 2023 dalam laporan BRIN, termasuk perbedaan cakupan sektor, sumber data, dan status angka sementara. Mulai siklus 2027, BRIN perlu melaporkan data belanja litbang dan peneliti ke UIS setiap tahun dengan kelambatan tidak lebih dari dua tahun. Dalam fungsi risetnya, BRIN perlu memprioritaskan kontrak riset bersama dengan industri dan pemerintah daerah, karena badan riset pemerintah yang tidak tersambung dengan pengguna pengetahuan hanya akan memindahkan masalah, bukan memecahkannya.
Bagi Bappenas, rekomendasi utamanya adalah menjadikan indikator sistem pengetahuan bagian dari kerangka pemantauan pembangunan, bukan sekadar lampiran sektor pendidikan. Dalam evaluasi paruh waktu rencana pembangunan jangka menengah, Bappenas perlu menetapkan target terpisah untuk belanja riset total, porsi belanja riset bisnis, dan porsi paten residen, karena target agregat tanpa komposisi akan mendorong kenaikan belanja pemerintah tanpa mengubah struktur. Bappenas juga perlu menguji opsi insentif fiskal riset dan pengembangan berdasarkan data penggunaannya, dan menyusun kerangka agar investasi asing di sektor strategis disertai komitmen kemitraan riset domestik yang terukur.
Bagi universitas, termasuk perguruan tinggi keagamaan, rekomendasinya adalah memisahkan jalur karier riset dari jalur pengajaran secara bertahap mulai 2027, membentuk kantor alih teknologi yang benar-benar dikelola tenaga profesional, dan memasukkan indikator kemitraan, paten, serta kontrak riset ke dalam rencana strategis institusi. Universitas juga perlu memanfaatkan dataset seperti ini sebagai bahan ajar dalam program studi kebijakan, manajemen pengetahuan, dan studi pembangunan, agar generasi berikutnya mampu membaca statistik iptek secara kritis. Bagi perusahaan besar, termasuk badan usaha milik negara, langkah minimum adalah melaporkan belanja riset secara sukarela dan konsisten kepada BRIN mulai 2027, karena kelemahan statistik sektor bisnis adalah salah satu penyebab utama ketidakpastian angka nasional.
Bagi Komite ASEAN untuk Sains, Teknologi, dan Inovasi (COSTI, sebelumnya dikenal sebagai ASEAN COST), rekomendasinya adalah menjadikan observatorium statistik iptek kawasan sebagai salah satu keluaran awal Rencana Aksi ASEAN untuk Sains, Teknologi, dan Inovasi 2026–2035, yang tahap pertama penyusunannya dikerjakan konsorsium yang dipimpin Cambridge Industrial Innovation Policy untuk Sekretariat ASEAN. Observatorium itu perlu menghasilkan tabel tahunan yang selaras dengan UIS dan WIPO, dengan indikator porsi paten residen dan neraca royalti, sebelum 2028. Bagi UNESCO dan UIS, rekomendasinya adalah dukungan teknis yang terfokus untuk Indonesia, Filipina, dan Kamboja agar observasi setelah 2020 tersedia paling lambat 2028, serta penandaan vintase yang lebih jelas dalam distribusi data. Laporan UNESCO Science Report 2021 telah menyediakan kerangka regional yang dapat diperbarui dengan data tersebut.
Bagi Bank Dunia, rekomendasinya adalah menerjemahkan diagnosis World Development Report 2024 ke dalam analisis tingkat negara untuk Indonesia, Filipina, dan Vietnam sebelum 2028, dengan menilai secara eksplisit di tahap mana masing-masing negara berada dalam urutan investasi, penyerapan, dan inovasi. Program pinjaman di sektor pendidikan tinggi dan inovasi perlu menyertakan komponen penguatan statistik iptek, karena tanpa data yang tepat waktu evaluasi program tidak dapat dilakukan. Bank Dunia juga perlu menandai vintase sumber dalam metadata World Development Indicators secara lebih menonjol, agar pengguna tidak membandingkan data 2020 dengan data 2024 seolah-olah keduanya sezaman.
Keterbatasan dan kesimpulan
Keterbatasan kajian ini perlu dinyatakan secara jelas. Pertama, dataset ini bersifat potong lintang dengan tahun observasi yang berbeda antarnegara, sehingga perbandingan antarnegara yang selisih tahunnya besar harus dibaca sebagai indikasi, bukan sebagai pengukuran presisi. Kedua, indikator formal seperti belanja litbang, paten, dan artikel terindeks tidak menangkap inovasi informal, inovasi organisasi, dan produksi pengetahuan di luar sains dan rekayasa, termasuk ilmu sosial, humaniora, dan pengetahuan lokal yang penting bagi kawasan. Ketiga, kajian ini tidak menguji hubungan kausal; pernyataan tentang mekanisme, misalnya tentang peran koalisi politik atau struktur kepemilikan perusahaan, adalah interpretasi yang konsisten dengan data, bukan kesimpulan statistik. Keempat, angka Brunei dan Kamboja untuk peringkat komposit berasal dari sumber sekunder, dan indikator royalti tidak tersedia untuk Vietnam, Brunei, dan Kamboja. Kelima, angka Indonesia bergantung pada dua sistem pengukuran yang belum direkonsiliasi.
Kesimpulannya dapat dinyatakan dengan tegas. Asia Tenggara telah membuktikan kemampuannya menyerap teknologi, dan hal itu adalah prestasi pembangunan yang nyata. Namun dataset ini menunjukkan bahwa kemampuan menyerap belum dikonversi menjadi kemampuan mencipta: porsi paten residen di seluruh ASEAN berada di bawah dua puluh persen, neraca royalti sebagian besar negara defisit lebih dari sepuluh kali lipat, dan lintasan Malaysia serta Thailand memperlihatkan bahwa intensitas riset sedang tidak otomatis berlanjut menjadi kepemilikan teknologi. Indonesia menghadapi celah ganda, dengan lapis publikasi yang tumbuh pesat tetapi tidak tersambung ke lapis teknologi, sektor bisnis yang hampir tidak meneliti, dan statistik iptek yang belum mampu memberi satu angka dasar yang disepakati.
Implikasi kebijakannya bukan sekadar menaikkan anggaran riset, meskipun itu diperlukan. Implikasinya adalah membangun simpul-simpul yang hilang dalam sistem: riset di dalam perusahaan, jalur karier peneliti yang nyata di universitas, sambungan antara investasi asing dan lembaga riset domestik, dan statistik yang tepat waktu dan dapat dipercaya. Korea dan Tiongkok menunjukkan bahwa peralihan dari penyerap menjadi pencipta dapat dicapai dalam satu generasi, sedangkan Vietnam menunjukkan bahwa sinyal politik yang konsisten dapat menghasilkan output di atas input. Bagi Indonesia, pertanyaannya bukan lagi apakah sistem pengetahuan nasional perlu diubah, tetapi apakah keputusan untuk mengubahnya akan diambil sebelum jendela demografi dan rantai pasok yang terbuka saat ini menutup.
◆ ◆ ◆
Referensi
Dokumen resmi dan data
- Badan Riset dan Inovasi Nasional. 2024. Indikator Iptek, Riset, dan Inovasi Indonesia 2024. Jakarta: Penerbit BRIN. DOI 10.55981/brin.1319.
- Bank Dunia. 2026. Research and development expenditure (% of GDP), GB.XPD.RSDV.GD.ZS. World Development Indicators, data UNESCO Institute for Statistics.
- Bank Dunia. 2026. Researchers in R&D (per million people), SP.POP.SCIE.RD.P6. World Development Indicators.
- Bank Dunia. 2026. Scientific and technical journal articles, IP.JRN.ARTC.SC. World Development Indicators, data NSF Science and Engineering Indicators.
- Bank Dunia. 2026. Patent applications, residents, IP.PAT.RESD; dan Patent applications, nonresidents, IP.PAT.NRES. World Development Indicators, data WIPO.
- Bank Dunia. 2026. School enrollment, tertiary (% gross), SE.TER.ENRR. World Development Indicators.
- Bank Dunia. 2026. High-technology exports (% of manufactured exports), TX.VAL.TECH.MF.ZS. World Development Indicators.
- Bank Dunia. 2026. Charges for the use of intellectual property, payments; dan receipts. World Development Indicators.
- Bank Dunia. 2026. Population, total, SP.POP.TOTL. World Development Indicators.
