How Did America Build the Arsenal of Democracy? (with Brian Potter)

Brian Potter: Hi, Russ. Thanks for having me.

Brian Potter: So, that’s a rough estimate, but the magnitudes are such that it seems like it very likely must be true. Today, we build for commercial transport–so, like, jet airplanes or whatever–we build on the order of a thousand, 1,500 airplanes a year. So, you just multiply that by the time that aviation has existed as an industry, the numbers are a huge stock[?]. And of course, aviation today is much larger than it was. Commercial aviation is much larger than it was in the 1950s, 1940s, and 1930s, before World War II. So, we just don’t build that many commercial aircraft. The amount the industry had to scale up to produce the wartime airplanes was just incredibly enormous.

Brian Potter: Yeah, it was more complicated than that. And, what’s interesting is that just the flipping the switch was sort of implicitly what the plan for wartime mobilization was–what they called M-Day [Mobilization Day] in a lot of this documentation, the day where we were going to switch over to a wartime economy–

Brian Potter: Yeah, Mobilization Day. And, they had sort of assumed that: ‘Okay, we’ll flip this switch. We will enter mobilization, and all our industry will now switch to sort of producing planes, tanks, and guns, and it will just be easy to just transfer over.’

It of course didn’t work like that, and there was no plan in place for how they could transition this to a wartime economy producing just huge amounts of munitions and airplanes and tanks in unprecedented numbers.

And so, and yes, I kind of break it down into a couple of different phases.

The first important phase of the scale-up was before the United States even entered the war. It was in the late 1930s. And, this was partially before really World War II even started. But, Britain and France were getting very nervous about Germany’s territorial ambitions in the late 1930s, and they suspected that war was coming, that war was on the horizon. And so, they started scaling up their own armed forces in anticipation of that. And so, they placed very, very large orders for airplanes and airplane engines with a lot of U.S. manufacturers–companies like Boeing, companies like Lockheed, companies like Pratt & Whitney. And, that initial injection of funding allowed these companies to start scaling up their operations.

And, even these orders–there was an order from Britain for, like, 250 bombers or fighters or something from Lockheed–which by the later standards of the war was very small, but that was, like, a huge order. Much, much bigger than Lockheed had gotten in its history.

And so, those initial orders allowed these companies to start scaling up their operations. And in some cases, rescued these companies from the brink of bankruptcy. I think Pratt & Whitney was on the cusp of going out of business before they got this huge order of airplane engines from France.

And then, so this was part of the initial scale-up. So, the industry itself had started scaling up its operations then, but the U.S. armed forces weren’t really preparing to build huge numbers of airplanes yet. As late as kind of early 1940, you have armed forces saying, ‘Well, maybe over the next five years we’ll build 5,000 airplanes, 10,000 airplanes,’ something like that. Very, very small numbers compared to what was ultimately built.

But then, when Germany invades Western Europe in later 1940, then all of a sudden this becomes an emergency. It’s like, ‘Oh crap, this is going to be huge. We’re really going to need to build a lot of airplanes, a lot of munitions.’

And so, this is–after this is when Roosevelt goes to Congress, and he says, ‘I want to be able to build 50,000 airplanes a year,’ which is, like, a preposterous number–roughly as many airplanes as the United States had built since the airplane had been invented by that point, including all small, little, tiny planes and things like that.

After that is when, kind of after that is when the U.S. mobilization started to scale up and get into gear.

And then, when Japan attacks Pearl Harbor, Roosevelt goes back to Congress and demands even larger amounts of airplane production, and mobilization kind of scales up further.

So, those are the three stages when the United States began to transform its aircraft industry.

Brian Potter: So they initially hoped–the initial hope in the scale-up–that the industry could scale up its operations itself. They could fund the construction of their new factories based on the new orders that they were getting from the government. And so, the initial phase of this aircraft mobilization was making it easier for the industry to finance its own scale-up.

And so, they did things like change the amortization period for factory construction and change the amount of profit allowed in government contracts to make it easier for these things to build their new facilities.

But, it eventually became clear that the size of the scale-up was so enormous, and also the fact that the methods of production were so different than what had previously been used. Where, previously these aircraft were essentially built by hand in a craft production process–basically the same way that cars were built before the invention of the Model T. So, you had these things assembled in place by hand, not using interchangeable parts. So, every aircraft was, like, a little bit different. But then you had to build the aircraft–when the numbers were necessary, you had to build in something called line production. It’s like an assembly line where you have material flow and a series of sequential operations, and the airplane gradually gets built together piece by piece in that fashion. The facilities that were built to make airplanes in this fashion were not the sort of facilities that were actually all that useful for building airplanes in non-wartime conditions. And so, we’re building all this infrastructure that the industry couldn’t really use in peacetime.

