Updated: May 20th, 2026
I moved from HBS to Kellogg last summer. Since then, I’ve been thinking a lot about how innovation reshapes work and labor markets. During my seven years at HBS, I mostly studied job design and the future of work, with a particular focus on inequality: how firms structure jobs, how managerial roles change, and how new work arrangements affect inequality inside organizations. That line of work was what got me tenure (see my tenure statement here).
My new work grows out of that earlier agenda, but it also moves in a few new directions. I am now more focused on innovation itself, including new technologies, new forms of knowledge, and new organizational forms. I am also more interested in long-run consequences, which is why many of my current projects use historical settings. And I am trying to write for a broader audience, including not only sociologists and management scholars, but also economists.
These days, the question I keep coming back to is simple: how does innovation reshape work? I use “innovation” pretty broadly. Sometimes it means a new technology, like factory production in watchmaking. Sometimes it means codified knowledge, like paper patterns in dressmaking. Sometimes it means a new organizational form, like department stores. I like historical settings because they let us watch these changes unfold over time. They also remind us that many of today’s debates about technology, skills, and displacement have much longer histories.
​
Here are examples of the type of work that I’ve been doing recently.
What happened to technologically displaced workers?
The usual story goes something like this. A new technology makes certain tasks obsolete, and the workers who used to perform those tasks have to find something else. That transition often comes with a displacement penalty: lower wages, worse jobs, or in some cases unemployment. We have seen versions of this pattern in many settings, from robots replacing assembly-line workers to automated dialing systems replacing telephone operators.
Sissy Gao, an extremely talented PhD student at Michigan Ross, and I had a lively conversation about this one day and quickly got pulled into the question. One thing we noticed is that much of the evidence on technological displacement comes from workers doing a relatively narrow set of tasks, often routine ones. When those jobs disappear, it can be hard for workers to recombine their skills in a way that leads to another job of similar quality.
But that made us wonder: is this always the right model for thinking about displacement? What about workers whose skills are broader, more craft-based, and more portable? If a technology destroys their occupation, do they necessarily suffer the same fate? Or can they move into nearby lines of work where some of their old skills still matter?
With this question in mind, we started reading a lot of historical work on different technologies and their effects on workers. The historical record is full of suggestive examples. There are stories of skilled artisans in the late nineteenth century who seemed to land on their feet after new technologies displaced their old occupations. They moved into nearby trades, used parts of their old skills, and sometimes found work that was just as good as what they had before.
But these stories were mostly anecdotal. They were fascinating, but they did not tell us how common this pattern was. Did many displaced skilled workers actually make successful transitions? Or were historians simply noticing the most memorable cases? We realized that there was surprisingly little systematic evidence on this question. That made it feel like a good research paper: take a historical story that people had noticed, but turn it into something we could measure carefully.
That search led us to watchmakers. In the late nineteenth-century United States, artisanal watchmakers were still common. They repaired watches, made parts, adjusted movements, and relied on a broad set of highly specialized skills. But the rise of American watch factories posed a serious threat to this world. These factories standardized production, divided work into specialized tasks, and produced watches at a scale that individual artisans could not match.
To study this transition, we spent a lot of time reconstructing the history of watch factories in the United States. We collected information on where these factories were located, when they were founded, and, when possible, how much they produced. We then linked this factory history to the linked historical U.S. census. This allowed us to estimate each watchmaker’s exposure to nearby watch factories.
​
The first result is exactly what one might expect. The arrival of watch factories did, in fact, drive out artisanal watchmakers. In places more exposed to factory production, traditional watchmaking declined. So far, no surprise. It looked like another standard story of technological displacement.
​
But the surprising part comes next. Unlike assembly-line workers or telephone operators in many standard displacement stories, watchmakers seemed to recover fairly quickly. We see this whether we measure recovery by occupational standing or by estimated wealth.
​
This was not because they simply moved into watch factories. Some did, but the factories mostly relied on less-skilled labor. It was also not because they moved away to places less exposed to factory production. Instead, what seems to have happened is that many watchmakers redeployed their skills into nearby occupations with similar earning potential and overlapping task content.
​
In other words, their old occupation was being destroyed, but their skills were not. The precision, mechanical knowledge, manual dexterity, and experience with delicate instruments that made someone a good watchmaker could also be valuable in related lines of work. Many moved into occupations such as jewelry, engraving, machining, and other skilled trades.
We then became curious about how these transitions actually happened. Did watchmakers simply close their shops one day and enroll in some nineteenth-century toolmaking boot camp? Probably not.