- WIPO. 2026. Global Innovation Index 2026, profil ekonomi: Indonesia, Malaysia, Singapura, Thailand, Vietnam, Filipina. Jenewa: World Intellectual Property Organization.
- WIPO. 2026. Global Innovation Index 2026, profil ekonomi: Republik Korea, Tiongkok, India. Jenewa: World Intellectual Property Organization.
Literatur ilmiah
- Cohen, Wesley M., dan Daniel A. Levinthal. 1990. “Absorptive Capacity: A New Perspective on Learning and Innovation.” Administrative Science Quarterly 35(1): 128–152.
- Doner, Richard F., dan Ben Ross Schneider. 2016. “The Middle-Income Trap: More Politics than Economics.” World Politics 68(4): 608–644.
- Freeman, Christopher. 1995. “The ‘National System of Innovation’ in Historical Perspective.” Cambridge Journal of Economics 19(1): 5–24.
- Kim, Linsu. 1997. Imitation to Innovation: The Dynamics of Korea’s Technological Learning. Boston: Harvard Business School Press.
- Lee, Keun. 2013. Schumpeterian Analysis of Economic Catch-up: Knowledge, Path-Creation, and the Middle-Income Trap. Cambridge: Cambridge University Press.
- Lundvall, Bengt-Åke, ed. 1992 (terbit ulang 2010). National Systems of Innovation: Toward a Theory of Innovation and Interactive Learning. London: Anthem Press.
Analisis kebijakan dan laporan lembaga
- Bank Dunia. 2024. World Development Report 2024: The Middle-Income Trap. Washington, DC: World Bank.
- Chen, Derek H. C., dan Carl J. Dahlman. 2005. The Knowledge Economy, the KAM Methodology and World Bank Operations. Washington, DC: World Bank Institute.
- UNESCO. 2021. UNESCO Science Report: The Race Against Time for Smarter Development. Paris: UNESCO.
- Cambridge Industrial Innovation Policy. 2023 (diperbarui 2026). ASEAN Plan of Action on Science, Technology and Innovation (APASTI) 2026–2035. Cambridge: University of Cambridge.
Media dan pemberitaan
- Wiratri, Amorisa, Betti Rosita Sari, dan Lamijo. 2025. “Budget Efficiency Contrast in ASEAN: Tales from Indonesia and Vietnam.” The Jakarta Post, Maret 2025.
- Bisnis.com. 2025. “Pendanaan Riset RI Lebih Rendah dari Rerata Asean, Butuh Dukungan Swasta.” 28 April 2025.
- Rizqiyah, Aslamatur. 2024. “Tahun 2023, Belanja Riset dan SDM Iptek Masih Didominasi Perguruan Tinggi.” Good News From Indonesia, 8 September 2024.
- VietnamPlus. 2026. “WIPO Innovation Report Viewed through Resolution 57 Perspective.” 3 Oktober 2026.
- Yicai Global. 2026. “China Retains 10th Spot in WIPO Ranking of World’s Most Innovative Economies.” 30 September 2026.
- Sumarno, Daniel Dimays. 2026. “Southeast Asia’s Most Innovative Economies in 2026.” SeasiaStats, 7 Oktober 2026.
Arsip & Pustaka
- Gunawan, Hendra. 2014. “Mengintip Masa Depan.” Kompas, 2 Mei 2014.
- Santosa, Budi. 2014. “Dari ‘Brain Drain’ ke ‘Brain Gain’.” Kompas, 2 Mei 2014.
- Makin, Al. 2018. Bisakah Menjadi Ilmuwan di Indonesia?: Keilmuan, Birokrasi dan Globalisasi. Pidato pengukuhan guru besar. Yogyakarta: UIN Sunan Kalijaga.
- Ahmad, Kamaruzzaman Bustamam. 2026. Rencana Strategis UIN Ar-Raniry. Paparan, Tim Renstra UIN Ar-Raniry Banda Aceh.
- Ahmad, Kamaruzzaman Bustamam. 2026. PKDP 2026: Memahami Paradigma Keilmuan di PTKI. Paparan.
Cara Mengutip
Ahmad, Kamaruzzaman Bustamam. 2026. “Menyerap Tanpa Mencipta: Kapasitas Produksi Pengetahuan Negara-Negara ASEAN: Dataset Komparatif Terdokumentasi dengan Tolok Ukur Korea Selatan, Tiongkok, dan India.” KBA13 Academy Research Center, Datasets, 1 November 2026. https://kba13academy.com/kapasitas-produksi-pengetahuan-asean/
Guru Besar Antropologi Agama di UIN Ar-Raniry Banda Aceh dan pendiri KBA13 Academy. Kajiannya meliputi Islam di Asia Tenggara, antropologi, kajian strategis, dan sistem pengetahuan.
Datasets · Knowledge Systems
Absorbing Without CreatingThe knowledge-production capacity of ASEAN states: a documented comparative dataset benchmarked against South Korea, China and India
Southeast Asia has become one of the world’s most effective absorbers of technology, yet it remains a thin producer of knowledge. This dataset compiles R&D spending, researcher numbers, scientific articles, patents, higher education and innovation rankings for eleven countries, complete with years, sources and a data dictionary, and reads the gap between absorbing and creating as a problem of national knowledge-system architecture.
Contents

Abstract
This study presents a documented comparative dataset on the knowledge-production capacity of eight ASEAN states (Indonesia, Malaysia, Singapore, Thailand, Vietnam, the Philippines, Brunei and Cambodia), with South Korea, China and India as benchmarks. The data are compiled from the World Bank’s World Development Indicators (which relay statistics from the UNESCO Institute for Statistics and WIPO), the Global Innovation Index 2026 economy profiles and BRIN’s 2024 Science and Technology Indicators report, with the latest observation year recorded for every value and “n.a.” for any value that could not be verified. The analytical framework distinguishes three layers: research inputs, knowledge outputs in the form of publications, and technology outputs in the form of patents and intellectual property receipts. The central finding is a systematic gap between the capacity to absorb and the capacity to create. Malaysia, the Philippines and Thailand export high-technology products in large shares, yet only 11 to 19 per cent of patent applications in their countries are filed by residents, and their royalty payments are eleven to twenty times their receipts. Indonesia has increased its scientific articles twenty-eight-fold since 2010, but its resident patent applications have fallen from their 2019 peak and its research spending by BRIN’s measure was only 0.10 per cent of GDP in 2023. The study closes with a method note, a data dictionary, scenarios to 2030 and recommendations for ministries, BRIN, Bappenas, universities, ASEAN COSTI, UNESCO and the World Bank.
Southeast Asia does not lack access to the world’s knowledge; it lacks the domestic nodes that turn that access into the capacity to create, and the gap is determined more by institutional architecture and political coalitions than by publication counts.
- Research inputsR&D spending relative to GDP, researchers per million inhabitants and tertiary participation.
- Knowledge outputsScientific and technical articles, absolute and per million inhabitants, with their trajectory since 2010.
- Technology outputsResident patents, the non-resident patent ratio, high-technology exports and the intellectual property royalty balance.
Keywords
- national innovation system
- absorptive capacity
- research and development spending
- ASEAN
- middle-income trap
- resident patents
- BRIN
- documented dataset
The question, the bottom line, and why it matters
How large, really, is the capacity of Southeast Asian states to produce new knowledge, rather than merely to consume, adapt or assemble knowledge created elsewhere? The question is often answered with a single number, usually a rank in a global innovation index or a count of indexed articles, and then debated as though that number described the whole system. This dataset refuses that shortcut. It compiles eleven indicators from multilateral sources actually retrieved at the end of October 2026, records the observation year of every value, flags gaps openly, and arranges them in a framework that separates inputs, knowledge outputs and technology outputs.
The bottom line can be stated in one sentence. The region has become a highly effective absorber of technology, but most of its states have not built the domestic nodes that convert absorption into creation. The sharpest evidence comes from the contrast between two sets of numbers. On one side, high-technology exports reached 60.4 per cent of manufactured exports in the Philippines, 59.4 per cent in Singapore and 58.6 per cent in Malaysia in 2024 according to the World Development Indicators, higher than South Korea (36.3 per cent). On the other side, Malaysian residents filed only 841 patent applications in 2024, Filipinos 873 and Thais 930, while Koreans filed 195,786, according to WIPO resident patent data relayed by the World Bank. The region ships chips to the whole world, yet patents almost nothing of what is inside them (confidence: high).
