The diffusion of soft technologies during and after WWII
01:29:23 Share

Welcome to the LSE Events Podcast by the London School of Economics and Political Science. Get ready to hear from some of the most influential international figures in the social sciences. So good evening. A very warm welcome to you all to the LSE Department of Economic History annual Epstein Lecture. This year, it's a great pleasure to welcome Professor Michele Giottelli as our Epstein Lecturer.

I'm Patrick Wallace. I'm a professor in the Department of Economic History and its current head of department. And as such, it's always my privilege to chair this lecture. Before we get on to business, just a couple of practicalities.

The event is recorded. It's going out live, but it will also be made available online afterwards. After the lecture, we will have time for questions, and I would encourage you very much to think about questions and be ready to present them.

And this extends to those of you who are joining us online. We do have a system whereby you'll be able to ask questions and they will be fed through to me through our digital, excellent digital infrastructure. So please ask questions and at the end I will invite and voice some of those questions. If you do want to ask a question from online, please include your name and affiliation and where you're coming from so that we can introduce you to the room. So the lecture.

This annual lecture series was established in the memory of Professor Larry Epstein. He was a dearly missed colleague and he was also one of the most exciting economic historians of his generation. Larry was a mentor and friend to many young economic historians. I benefited personally from his support, his criticism and robust engagement with ideas that he disagreed with vigorously. And so when

we had to think about how to remember Larry. This was a very natural way to do so, to invite an outstanding young economic historian who's made some progress in their career, enough to show quite what they're able to do, but to bring them in at a point when we wouldn't necessarily have this kind of forum and space.

So this is about showcasing rising stars in economic history. Recent lecturers have included Mariam Wanameca, Mohammed Saleh, who's somewhere in the room, Marcela Alsa, Mark Kayama, and last year, James Feigenbaum. And this is the 17th lecture in its series.

And our lecturer, for our 17th lecturer, Michaela, is an associate professor at UCLA in the Department of Economics. We're extremely pleased that she's able to join us here today. And we're very pleased because in what really is a very short few years, she's established herself at the forefront of what is one of the hardest questions in economic history. That is, what drives productivity growth?

Explaining productivity in detail is one of the great dreams of our discipline and several others. The level that Michaela's engaged with in her work is, I think, one of the most relevant levels of detail that one could imagine. That's the reason why firms achieve different levels of productivity, right down at the actors of the economy. That this matters is not just a question of historical interest, although, obviously, those of us in the room are happy with those questions.

It's a question that matters enormously to society at large, particularly as we wrestle with slowdowns in productivity in the UK and other parts of the developed world in recent decades. And some of the answers to this puzzle, they surely lie in the hands of some of the people I think Michele will be talking about tonight. These are the managers, the policy makers, the scientists, the politicians that she studied. And it's partly this, it's the importance of her research

of bringing real empirical rigor, bringing real social science to our understanding of what this group of critical actors do in the modern economy and what they do to shift productivity, or maybe sometimes not shift it, that I think is really so exciting in her work.

So today she's going to talk about some of these issues. There's a long list of things if you really want to read all that she's done. Some of these issues in the context of the Second World War in her lecture on the diffusion of soft technologies during and after World War II. So please join me in welcoming Michaela to the stage. Hi everyone.

I am very excited to be here today. I would like to thank you, Patrick, for the very generous introduction and all of you being here. So before getting started about talking about productivity, why productivity is important and we care about productivity so much, I would also join Patrick in sharing a few words about the

Larry. I didn't have the pleasure of meeting him in person, but of course I did know his words very well. So I don't want to go through all the numerous contributions that he did and made to the economic history discipline, but I would just...

I'd just like to underscore the legacy that these words have for the field and of course also for today's lecture. In particular the fact

that knowing and understanding the past is essential for understanding the present. How different disciplines are complementary, history, economics, sociology all play together in our studies and I will touch during the lecture on some of these aspects and

and why they combine together in determining productivity and soft technologies as the title of the talk. But I think that the part of his research that is more related to what we're going to talk about today is the importance of technology in shaping human capital. In particular today we're talking about how soft technologies could shape human capital and sustain productivity growth.

So the productivity slowdown is a hot topic in almost all developed countries, as Patrick said. So here we can look at labor productivity in Britain and

and see how the rates slowed down since 2007. So productivity grew substantially between 2000 and 2007. There was an increasing productivity during the 2008 crisis, but then productivity, even if it started growing again, did so at a pace that is substantially lower than the first few years of the 2000.

If you are familiar with the Draghi reports of the European Union released this past September, a lack of productivity growth, like an increasing productivity gap between Europe and the US, has seen one of the most urgent issues to face in order for European firms to gain and remain competitive.

But the natural question is why do we care about productivity so much? And the answer is that it is arguably the most important determinant of the living standard. So in terms of productivity, we can see productivity in two different ways.

First one is just how much output could be produced given a set of inputs. So if we think about firms or countries that can produce a higher amount of output using the same amount of input, one country, one firm is more productive than the other. This is the definition that I will mostly use during the talk today.

But we can also think more simple in terms of labor productivity, either the output of the revenues that workers are able to produce. So once again, more productive workers can produce a higher level of output.

differences in productivity according to both the definition could explain up to 30% of differences in GDP per capita across countries so an important determinant of living standard wealth of nations goes back to productivity

But on top of this comparison across different countries, it has been documented that there are large and persistent productivity spread even within countries.

within the narrowly defined sectors. So for instance in the US even if we consider their sectors producing their homogeneous goods like concrete, the most productive firm is twice as more productive as the least productive firms. Meaning that the most productive firms could produce twice as much as output as the least productive firms.

This is for the US, if we take other countries, for instance India and China, productivity spread seems even larger in the ratio of 5 to 1.

but also productivity is very persistent over time, like 80% of current level of productivity of a firm is determined by its productivity in the past years. So it looks like that some businesses are able to understand in which way they could be productive and maintain it while some other businesses systematically fail to do so.

So the next important question is what determines productivity? Is productivity so important for living standards? What are the determinants?

