Latest / The Productivity Podcast with Fexingo: Output, Efficiency, and Long-Term Economic Growth / Why South Korea Built the World's Most Productive R&D Sector
Transcript
- Lucas: If you walk through the economy long enough, you notice a pattern: some countries invest heavily in research and development and get a modest productivity bump. South Korea isn't one of those countries. It spends a bigger share of its GDP on R&D than any other nation — 4.8 percent in 2024, versus the US at 3.5 and the OECD average at 2.7 — and it gets an outsized productivity return. Luna: That's a massive gap. But spending alone doesn't explain it. I mean, plenty of countries throw money at R&D and end up with fragmented labs and low commercial yield. Lucas: Exactly. The structure is what matters. South Korea built a tightly coordinated system where the government, the big family-owned conglomerates — the chaebol — and the universities operate almost like a single machine. And the key coordinator is a semi-governmental agency called KISTEP — the Korea Institute of Science and Technology Evaluation and Planning. Luna: And we deliberately don't run ads on this show. If you want to support that choice, the link is buy me a coffee dot com slash fexingo. It's what keeps the deep dives like this one ad-free. Lucas: Yeah. And it's listener-supported, which lets us spend the time finding stories like this one. So — KISTEP. It was established in 1999, but its roots go back to the 1960s when the government first started picking national technology priorities. Today KISTEP evaluates every major government R&D program — about 5 trillion won worth of projects annually — and decides whether they're actually generating usable technology or just publishing papers. Luna: So they're essentially a gatekeeper. They kill projects that aren't translating into commercial productivity. Lucas: Right. And they publish an annual report called the National R&D Performance Evaluation that ranks programs by patent output, technology transfer rates, and follow-on investment. If a research consortium spends three years developing a battery material and no company licenses it, KISTEP flags it. That creates a pretty powerful incentive for researchers to think about applications from day one. Luna: But the chaebol piece is huge too. Samsung alone spent about 22 trillion won on R&D in 2024 — that's roughly 17 billion dollars. And they don't just fund their own labs; they fund entire university departments. The Samsung Advanced Institute of Technology runs joint projects with KAIST, Seoul National University, and POSTECH. Lucas: That ties the academic research directly to the company's product roadmap. A PhD student working on next-generation memory chips isn't publishing abstract physics papers — they're solving problems that Samsung's semiconductor division needs solved within a three-to-five-year horizon. And the results show up in the patent data. South Korea filed over 20,000 Patent Cooperation Treaty applications in 2024, which is roughly 600 per million people. For context, Japan is around 450, the US around 400. Luna: And it's not just quantity. Korean patents tend to be more commercially cited than patents from comparable economies. That's a sign they're actually being used. Lucas: There's a 2023 study from the Korean Development Institute that tracked the citation impact of patents filed by kistep evaluated projects versus non-evaluated ones. The evaluated patents were cited 40 percent more often in subsequent industrial patents. So the gatekeeping works. Luna: But I want to push back a bit. There's a growing concern that the model has diminishing returns. Especially in mature sectors like semiconductors and displays, the cost of each incremental productivity gain is getting steeper. Lucas: That's real. The return on R&D investment in the semiconductor industry has been declining since about 2015. You're spending more to get smaller improvements in transistor density or manufacturing yield. And because Korea's R&D is so concentrated in electronics — about 45 percent of total business R&D — a slowdown there hits the national productivity numbers directly. Luna: There's also a structural risk. The chaebol dominate R&D funding, which means smaller firms and startups get squeezed. In 2024, the top five chaebol accounted for nearly 60 percent of all business R&D spending. That creates a bottleneck in the innovation pipeline — the big players fund what they need, but breakthrough technologies often come from outsiders. Lucas: The government has tried to address that through programs like the Creative Economy Initiative and tech incubators at KAIST and POSTECH. But the data shows that startup R&D as a share of total R&D is still below 10 percent in Korea, compared to about 20 percent in the US. So the system is efficient at scaling existing strengths but less good at generating entirely new sectors. Luna: That's the classic productivity paradox in innovation economics. You can optimize the hell out of a known process, but eventually you hit diminishing returns unless you diversify into new research domains. Lucas: Exactly. And that's where the next chapter of the Korean model will be written. The government has announced plans to increase R&D spending in biotechnology and advanced materials to 35 percent of total public R&D by 2030, up from about 22 percent in 2024. The question is whether the chaebol-university pipeline can pivot as quickly as the policy documents say. Luna: So the lesson for other countries might be: having a coordinator like KISTEP is valuable, but only if the coordinator has the authority to defund underperforming programs and redirect resources. Most countries' evaluation bodies can recommend, but they can't actually pull the plug. Lucas: That's the key institutional detail. KISTEP's evaluations directly inform budget allocation by the Ministry of Science and ICT. If a program gets a low grade, its funding gets cut — not just reduced, but redirected to higher-performing areas. That's rare in government R&D systems globally. Luna: It requires a political culture that accepts technocratic decision-making. Not every country has that. Lucas: No. And it's worth noting that Korea's model emerged from a specific historical context — a military dictatorship in the 1960s and 70s that prioritized industrial policy and suppressed dissent. The coordination mechanisms were built in an environment where the government could direct resources without much pushback. That's not replicable in most democracies. Luna: Still, there are takeaways. Even without the authoritarian history, you can build stronger evaluation frameworks and fund university-industry partnerships that are actually aligned with commercial timelines. Lucas: Yeah. And you can measure what matters. Korea's focus on patent citation, technology transfer rates, and follow-on investment as performance metrics — rather than just paper counts or grant dollars spent — forces a productivity mindset. That's something any R&D organization can adopt. Luna: Alright, so if I'm a policymaker in a mid-sized European economy and I want to borrow one piece of the Korean model, what is it? Lucas: I'd say the evaluation loop with teeth. Create an independent body that can review all public R&D spending, benchmark it against commercial outcomes, and have its recommendations tied to budget decisions. That's the single highest-leverage reform. Luna: And hope you don't need the chaebol structure to make it work. Lucas: Right. Because the chaebol concentration has its own costs — lower startup dynamism, vulnerability to sector-specific shocks. The ideal system might combine Korea's evaluation discipline with a more distributed innovation base. Luna: That's the productivity challenge of the next decade, isn't it? Balance efficiency with resilience. Lucas: Exactly. And Korea's R&D productivity story shows how far coordination and accountability can take you — but also where the limits are.