Indonesia occupies a distinctive position within that pattern. It is not a strong absorber of high technology, since its high-technology exports were only 8.7 per cent of manufactured exports in 2024, nor yet a producer of technology, since its national research spending according to BRIN’s Indikator Iptek, Riset, dan Inovasi Indonesia 2024 was only about Rp20.39 trillion, or 0.10 per cent of GDP, in 2023. The only layer that has grown dramatically is publication: the number of scientific and technical articles rose from 1,391 in 2010 to 38,992 in 2023. Indonesia therefore faces a double gap, absorbing too little and creating too little, and publication growth has not yet connected to either of the other layers (confidence: high on direction, moderate on magnitude because of the measurement differences set out in Section V).
This matters globally for three reasons. First, the World Bank’s World Development Report 2024 treats the shift from investment to infusion and then to innovation as the condition for escaping the middle-income trap, and most ASEAN states now sit precisely at the junction of the second and third stages. Second, the map of technology supply chains is being redrawn by US-China rivalry, so countries that serve only as assembly locations risk being left behind when the next relocation decisions are made. Third, the region’s most populous countries, Indonesia, the Philippines and Vietnam, will determine whether Southeast Asia as a whole becomes Asia’s third pole of knowledge or remains a productive periphery between Northeast and South Asia.
The dataset was built as teaching material for the School of Knowledge Systems & Knowledge Management, and so it is designed to be checked, updated and contested. Every figure in the tables carries a year and a source; every indicator code is recorded in the data dictionary; and every harmonisation decision is explained. This study deliberately does not revisit the debate on publication metrics or on the theoretical poverty of the social sciences, two themes treated elsewhere. Its focus is national system capacity: who funds research, who performs it, what it yields, and whether those results return to the economy as technological capability that is owned at home.
Analytical framework: national innovation systems and three layers of capacity
The framework of this study rests on the national innovation systems tradition. Christopher Freeman, in his 1995 article “The ‘National System of Innovation’ in Historical Perspective” in the Cambridge Journal of Economics, traced the idea back to Friedrich List and showed that a nation’s technological performance is determined by a network of institutions, namely firms, universities, government agencies and the education system, and by the way that network learns and imports technology. Bengt-Åke Lundvall, in National Systems of Innovation: Toward a Theory of Innovation and Interactive Learning (1992; reissued by Anthem Press), added that the core of such a system is interactive learning between users and producers of knowledge. From these two sources this study takes one methodological principle: knowledge-production capacity cannot be measured at a single node, because what matters is the connection between nodes.
The second principle comes from Wesley Cohen and Daniel Levinthal, who argued in their 1990 article “Absorptive Capacity: A New Perspective on Learning and Innovation” in Administrative Science Quarterly that an organisation’s ability to recognise, assimilate and exploit outside knowledge depends heavily on the related knowledge it already holds, and that in-house research is one of the main ways of building that ability. The implication for policy analysis is large. Absorption and creation are not two strategies of which one may be chosen; they are two faces of the same investment. A country that stops investing in its own research, on the grounds that imported technology is enough, eventually also loses the ability to absorb that imported technology deeply.
On this basis the study builds four measurement dimensions. The input dimension covers R&D spending as a percentage of GDP, researchers per million inhabitants and the gross tertiary enrolment ratio. The knowledge-output dimension covers scientific and technical articles. The technology-output dimension covers patent applications by residents. The absorption dimension covers high-technology exports, patent applications by non-residents, and payments and receipts of charges for the use of intellectual property. The Global Innovation Index 2026 is used as a comparator composite, in the knowledge that it is itself an aggregation of many indicators, some of which overlap with this dataset.
Four dimensions of knowledge-production capacity
- 1InputsresourcesR&D spending, full-time-equivalent researchers, tertiary participation.
- 2Knowledge outputspublicationsScientific and technical articles, absolute and per million inhabitants.
- 3Technology outputsownershipResident patents and intellectual property royalty receipts.
- 4AbsorptionuptakeHigh-technology exports, non-resident patents, royalty payments.
The most important analytical distinction in this framework is the ratio between the fourth and third dimensions, which this study calls the absorption-production gap. It is measured in two mutually reinforcing ways. The first is the resident share of total domestic patent applications: the smaller the share, the greater the system’s dependence on foreign inventors protecting their technology in the domestic market. The second is the ratio of royalty payments to receipts: the larger the ratio, the larger the outflow of knowledge rent abroad relative to the inflow from knowledge owned at home. Neither measure is perfect, since multinational firms can shift royalty income across countries for tax reasons, but both capture what input and publication indicators miss.
The framework also borrows from the knowledge-economy index tradition. The World Bank’s Knowledge Assessment Methodology, described by Derek Chen and Carl Dahlman in their 2005 paper The Knowledge Economy, the KAM Methodology and World Bank Operations, divides the knowledge economy into four pillars: the economic incentive and institutional regime, education, the innovation system, and information technology. This study does not build a new composite index, because such indices tend to hide contradictions between dimensions behind a single score. Instead it lets the indicators stand side by side so that their contradictions are visible, because those contradictions are precisely the object of analysis.
The short method note is as follows. The main source is the World Bank’s World Development Indicators as updated on 8 October 2026, which for R&D spending and researchers relays data from the UNESCO Institute for Statistics, for patents relays WIPO data, and for publications relays the US National Science Foundation’s Science and Engineering Indicators, which count Scopus-indexed articles. For each country and indicator the latest available observation was taken, so observation years differ across countries and are recorded in each cell. Innovation ranks are taken from the Global Innovation Index 2026 economy profiles published by WIPO at the end of September 2026. Indonesia’s national data are compared with BRIN’s 2024 report. The trajectory period analysed runs from 1996 to 2025, depending on series availability. The approach is structured descriptive comparison, not causal estimation, and every main judgment carries a confidence level.
Core dataset: inputs, outputs and composite position
Table 1 presents the input dimension for eleven countries. R&D spending and researcher figures come from the series GB.XPD.RSDV.GD.ZS and SP.POP.SCIE.RD.P6, while tertiary participation comes from the series SE.TER.ENRR. Each value is the latest observation available in the World Development Indicators as of 8 October 2026, rounded for readability, with the observation year in brackets.
| Country | R&D spending (% of GDP) | Researchers per million inhabitants (FTE) | Tertiary gross enrolment ratio (%) |
|---|---|---|---|
| Indonesia | 0.28 (2020) | 395 (2020) | 44.9 (2023) |
| Malaysia | 1.01 (2022) | 1,218 (2022) | 38.5 (2024) |
| Singapore | 1.81 (2022) | 8,782 (2022) | 97.3 (2023) |
| Thailand | 0.94 (2023) | 1,592 (2023) | 49.1 (2025) |
| Vietnam | 0.41 (2023) | 836 (2023) | 37.6 (2024) |
| Philippines | 0.28 (2020) | 320 (2020) | 47.4 (2024) |
| Brunei Darussalam | 0.28 (2018) | 506 (2022) | 36.4 (2023) |
| Cambodia | 0.09 (2015) | 30 (2015) | 21.4 (2024) |
| South Korea | 4.94 (2023) | 9,472 (2023) | 111.9 (2024) |
| China | 2.58 (2023) | 2,107 (2023) | 76.9 (2024) |
| India | 0.65 (2020) | 259 (2020) | 34.4 (2025) |
Table 1 already reveals three clear groups. The first is Korea and Singapore, with more than 8,700 researchers per million inhabitants and research spending above 1.8 per cent of GDP. The second is Malaysia, Thailand and China, with between 1,200 and 2,100 researchers per million, although China’s spending intensity is far higher than that of the two ASEAN states. The third is Indonesia, Vietnam, the Philippines, Brunei, Cambodia and India, with fewer than 850 researchers per million. Vietnam sits at the top of the third group with 836 researchers per million in 2023, more than double Indonesia’s 2020 figure, even though the two countries’ research spending is relatively close. The difference shows that researcher numbers can grow faster than spending when a country prioritises building a research workforce, but it is also a reminder that underfunded researchers easily become full-time teachers who hold the title of researcher.
The tertiary participation column reveals a paradox that should not be passed over. Indonesia (44.9 per cent) and the Philippines (47.4 per cent) have higher tertiary participation than Malaysia (38.5 per cent) and Vietnam (37.6 per cent), yet their researcher intensity is far lower. In other words, the expansion of access to higher education in the two archipelagic states has not been converted into research capacity. Their higher education systems have grown mainly as systems of mass teaching, not as systems of knowledge production, and this is a structural finding more important than any rank difference (confidence: high).