And there are two major determinants of productivity that past literature has brought up. The first one, the more traditional one, is differences in hard technological innovations, as this that is embedded in patent information communication technologies, and so on.

But more recently, also other technologies and soft technologies have been studied and systematically measured. We're talking about management and management is the focus of today's talk. In particular, we will discuss management and productivity growth in the context of World War II.

What is management? This is probably the hardest question, I wish I had an answer, it's probably the hardest question that we can ask in the sense that we can think management as a set of practices, as the management system that the firms put in place. Recent work has

have tried to establish a set of good managerial practices and systematically measured them across firms and across countries. But management is, of course, also the people that put in place these practices, CEOs, managerial talents that firms could acquire, and in the constructed form, the so-called managerial capital, like the human capital that is specific to the implementation of managerial

material practices. How can we incorporate management in a production function? So we can think that management could affect productivity through three main channels. The first one is technology. We can think that management acts as a technology in the sense that it increases the productivity of other inputs, mostly labor and capital. Said in another way,

We can also think that as long as the level of management remains low, even increasing labor and capital does not increase productivity that much. And in essence, it could explain why some countries remain poor, some firms remain little productive, no matter how much inputs they employ in the production process.

Then the second channel is efficiency. Management, managers could better use the inputs and the resources of firms. So we can think about a better allocation of resources within the firm.

And finally, management could affect human capital directly in a bunch of different ways, mostly through the allocation of workers to the most appropriate tasks on the one hand, but also in terms of acquisition of the best workers available on the market or by the retention of the workers that are more important for determining firm performance.

This goes back to the fact that replacing key employees is particularly challenging and good managers are able to avoid that these talents are lost by the firm. What we're going to talk about today is an historical period of huge productivity growth, that is, World War II.

And WUWU2 has been defined as an extraordinary surge of growth, mostly by Chandler and scholars since 1977, in references to the technological advancement, the development of new products that occurred between 1940 and 1945.

In particular, the US government assigned more than 200,000 contracts for research and development to academia, private industries, with the goal of increasing innovation. And this was reflected in a number of new products, some used directly after the war, some that had huge spillovers in terms of everyone's life,

Just a simple example, Internet was first developed in the context of World War II and now is affecting our life deeply. What I will argue in today's talk is that World War II was also a major turning point for US businesses.

in the sense that the implementation of key managerial practices increased the productivity of firms in the short term, but was also able to put them on a higher growth path that lasted at least a decade. A set of good managerial practices that after the war were exported in Western Europe and Japan contributing to creating the Toyota way of management.

In doing so, my goal is also to bring together two distinctive literatures. The first one, that kind of scored the importance of managers and managerial talent in shaping US large enterprises. But also more recent literature, pioneered notably by John Barrinen and Inko Blum,

the implementation of good managerial practices can have large effects on firm productivity and performance.

But the bridge between these two literatures relies on thinking that there was no quantitative assessment of the importance of the implementation of managerial practices in the context of World War II, which was also a turning point for the importance of management, at least in the Western world.

So just to give you a brief overview of what we are going to talk about today, we will start with a brief history of development of management. We will talk about World War II, the development of the training within the industry program in the US, how the principles of the training within the industry were exported after World War II, and then some conclusive remarks.

So to get started, the concept of management as we know today goes back to the industrial revolution. Before the industrial revolution, the scale of production was very limited. Businesses were averaging between one to four employees, so there was no real need of coordinating production. The owner could observe what his employees were doing.

The situation started gradually changing since 1840s, mostly in the US, due to the rise of big businesses that needed managerial hierarchy that could coordinate the effort of different plants maybe located in different areas of the countries. Huge infrastructure constructions, notably the railroads and the telegraphs, rely mostly on management skills in coordinating the effort.

Between the first and the second industrial revolution there was also a change where the leading figures in firms. We can think that the first industrial revolution was mostly led by entrepreneurs who were able to think outside the box, having new ideas, developing new projects that lead to the great extent of innovation during this period.

The second industrial revolution saw a prominent feature of managers who think within the box by solving coordination problems and allocating resources within the firm.

that management, managerial theories, managerial organizations are becoming more and more important. We are towards the end of the 19th century, beginning of the 20th century. The most famous one are probably the scientific management theory that were mostly focused on increasing labor productivity.

And in particular, there were a lot of systematic studies that also were across fields: sociology, psychology, really start analyzing management and how, for instance, workers were doing a lot of unnecessary movements, particularly in the car industry. So while I said like, car,

And the reduction of these movements, for instance, for the factory allowed a drop in the production time from 12 hours to 2.5 hours, de facto allowing the mass production of cars and the start of mass production. The moment in which the demand for goods largely increased was

and in particular in shipbuilding, the demand increased by 10 times in 1917 relative to the pre-war period. And so US firms were in need of increasing production, productivity, and trained workers. And they created a four-step method of learning by doing to do in some.

The results were pretty amazing: 500,000 workers were trained and a new vessel was launched every five days between 2017 and 2018. However, this productivity increase remained mostly confined in shipbuilding and some of these techniques were completely abandoned with the end of World War I.

During the 20s and the 30s, however, the idea of productivity and in particular what was defined at the time optimal productivity remained in the management debate quite a bit. So compared to the early theories of Taylor, for instance,

Now the idea is no longer to just increase labor productivity but increase productivity more in general so we can think that this concept is closer to the total factor productivity we discussed today.

and there were even attempts to run quote-quote experiments. The most famous one are the Ultron experiments running different ways between 1924 and 1933 whose goal was to test what could increase firm productivity for instance improvement in working condition or try to

try different combination of teamwork management in order to see in which conditions workers could perform better. These experiments were quite innovative for the time. There was an attempt to change conditions in some plants, but not in some others. So having, using modern terms like treatment or control group, and so have a systematic data collection, including interviews.

So these studies produced an incredible amount of work of material whose main results were mixed in the sense that on the one hand these studies confirmed that matter management increased workers' productivity on the one hand but on the other hand the tools of psychological aspects are very important determinants of workers' productivity. The so-called Ultron effects

stemming from this experience, the fact that even in plans in which conditions were not changed, workers report better working conditions just because they were taught that something would have changed and probably put more effort in light of this perceived change that didn't really happen.

