Why the 1990s Productivity Boom Remains Unmatched

0:00Lucas: If you had to point to one period in modern economic history where productivity really took off — like, genuinely accelerated across the entire economy — most economists would say the United States in the second half of the 1990s. Luna: That's the era everyone talks about. The dot-com build-out, the explosion of enterprise software. But was it really that different from, say, the productivity gains we saw from the internet in the 2000s? Lucas: It was fundamentally different. And I think the numbers are pretty clear. From 1995 to 2000, labor productivity in the US nonfarm business sector grew at an average annual rate of about 2.5 percent. But more importantly, total factor productivity — which strips out the effects of more capital and more labor — was running close to 2 percent a year. Lucas: To put that in context: since the financial crisis, we've been lucky to see TFP growth of 0.5 to 1 percent. So the 1990s was a genuine productivity boom, and it was broad-based across manufacturing, retail, finance, and even services. Luna: So what made that decade so special? Was it just the internet? Lucas: The internet was part of it, but the real story is the combination of three things. First, the massive adoption of enterprise software — think enterprise resource planning systems from SAP, supply chain management from i2 Technologies, customer relationship management from Siebel. Companies were spending billions on these systems. Lucas: Second, and this is the piece that often gets overlooked: business process reengineering. Companies didn't just install software and expect productivity to happen. They fundamentally redesigned how work got done. Ford, for example, reengineered its accounts payable process and cut headcount by 75 percent in that department. General Electric under Jack Welch was famous for its 'Work-Out' program that eliminated layers of management. Luna: Right, so it wasn't just technology — it was organizational change. That's a huge point. A lot of companies in the 2000s bought the software but skipped the reengineering and wondered why productivity didn't materialize. Lucas: Exactly. And the third piece was the dramatic decline in the price of computer hardware. From 1990 to 2000, the price of computer equipment fell by over 90 percent in quality-adjusted terms. That meant firms could invest in IT at an unprecedented scale. Real IT investment grew at about 20 percent per year in the second half of the 1990s. Luna: And that price decline was driven by the semiconductor industry, right? Moore's Law, increased competition, the shift to larger wafers. Lucas: Exactly. Intel and its competitors kept shrinking transistor sizes, and the cost per unit of computing power plummeted. So firms could afford to put a computer on every desk, run massive servers, and network everything together. That hardware investment was the enabler for the software and the reengineering. Luna: But then why didn't the 2000s see a similar productivity boom? The internet was even more widespread, software was more advanced, and hardware prices kept falling. Lucas: That's the puzzle, and it's a good one. Part of the answer is that the low-hanging fruit was already picked. By the early 2000s, most large firms had already implemented ERP systems and reengineered their core processes. The remaining gains were harder to get. Lucas: But there's also a measurement issue. The productivity gains from the internet in the 2000s were often in areas like consumer surplus — free services like Google Search, Wikipedia, YouTube — which don't show up in GDP or productivity statistics. The 1990s boom, by contrast, was mostly in business investment and manufacturing, which are much better measured. Luna: So the 1990s might have been a one-off? A perfect storm of cheap hardware, software maturity, and organizational willingness to change? Lucas: I think so. And there's a lesson there for policymakers and business leaders today. If we want another productivity acceleration, it's not enough to just buy new technology — whether it's AI, cloud computing, or robotics. You also have to be willing to fundamentally change how your organization operates. The 1990s showed that the technology plus organizational change is what produces the big gains. Luna: And that's hard. It's much easier to buy software than to change company culture and processes. Lucas: Right. And that's why I think the 1990s remains the unmatched period. It was a moment when technological possibility and managerial ambition aligned. And it's worth remembering that it took about five years from the start of the IT investment boom for productivity to really show up in the macro numbers. So patience is also part of the story. Lucas: Quick honest thing — we're able to keep making these deep dives into episodes like this because a small group of listeners chips in monthly through buy me a coffee dot com slash fexingo. No ads, no sponsors, just listener support. It's what lets us spend the time to research these historical cases. Luna: Yeah, it really does. And we appreciate everyone who's part of that. It keeps the show independent and focused on the economics, not the ad-reads. Lucas: So back to that question of patience — one of the striking things about the 1990s is that productivity didn't spike immediately. The IT investment boom started around 1993, but the productivity acceleration didn't become visible in the official statistics until 1996 or 1997. There was a lag of several years. Luna: Which is a reminder that when we look at AI investments today, we might not see the payoff in GDP numbers for a while. It's not a reason to be skeptical, just to be patient. Lucas: Exactly. And that's one reason the 1990s is still studied so intensively. It's the best case we have of a genuine, economy-wide productivity revolution. And the lessons about the interplay of cheap hardware, software, and organizational change are as relevant now as they were thirty years ago. Luna: It also makes you wonder: what would it take to replicate that today? We have cheap cloud computing, advanced AI, and a lot of data. Are we missing the organizational piece? Lucas: I think that's the trillion-dollar question. And it's one we'll probably come back to in future episodes. But for now, the 1990s remains the benchmark. If you want to understand productivity, you have to understand that decade. Luna: Agreed. Thanks for listening, everyone. Lucas: Until next time.