Table 2 presents the output dimension and composite position. Articles come from the series IP.JRN.ARTC.SC for 2023, resident patents from the series IP.PAT.RESD for 2024, and per-million values are calculated by this study by dividing absolute numbers by 2023 population from the series SP.POP.TOTL. Global Innovation Index 2026 ranks come from WIPO’s economy profiles for Indonesia, Malaysia, Singapore, Thailand, Vietnam, the Philippines, Korea, China and India; ranks for Brunei and Cambodia are taken from a SeasiaStats summary because their WIPO profiles were not retrieved for this study.
| Country | Scientific and technical articles (2023) | Articles per million inhabitants | Resident patents (2024) | Resident patents per million inhabitants | GII 2026: rank (input/output) |
|---|---|---|---|---|---|
| Indonesia | 38,992 | 139 | 1,931 | 6.9 | 55 (60/60) |
| Malaysia | 23,691 | 674 | 841 | 23.9 | 34 (29/40) |
| Singapore | 11,894 | 2,010 | 1,898 | 320.7 | 5 (1/9) |
| Thailand | 16,656 | 232 | 930 | 13.0 | 44 (46/41) |
| Vietnam | 10,644 | 106 | 1,226 | 12.2 | 43 (54/35) |
| Philippines | 4,752 | 41 | 873 | 7.6 | 52 (69/49) |
| Brunei Darussalam | 502 | 1,093 | 1 | 2.2 | 98 (n.a.) |
| Cambodia | 201 | 12 | 1 | 0.1 | 93 (n.a.) |
| South Korea | 71,724 | 1,387 | 195,786 | 3,786 | 4 (4/6) |
| China | 932,712 | 661 | 1,672,001 | 1,185 | 10 (22/5) |
| India | 228,174 | 159 | 63,217 | 44.0 | 38 (49/25) |
The last column of Table 2 holds information that is seldom read: the difference between input and output ranks in the Global Innovation Index. Countries whose output rank is better than their input rank convert resources into results relatively efficiently; countries with the opposite pattern have relatively good resources and institutional environments but lagging results. Vietnam (input 54, output 35), India (input 49, output 25) and China (input 22, output 5) belong to the first group. Malaysia (input 29, output 40) and Singapore (input 1, output 9) belong to the second, while Indonesia sits exactly in the middle at 60 for both. The WIPO profile also notes that Vietnam and India are the two economies that have outperformed expectations for their level of development for sixteen consecutive years, as VietnamPlus reported on 3 October 2026.
Comparing publications with patents in Table 2 yields the sharpest reading. Indonesia produces more articles than Malaysia, Thailand and Vietnam in absolute terms, but per million inhabitants it produces only a fifth of Malaysia’s output, and its resident patents per million inhabitants (6.9) are lower than those of Thailand, Vietnam and Malaysia, and only slightly below the Philippines. Singapore, with about 5.9 million people, files almost as many resident patents as Indonesia with more than 281 million. Korea files about 2.7 resident patents for every scientific article, while Indonesia files about five patents for every hundred articles. That ratio is not a measure of scientific quality, but a measure of how far scientific work is connected to technological work within the system (confidence: high).
Data dictionary
The following data dictionary defines every variable in the dataset, its unit, its original source, its series code where one exists, and its country and year coverage. Derived variables are marked as calculations by this study, so that readers can distinguish figures quoted directly from a source from figures that result from arithmetic operations on source figures. The dictionary is meant to allow the dataset to be updated each year by students or other analysts without having to guess at decisions taken in the first version.
| Variable | Definition | Unit | Source | Code | Coverage |
|---|---|---|---|---|---|
| R&D spending | Gross domestic expenditure on research and development, capital and current, in the business, government, higher education and private non-profit sectors | % of GDP | UNESCO UIS via WDI (UIS release February 2025) | GB.XPD.RSDV.GD.ZS | 11 countries; 2015–2023; 1996–2023 series for six countries |
| Researchers | Professionals engaged in the creation of new knowledge, products, processes or methods, counted in full-time equivalents | per million inhabitants | UNESCO UIS via WDI | SP.POP.SCIE.RD.P6 | 11 countries; 2015–2023 |
| Tertiary gross enrolment ratio | Total tertiary enrolment relative to the population of official tertiary age; may exceed 100 | % | UNESCO UIS via WDI | SE.TER.ENRR | 11 countries; 2023–2025 |
| Scientific and technical articles | Scopus-indexed journal and conference articles in science and engineering fields | count | NSF Science and Engineering Indicators via WDI | IP.JRN.ARTC.SC | 11 countries; 2010–2023 |
| Articles per million inhabitants | 2023 articles divided by 2023 population | per million | Calculation by this study | derived | 11 countries; 2023 |
| Resident patents | Patent applications filed by residents at the national patent office | count | WIPO via WDI | IP.PAT.RESD | 11 countries; 2010–2024 |
| Non-resident patents | Patent applications filed by foreign parties at the national patent office | count | WIPO via WDI | IP.PAT.NRES | 9 countries; 2024 |
| Resident patent share | Resident patents divided by the sum of resident and non-resident patents | % | Calculation by this study | derived | 9 countries; 2024 |
| High-technology exports | Exports of products with high R&D intensity as a share of manufactured exports | % | WDI | TX.VAL.TECH.MF.ZS | 9 countries; 2023–2024 |
| IP royalty payments | Charges for the use of intellectual property paid abroad, balance of payments | current USD | WDI | BM.GSR.ROYL.CD | 8 countries; 2025 |
| IP royalty receipts | Charges for the use of intellectual property received from abroad | current USD | WDI | BX.GSR.ROYL.CD | 8 countries; 2025 |
| Royalty ratio | IP royalty payments divided by receipts | times | Calculation by this study | derived | 8 countries; 2025 |
| GII rank | Global Innovation Index rank, with input and output sub-index ranks | rank out of 139 | WIPO, GII 2026 economy profiles | none | 11 countries; 2026 edition |
| National research spending (BRIN) | Sum of government, higher education and business research spending under the Frascati Manual 2015 | Rp and % of GDP | BRIN, Indikator Iptek 2024 | none | Indonesia; 2023 |
| Population | Mid-year total population | persons | WDI | SP.POP.TOTL | 11 countries; 2023 |
Two decisions in this dictionary need explaining. First, scientific and technical articles exclude the humanities and most of the social sciences, because the National Center for Science and Engineering Statistics taxonomy focuses on science and engineering, so this indicator measures science-and-engineering capacity, not all academic production. Second, resident patents count applications, not granted patents, and count them at each country’s national patent office. This measure is sensitive to incentive policy: a surge in applications can arise from administrative requirements placed on lecturers or researchers rather than from a surge in invention. For that reason this study reads patent figures together with their trajectory, not as single points.
Method and data quality: selection, harmonisation, vintage and limits
Country selection follows three criteria. The six core ASEAN states were chosen because they are the largest economies in the region and have the most complete data series. Brunei and Cambodia were included because data exist for most indicators, albeit with old years, so their presence shows the range of capacity within ASEAN. Laos and Myanmar, which according to the SeasiaStats summary rank 108th and 125th respectively in the Global Innovation Index 2026, were not included because this study did not retrieve R&D and researcher series for them, so including them with composite ranks alone would produce a lopsided comparison. South Korea was chosen as the most successful example of the transition from absorber to creator, China as an example of the transition at continental scale, and India as a comparator large lower-middle-income country with output strength out of proportion to its inputs.
Harmonisation is minimal and explicit. This study does not impute missing values, does not interpolate between years and does not adjust values to a common base year. As a result, cells in a single column can come from years up to eight years apart, as with Cambodia’s R&D spending (2015) compared with Korea’s (2023). This choice is deliberate because imputation would give an impression of precision that the data do not possess. Per-million variables are calculated with 2023 population for all countries; for 2024 patents this produces a slight overestimate, but one year of population growth in these countries is too small to change rankings or magnitudes meaningfully.
The vintage problem is far more serious than is usually acknowledged. The World Development Indicators metadata state that the R&D spending and researcher series are drawn from the UNESCO Institute for Statistics bulk download service published in February 2025, even though the World Bank database itself was updated in October 2026. The Global Innovation Index 2026 profile for Korea lists R&D spending of 5.13 per cent of GDP for 2024, while the World Development Indicators still list 4.94 per cent for 2023. The two official figures do not contradict each other; they come from different vintages. But a reader who takes one figure without its year and source will produce a faulty comparison. This study chooses the World Development Indicators figures for cross-country consistency and records the newer WIPO figure in the text.