There were also some major shortcomings in the data collection that cast doubt about the results in the sense that for instance worker productivity was measured pre-experiments on Friday and post-experiment on Monday. So it's hard to know if the increased productivity was the effect of better managerial practices or the effect of weekends and some rest.

With the ultra-earned experiments and some other studies in the turges, World War II was about to start. And the results that I'm going to present in the next few slides are mostly based on a paper co-authored with Nicola Bianchi back in 2022.

And what happened during World War II was that there was a surge in the demand for US products as soon as 1939, even before the US officially entered World War II, that happened a couple of years later in December 1941.

So these orders were largely exceeding the US productive capacity. Moreover, since the US were thinking about entering the war at some point, men in working age were expected to serve, with the result of making the productivity issue even more salient.

So in an effort to increase production and productivity, the US government decided to sponsor the training within the industry program.

that was a management training program for US war contractors, so for firms working in the defense industry, that ran from August 1940 to September 1945 and was one of the first emergency programs after the fall of France.

So the program had very ambitious goals. The underlying purpose was to assist all firms in the defense industry to meet their main power needs by training within the industry each worker. This is something that of course didn't happen.

World War II was a moment in which the entire US economy was working for the production effort, for the war effort, resources were scarce and there was still not a complete agreement between the government and the Labor Division, the Manpower Commission, that management was really in

So the question was whether management training should be implemented in the moment in which production had to increase as much as possible. At the end, some agreements were reached and the training within the industry program was launched.

So the training within the industry program was based on three sets of practices that we can see very related to what we think is good management today. The first practice was job instructor that had the goal of establishing standard procedure for operation, improving of working conditions, for instance, lightning, job safety,

regular maintenance of the machines, recording of reasons to form breakdown in order to create a firm specific memory. Every time there was an issue, a solution had to be found but if all the issues were recorded and the solutions as well, it could have been easier to fix them since most of the issues could be quite repetitive.

The second model was focused on job relations. We can think about human resources management in today's terms that aim at improving the relationship with the workforce, training workers and also implement performance-based incentives.

Why was that important? Well, of course, in the context of the war, avoiding any type of production disruption due to workforce strike or injuries or general mismatch between the workers' skills and the jobs that we're assigned were particularly important. And finally, job methods.

that had the goal of simplifying and improving the methods for doing jobs. So having

like using inventory, production planning, prioritization of orders, that were all aspects particularly important in the context of the war to not having a production disruption but also to make sure that the most important orders were the first to be. Just to give you an idea of how the program worked, so the

contractors that were eligible to participate in it were roughly 23,000. It happened in business training program not all of them wanted to do so. So most programs during World War II or even after World War II were implemented on a voluntary basis.

More or less half of our contractors applied for the program, the other half did not. We can see that there is no evidence of differences in the geographical distribution between the archetype of firms. Of course most of firms at the time were located in the north-east and

Midwest part of the country, some in the South in an effort to industrialize the South the US government undertook during World War II and some were on the West Coast. Just to give you a sense of how these applicant firms were,

Because I think it's an important question is to understand whether these firms really need the managerial training or maybe they were already very productive and so the training was not needed.

So these plants were huge. They were among the largest firms of the country. They have an average of six plants. They were employing 1,000 workers. Most of them were in the manufacturing sector, but as a more effort, of course, involved in our economy, also transportation services and agricultural firms were able to participate in this program.

In terms of representation of LUS firms in terms of employment, these firms accounted for roughly 40% of the US labor force working in manufacturing in 1940. So we can think that they were a good representation of the US labor market.

Despite that, however, these firms do not appear to be very productive. If we look at the measure of productivity, we are below two, which is a pretty low number. Of course, it's always very hard to make comparison about productivity estimates across different countries of different historical periods.

So we can also look at other indicators, for instance, return on assets that were only 3%. Inventory was massive. It accounted for 80% of current sales. Importantly, these firms were facing major destruction related to World War II. On average, more than 20% of the workforce was drafted at some point between 1940 and 1945.

and half of them had to switch production in order to meet the work. So even if these firms were large and higher sales and higher profits, they received a substantial margin of improvements thanks to implementation of new practices.

So the idea of the implementation of the training within the industry program at the firm level was to have a highly standardized in-plant training that was 10 hours in five sessions in a very simple way. And most of the teaching methods, so the learning by doing, goes back to the training program in shipbuilding that was implemented during World War I.

The firms which receive

this management training and we're interested in order to assess the success of the program, we're particularly excited about it. Manufacturing companies were reporting a 50% decrease in defective product after the implementation of systematic factory operation. Steel companies reported a 50% drop in complaint cases after better relationship with their labor force.

textile warehouse reported 48% decrease in man-hour to rock operation after the dome method's implementation. So overall it looks like that at least at the time the program was perceived as a large success.

The major problem that the US government had to face, on top that not all the firms want to participate, was that there was lack of trained personnel to deliver the teaching models. So the instructors in the training within the industry program were prepared through 50-hour class by the TWI Institute

They were hired either full-time or part-time. Some of them were industry men that wanted to preserve and continue their jobs. And there were too few in comparison to the applicant firms. So they specialized only in 1J models and they were not employed for the entire duration of the period.

So these are two important consequences in terms of the delivery of the program. First one is that there were too few instructors, so among all applicants, only roughly half of them eventually received training. And there was also a difference in the combination of training that firms received.

The very lucky one that we see on the left, we see training in all three models. There are 7% of firms. Most receive a combination of two models, while some only receive training in one model.

And I think this is an interesting feature because it allows to study not only whether management improves from performance, but also if there are differences in the managerial practices that affect firm outcomes.

Moving to some regressions and I'll try to be less technical as possible here. What we're doing is estimating the effect of participating into this program on firm outcomes. The outcomes of firm high in time period T. The three main outcomes that we look are sales, productivity, return on assets. Then we also have some major managerial practices that we can use

And we look at indicators for whether firms participate, you don't know, in the program and the years before and after the implementation of the program. We collect the data, the firm level that allows us to follow all firms from five years before the program to 10 years after the program.