So, for reasons like this, it basically became clear that the government needed to sort of finance the construction of all this infrastructure. So, they ended up building all of it, and the government [?owned it?], and that these contractors would come in and operate it.

And, you see this across the board in wartime mobilization–that most of the facilities, manufacturing plants, and things like that for building the armaments were this government-owned, contractor-operated model.

Brian Potter: Yeah, these really huge, enormous plants, millions of square feet. The factories that they were building were the largest factories in the world at the time when they were building. I have one image in the essay that shows one–I think it’s a Dodge plant outside of Chicago–and it’s roughly the size of the lower part of the island of Manhattan. It’s this truly, truly enormous factory, millions of square feet, just cranking out huge, huge, huge numbers of airplanes. I think that factory or another similar factory actually later became O’Hare International Airport because the facilities were that large.

Brian Potter: Yeah. The output was truly extraordinary. And it ended up resembling–Ford with its Willow Run plant ended up resembling car manufacturing where you have this assembly line: you have this continuous stream of products being produced in these factories.

Getting to that point with aircraft manufacturing was incredibly difficult, because an aircraft is way, way, way more complicated than a car. It’s much larger. It has many, many, many more parts. I think a mid-sized car at that time had something on the order of 5,000 parts. These aircraft had closer to 150,000 parts, plus hundreds of thousands of rivets, plus all the parts in the engine, and all the equipment that they had to attach to the plane. So, it’s much, much more complicated.

And, a car could kind of be built with these fairly heavy, robust pieces of metal that you stitch together. But, an airplane was a very, very high-performance piece of equipment. So, it had to be sort of, A). operating at a much higher level of performance. The horsepower of an aircraft engine was much, much higher than, say, a car engine–a thousand to 2,000 horsepower compared to, say, 70 or 80 horsepower for a car engine–all while having to be much, much lighter proportionally than the car engine. And, it was operating at close to its maximum performance level for almost the whole time that it was flying. Unlike a car engine where it only rarely reached its maximum level of output.

And because it had to be–this thing had to fly through the air. These airplanes had to be as light as possible. So, it’s all stitched together using these, like, really, really thin pieces of metal bent into complex curves, which needed to be produced with a lot of precision–much more precision than in car manufacturing. And, stitched together using these kind of complicated assembly methods.

And so, achieving that level of high output and continuous production with these really complex, precise products–large and complicated products–was very, very difficult for these manufacturers. And it took them a long time to be able to do it.

Brian Potter: Yeah. It was incredibly difficult. As I understand it, what you see is that for these aircraft manufacturers–these companies started as pretty small operations. I think of the five largest aircraft companies that ended up making the lion’s share of U.S. planes during World War II, I think they started the war with something like an average of a thousand or 1,500 employees apiece. Some of them were even smaller. Some of them had only a couple hundred employees. And, these grew to enormous organizations of hundreds of thousands of employees.

And so, what you saw is that the executive team that ran these companies, they tended to be experienced, as group of people who were very experienced, who had a lot of knowledge of how to build these aircraft. But, the actual factory line-workers were just huge numbers of new people, lots of them women. There ended up being almost half a million women working in airframe and aircraft engine assembly plants.

Overall, it was people who had never really worked in aircraft assembly before. There was just another big and difficult transition, where you were moving from a process in which the knowledge was embedded in the factory workers’ heads and the factory foremen and the people actually doing the assembly, to a situation where the knowledge is embedded in the process itself. And you have these things being put together by relatively unskilled workers who did not have a lot of experience building airplanes, or building anything at all, or having traditional manufacturing skills. And so, that was another very big, difficult transition that these companies had to go through to be able to produce these airplanes in huge numbers.

One of the sources that I read said that to a very large extent, the workforce that was available basically chose the production system, because the workers who built these airplanes were basically unskilled. It was lots of women, and they weren’t as physically strong as men. So they basically had to design a production process around the workers that you had available to staff it.

Brian Potter: Yeah. It goes back to what I was saying previously, that these are really high-performance, very precisely machined, precisely built pieces of equipment operating under very, very tight tolerances.

So, you had situations where, if there was even a scratch on, like, a gear, or the wrong piece of equipment or something, that thing might have to be thrown out because that scratch might create a place where stress could concentrate. And these things were operating so close to their limits of material strength that you couldn’t have these sources of, like, stress concentration.

And so, another example of how difficult it was to get these things built was that for something traditional, like a car engine, once the engine was assembled, you would turn it on and run it and do a sort of brief breaking-in period.