UIS reporting lags affect the countries that matter most for this analysis. The latest R&D spending and researcher observations for Indonesia, the Philippines and India date from 2020, a pandemic year in which GDP fell and spending-to-GDP ratios could be distorted. WIPO’s profile for Indonesia explicitly flags sixteen indicators as outdated, including R&D spending, researchers and business-sector R&D spending, while India’s profile notes that its R&D spending figure of 0.65 per cent and 259.3 researchers per million are 2020 data. This means that the position of three populous countries in every global innovation index is partly determined by data that are six years old when the index is published (confidence: high).
Indonesia’s case requires a separate note, because the institutional reorganisation of research since BRIN’s creation has changed how data are collected and reported. BRIN’s 2024 report puts 2023 national research spending at Rp20,394,979,384,006, or about 0.10 per cent of GDP at current prices of Rp20,892.4 trillion, and states that the ratio was unchanged from the previous year. That figure is less than half the 0.28 per cent for 2020 relayed by UIS and used in the Global Innovation Index. BRIN’s report itself explains that its data come from several agencies, namely BRIN for government, the education ministry’s data centre for higher education, the Ministry of Industry’s industrial information system for business, and the Ministry of Home Affairs for regional governments, that some figures are still provisional, and that collecting business-sector data is its main challenge. This study cannot determine whether the difference between 0.28 and 0.10 per cent reflects a real decline, a difference in survey coverage, or both, so the two figures are reported side by side (confidence that Indonesian research spending is below 0.3 per cent of GDP: high; confidence about its exact magnitude: low).
Definitional differences also affect researcher figures. UIS reports 395 full-time-equivalent researchers per million inhabitants for 2020. BRIN’s 2024 report records 439,349 researchers by headcount for 2023, which, divided by the population of 279,179,042 used in that report, gives about 1,574 researchers per million inhabitants. The same report presents no full-time-equivalent figure and notes that 96.62 per cent of researchers are in higher education, a category that includes lecturers and registered doctoral students. The fourfold gap between the two figures is not an error; it is a measure of how much of Indonesian lecturers’ time is effectively available for research. Headcount captures the number of people administratively able to research, whereas full-time equivalence captures the research work actually done.
The final limit concerns the absorption indicators. Royalty payment and receipt data for Vietnam, Brunei and Cambodia were not available in the retrieval carried out, so they are written as “n.a.”. Vietnam’s high-technology exports are available up to 2023, not 2024. Royalty payments may also reflect multinational transfer-pricing practices rather than knowledge flows alone. For these reasons this study uses absorption indicators to read broad patterns, not to compare small differences between countries.
The absorption-production gap: reading the region’s contradiction
Table 4 places absorption indicators alongside indicators of technology ownership. High-technology exports come from the series TX.VAL.TECH.MF.ZS, non-resident patents from the series IP.PAT.NRES, and royalty payments and receipts from the series BM.GSR.ROYL.CD and BX.GSR.ROYL.CD. The resident patent share and the royalty ratio are calculations by this study.
| Country | High-tech exports (% of manufactured exports, 2024) | Non-resident patents (2024) | Resident patent share (%) | IP royalty payments (USD billion, 2025) | IP royalty receipts (USD billion, 2025) | Payments/receipts ratio |
|---|---|---|---|---|---|---|
| Indonesia | 8.7 | 8,971 | 17.7 | 2.49 | 0.22 | 11.4 |
| Malaysia | 58.6 | 6,356 | 11.7 | 2.66 | 0.24 | 11.2 |
| Singapore | 59.4 | 11,523 | 14.1 | 20.69 | 15.68 | 1.3 |
| Thailand | 28.3 | 7,797 | 10.7 | 6.97 | 0.35 | 19.9 |
| Vietnam | 44.3 (2023) | 8,678 | 12.4 | n.a. | n.a. | n.a. |
| Philippines | 60.4 | 3,698 | 19.1 | 0.38 | 0.02 | 16.1 |
| South Korea | 36.3 | 50,459 | 79.5 | 14.14 | 9.98 | 1.4 |
| China | 26.3 | 156,053 | 91.5 | 46.63 | 13.07 | 3.6 |
| India | 18.6 | 41,940 | 60.1 | 18.76 | 1.74 | 10.8 |
The pattern in Table 4 is consistent and striking. In every ASEAN state in the table, the resident share of total patent applications lies between 10.7 and 19.1 per cent. In Korea it is 79.5 per cent, in China 91.5 per cent, and even in India, with research spending no higher than Thailand’s, it is 60.1 per cent. In other words, patent offices in Southeast Asia function mainly as places where foreign firms protect their technology in the regional market, not as places where local inventors register inventions. This study judges that a resident patent share below twenty per cent across the whole of ASEAN is the most concise indicator of the region’s absorption-production gap (confidence: high).
The royalty balance confirms that reading from the side of money flows. Thailand paid almost twenty times more in intellectual property royalties than it received in 2025, the Philippines sixteen times, and Indonesia and Malaysia about eleven times. Korea, by contrast, paid only 1.4 times its receipts, and Singapore 1.3 times. For Singapore, the low ratio partly reflects its role as a regional hub for multinational ownership of intellectual property, so the figure cannot be read directly as an advantage of local inventors. But for Korea, the ratio is the result of four decades of investment that made domestic firms owners of technology rather than mere users of it.
Malaysia and the Philippines are the clearest illustrations of absorption without production. Both export high-technology products in larger shares than Korea and China, yet their resident patent shares are among the lowest. WIPO’s profiles rank Malaysia first in the world for high-technology exports and imports as a share of total trade, and the Philippines fourth for high-technology exports. Such a structure is common when high-technology exports consist mainly of assembly, testing and packaging of electronic components by foreign subsidiaries, with research and design carried out in the parent firm’s home country. WIPO’s profile for the Philippines also records that the country’s human capital and research pillar ranks 100th, a sharp contrast with its fourth place for high-technology exports.
The region ships chips to the whole world, yet patents almost nothing of what is inside them.
Vietnam shows a different and more dynamic variant. Its resident patent share (12.4 per cent) is still low, but its absolute number of resident patents rose from 306 in 2010 to 1,226 in 2024, a fourfold increase with almost no interruption, and its output rank in the Global Innovation Index (35) is far above its input rank (54). Vietnam is also the only core ASEAN state with an explicit quantitative target for raising research intensity. An opinion piece by three BRIN researchers in The Jakarta Post in March 2025 noted that Politburo Resolution 57 targets a rise in R&D spending from 0.4 per cent to 2 per cent of GDP and 12 researchers per 10,000 inhabitants by 2030. The target is highly ambitious and will very probably not be met in full, but it shifts the policy conversation from absorbing foreign investment to building domestic capability.
The analytical lesson of this section is that high absorption indicators do not automatically signal a strong knowledge system, and that moderate input indicators do not automatically signal a weak one. A country that looks advanced in high-technology exports can be very fragile if decisions on production location are made elsewhere, while a country with moderate inputs can show strong conversion capacity if its nodes are connected. In Lundvall’s language, what matters is the intensity of interactive learning between producers and users of knowledge within the system, and absorption indicators alone do not tell us whether that learning occurs at home or merely passes through.
Indonesia’s position: publications surge, the system lags
Indonesia’s trajectory since 2010 is an unbalanced one. The number of scientific and technical articles rose from 1,391 (2010) to 2,652 (2013), 5,850 (2016), 17,156 (2019), 23,654 (2021) and 38,992 (2023). That twenty-eight-fold increase in thirteen years far outpaces Vietnam (almost tenfold, from 1,096 to 10,644) and Malaysia (about twofold, from 10,816 to 23,691). Over the same period, Indonesia’s resident patent applications rose from 508 (2010) to 1,058 (2015) and peaked at 3,093 in 2019, then fell to 1,549 in 2022 before recovering partly to 1,931 in 2024. The publication layer grew exponentially, while the technology layer fluctuated without a convincing trend.
This study does not discuss the quality of or incentives behind that publication growth, which is a separate debate. The question relevant to system capacity is why publication growth has not connected to the technology layer. The answer best supported by the data is the structure of research performers. BRIN’s 2024 report, as also summarised by Good News From Indonesia in September 2024, records that of total 2023 research spending, higher education contributed 46.02 per cent, government 32.42 per cent, and business or industry only 21.56 per cent. On the human-resource side, 95.28 per cent of science and technology personnel are in higher education and only 1.40 per cent in business. In Korea, the WIPO profile records business-performed R&D of 4.2 per cent of GDP, first in the world, while in Thailand business spending is about 0.6 per cent of GDP. Indonesia’s is a research system run almost entirely by universities, in an economy whose firms hardly do research.
