America’s manufacturing sector isn’t what it used to be. In 1950, it accounted for one in every four jobs in the U.S. Now it is less than one in 10. But Willy Shih, professor of management practice at Harvard University, argues that these jobs are critical to the economy and innovation in technology markets.
Shih was one of a group of Harvard Business School professors who came to Silicon Valley this week to convince the region’s leaders that America’s competitiveness is slipping. Sure the region is creating plenty of jobs in social networking and apps, but that’s not going to save the country, says the self-described “hardcore manufacturing guy.”
Fifteen professors created reports on the subject and found in an October 2011 survey of 10,000 HBS alumni across the economy that 71 percent of respondents expect U.S. competitiveness to decline over the next three years. About 1,700 of the respondents had to personally make decisions about whether to leave the U.S. for jobs overseas. That’s pretty bleak. And the causes are numerous — from education to the U.S. tax code, to the political system, structural deficits, healthcare costs, manufacturing policies, and immigration issues. But Shih believes that getting an understanding of the complex issues is critical to fixing them.
We caught up with him for an exclusive interview at an event at Electronic Arts in Redwood City, Calif. Here’s what Shih has to say about Silicon Valley, competitiveness, and the importance of manufacturing in providing a path for technological innovation and the upward mobility for the middle class.
Willy Shih: What we have done is combined the research of faculty members on a wide range of topics to frame the issue and present it, to really try to change the discussion in this country about competitiveness. Each of the faculty members who’s participating — you’ll hear three of us today — has been working on different aspects of it. For example, I’ve been working on this notion of the industrial commons. The supplier networks, the skills embodied in your employees, the components suppliers, your tool suppliers, the universities. All the infrastructure that is required to support innovation in this country. One of the things that Professor Gary Pisano and I have found is that a lot of companies have moved some of their manufacturing operations offshore, and in some cases, in many cases, outsourcing makes sense.
But in those cases where manufacturing is closely tied to a process development of the technology, or your products are not modular yet, it’s relatively important to keep those things close. We’ve let the commons deteriorate in some of those areas. The country is losing the ability to innovate. A good example would be in the consumer electronics arena. In the 1970s and early 1980s, most of the consumer electronics assembly was moved offshore. That was mostly chasing after labor costs. But one of the consequences of that was, the development of rechargeable batteries then went offshore as well. So even though most of the primary inventions around lithium-ion batteries, for example, took place in the United States, the primary battery makers in the U.S. weren’t really interested, because it’s a kind of complement to consumer electronics devices. Cameras or camcorders or portable music players. It was viewed at the time as an unimportant complement, but what we’ve now found — as portable energy storage, for example, for automobiles is becoming more important — is there’s no commons, if you will, or no technological capability left in the U.S. to support that. I saw two Chevy Volts parked in the parking lot. Half of the manufacturing value-add in those vehicles is rechargeable batteries, and you couldn’t even make those in the U.S. anymore. So that’s one of the areas we’ve been looking at. A number of other faculty have looked at the implications of tax policy, the implications of immigration policy, and so on. We’ve really been looking towards the future in terms of what kinds of things are going to be important for a competitive economy.
VentureBeat: I haven’t had the benefit of a lot of the research here, but I have been following a lot of these issues closely. Partly just by reading some of the New York Times stories. One was about Apple taking its manufacturing jobs to China The other one was about how robot manufacturing could bring jobs and factories back.
Shih: Right, the [John] Markoff story. There are two schools of argument on this. One school says, “Well, this is a normal progression in the economy. The exact same thing happened in agriculture.” And then there’s another school of people who say, “Well, actually, manufacturing and the capability that it embodies is very important for the future of innovation in this country.” What we are really trying to highlight is, in some of those cases, especially where you don’t have mature process technologies, an enormous amount of the value-add is actually in the process development and the manufacturing.
Historically, people tend to think of manufacturing as hand work, not knowledge work. I think modern manufacturing is very much knowledge work. If you go inside factories these days, especially with some of the more sophisticated tools, there is a pretty high bar in terms of requirements for operating some of these technologies and some of these tools. That also says we’re going to see a much higher educational requirement for the workforce. With a lot of this push to bring manufacturing back into the U.S., I’m not sure that people realize what the implications are yet, as far as the skills of the workforce. Traditionally, manufacturing has been a pathway to the middle class for people who’ve had less education. If you had less education, then you got a job in the auto industry, for example. That was a good way to a stable income and an avenue into the middle class. It’s not clear to me that that’s going to continue to be true, because the level of know-how that’s going to be required is constantly increasing.