Importantly, this analysis is performed on firms that applied to the program and get it or didn't get it based on the scarcity of instructors. It's not a comparison between firms that apply and firms that did not apply to the program, which of course may have been different in terms of many dimensions.

their characteristics that we can observe but also an observable characteristics like for instance the desire of improvement or the desire of success of the own managers. So the results indicate that the effect of attending the training within the industry program were quite big.

If we look at the five years before the implementation of the program, we can see that there is no differences between firms who eventually got it and firms which eventually didn't get it, which is reassuring. These firms did not appear to be different to begin with. So there was no selection into the program based on the performance of the firms.

And then there was a systematic increase in sales, pretty sizable, 5% increase within a year, 22% increase in eight years, and the cumulative effect of 16% increase after 10 years. Here I put two red lines that correspond to the moment in which firms were trend, which is year zero,

Since the training happened between 1940 and 1945, but most firms were trained between 1941 and 1944, we can think that three to four years after year zero is when World War II was finishing. And so we can see that this increasing sales is definitely not driven by the work per se, but continues also in peace time.

Look at productivity, we can see that the results are even larger than sales. Again, there is not that much going on in the pre-period. The sales increased by, sorry, productivity increased by 5% in a year and reached 36% cumulative effect in 10 years.

Now you may wonder why sales increase not as much as productivity. Usually if you think that sales are a good proxy for output of the firm.

and productivity is driven by increasing output not increasing inputs, the two should move more or less at the same pace. What we find in this case is that there was a substantial reduction in the usage of raw materials by firms.

who attended the program. So that's why privativity is increasing even more than sales because firms could produce using fewer, could produce both more output and they were using fewer.

And finally, we can look at return on assets as measured by the measure of profitability of the firms. Once again, there is not much going on the years before the program. And then we had a 10% increase in profitability over 10 years by firms that received the training within the industry. The next question is trying to understand a little bit better why there was these

extraordinary growth in productivity. Is it just related to the implementation of specific practices? Can we think that the government helped this firm more just because they were trained, which by the way would have made a lot of sense. The question in this, not just in this paper, but in this literature is about management versus managers.

Was management to improve, like a better implementation of managerial practices or just managers got better? And finally the supply chain. So were there spillovers on the supply chain of these firms that could have made more production?

So let's start with implementation of managerial practices. We find that the short answer is yes. There was a substantial increase in the implementation of managerial practices. For a few of them selected the sample of practices, we can follow them over time.

And we can see that the effects were quite persistent and importantly related to the training that firms received. For instance, we have a systematic reduction in worker injuries for firms that received the factory operation training that continued up to the end of the sample. For the other two groups of firms, not much was going on.

Once firms improved their relationship with the workforce, there was a higher probability of training the workers on the job. Of course, this makes a lot of sense during the war years,

But we can see that the training continued, even increased after the war. So it looks like the firms realized the importance of work training and the sharp reduction in the number of strikes again during the war, but the effects persisted quite a bit after the war itself. And finally, if you think about the production planning, we have a substantial reduction in

the inventory of the firms and so one of what we saw that inventory was 80 percent of access of assets uh so was huge uh before uh the worst reduction in inventory definitely pretty almost quite important even uh after the end of the government

granted more contracts to these firms after they received the training, it looks like it was not the case. The value of the contract, the numbers of the contract, even the refunds that firms got after the war that we can see as a proxy for

the government is giving to these firms to resume their normal operations after the disruption caused by the war are not differential if anything they're negative between firms that received or not received the training within the industry program. So probably the effort in terms of management was not exploited well,

by the government. I think that at some point there was lack of coordination, good management in a way between the different units who was allocating the contract, who was managing the program. And so at the end, we do not find evidence of better relation with the government. So this is definitely not driving our results. We can also think about the effects of management versus managers.

This is something that we cannot observe perfectly in the sense that we don't know exactly who got the training within the firm. What we can observe is what happened when after the program managers left the firm, we have three terms based on the share of managers that were working there during the war and left afterwards

afterwards and we can see that no matter what the effects on sales and productivity remain pretty large, persistent and also they were following kind of a similar pattern. So overall it seems to suggest that it's not the role of specific managers, so just in a few firms that is driving all our results.

Finally, we can look at the effect in the supply chain and we can find that indeed almost all firms that were related or is vertically related to firms that received the training increased their productivity across a different type of training that the major firms receive which is also associated to some implementation of managerial

So we can think about a knowledge spillover channel within the supply chain that is going on here. So to sum up, were the results driven by the implementation of managerial practices? It looks like it was the case. The relation with the government, not much. Was more management or managers? I think it was more management, even if of course

higher managers left determined like lower improvement in firm performance and also increase in productivity of firms in the supply chain that could have in turn increased even more the productivity of the firms themselves.

LSE IQ asks social scientists and other experts to answer one intelligent question. Like, why do people believe in conspiracy theories? Or, can we afford the super-rich? Come check us out. Just search for LSE IQ wherever you get your podcasts. Now, back to the event. Another thing that is worth mentioning is

is that why we take as spillovers, so we're positive and quite sizable, or sometimes spillovers, we're not. If we look at firms that we're nearby,

that apply to the program, but they decide not to apply, so these are firms that were eligible to apply but decide not to do so, we can see that there is almost a zero effect. And the zero effect is persistent both if we look at firms in the same sector or in different sectors.

Why this distinction is important because we can think that firms in the same sector could be either exposed to competition effect, trading firms are doing much better, they could be somehow harmed. And we can also think that the sharing of knowledge between traded firms and potential competitors was somehow limited or maybe even avoid. But even if we look

at the firms in different sectors. So firms in different sectors should not be affected by competition, but they will also not be affected by being close to firms that were doing so well in the market.

Which leads to another question, which is why the firms that decided not to apply for the training within the industry did not adopt these principles even after World War II when the other firms had proven to be so successful. And these principles were actually exported mostly in Western Europe and Japan.