The industry share figure itself is contested. At the Business Gathering 2025, as reported by Bisnis.com on 28 April 2025, the Minister of Higher Education, Science and Technology, Brian Yuliarto, cited Indonesian research spending of 0.28 per cent against an ASEAN average of 0.70 per cent, and the same report quoted an industry funding share of 7.3 per cent, far below BRIN’s figure. The difference between 7.3 and 21.56 per cent probably stems from the difference between who funds and who performs research, or from differences in year and source, but the report does not name its source. The fact that the minister and the national research agency use different figures for a basic measure of the system is a finding in its own right about Indonesia’s science and technology statistical capacity (confidence: high).
Institutional reorganisation adds a layer of uncertainty. The merger of government research institutes into BRIN consolidated budgets and researchers, but it also placed most government research in a single budget line vulnerable to fiscal policy. The Jakarta Post piece already cited notes that under the 2025 budget-efficiency policy, BRIN initially proposed a cut of about Rp2 trillion from a total proposal of about Rp5.8 trillion, and the House of Representatives finally approved a cut of about Rp1.4 trillion. For a system whose national research spending is only about Rp20 trillion, a cut of that size at a single node is a meaningful shock. On the other hand, the creation of the Ministry of Higher Education, Science and Technology opens an opportunity to reconnect higher education and research policy, which had previously been separated.
Indonesia’s problem is not new, and several documents in the Arsip & Pustaka collection show that its diagnosis was formulated more than a decade ago. In 2014, Hendra Gunawan, writing in the daily Kompas under the title “Mengintip Masa Depan” (Peeking into the Future), called universities and research institutes the nation’s seed barn whose investment was very small, and compared Indonesia’s 2014 higher education budget of about Rp3.2 trillion with about Rp18 trillion for the National University of Singapore in 2013. On the same day, Budi Santosa, writing in Kompas under the title “Dari ‘Brain Drain’ ke ‘Brain Gain’” (From Brain Drain to Brain Gain), argued that concern about losing scholarship talent was overstated, pointing to China and Korea, which drew their scientists back as their economies improved, and proposed funding foreign professors and postdoctoral researchers at Indonesian universities. Al Makin’s 2018 professorial inaugural lecture Bisakah Menjadi Ilmuwan di Indonesia?: Keilmuan, Birokrasi dan Globalisasi (Can One Become a Scientist in Indonesia? Scholarship, Bureaucracy and Globalisation) added the micro-diagnosis that bureaucratic and administrative burdens erode lecturers’ research time, a diagnosis that can now be read quantitatively in the fourfold gap between headcount and full-time-equivalent researchers.
The diagnosis also applies to religious higher education institutions, which house a large share of lecturers outside science and engineering. A 2026 presentation entitled Rencana Strategis UIN Ar-Raniry (Strategic Plan of UIN Ar-Raniry) places the research ecosystem among its transformation strategies, alongside internationalisation, lecturer capacity and digital transformation, while a 2026 presentation entitled PKDP 2026: Memahami Paradigma Keilmuan di PTKI (Understanding Scholarly Paradigms in Islamic Higher Education) maps four scholarly paradigms that have developed in Islamic religious higher education. Both documents show that at institutional level, the question of research capacity has already become part of strategic planning. But this dataset shows that institutional plans will not shift national figures as long as research continues to be treated as an add-on function for teachers, rather than as work with its own time, funding and career track.
Comparator trajectories: Korea, China, Malaysia, Thailand and Vietnam
Single points can mislead, so Table 5 presents the trajectory of R&D spending relative to GDP for six countries since 1996, based on the series GB.XPD.RSDV.GD.ZS. An “n.a.” cell means the World Development Indicators hold no observation for that year.
| Country | 1996 | 2000 | 2005 | 2010 | 2015 | 2018 | 2020 | 2022 | 2023 |
|---|---|---|---|---|---|---|---|---|---|
| South Korea | 2.14 | 2.05 | 2.43 | 3.18 | 3.79 | 4.27 | 4.52 | 4.85 | 4.94 |
| China | 0.56 | 0.88 | 1.29 | 1.68 | 2.02 | 2.10 | 2.36 | 2.49 | 2.58 |
| Malaysia | 0.22 | 0.47 | n.a. | 1.04 | 1.28 | 1.04 | 0.95 | 1.01 | n.a. |
| Thailand | 0.12 | 0.24 | 0.22 | n.a. | 0.62 | 1.11 | 1.33 | 1.16 | 0.94 |
| Vietnam | n.a. | n.a. | n.a. | n.a. | 0.36 | n.a. | n.a. | n.a. | 0.41 |
| Indonesia | n.a. | 0.07 | n.a. | n.a. | n.a. | 0.23 | 0.28 | n.a. | n.a. |
Korea is the reference case because its research intensity was already above two per cent of GDP in 1996 and kept rising to almost five per cent in 2023, and even 5.13 per cent for 2024 according to the WIPO profile. Linsu Kim, in his 1997 book Imitation to Innovation: The Dynamics of Korea’s Technological Learning (Harvard Business School Press), described how Korean firms moved from imitating mature technologies towards innovation through deliberate learning, often forced by crisis. Keun Lee, in his 2013 book Schumpeterian Analysis of Economic Catch-up (Cambridge University Press), argued that successful catching-up countries tend to enter technology fields with short cycles, where dependence on old knowledge is smaller and the chances of leapfrogging greater. For this dataset, Korea’s lesson lies not in the size of spending but in its composition: the business sector became the main research performer, and a resident patent share of almost eighty per cent is the consequence.
China shows that the transition can be made at far larger scale and in a shorter time. Its research intensity rose from 0.56 per cent in 1996 to 2.58 per cent in 2023, and its resident patent share reached 91.5 per cent in 2024. WIPO’s profile ranks China 10th in the Global Innovation Index 2026, with an output rank of 5 and an input rank of 22, and Yicai reported that the Shenzhen-Hong Kong-Guangzhou cluster retained first place in the world as an innovation cluster. For Southeast Asia, China’s case matters not as a model to be copied, since its scale and state capacity are not available in the region, but as proof that the resident patent share can shift dramatically within a generation when business spending and policy targets move together.
Malaysia is a warning. Its research intensity rose rapidly from 0.22 per cent in 1996 to 1.28 per cent in 2015, then fell and held at around one per cent. Its resident patent applications actually declined from 1,231 in 2010 to 841 in 2024. Malaysia has one of the most integrated electronics export sectors in the world, but this trajectory shows that integration has not produced a comparable deepening of domestic technological capability. Richard Doner and Ben Ross Schneider, in their 2016 article “The Middle-Income Trap: More Politics than Economics” in World Politics 68(4), argued that upgrading productivity through skills and innovation requires heavy investment in institutional capacity, while the political capacity to build those institutions is weakened by fragmented support coalitions, especially among business and labour, and by inequality. This dataset cannot test that argument causally, but Malaysia’s plateau since 2015 fits its prediction: a country can reach a moderate level of spending through government programmes, but going beyond it requires a coalition with domestic firms willing to do research (confidence: moderate).
Thailand shows a more undulating trajectory. Its research intensity rose from 0.12 per cent in 1996 to 1.33 per cent in 2020, one of the fastest increases in the region, then fell to 0.94 per cent in 2023. WIPO’s profile records Thai business-performed R&D at about 0.6 per cent of GDP on 2023 data, indicating that much of the earlier rise was indeed driven by firms. Yet Thailand’s resident patent applications moved from 1,214 in 2010 to 930 in 2024, so the rise in business spending has not yet been reflected in technology ownership. The most likely explanation is that part of business research in Thailand consists of process development and product adaptation that is not patented, or is performed by foreign subsidiaries that patent in the name of their parent, but this study lacks the data to confirm it.
Vietnam is the last case and the most relevant for Indonesia. Its research intensity barely changed, from 0.36 per cent in 2015 to 0.41 per cent in 2023, but researchers per million inhabitants reached 836, up 8.88 per cent from the previous year according to the WIPO profile, its resident patents rose fourfold since 2010, and its Global Innovation Index rank rose to 43rd in 2026. Vietnam is also classified as a lower-middle-income economy in that profile and ranks second of 36 in its group, while Indonesia ranks seventh of 35 upper-middle-income economies. With similar inputs and lower per capita income, Vietnam produces higher innovation outputs. This study judges that the difference stems mainly from deeper industrial integration into electronics supply chains and from consistent political signals at the highest level, not from an already proven superiority of its research system (confidence: moderate).