With an airplane engine, you would run it for a very long time to make sure that it worked. And then, after it had run, they would have to disassemble the entire engine and inspect every single part to make sure that it hadn’t worn unnecessarily. And if it had worn unnecessarily–if there was too much wear on a single part–they would replace that part, reassemble the engine, run it again, and then keep checking to make sure there was not any undue wear or any sort of problems with any of these engines.

Because they are operating–this was basically at the limit of piston engine technology. And so, you were at the limit of what this technology could do. So, your margin of error was very, very slim. And so, yeah: inspections and quality control became incredibly important to get these things produced in the battlefield in working order.

Brian Potter: Yeah. Another thing that was very, very difficult.

So, in traditional mass production manufacturing, you try at some point to freeze your design. So, you figure out what it is you are going to build, and okay, at some point, we’re not going to make any more changes. This is the thing that we’re going to build until the next major revision, which may be a year, two years, five years, whatever. And, they would just build that thing over and over again.

It didn’t always strictly work like that. You would sometimes make small changes to models if problems revealed themselves, but in general, you’re trying to lock your design and not introduce any more changes.

And they tried to do that with aircraft manufacturing during the war, initially; but it basically proved not feasible because they had to be constantly improving their aircraft. Because the enemy, of course, was constantly fielding better and better airplanes. And so, they had to make sure that they could field competitive equipment and weapons. They had to be improving their own aircraft to keep up.

And so, yeah, they were constantly introducing changes–problems that they’d identified in the field that they had to go and fix–performance enhancements and specialized versions that were adapted for specific missions, and things like that.

So, you had this huge flood of constant changes coming into the factory as you were trying to produce these things in huge numbers. As you were trying to scale up your manufacturing, you were also constantly changing what it was you were actually building.

To do something like that, you need what’s called a very good Production Control System, which is basically a system of information management that tells you what to build, when to build it, and who exactly is going to do what task; and what that task is, and how the material needs to move through your manufacturing operation.

And that was another thing that these companies had to develop over time. Because, when you are building really, really small numbers of airplanes very, very slowly, you don’t need very good production control. Everyone can just keep the knowledge in their heads. But, when you are building hundreds of thousands of these things, and what you are building is constantly changing, you need a much, much more robust system. So, these airplane manufacturers had to invent that along with everything else they were inventing about the production process.

Brian Potter: I’m just very interested in manufacturing and what it takes to produce things and how we get better at producing things over time. That’s kind of a through-line to a lot of the essays and topics that I write about. A very broad summary is how to build things, and how to build things better over time.

And so, this was just a topic that I came across, and I realized I didn’t know that much about it. I wanted to dive in more deeply.

I don’t remember how long, specifically, this one took. Usually it takes me a few days, a week or so, maybe of research, and then a few days to actually write the post. Sometimes, if it’s a very, very involved one, it takes longer. This one I think is about 5,000 words, so it probably didn’t take that much longer than that. So, probably on the order of two to three weeks, I would say–not maybe working solely on that, but roughly that amount of time.

Brian Potter: Yeah, it’s interesting that you mention stress, and that’s actually something that gets mentioned in some of these documents about this effort–is that, this scale-up was so incredibly stressful that it was very difficult for these managers to manage these new, growing manufacturing operations that were so much larger and so much more complicated than what they had managed previously.

And some of them weren’t cut out for the task. It was beyond their capabilities to do. Some of them found it so stressful dealing with this that they had to end up leaving because of health reasons. I think I read one statistic that at one factory–I think this was Lockheed, for one of their aircraft, the P-80–this was developed in record time, but they also mentioned offhand that something like 30% of the designers ended up making themselves sick because they were working so hard to get this plane ready in time. And you just kind of see that across the board: There was a huge willingness to sacrifice and take on what needed to be done to get these things built and to try to win the war.

I only mention him very briefly in the essay, but a figure who was very, very important in U.S. wartime mobilization was this guy, Bill Knudsen. He was an immigrant from Europe. He had started out working in Ford’s factories and had been one of the people who helped work on the initial assembly line systems. And then he became the president of General Motors [GM]. And then he, at the request of President Roosevelt, he basically left his very high-paying job at General Motors–he took a leave of absence–and worked for $1 for the government to basically help them do their wartime mobilization. I think this was in 1940, is when he went and joined.

A lot of the reason that U.S. mobilization wasn’t slower than it was is because of Bill Knudsen, who basically–he knew all the manufacturers and knew what needed to be done to get systems in place and how to build huge volumes of stuff. And so, he set up the government systems that ended up managing this whole mobilization. I think he later left that role before the war was over.