So I think that we can have a different explanation for this. One for sure is that the US economy had a large economic expansion in the 50s and so probably the training was not considered that important anymore and that probably the fact that the government could offer these services even after the war could seem to some extent as a limitation of their autonomy.

So the program was eventually dismissed in 1945. Even if you're thinking about the cost-benefit analysis, the higher profit that firms made would just be equivalent to the cost of 21 months of training. So in terms of the cost-benefit analysis, the program was definitely a big success.

Now, the US government, in light of the lack of interest for the practices in the US, decided to export them after World War II. Most of the results I'm going to present are based on papers that were published in 2019

And these, at least for Western Europe, these exports of managerial practices was part of the Marshall Plan, a plan of economic and financial aid that the US implemented between 1848 and 1852 to help firms recover afterward.

The aid of the Marshall Plan were overall successful. The GDP, Industrial Indices, reached the 1938 level in a great majority of Western Europe by 1952 or a few years after that.

However, the investor expert noted how there was an increasing productivity gap between US and European firms. Actually, this gap was not caused by more than 2%. The gap was increasing since the 30s, probably mostly related to the changes in managerial practices that we already described in the 20s and the 30s, but of course the war didn't help to close the gap.

Silverman was one of the prominent experts visited Europe after the war, actually argued that the lack of management in European firms was a more severe problem than the damages caused by the war itself. It's probably an extreme statement but I think it speaks to the importance and the attention that the US gave to managerial practices during the war.

So in order to close this productivity gap, the US decided to implement another program, the productivity program between 1852 and 1858 that involved the transfer of US management to European firms. In particular, European managers were sent to the US

for a period of 3-4 months in order to work and learn from European managers and the European firms could also buy technological advanced machines that were built in the US and had higher technology than those available in Europe.

We so far studied the effect of the productivity program in the context of Italy. Just to give you a sense of how the program worked, the kind of comparison that we're making here, the firms that were eligible to participate in the program were small and medium-sized firms in the manufacturing sector, roughly 6,000 firms coming from a few regions.

And the idea of this phase of the program was to test its effectiveness before initial implementation. So the regions were chosen to be representative of different areas of Italy, Northwest, Northeast, Central, South, and then Island.

Okay, so these were the firms that were eligible to participate in the program. Also in this case, we find that not all of them applied. Like 60% applied, 40% did not apply. So probably not all the firms believed that management was important or did have this desire of improvement. And among the applicants, they could decide whether they want to send their managers in the US, purchase the new machines, or both.

Rejection was almost non-existent only during the applications were rejected the firms were either in very bad shape or that made huge mistakes in the applications.

However, the US at some point ran out of budget and so they moved from the regional level implementation of the program to a smaller province level. In particular, they selected one province for each of the micro areas that were initially eligible to participate in the program and only firms in these areas eventually received the training that they applied for.

So I think this is very interesting because we could observe which training firms wanted to receive and which training they eventually received.

Okay, so the empirical specification is very similar to what we saw before. We're looking at outcomes of a given plant in a given year and we request it on an indicator of whether you receive a particular type of treatment considering the comparison group of firms that apply for the same treatment but did not receive it.

So first of all, I would like to look for a second at the survival probability that these firms have. If we want to see what is the long run effect of management, we want to follow these firms over time, it's important to understand how many firms survive. We are talking about small firms, not firms as large as the US, and we all know that small firms have a pretty high failure rate.

relative to larger firms. So if we look at the failure rate in the five years after the implementation of the program, it is quite comparable. 97% for firms that were treated, 94% for firms that were not treated is not a huge difference. But if we look at the effects after 15 years, 90% of firms that got the training survived versus 68% of firms that didn't get the training. And the thing

And I think this is a quite striking result that suggests why it's important and why economic history could be used to study long-term outcomes. Very often when we do policy evaluation, we just focus on the short term, but sometimes the results could take time to materialize, as shown in this case.

So if you look at the effects on the implementation of such practices, we can see also in this case in the five years before the program there was not that much going on. Firms that eventually received the training do not appear to be better than firms that did not.

We have a 15% increase in productivity one year after the intervention and the cumulative effect of almost 50% after 15 years. So these are for sure very large results. And in order to see whether they were comparable to some recent findings on

on management implementation in modern setting, you're looking at some other works. So the baseline is the paper that I presented before, large firms and small firms. So we can see that year and year, for instance, the effects are quite in line. So it looks like that would be between the 40s and the 50s. The good implementation of managerial practices was a kind of

comparable between small Italian firms and large US firms. If we look at short term results of randomized controlled trial about consulting services in India, we can see that the results are even larger. So the increasing performance of Italian or US firms was quite

substantial but definitely we can think about even larger margin of improvement of firms today when these managerial practices programs are implemented.

Now, we also saw that another component of the program was the technology transfer. So firms could purchase machineries from the US. If we do so, we can see that the effects are much smaller in magnitude and not that persistent.

not that much going on up to three years after the implementation of the program and then only 10 percent increase in productivity after 15 years which is way lower than the 50 percent increase that we documented for firms that received the management intervention and we can also look at that complementarity for firms that apply for both uh for both interventions uh um here there are a lot of numbers but the um

takeaway is that there was a 14% additional increase in productivity for firms that got both training relative to the sum of the productivity increase of firms that received the two programs separately. So overall there was more 60% productivity increase after 15 years.

So the program was particularly successful in Italy and even if I didn't collect the data, the reports at the time indicate that all Western Europe was particularly happy about the transfer of these managerial principles from the US. But the US also transferred

these managerial technologies to Japan. And the Japanese case is particularly interesting because Japanese firms were able to merge together two components of the US managerial practices, the training within the industry principles, but also quality control and statistical methods that were also introduced during the war. They were not part of this training, but they were quite important during the war.

combined with local means, in particular learning by doing as a way to solve the language, the difference in languages like the communication barrier between US experts that visit Japan and Japanese managers that have to implement these practices.