The counter-argument: do catching-up countries really need to create?
The strongest counter-argument to this study deserves to be stated in its best form. For a catching-up country, the argument runs, absorbing existing technology is far cheaper and faster than creating new technology, so the rational strategy is to maximise foreign investment, trade and mass education, and then let innovative capacity grow by itself as income rises. Singapore is often cited as proof: the city-state became rich mainly through multinational firms. For Indonesia, the argument is reinforced by the WIPO profile, which ranks domestic market scale 7th in the world, the state of cluster development 7th, entrepreneurship policies and culture 9th, and university-industry research collaboration 10th. Moreover, R&D and patent indicators are notorious for failing to capture informal innovation, business-model innovation and digital innovation, which are growing fast precisely in large economies such as Indonesia. From this angle, concern over low research spending figures is statistical fetishism.
The argument contains important truths, and this study accepts three parts of it. Absorption is indeed a necessary stage, not a failure; formal indicators are indeed incomplete; and Indonesia’s domestic market scale is indeed a strategic asset that Malaysia or Thailand do not have. But the argument fails at three points that can be checked against the data. First, Cohen and Levinthal showed that the capacity to absorb depends on related knowledge already held and built, among other ways, through in-house research. A country that does not do research does not merely fail to create; it also absorbs shallowly, limited to operating technology it does not understand. The patent trajectories of Malaysia and Thailand since 2010 are an empirical illustration of absorption that does not deepen.
Second, Singapore’s case actually refutes the argument when read with data. Singapore has 8,782 researchers per million inhabitants, research intensity of 1.81 per cent of GDP, and first place in the world on the input sub-index of the Global Innovation Index 2026. The city-state did not simply wait for multinationals; it built research universities and public research institutes on a scale that makes it one of the world’s most intensive knowledge producers per capita, with 2,010 articles per million inhabitants. Third, the World Development Report 2024 does not say that absorption is enough; it says that middle-income countries must move from investment to technology infusion and then to innovation, and that this shift does not happen by itself. Evidence that rising income automatically gives rise to innovation is not found in this dataset: Malaysia and Thailand have been upper-middle-income countries for years, yet their resident patent shares remain around eleven per cent.
Indonesia’s strengths in the WIPO profile must also be read carefully. Market scale, clusters, entrepreneurship and university-industry collaboration describe potential and institutional conditions, not technology outputs. The same profile ranks education spending as a share of GDP 131st, scientific articles per billion PPP dollars of GDP 125th, and greenfield foreign direct investment in R&D and high technology 112th. Potential that is not converted will not last long as an advantage, because digital firms growing in a large domestic market will buy technological capability from abroad if it is not available at home. The counter-argument is therefore right about sequence but wrong about automaticity. This study’s conclusion stands: without deliberate investment in the nodes of knowledge production, especially in the business sector, the absorption-production gap will persist (confidence: high).
Scenarios to 2030 and signposts to watch
The following scenarios focus on Indonesia in its regional context, because this dataset shows that Indonesia is the country with the largest gap between scale and capacity. The 2030 horizon was chosen because it coincides with the target year of Vietnam’s Resolution 57 and with the first half of implementation of the ASEAN Plan of Action on Science, Technology and Innovation 2026–2035. The probabilities given are qualitative assessments based on the data trajectories in this study, verified policy signals and known fiscal constraints.
Scenarios to 2030
- AAbsorptive stagnation45%
- BGradual consolidation30%
- CFiscal regression15%
- DDirected leap10%
Probabilities are the author’s assessment as of 1 November 2026, not statistical forecasts.
Scenario A, absorptive stagnation, is the continuation of the current pattern and therefore the most likely. In this scenario, national research spending by BRIN’s measure stays below 0.2 per cent of GDP, publications keep growing but more slowly, resident patents fluctuate around two thousand applications a year, and the Global Innovation Index rank moves within the 50s. Indonesia keeps attracting investment in mineral processing and manufacturing, but the technological capability within those value chains remains owned by foreign partners. Scenario B, gradual consolidation, occurs if the government aligns its science and technology statistics, sets business spending targets and offers fiscal incentives that medium and large firms can actually use. Research spending rises towards 0.3 per cent by BRIN’s measure, the industry share of spending moves towards a third, and resident patents surpass their 2019 peak.
Scenario C, fiscal regression, occurs if budget pressure drives further cuts to BRIN and higher education without compensation from the private sector, so that senior researchers move abroad or into non-research sectors and even the publication layer begins to weaken. The 2025 cuts to BRIN’s budget show that this scenario is not merely hypothetical. Scenario D, the directed leap, is the Vietnamese variant: political commitment at the highest level with quantitative targets, a rapid rise in public spending and mandatory partnerships between foreign investors and domestic research institutions. Its probability is low because there is no policy signal of that magnitude yet, but it is not zero, since the Indonesian government has the fiscal and market capacity to change course if the political decision is taken.
The signposts to watch can be ordered by how soon they appear. In the short term, the most important is the next edition of BRIN’s Science and Technology Indicators report: whether the research-spending-to-GDP ratio moves from 0.10 per cent, whether the industry share is calculated with a stable method, and whether full-time-equivalent researcher figures begin to be presented. The second is whether UIS receives observations for Indonesia, the Philippines and India after 2020, so that global indices stop relying on pandemic-year data. In the medium term, the most informative signpost is the resident patent share: if that figure in Indonesia or Vietnam exceeds a quarter by 2030, the absorption-production gap will have begun to narrow. The regional signpost is whether implementation of ASEAN’s science and technology plan of action produces a shared statistical system that shortens data lags. The final signpost is Vietnam’s achievement of its 2030 targets, since its success or failure will become the strongest argument in research policy debates across the region.
Time-bound recommendations by actor
For the Ministry of Higher Education, Science and Technology (Kemdiktisaintek), the first recommendation is to change how research is treated in lecturers’ workloads. Before the 2027/2028 academic year, the ministry should issue verifiable guidelines on research time allocation and conduct a full-time-equivalent researcher survey consistent with the Frascati Manual, so that the fourfold gap between headcount and full-time equivalence can be measured and then reduced. By 2028, the ministry should tie part of university research funding to genuine industry partnerships, such as industrial doctorate programmes and co-funded joint projects, rather than to publication counts. The ministry should also ensure that the figures its officials use in public forums come from the same official statistics as those reported to international agencies.
For BRIN, the most urgent priority is statistics. By mid-2027, BRIN should publish a reconciliation note explaining the difference between the 0.28 per cent of GDP for 2020 relayed by UIS and the 0.10 per cent of GDP for 2023 in BRIN’s report, including differences in sector coverage, data sources and the status of provisional figures. From the 2027 cycle, BRIN should report R&D spending and researcher data to UIS every year with a lag of no more than two years. In its research function, BRIN should prioritise joint research contracts with industry and regional governments, because a government research agency that is not connected to users of knowledge will only relocate the problem, not solve it.
For Bappenas, the main recommendation is to make knowledge-system indicators part of the development monitoring framework, not merely an annex to the education sector. In the mid-term review of the medium-term development plan, Bappenas should set separate targets for total research spending, the business share of research spending and the resident patent share, because aggregate targets without composition will push up government spending without changing the structure. Bappenas should also test R&D fiscal incentive options on the basis of their take-up data, and design a framework so that foreign investment in strategic sectors comes with measurable commitments to domestic research partnerships.
For universities, including religious higher education institutions, the recommendation is to separate research career tracks from teaching tracks gradually from 2027, to establish technology transfer offices that are genuinely run by professional staff, and to build partnership, patent and research contract indicators into institutional strategic plans. Universities should also use datasets such as this one as teaching material in policy studies, knowledge management and development studies programmes, so that the next generation can read science and technology statistics critically. For large firms, including state-owned enterprises, the minimum step is to report research spending voluntarily and consistently to BRIN from 2027, because the weakness of business-sector statistics is one of the main causes of uncertainty in the national figures.
For the ASEAN Committee on Science, Technology and Innovation (COSTI, formerly known as ASEAN COST), the recommendation is to make a regional science and technology statistics observatory one of the early outputs of the ASEAN Plan of Action on Science, Technology and Innovation 2026–2035, the first phase of which was developed by a consortium led by Cambridge Industrial Innovation Policy for the ASEAN Secretariat. The observatory should produce annual tables aligned with UIS and WIPO, including the resident patent share and royalty balance, before 2028. For UNESCO and UIS, the recommendation is focused technical support for Indonesia, the Philippines and Cambodia so that post-2020 observations are available by 2028 at the latest, together with clearer vintage labelling in data distribution. The UNESCO Science Report 2021 already provides a regional framework that can be updated with such data.