But, you know, he was a Republican; the whole Administration was Democrats. People asked him, ‘Why are you going to do this? Why are you leaving your job at GM to go work for the government and basically make no money?’ And he’s, like, ‘Well, this country has been very good to me, and I need to return the favor.’

That’s just one example of many, many examples up and down the line of people who thought it was important to do these jobs and figure out how to make all this stuff to win the war.

Brian Potter: Yeah. Yeah, it’s–the whole thing is just incredible. And, you often wonder to what extent that the United States could accomplish something like that today if it needed to.

Willow Run, to talk about that a little bit, is interesting because–we talked about this sort of adoption of mass production-style methods that were used in things like car manufacturing and what was needed to sort of modify them to work in the complexities of aircraft manufacturing.

What most manufacturers did–even the ones–most of the airplanes were built by existing airplane manufacturers, but even the ones that didn’t, most of them started by using traditional aircraft manufacturing methods or the craft methods that they’d already been using to build airplanes. And then once they had learned those methods and learned what it took to build an airplane or an airplane engine, only then would they say: ‘Okay, now that we know how to build these things, now that we know what it takes to produce something with this level of precision and performance, now we will start to bring in our knowledge of automobile manufacturing and mass production and adapt our manufacturing processes to these high-volume productions that we need.’

But, Ford was almost the only one that just started to go right into sort of car-style mass production methods–just jump into that immediately without learning basically what it took to build an airplane or an airplane engine first. And a result of that was they struggled for a very, very long time to build anything in appreciable numbers. Yeah: it was going very, very poorly for quite a long time; and the government was getting more and more uncomfortable with how it was going. They eventually figured it out, but it took them quite a long time to get their production operations running along nicely.

Brian Potter: Yeah. You just look at–the arc of it is summarized by the statistics. You feel like you can get a sense of it. But when you’re in it day in and day out, and there’s just one problem after another, it’s, like, ‘Oh, we can’t get enough aluminum. Oh, there’s a huge shortage of airplane gasoline. Oh, we’re bringing in 10,000 more workers to this plant, and none of them have worked in a plant before, and none of them know how to operate a rivet gun or a drill press or any of these tools. And, all our workers who knew how to operate these things are off fighting in Europe,’ or whatever. Just the tidal wave of problems that was washing over everybody constantly–it’s just difficult to imagine how hard it must have been.

Brian Potter: Yeah. Traditionally, riveting had been done as a team process where one person would hold the rivet gun and one person would be on the other side. These were big, heavy tools that needed a lot of strength and skill to work. To allow women or unskilled workers to operate these things, they had to redesign them. They put counterweights on them so they were easier to move around, and they changed it in such a way that it could be done by one person.

So that was another big trend you saw more broadly with these manufacturers. I talked earlier about how the manpower chose the production process. They had to redesign a lot of these tasks, the actual specific things that the workers were doing, such that the workers they had could actually do them. So: make things easier to do, make things so that didn’t require five years of experience with a drill press to do the operation.

So, totally new tools, totally new jigs, fixtures, and support systems made it easier to do all this stuff. Almost every job that needed to be done to put an airplane together was redesigned to make it easier for unskilled, physically less capable workers to do.

Brian Potter: Yeah. Like you said, this production machine had finally ramped up, producing airplanes in enormous numbers–like you talked about earlier–B-24s coming off the line every hour or whatever. And, it was such an enormous flood of airplanes that, yeah, it had reached the point where they kind of didn’t know what to do with them all. It was more than they could find immediate uses for.

But they didn’t want to–they had this big, beautiful production machine that could crank out huge numbers of airplanes, and they didn’t want to turn it off because they were still fighting the war. They didn’t want to lose all this capability that they had to build these huge numbers of airplanes. So, they basically just started finding ways to make use of them.

And one of the things you see is that if an airplane was slightly damaged, they would just, like, ‘All right, get rid of it. We’re going to bring in a new airplane,’ because huge numbers were coming off the lines constantly. An aircraft comes back from battle, lands on an aircraft carrier, it’s slightly damaged: ‘Push it over the side of the carrier. We have new planes coming in all the time, and we can just land a new one and just forget about those damaged ones.’ I think there were islands in the Pacific where they dug huge pits and pushed large numbers of airplanes into them, and just compacted them down with tanks or whatever. Because, again, these airplanes were coming, and they didn’t have any place to put them or anything to do with them.

And, at the very end of the war, you had the situation where, in some cases, apparently airplanes were flying directly from the factory to junkyards, to boneyards, to be disassembled or put into storage. They didn’t even make their way anywhere near resembling a battle. They just immediately went from the factory to be junked. This machine to build these airplanes had taken so long to spin up, and it didn’t take quite as long but it took a while to spin down, as well.