VentureBeat: I’ve covered the chip industry for about 20 years as well. Intel’s own policies here seem to ring true as far as what you say about process development in their chip factories.
Shih: Intel is a great example. If you look at what they’ve done on 22-nanometer FinFETs, it’s because they preserve the manufacturing capability on shore, close to R&D. A lot of people would argue that they’re one or two years ahead of everybody in terms of that type of vertical FinFET construct. I think it’s because they’ve held on to the process technology. That’s a great example. Since you cover semiconductors, a good of example of this commons. You can ask, why was it that, in Taiwan and Korea and Japan and east Asia in general, they were able to move into flat-panel displays and solar and LED with such facility? It’s because a lot of that capability embodied in the commons, specifically tool manufacturers, industrial engineering, process engineering, being able to do metal organic chemical vapor deposition and things like that. That transferred directly from semiconductors into flat panels. A lot the factory industrialization processes. That transferred directly. Certainly the same thing happened with LEDs. That’s another great example of the commons.
VentureBeat: How do you get above what is very specific to one industry, or one company even, to get ahead of some of the larger problems that exist here? Like, the chip industry is very different from what Apple is doing when they’re assembling things. You almost assume that the assembly just doesn’t ever come back here anymore…
Shih: The assembly, especially the type of work that Apple does, that’s not going to come back. But most of Apple’s hardware products are pretty mature, from a mechanical assembly standpoint. They do use some leading-edge process technologies, but it’s pretty cleanly separable, and they’re very labor-intensive. They’re making a classic substitution of labor for capital. In other words, the best automation you can get, the most flexible automation you can get, is really human beings. That type of work isn’t going to come back. The things that we worry more about are when you look at things like biopharmaceuticals. If you’re making biologics, today there’s still a tremendous amount of know-how in the industrialization. From when you first identify the compound you want to make, and identify a synthetic path, to actually being able to produce it in commercial quantities and manage all the materials flow, including waste, there’s a great deal of innovation associated with that. I think in the last several decades, we’ve discounted that very knowledge-intensive piece of the work. Kevin [Sharer] would know a lot about this.
VentureBeat: To finish that thought, once you lose your supplier base around an industry, it’s all over, and you never get that back?
Shih: Well, you can it get it back. It’s expensive. For example, as the semiconductor industry has migrated to Asia, the tool manufacturers are going there too. As the tool manufacturers go over there, there’s less precision machining over here. Then the tool manufacturers go away. If you lose the precision machining capability, that’s a fairly generic capability. And when we say “capability,” that can combine things like scale, the ability to … I worry about the number of five-axis machines in the U.S., because that’s a critical capability. But as that stuff moves away, then you say, “Well, I have to go spend more time in those regions where the capabilities are.” And of course they’re all competing with us, so they’ll develop the capabilities as well.
VentureBeat: How much subsidy should we grant to someone like Intel to keep that supplier base here? How do we change the mathematical equation for them so that it’s worth it to manufacture here?
Shih: Well, I’m less keen on subsidies, although certainly in the semiconductor business, no major plant gets built without a subsidy. Even Intel’s in the U.S. I think the more important thing to focus on is, are we a competitive location choice for these facilities? Craig Barrett before him, and now Paul Otellini (chief executive of Intel), have both testified in Congress that the incremental cost of operating a leading-edge fab in the United States over the 10-year life of the fab is about a billion dollars. That’s tax and factor costs. Labor isn’t a big part of it. If you look inside a fab, you don’t have a lot of labor. It’s a relatively small part of the total costs. But obviously tax is a cost. There are hidden costs, for example, in how long it takes to get siting approval and how long it takes to build the fab. I was talking to a colleague in Asia. We did a quick calculation, saying, “Okay, I’m going to build a 22-nanometer leading-edge fab. What’s the capital cost of that?” I don’t know, four, five, six billion dollars? Maybe eight? More, if you depreciate that over the normal life of the fab, or whatever depreciation schedule you’re looking at. Then he asked, “How much is that depreciation per day?” It’s a large number. You go over to Asia, they’ll bring on something like [Taiwan Semiconductor Manufacturing Company] Fab 15. These guys work 24 hours a day, seven days a week, and they bring these things on incredibly fast compared to the U.S. And there’s a reason for that. There’s a lot of those other subsidiary things as well. But certainly taxation, factor costs, those are all important. Can I get an educated workforce to work in that facility? And so on.