For the World Bank, the recommendation is to translate the diagnosis of the World Development Report 2024 into country-level analysis for Indonesia, the Philippines and Vietnam before 2028, explicitly assessing at which stage of the investment, infusion and innovation sequence each country stands. Lending programmes in higher education and innovation should include a component for strengthening science and technology statistics, since programme evaluation cannot be carried out without timely data. The World Bank should also flag source vintages more prominently in the World Development Indicators metadata, so that users do not compare 2020 data with 2024 data as if they were contemporaneous.
Limitations and conclusion
The limitations of this study need to be stated clearly. First, the dataset is cross-sectional with observation years that differ across countries, so comparisons between countries with large year gaps should be read as indications, not precise measurements. Second, formal indicators such as R&D spending, patents and indexed articles do not capture informal innovation, organisational innovation and knowledge production outside science and engineering, including the social sciences, humanities and local knowledge that matter to the region. Third, this study does not test causal relationships; statements about mechanisms, for example about the role of political coalitions or the ownership structure of firms, are interpretations consistent with the data, not statistical conclusions. Fourth, the composite ranks for Brunei and Cambodia come from a secondary source, and royalty indicators are not available for Vietnam, Brunei and Cambodia. Fifth, Indonesia’s figures depend on two measurement systems that have not yet been reconciled.
The conclusion can be stated firmly. Southeast Asia has proved its capacity to absorb technology, and that is a real development achievement. But this dataset shows that the capacity to absorb has not been converted into the capacity to create: the resident patent share across ASEAN is below twenty per cent, the royalty balance of most countries is in deficit by more than tenfold, and the trajectories of Malaysia and Thailand show that moderate research intensity does not automatically carry on into technology ownership. Indonesia faces a double gap, with a fast-growing publication layer that is not connected to the technology layer, a business sector that hardly does research, and science and technology statistics not yet able to provide a single agreed baseline figure.
The policy implication is not merely to raise research budgets, though that is necessary. It is to build the missing nodes in the system: research inside firms, real researcher career tracks in universities, links between foreign investment and domestic research institutions, and statistics that are timely and trustworthy. Korea and China show that the transition from absorber to creator can be achieved within a generation, while Vietnam shows that consistent political signals can produce outputs above inputs. For Indonesia, the question is no longer whether the national knowledge system needs to change, but whether the decision to change it will be taken before the demographic and supply-chain windows now open begin to close.
◆ ◆ ◆
References
Official documents and data
- Badan Riset dan Inovasi Nasional. 2024. Indikator Iptek, Riset, dan Inovasi Indonesia 2024. Jakarta: Penerbit BRIN. DOI 10.55981/brin.1319.
- World Bank. 2026. Research and development expenditure (% of GDP), GB.XPD.RSDV.GD.ZS. World Development Indicators, UNESCO Institute for Statistics data.
- World Bank. 2026. Researchers in R&D (per million people), SP.POP.SCIE.RD.P6. World Development Indicators.
- World Bank. 2026. Scientific and technical journal articles, IP.JRN.ARTC.SC. World Development Indicators, NSF Science and Engineering Indicators data.
- World Bank. 2026. Patent applications, residents, IP.PAT.RESD; and Patent applications, nonresidents, IP.PAT.NRES. World Development Indicators, WIPO data.
- World Bank. 2026. School enrollment, tertiary (% gross), SE.TER.ENRR. World Development Indicators.
- World Bank. 2026. High-technology exports (% of manufactured exports), TX.VAL.TECH.MF.ZS. World Development Indicators.
- World Bank. 2026. Charges for the use of intellectual property, payments; and receipts. World Development Indicators.
- World Bank. 2026. Population, total, SP.POP.TOTL. World Development Indicators.
- WIPO. 2026. Global Innovation Index 2026, economy profiles: Indonesia, Malaysia, Singapore, Thailand, Viet Nam, Philippines. Geneva: World Intellectual Property Organization.
- WIPO. 2026. Global Innovation Index 2026, economy profiles: Republic of Korea, China, India. Geneva: World Intellectual Property Organization.
Scholarly literature
- Cohen, Wesley M., and Daniel A. Levinthal. 1990. “Absorptive Capacity: A New Perspective on Learning and Innovation.” Administrative Science Quarterly 35(1): 128–152.
- Doner, Richard F., and Ben Ross Schneider. 2016. “The Middle-Income Trap: More Politics than Economics.” World Politics 68(4): 608–644.
- Freeman, Christopher. 1995. “The ‘National System of Innovation’ in Historical Perspective.” Cambridge Journal of Economics 19(1): 5–24.
- Kim, Linsu. 1997. Imitation to Innovation: The Dynamics of Korea’s Technological Learning. Boston: Harvard Business School Press.
- Lee, Keun. 2013. Schumpeterian Analysis of Economic Catch-up: Knowledge, Path-Creation, and the Middle-Income Trap. Cambridge: Cambridge University Press.
- Lundvall, Bengt-Åke, ed. 1992 (reissued 2010). National Systems of Innovation: Toward a Theory of Innovation and Interactive Learning. London: Anthem Press.
Policy analysis and institutional reports
- World Bank. 2024. World Development Report 2024: The Middle-Income Trap. Washington, DC: World Bank.
- Chen, Derek H. C., and Carl J. Dahlman. 2005. The Knowledge Economy, the KAM Methodology and World Bank Operations. Washington, DC: World Bank Institute.
- UNESCO. 2021. UNESCO Science Report: The Race Against Time for Smarter Development. Paris: UNESCO.
- Cambridge Industrial Innovation Policy. 2023 (updated 2026). ASEAN Plan of Action on Science, Technology and Innovation (APASTI) 2026–2035. Cambridge: University of Cambridge.
News media
- Wiratri, Amorisa, Betti Rosita Sari, and Lamijo. 2025. “Budget Efficiency Contrast in ASEAN: Tales from Indonesia and Vietnam.” The Jakarta Post, March 2025.
- Bisnis.com. 2025. “Pendanaan Riset RI Lebih Rendah dari Rerata Asean, Butuh Dukungan Swasta.” 28 April 2025.
- Rizqiyah, Aslamatur. 2024. “Tahun 2023, Belanja Riset dan SDM Iptek Masih Didominasi Perguruan Tinggi.” Good News From Indonesia, 8 September 2024.
- VietnamPlus. 2026. “WIPO Innovation Report Viewed through Resolution 57 Perspective.” 3 October 2026.
- Yicai Global. 2026. “China Retains 10th Spot in WIPO Ranking of World’s Most Innovative Economies.” 30 September 2026.
- Sumarno, Daniel Dimays. 2026. “Southeast Asia’s Most Innovative Economies in 2026.” SeasiaStats, 7 October 2026.
Arsip & Pustaka
- Gunawan, Hendra. 2014. “Mengintip Masa Depan.” Kompas, 2 May 2014.
- Santosa, Budi. 2014. “Dari ‘Brain Drain’ ke ‘Brain Gain’.” Kompas, 2 May 2014.
- Makin, Al. 2018. Bisakah Menjadi Ilmuwan di Indonesia?: Keilmuan, Birokrasi dan Globalisasi. Professorial inaugural lecture. Yogyakarta: UIN Sunan Kalijaga.
- Ahmad, Kamaruzzaman Bustamam. 2026. Rencana Strategis UIN Ar-Raniry. Presentation, Tim Renstra UIN Ar-Raniry Banda Aceh.
- Ahmad, Kamaruzzaman Bustamam. 2026. PKDP 2026: Memahami Paradigma Keilmuan di PTKI. Presentation.
How to Cite
Ahmad, Kamaruzzaman Bustamam. 2026. “Absorbing Without Creating: The Knowledge-Production Capacity of ASEAN States: A Documented Comparative Dataset Benchmarked against South Korea, China and India.” KBA13 Academy Research Center, Datasets, 1 November 2026. https://kba13academy.com/kapasitas-produksi-pengetahuan-asean/
Professor of Anthropology of Religion at UIN Ar-Raniry Banda Aceh and founder of KBA13 Academy. His work spans Islam in Southeast Asia, anthropology, strategic studies and knowledge systems.