So, for a while, you were producing huge numbers of airplanes that weren’t really needed, at the same time when countries like Germany and Japan were having a very, very, very hard time fielding any aircraft at all. At the same time, the United States was producing such huge numbers that it could basically trash the ones that they don’t need.

Brian Potter: Apparently. This isn’t one specific source that talked about the pushing-it-over-the-side. So, it’s not incredible. It’s not a sort of anecdote that shows up over and over again in these descriptions of the historic manufacturing. But, in general, it tracks to me, just based on the huge numbers of airplanes produced and the time and effort it would have taken to fix them in the field. It doesn’t seem crazy to me.

Brian Potter: Yeah. And, this is a–I think, a larger debate among historians and economic historians is: Yeah what extent? It’s sort of a flavor of, like: the Great Man theory of history versus to what extent are you driven by specific decisions of specific important people in key roles? To what extent is history driven by large-scale economic trends? So: How much of the war was dependent on the United States’ manufacturing output, versus–and the rest of the Allies versus the Axis? And of course, on the eve of the war, the United States was the largest manufacturer in the world. I think it made more cars and more steel than the rest of the world combined, and more machine tools than anybody else.

So, yeah, you look at something like that, and maybe you can argue–or maybe the data show–that it could have bumps along the road, but the 30,000-foot view, it was inevitable the United States would scale up its operations. It was inevitable the output of those munitions, and weapons, and things like that would be so huge that it would overwhelm the Allies. The Axis–excuse me. The economic disparity was so great.

And so, I don’t have a dog in that fight. I’m not well-informed enough about details of all of World War II; and the various production operations and the decisions made in those contexts. So, it’s not necessarily obvious to me whether winning or losing the war turned on some of these lucky breaks from the United States. It could be true; it could not be true. And maybe the war just ends up getting delayed. Winning the war ends up getting delayed. Maybe casualties are much higher than they would have been. Maybe the outcome is slightly different on some margin, rather than the the war itself–the outcome of the war itself being different. So I don’t know that, per se; but I think it could obviously have not gone the way it did, with the United States winning in 1945, basically.

Brian Potter: So, yeah, I have a few scattered thoughts on that. One, I do agree that U.S. manufacturing capabilities have been offshored, hollowed out to a very large degree. And so, we lack a lot of the fundamental capabilities that would be needed to scale up manufacturing quite quickly.

An example that I’m a little bit familiar with is injection molding, which is basically a way of making plastic parts. You have these big metal molds, and they basically get filled with hot plastic into the shape of whatever it is that you’re making. For plastic things, that’s a standard way of making anything in very, very large volumes. The United States does do some injection molding domestically, but the actual making of the molds for those machines–which is a key bottleneck for manufacturing, to actually make the mold itself for whatever part you are producing–is almost entirely done overseas. Most of it, a very, very large fraction of it, is in China. Even companies that do injection molding in the United States will buy the molds or get their molds designed by overseas companies.

So, that’s a capability that has really, really atrophied. It’s not something we could scale up quickly at a moment’s notice. And so, if you need to make large volumes of plastic parts for something like drones, that would be a pretty big challenge.

And you see similar things in other areas of manufacturing. It took the United States a pretty long time to scale up artillery shell production. At one point that I looked–I don’t remember the exact numbers–but the United States was supplying fewer artillery shells to Ukraine than North Korea was supplying to Russia. So, we were being outproduced in artillery shell production by North Korea, which I think is a pretty stark illustration.

Brian Potter: Yeah, it shows how difficult scaling up certain manufacturing operations. I think we do lose these abilities, and it’s not something where you can just say, ‘Oh, we still make a lot of stuff. If we need to, we could scale up our manufacturing capabilities quite quickly.’ 


I think the United States in World War II was drawing on both its huge manufacturing output in other areas and its muscle memory–its ability to build things quickly. In the late 19th century and the early 20th century, the United States was very, very good at building very large amounts of infrastructure very, very quickly. That’s when you had the construction of hundreds of thousands of miles of railroad, going from nothing to having a huge electrical grid, going from nothing to having millions of cars on the road, building things like the Hoover Dam, building things like the Empire State Building, mass urbanization in a lot of cities.

The United States was just very, very good at building things quite quickly. And it’s not something the United States is very good at today. We struggle with building very, very large numbers of things quickly, or very, very large amounts of infrastructure quite quickly. In some ways, that is what worries me more than the actual building of the physical stuff, is that we’ve lost the muscle memory of how to build this stuff quickly if we need to. I think that’s what kind of gives me more pause. [More to come, 1:00:47]