VentureBeat: I think there are some people who assume that it’s a losing battle to try to keep manufacturing, and it’s just going to continue to dwindle. You sound like you have some hope that we can hang on to a fair amount of it.
Shih: I don’t think so. I think some of the labor-intensive stuff, which is very modular and very mature, we’re not going to be able to hold on to it. I think some of the advanced things… People are beginning to recognize the importance of location choices. One of the purposes of this initiative is to elevate the level of discussion in this country about that. We’ll see some advanced manufacturing come back the United States, I believe, because people are beginning to realize there’s a cost associated with it being far away. If you talk to people in Silicon Valley who manufacture in Asia, they spend half their lives over there. You go down to Cupertino, trust me, those guys are spending a lot of time over there. So there are those benefits. But I think one of the things to highlight is, it’s not going to solve our jobs problem.
Manufacturing as it is in this country only employs nine percent of the work force. Nine percent… You’re not going to solve the unemployment problem simply by doing that. What Gary and I have been talking about is the importance of keeping some of those manufacturing capabilities, because they’re closely tied to preserving your ability to innovate. Those are the things, I think, that are important over the long term. If you think about the number of people in the U.S. who are employed in R&D, it’s around one percent of the population. You don’t necessarily think about… There’s no criticism of that in terms of the ability to drive future job growth. I think there’s a parallel there.
VentureBeat:The New York Times series on Apple sort of made me somewhat pessimistic about hanging on to jobs here in manufacturing, when Apple was saying it would take them nine months to hire 15,000 people here. In China, they could do that so quickly. The robot manufacturing story, though, made it seem more optimistic. Instead of a million workers in China, maybe you would build a bunch of these factories here and bring a lot of that manufacturing back here.
Shih: I think there’s some truth to that, but every time you embody capability in tools… Then it only becomes a matter of capital and factor costs. One of the things we’ve seen is, as you embody know-how, for example, in semiconductor tools… Not to minimize this, but then all you need is money. You can grow your capabilities as well. Then the key question becomes, is the U.S. an attractive location to locate a manufacturing facility using those things? Closeness to the world’s largest market, certainly, is a major factor. An educated workforce, capable suppliers, close to R&D, all those things play into that as well.
But simple automation, I would argue, is not sufficient. You still have to make yourself an attractive location. There’s some parts of the country that have done a very good job of making themselves an attractive location. There’s really a kind of manufacturing boom in the Carolinas and Alabama and Mississippi. There are lots of foreign manufacturers who’ve come in there because it is a competitive location. I think the challenge for us as a country is, how do we make ourselves the most competitive location? You’ll see some data on that in terms of the survey tonight. We’ve surveyed a lot of HBS alumni about location choices. All of those things are things where we’re trying to get the discussion elevated so that, as a country, we talk about them.
VentureBeat: This is kind of an aside from manufacturing, but in the video game industry, Canada has had a lot of success attracting studios.
Shih: The Vancouver area in particular. Location is important, right? In terms of computer graphics, historically there’s also been this cluster around the Toronto area as well. But Vancouver, certainly, has attracted a lot. We have the people who talk about “the world is flat,” everything is tradable and you can fly all over the place. But the key thing you point out, which I think is exactly right, is that location still matters. That’s Michael Porter’s claim to fame, around clusters. I think that’s certainly true.
GamesBeat: Is there another piece of the manufacturing findings that you’re going to be talking about here too? Something we haven’t touched on yet?
Shih: We’re going to talk about location choices… We’ve only brought three of 18 faculty total. 18 projects. It depends on how you count, but we have a whole range. Tonight, I think I’m the manufacturing guy, the hardcore manufacturing guy. [laughs]
VentureBeat: Is there a success story that you can talk about yet?