But also the idea of production just in time solved the space constraint problem that most Japanese firms had. So we saw that one of the key principles of training in the industry was producing, prioritizing the delivery of the orders and also

try to reduce the inventory as much as possible that we can see somehow related to the development of the Toyota management system that in fact became prevalent in the 60s and mostly in the 70s. So at some point the Japanese firms overtook US companies, they became more productive, US market was flooded by Japanese cars in the 80s and at that point

At that point, the US organized training trips for Japanese managers to go to the US, teach the US managers what they learned based on the Toyota management system. So in a way, the US explored the practices that they have themselves developed during the war in the 80s.

But I think that this is an important example on how productivity can vary substantially across countries and across time. And even a big push in increasing productivity may not have long lasting effects if it is completely abandoned, while other countries, for instance, the Japanese example, could catch up quite quickly. So I think I'm out of time.

but I'm also ready for the conclusions. So I think that what this lecture

discuss was how World War II dramatically changed firm management that had a long lasting effect on production and productivity of firms that were trained in the work context, complementarities with technologies, but also these results showed that managerial practices could be adapted and adopted

adopted and worked in different contexts and in firms of different sizes. What can we learn by the experience of World War II? So first of all, World War II, TWA principles are still taught today. And if you think that production had dramatically changed, production methods have dramatically changed between the war and today, managerial principle, what we think is good management today, has not changed that much.

So I think that the main policy implications are that managerial practices could work in different contexts. So we can definitely learn from the past in order to try to understand how the diffusion of these practices could affect for production productivity in a modern setting. But I think it also speaks about the importance of evaluating the long-term impact of

of such policies as we saw, most of the effect took a few years to get at scale and

persisted in the long run. So sometimes when some policies are considered ineffective, it could also be because they are evaluated in a short time period. But I think that it also speaks about how economic history could be important and useful to understand the presence and its implication for the

So, I'm going to stop here. Thank you so much for your attention. It was a great pleasure and honor to be here. And we're then happy to take any questions or comments or suggestions that you may have. Thank you so much. Thank you very much indeed. We do indeed have time for questions. I find it tremendously reassuring as a talker, someone who's spent four years as an academic manager. I worry slightly I may have done the wrong course, but, you know,

We can find that out. Can I just open a little bit? I'm going to--there's time obviously for people to put their hands up in the room and if you are joining us online, please enter questions. They will be fed through to my trusty iPad. But let me just ask one question that struck me listening to you, which is these were incredibly impactful training, yet they didn't spread. You know, and all these other companies, they kind of--they were

millions of dollars left on the sidewalk, so to speak. And you mentioned kind of, you know, maybe they felt they were growing fast enough, but

Is there anything else going on? I mean, why are companies missing out on what's such an amazing opportunity? Yeah, I totally agree. And I think that this is the big open question in the management literature in the sense that, yes, spillovers affected the time, we're overall limited. But still today, management spillovers remain pretty limited. I think it's

it's hard to answer your question. My understanding could be that either firms don't realize how important management is until they are forced to adopt these practices. Some of the practices that we discussed or they're still taught today are pretty simple, pretty basic, right? So, oh, we should record the reason for breakdowns for the machines. It's not something particularly complicated, but kind of

huge effects on productivity. So this could be something that firms may not realize until they see that it actually works. The other problem could be that firms could observe the competitors becoming more successful but they could not understand why this is the case. Maybe they think that it's something else rather than implementation of managerial practices and

For instance, in the worst setting they could have thought that because these firms were chosen to go the training, they were also doing better due to connection with the government or whatever.

And the other thing that I think applies mostly to small and medium-sized firms is that not all firms have desire to grow and improve. Like the implementation of managerial practices requires some changes in the production. They could even become structural changes if firms become more productive at some point today.

could or should expand their volume of production and not all firms have the desire to do so. And this is again a case that is observed in the context of developing countries. That's really interesting. So some of us want to leave the money on the sidewalk in some sense. I've got a bunch of questions in the room. There's a few coming in. So I'm going to just work through the list a little bit. Let me start with Chris because yours was the first hand that I saw. And I will work on from there.

Thank you for a great talk. I'm looking at the poster behind you of Rosie the Riveter and want to ask about feminization.

There's going to be a big temporary influx of women through the labor market. That's going to have management implications. I can imagine if you're J, human resources policy, on the job training, you might think very differently. You're expecting a woman also to leave at the end of the war. So how does that match on with your outcomes? Are these very feminized sectors that take the training? Do you know something about that?

So, I have to say that there was a substantial heterogeneity in the sense that some of these firms could maintain their workforce or they could recover their workforce after the war.

Others could not. In general, we look at different dimension of heterogeneity in the papers in terms of the number of, like the share of managers who left the firm.

how many, like how big the disruption caused by the war was in terms of workers draft and also in terms of change in the production. And

In almost all these cases, we find that yes, firms that had fewer managers that left, fewer workers that were drafted, it didn't change their production much, even higher performance. But the effects of management seems to be positive, large, and long-lasting almost along all these firms.

It's not that some specific issues caused by the war seems to explain whether the results are there or not. OK. Let me move on. Why don't we go with Tim, because he's right beside Chris, and that's efficient. And then we'll go to Mohammed, because he's done this before. We like him.

I wanted to ask about the initial conditions of the firm. You showed the productivity rose. I wanted to know whether that's caused by the initial level of productivity. So is it the bad firms become average, average firms become good and everybody moves up the same? Or is it the bad firms become good but good firms remain where they are? Or is it the superstars become even better?

So all firms improve to some extent, but the firms for which the improvements were the largest are those who were on the bottom part of the productivity distribution before the war. So we can think that the best firms become quite good and also better firms become even better, but the change is not that large.

I've got some questions coming online. If you're posting a question to us online, please do include your name so that we can call you out. Let's go with Mohammed and then we'll bring someone in from online. Can you hear me? Yes, sir. Thanks a lot, Miguel. That's a great presentation. I was thinking, I have two questions. One is about opening a black box of the management practices here.

and trying to understand exactly what kind of practices were working and what were more effective, essentially increasing productivity. Because I was also thinking that sometimes you can also observe that management can be subservient. It can serve the workers. It can be very, very bad.