One clue comes from newspaper ads. A nice feature of this period is that workers often posted ads advertising their own services. By looking at these ads, we can see how watchmakers described themselves to the market. Watchmakers exposed to nearby factories were less likely to present themselves only as watchmakers. Instead, they increasingly added related lines of work: “watchmaker and jeweler,” “watchmaker and engraver,” or other combinations that linked watchmaking to nearby trades.
This gives us a plausible picture of the transition. As factories began to take over more of the watch market, many artisanal watchmakers did not immediately abandon their trade. They first added sideline occupations. They broadened what they offered. Over time, some of these sidelines may have become their main line of work. So the transition was probably gradual. They did not jump from one occupation to another overnight. They stretched their existing skills into nearby markets until, for some, the nearby market became the new occupation.
This paper is still a work in progress. We are continuing to add analyses and run new studies, but a working draft is available here. The core message, though, is already clear: technological change can wipe out an occupation, but the consequences for workers depend on whether their skills have somewhere else to go.
We do not want to overstate the comparison, but it is hard not to think about AI. Many of today’s AI debates are really debates about skill portability. If AI replaces some tasks that high-skilled workers currently perform, what happens next depends partly on whether those workers can redeploy their skills elsewhere. The watchmakers case suggests that displacement is not only about whether a technology can do a task. It is also about whether workers have adjacent skills, nearby occupations, and alternative markets where their expertise still matters.
Spread of innovation and work
When we think about innovation and labor markets, we often picture the big invention: a new machine, a patent, a scientific breakthrough, or a dramatic new technology. Those moments are exciting, and of course they matter. But my hunch is that the bigger labor-market effects often come later. An invention becomes truly powerful only when it becomes easier to use, easier to teach, easier to copy, or easier to organize around. In other words, the key moment may not be the original invention itself. It may be the follow-up innovations that help it spread: standardization, templates, training systems, new distribution channels, new organizations, or new ways of packaging expert knowledge.
When I was at HBS, I had the good fortune of advising Yoon Jae Shin, one of the most creative PhD students I’ve worked with. Yoon Jae has always been interested in this broader question: not just where innovations come from, but what makes them spread. One of his ideas was to study paper patterns and their impact on dressmakers. This may sound like a niche topic, but I think it gets at something very general. Dressmaking in the nineteenth century was hard. It required years of apprenticeship and a great deal of tacit knowledge. If you have ever looked closely at a nineteenth-century dress, you know how complicated these garments could be. Paper patterns changed this world by turning part of that expert knowledge into a portable template. They made it easier to copy styles, standardize garment construction, and produce clothing without the same level of individualized training. In that sense, paper patterns were not just a tool for home sewing. They were an innovation in codified knowledge.
That distinction matters. Many innovations do not reshape work simply by replacing workers. They reshape work by making expert knowledge easier to package, circulate, and use. Once a dress pattern could be printed and sold, some of the knowledge that used to live in the hands of trained dressmakers became available to a much wider market.
That is what makes the case interesting to us. It is a story about what happens when craft knowledge becomes standardized and easier to access. The project is still in progress, and we are working toward a draft soon.
[more to come]​
​
Replication materials for recently published papers
​
Zhang, Letian and Simeng Wang "The Value of Internal Labor Markets: Evidence from LinkedIn Profiles and U.S. Inventors." American Journal of Sociology: Forthcoming. Link Replication Code
​
Zhang, Letian and Shinan Wang (2024) "Trusting Talent: Cross-Country Differences in Hiring." Administrative Science Quarterly: 69, no.2 : 417-457. Link Replication Code Online Appendices
​​
Hosseinioun, Moh, Frank Neffke, Letian Zhang, and Hyejin Youn. (2024). "Skill Dependencies Uncover Nested Human Capital." Nature Human Behavior: 9(4), p.673. Link Replication Code Online Appendices
​
Zhang, Letian (2023). "The Changing Role of Managers." American Journal of Sociology: 129, no.2 : 439-484. Link Replication Code
​
Zhang, Letian, Mitali Banerjee, Shinan Wang, and Zhuoqiao Hong (2023). “The Fragility of Artists’ Reputations from 1795 to 2020.” Proceedings of the National Academy of Sciences: 120(35). ​Link Replication Code Online Appendices
​​​
Wilmers, Nathan and Letian Zhang (2022). "Values and Inequality: How Prosocial Pay Discounts Offset the College Pay Premium." American Sociological Review: 87(3), 415–442. Link Replication Code​​​​​​