Shih: Well, I’d say, in terms of good things that I’ve seen in the U.S., we’re seeing aerospace doing very well in the U.S. Airbus is setting up a factory in Alabama. There’s a big cluster in the central part of the U.S. There’s a big turbo machinery cluster that goes to power generation turbines, as well as turbine blades for commercial aircraft engines, in the Carolinas. There are good things there. Obviously, Cambridge, Massachusetts has been a hub for biotechnology, as well as New Jersey and of course the San Francisco bay area and San Diego. There are some of those areas that the U.S. is still very good at. I think the U.S. is still much better at systems than most of the rest of the world. A lot of the offshoring has been in standardized components, exemplified by Apple. That type of systems work — the Apples, the Amazons, the Googles of the world…
VentureBeat: I saw that Google said it was going to manufacture Glass products in the U.S.
Shih: We’ll see. I’m guessing that they would like to. They own Motorola Mobility now. But for those types of things that are mostly assembly labor, that’s very hard. You look at somebody like Apple. They’re going to announce their iPhone 5 or whatever it is. They will have this incredible ramp in terms of how many units they’re going to ship in the first quarter. You wouldn’t be able to do that type of ramp here. If you talk to the Apple guys doing the channel loading for stuff like that… It’s really tough. I’ve been in a mobile phone factory in Asia that ramped from zero to 450,000 units a day over a four-week period. You couldn’t do that in the U.S.
Shih: I think great things happen in Silicon Valley. A huge amount of innovation happens in Silicon Valley. But what I was referring to earlier is the fact that a huge amount of innovation happens on a manufacturing floor, especially for some of those less mature processes. I’ll give you an example tonight. If you’re in the display materials business, one of the questions in my mind is, does it make sense to invest in R&D if your only path to commercialization relies on capabilities that can only be found in Asia? They’re working on it too, and they have this huge advantage of being close to where the work is done. I think it’s really problematic. The question is, as a result of letting go of those things, are there fields that you can no longer invest in that might be general-purpose and generically important? That would be my concern.
VentureBeat: How do you look at something like, say, Solyndra (the solar manufacturing company that went belly up after receiving government loans? And then Tesla, manufacturing cars here [in Fremont, Calif.]
Shih: Solyndra… They’ve had their share of criticism. The debate tends to be government-driven versus private industry-driven. I think the role for the government is in addressing basic R&D. That’s the type of stuff the government is very good at, which private industry often has a lot of difficulty doing. That’s in the definitions of public good, if you will. There’s been a lot of controversy over that on the political circuit, but if you look at what has historically served the U.S. tremendously well, it’s been investment in basic science R&D. Going back to the space program, which led to integrated circuits, computers… Defense and the space program drove a lot of things like that. The Human Genome Project really drove the biotech industry and investments there. I think government investment in basic science and R&D is really important. It’s hard as you get downstream, because then you always get accused of picking winners. It’s really hard to get that right. Governments don’t get that right very often. The truth of the matter is, venture capitalists know that they’re going to get a lot of those wrong as well, and that’s why you have to have a portfolio of these things. It’s really hard to do that. I think sticking with the basic science and the things that support the innovative environments, like education, would be a better choice in my view.
VentureBeat: Tesla seems like it should not exist here. But it somehow made things happen…
Shih: Again, auto plants also get a lot of incentives. It’s very hard for me to judge that. You have to judge the success of these things on a very long cycle. Especially something like Tesla. It’s hard to judge that after a year or two, because you’ll really see the outcome of that 10 years from now. My bias tends to be, where the government is really successful is where they do the basic R&D programs. I’ll give you one example in commercial jet engines. NASA had this program called the Energy Efficient Engine. Basic technology, increasing the combustion efficiency in the hot section of a jet engine. The world leader in large turbofans is General Electric Aviation, and then you have Rolls-Royce, both of whom benefit from government-sponsored research in the basic technology, metallurgy and things like that. But the government leaves the manufacturing and the design of the engine to those companies. I think it’s one of those tricky boundaries.
VentureBeat: Your conclusion tonight, do you know what it’s going to be?
Shih: Simple conclusions… Manufacturing is knowledge work. It’s important to preserving capabilities and preserving our ability to innovate. That’s why I think it’s so important.
CloudBeat 2012 is assembling the biggest names in the cloud’s evolving story to learn about real cases of revolutionary cloud adoption. Unlike other cloud events, customers — the users of cloud technologies — will be front and center. Their discussions with vendors and other experts will give you rare insights into what really works, who’s buying what, and where the industry is going. Register now and save 25 percent! The early-bird discount ends September 14.