And then the management can have-- because it wants to survive, then it may have different objectives from the firm and so on, which essentially leads to this principle Asian kind of conflict. So do you have any evidence on that? Are there some practices that were maybe detrimental or negative in some aspects that were more positive?

And my second question is about when you think of management within the production function and its relationship to technology in particular. I was thinking to what extent is it constrained by technology or shaped by technology? And it struck me like this interesting example that you brought up about

between the first industrial revolution and the second. And I was thinking, is this because of the nature of technology, essentially? Is it dictated as a type of management that could have been implemented in each? Maybe the electrification, the shift to different type of engines and so on, this might have made it more difficult, essentially, to be innovative or thinking outside the box and so on. And then this relationship between innovation and management,

Yeah, no, thank you for the question. So I start from the second one. So if we think about the production function, so we have some combination of capital and labor and then we have an A in front of the production function, we can think that management affects the A in the production function. And so that

higher level of management makes the inputs more productive. And I totally agree with you that the change between the first and the second industrial revolution could be due to the differences in production and in particular the fact that jobs became

mostly standardized and over-specialized during the second industrial revolution. And so there was more need of coordination across workers that were doing just one piece of the production rather than during the first industrial revolution, which the workmen was also producing a piece

on its own. So that's for sure. In terms of which practices were more effective, if we do a heterogeneity analysis, definitely human resources management was

the practice that affected productivity in general firm performance the most, even if to some extent all practices had a positive effect on firm performance. There is also evidence, I didn't have time to show it today, of complementarities among practices. That goes back to the Toyota hypothesis that

good management is a beautiful cycle. So we start with some practices, we get better, we implement even more practices, we get even better and so on. We can also think on the other way around. We don't implement any practices, we become worse and we continue not to implement practices, we become even worse and eventually we fail for sure.

And so in this case, I don't think there is a clear evidence, at least during the war, in terms of implementation of self-serving practices. And I think it could be due to two reasons. So first of all, we're talking about a war economy. And so these are firms in the defense industry. We can think that they were, at least to some extent, constrained by what the government, what they had to do.

but also that the training within the industry had the goal of training the entire management, not just the top executive or the CEOs of the companies, but train some key figures within the company that should have trained middle managers, that should have trained production supervisors, that should have trained the rest of the work.

So in this case, I think that there is less margin for the implementation of these self-serving practices, which is more common when there are CEOs or top executives running the business. So we can think about are your scale managerial practices that could shape the financial structure, strategic investment of the firms and practices more related to the R core production like in this case.

So I'm going to voice a question from the online, but then we're going to come here in the front row and have a student question and then to Greg, okay? So this is from an anonymous user. So my request for names didn't quite work, did it? That's the management training that I received. It's enough to try. No, I know. But I think they're asking an interesting question because they're asking about how World War II era management practices changed

influence even longer term productivity in the US? To some extent, is there a kind of ceiling of an effect, I guess is one of the reasons I want to put this. You show us this amazingly lovely long term effect and it seems to taper, right? Is it because all the potential is fulfilled from this kind of batch of practices or is there a kind of even longer term story that

Yeah, so here there is a data limitation in the sense that we are able to follow firms for 10 years after the limitation. Which is impressive, right? Yeah, yeah, but there could be even more going on. And the fact that the effect settled down at the end of the period is mostly ensured by the fact that firms in

control group at some point started catching up. So they didn't much for like seven, eight years, and then they do a little bit better, at least in terms of sales, not necessarily in terms of productivity.

If you look at the Italian case, the effects are longer because we have 15 years of data after the implementation of the program. It looks like the productivity effect continues to increase. I also searched how many of treated and comparison firms survived up to 2010, and I still find that firms that

that got the training program survived, the higher rated firms that did not. Up to 2010, of course, we can think about it as a super selective sample in the sense that they only have a handful number of firms, but it looks like there is something going on even farther than the 10, 15 years that we're looking here. Thank you. On behalf of the Lawrence user. Let's come to the front row here, and then we'll go over to Greg afterwards. Sorry, I forget. I don't know.

I had two questions, but they're very brief. The first question is, I think it's quite clear, especially with the war effort and the war reconstruction in Europe through the Marshall Plan,

We can't quite hear. Sorry. Is that better? I said it's quite clear with the war effort and then the war reconstructions through the Marshall Plan that these productivity increases are quite state-driven in a way. And I was wondering about how, whether there was a large role of the state in...

Japanese productivity increases and then the reactionary kind of productivity increases in the US and what the role of the state was there. And then the other question was related to the spillover effects in the years after the

the war in the US and I was just wondering could we look at kind of firm level data using some sort of census data to track the movement of managers from firms that saw productivity increases to firms that perhaps didn't and see if there was is that is that I know that would kind of require quite a large amount of data but is there kind of a way there to track further

Yeah, so to answer your first question, yes, most of the intervention were state-driven, both in the US and Europe, and I would say also in Japan. Like this train trip for Japanese managers were organized mostly by the US that were coordinating and helping the reconstruction effort, even in Japan. Definitely the role of the governments in fostering this program

decreased over time. I don't think that in the 70s or in the 80s there was such a big role of the state, at least in relation of this program anymore, but it was more the private initiative of seeing Japanese firms doing better and realizing that it could be related to management and try to learn out of it.

In terms of your questions about spillovers, yes, definitely, that would be a great analysis to do in order to see which kind of knowledge managers bring with them when they change firms, which also speaks to how easy it is to implement a set of practices from one firm to another.

One thing that I didn't specify during the presentation, but I think it's important, is that we do not observe all the firms in the economy. So this analysis is only made on firms, on other world contractors for which we have the performance data.

that were closed to firms that were traded. So there could definitely be spillovers, the fact that we are not able to capture because we don't have the full market structure of the economy, maybe in smaller firms there was more spillovers. This is something that our analysis is not able to document. Thank you. We're going to go to Greg next in the middle there, and after Greg to Matt because that's an easy pass. Thank you.

I'm going to cross over slightly to a couple of questions that you've had, but with a more specific example. It seems to me that the period that you're covering, the wartime and the reconstruction, is the period that delivers a really clear objective function. Do this and do it better. And that's the high point of the introduction of mathematical quantitative techniques to management, like linear programming, which is teachable and relatively transferable.

but once you've done it, it just becomes standard practice and it may well explain why some of the graphs flatten out. But once you've done that, and I think this is possibly where management then faces actually a completely different set of challenges, you then have to deal with much more, as it were, softer or wickeder problems around, for example, corporate culture or labour relations and the relative returns to labour and capital. Yes, so...

No, this is definitely true. I have to say that in general the type of practices that are taught in this stream program have not changed very much. So even if some practices, especially those related to factory operations or I know wasted inventories and so on, should be by now pretty straightforward, there is still a substantial variation on the number of firms that are systematically implemented these practices. Or maybe

To be more precise, it's not that they do not want to implement these practices, it's that they're not aware of how

how many inefficiencies their firms have. So it's definitely clear that the production function during the war was very clear, during the reconstruction was very clear, now challenges are more, there is also more complex implementation of practices in relation of human capital, this is more and more specialized, machineries, and in general inputs.

But the fact that the productivity spreads and management spreads are so large and so persistent seems to indicate that firms are not able to catch up that well, even in modern times. So it's not that once a managerial practice becomes the standard, all firms systematically implement it. - I'm gonna go to Matt in the middle there, it's a two alumn. - Thank you for the talk, that was really good.

So I was wondering how connected the TWI applicants that get selected, their location is to pre-military bases. Basically how connected, how integrated is the program with military interests? And do you see that as part of the story given how much internal migration there is in the U.S. during this period? Dragging people from agriculture maybe into industry, they need more training.

around the workers that way and then on the follow-up those boom towns kind of persist afterwards so does that feed into kind of their better performance yes so in terms of which firms received eventually received training we do not find evidence of any special correlation so the fact that instructors were not enough to train everyone seems the main constraint that the program faced

And firms were kind of put in a waiting list. So even firms that eventually were not trained were taught that they would have trained at some point. So the program was shut down all of the sudden in 1945, in August 1945, which is when World War II ended.

but they were still having open application for firms. So I think there was at least from the program administrator the willingness to continue and to eventually train all firms. So there is not this kind of choice, at least in the years in which we observe the production.

In terms of migration, it could definitely be that there was a reallocation of workers or in general migration, in particular in the 50s, in the 60s. I

don't think since the distribution of firms that we're trying is quite homogeneous i don't think that uh um that affects our results much uh we do see larger effects in the northeast and in the midwest but also the most industrialized areas of the country but uh

the effects are pretty large, even if we look at the other areas, in particular the south to the west. We still find pretty sizable effects there too. - We're gonna go over to that side of the room. So Eric and then Andrea, so you're beside each other. So you can hold the microphone in the spirit of efficiency.

Thank you very much. That was fascinating. I guess, can I ask you to zoom out a little bit? And so, you know, we're thinking about policy implications, like this seems to be targeted at manufacturing firms. And of course, we've experienced deindustrialization now. So do you think these practices would be as, you know, as effective in the service sector? Because that's kind of, I think, what European countries is, you know, efficiency beyond manufacturing.

Yeah, so yeah, it's true that most firms were in manufacturing sectors and in general when we're thinking about managerial practices increasing productivity, we are mostly thinking about the manufacturing sector. So there were some firms that were operating other industries even back then, in particular in services, and we do find an effect also for these firms.

There is an increasing body of work talking about efficiency in the service sector, and it looks like even in service sectors, the variation in practices is quite outstanding. The effect there seems, however, more driven by who are the managers, who can get different names across different sectors, of course. But it looks like that having good managers make the office, the services offered,

much more productive than having a bad manager, more than a set of practices as those that we discussed today.

Thank you, Michaela. It was fantastic. I want to ask you to zoom in, perhaps, and think about the TWI intervention much more closely. I was certainly intrigued by how these 50 hours of training caused persistent change in behaviors and practices within these firms. One could have this very optimistic view of training and having this very direct impact on behaviors, but I'm wondering if perhaps it's

were mediated perhaps by organizational changes within these firms? How were these principles and practices codified is the question I would like to ask you.

and perhaps connected to this as well, I was also intrigued by the fact that this was more due to changes in management than the role of managers themselves, right? So how does this story intersect with formal education in management and perhaps this experience attracting those people who believed in these principles to begin with? So can you speak about these two issues?

Yeah, definitely. So for sure during the war there was a huge effort from the government and in particular the training within the industry administrator to convince firms that management was important. This was done through a lot of advertising even for firms that weren't part of the training. So this idea was definitely in the air, if we can say it this way.

In terms of why relatively short training had such long lasting effects, well first of all, we have to think about the implementation of practices at the baseline. That in those years was very, very low. You're talking about less than 5% of applicants implementing any of these practices. So even if the training was short,

the baseline was so low that the effects could be that big, at least in part, due to that, on the one hand. On the other hand, the principle of the training within the industry was to offer the training to the management

but then have this training to be scaled up. So 50 hours of training is for the top management, but then there was training of the job for all the workers that is definitely much more and much more intensive than what the instructor gave. And in fact,

One of the reasons why this program was implemented during the war was exactly the idea of this kind of amplification effect. We train a few people and if they can transmit the knowledge within the firm then we can see even bigger results. I'm afraid that we are at the close. I know there are a number of people who are disappointed by not having had the opportunity to ask their questions.

Those of you online, I'm heartily sorry. Those of you in the room, please join us afterwards to grab Michaela and ask your question outside where we'll have a brief reception to mingle and say thank you. But for now, let us all join together and thank Michaela once again for a fabulous... Thank you so much.

Thank you for listening. You can subscribe to the LSE Events podcast on your favorite podcast app and help other listeners discover us by leaving a review. Visit lse.ac.uk forward slash events to find out what's on next. We hope you join us at another LSE event soon.