STRIKING TO WIN IN HIGHER ED:
Lessons from the 2023 Grad Strike
at the University of Michigan

Organizers from the UMich 2023 strike

15 August 2025

In negotiating a new contract between the University of Michigan and graduate student instructors (GSIs) in 2022-23, grad workers entered a pitched battle that lasted longer than many could have expected or hoped. Like many large universities in the United States, the University of Michigan relies on cheap, increasingly precarious labor to fulfill one of its main functions — teaching students. GSIs belong to a union, the Graduate Employees’ Organization (GEO), and comprise the vast majority of the approximately 2,300 bargaining unit workers.¹ Now, University of Michigan graduate workers insisted that we deserved a living wage while performing essential labor for the University. It took a multi-semester strike to resolve this antagonism. 

A driving issue for graduate workers was our unlivable wages. Before the campaign, the average graduate worker made $24,000 per year, some earning even less than this, and often with uncertain job prospects the next semester. We demanded a salary of $38,500 for all graduate student workers, a living wage for all.² In addition to this baseline living wage, graduate workers were fighting for financial resources and institutional support for parents, disabled workers, and international workers. 

Upon settling the five-month strike, all Ann Arbor PhD students within their guaranteed years of funding won our annual salary demand of $38,500 through a combination of raises and a 12-month funding model replacing the previous eight-month model, representing up to a 60 percent raise.³ Moreover, the University agreed to place PhD students on the Dearborn campus on a 12-month funding model by 2025. We also won six additional weeks of parental leave. Another significant win was UM’s stated commitment to implement a program for all PhD students (not only active workers or members of the bargaining unit) that provides one semester of full funding for workers who need to change advisors due to harassment, a significant counterbalance to the unequal power between graduate students and their advisors.  

The material gains won at the bargaining table were the result of an activated rank-and-file engaging in a protracted strike. Recent struggles by other university workers have suggested that winning significant demands necessitates prolonged strike action, a particular characteristic of this sector. But for workers to take on the risk and uncertainty involved in engaging in an indefinite strike, the demands need to feel both worthwhile and achievable through striking. When given full and ultimate control over both the moves at the bargaining table and the duration of the strike, graduate workers at the University of Michigan decided to stay out on strike for five months until we thought we had won all we could with the leverage we could wield.

Before negotiations began, union members decided collectively at a mass meeting for decisions to be open to all members that attended bargaining, rather than just the bargaining team. We also agreed never to engage in sidebars. As negotiations progressed, workers collectively decided, week after week, to maintain our position on our key proposals rather than negotiating down. When, after months of negotiations, we went on strike, grad workers had deeply and widely felt demands to strike for, and control over the terms of the strike. After striking through the final month of the winter 2023 semester, holding onto final grades to the extent possible, refusing to settle through the summer, and threatening to continue the strike that next fall semester, the University finally gave us a real offer that established the conditions for real negotiations. At that point, University of Michigan graduate workers and the University’s representatives in bargaining started having serious conversations about what could settle the strike. We reached a deal just before the fall semester began, having fought as long and as hard as we felt we were able.

This essay details what we take to be the key lessons of the campaign, which we found frequently challenged the standard wisdom of higher ed labor “experts”: from assumptions about how unions should bargain and relate to management, to what power workers have, and more. Fighting for and practicing an open bargaining format, where rank-and-file workers controlled the terms of bargaining (and consequently refused to make concessions unrooted in a shared evaluation of strike leverage), played a key role in building power. Our campaign demonstrated that the path to victory came not from what happened at the bargaining table, but from the activity of workers within their departments as they navigated the strike collectively and decided how to respond to challenges – from the struggle over bargaining to a legal injunction, other campus workers unprepared to move in solidarity with us, and pay docking. After months of an apparent stalemate in bargaining and management intransigence, UM finally caved, an outcome only explainable by the cumulative pressure built by workers over the course of the strike. We hope that our analysis here, as well as in previous essays, will prove useful to other labor organizers in higher ed grappling with similar challenges, and that others might avoid mistakes we made.⁴

THE FIGHT FOR OPEN BARGAINING

It took ten weeks of struggling against the University’s bargaining team before they finally conceded to conduct transparent, open negotiations. Over the course of those ten weeks, we questioned whether it was worth it to keep deferring discussions of matters of substance for week after week, especially as we worried that workers would lose interest or patience. We held our position that negotiations needed to be open to all grad workers, held in spaces large enough to accommodate the crowd, and be accessible virtually longer than management was willing to hold their stance that only a comparable small number of workers could enter the room at any given session. Ultimately, fighting for and using open bargaining developed a deep and wide commitment to many planks of the bargaining platform, a shared sense of ownership over bargaining, and a collective will to struggle and to win.

Before official negotiations began, management presented a set of ground rules that strictly limited who could be in the room with the negotiating teams and how we could communicate about negotiations with members and the media. In response, we made our position clear: any member of our union should be able to attend negotiations, and bargaining rooms should be large enough to accommodate all members who wished to attend. It was an eminently reasonable stance, one with ample precedent in the university’s negotiations with other campus unions, and it made immediate sense to the membership – it was their contract, after all. Management fought us tooth and nail, refusing to bargain over the substance of our contract until the bargaining team agreed to partially or totally exclude members from the room. Conversations around “bargaining logistics” absorbed the first two and a half months of negotiations. Management skipped scheduled bargaining sessions, scolded the bargaining team in front of the members, issued press releases framing our position as absurd, and lied about precedent in other negotiations. At one point, they brought in a state-appointed mediator, who made the ridiculous claim that she had the authority to impose closed-door sessions if we refused to come to heel.

All of this played out in direct view of the membership, and these experiences cultivated an oppositional, antagonistic mindset among a sizable segment of the union that included officers, stewards, and rank-and-file members. Members saw that the rhetorical skill and negotiating prowess of the negotiating team would never be enough to win the contract they wanted and needed. Moreover, they saw that management, and even our state labor board, did not seem all that concerned with adhering to the letter of the law. We had a choice: we could accede to management’s position, or we could refuse and find ways to exert our collective power outside the bargaining room. 

Our resistance to management took two forms: collective action and appeals to position holders. At a November general membership meeting, members wrote over one hundred letters to management’s lead negotiator, communicating that they had just as much a place in the bargaining room as she did. We flyered campus events, including a DEI panel featuring several high-level campus administrators who were clearly embarrassed by our presence, and we packed the room at regents meetings. More than one hundred members attended every discussion with management regarding ground rules. These exertions of collective power helped ground and focus members at a time when many of us were anxious about “falling behind” on our self-imposed schedule of negotiations. We also sought to leverage our parent union’s relationships with Democratic members of the University’s Board of Regents and other University officials to change management’s course. We asked for meetings with particular regents and gathered letters of support from elected officials and labor organizations. 

In January, management finally yielded. They continued to insist on ground rules, but their proposed rules conceded our key demands by allowing members into all bargaining sessions and securing three bargaining sessions open to the broader public. Once secured, open bargaining became a well-spring for organizing. Members who attended bargaining heard exactly what U-M’s bargaining team thought of graduate workers, and saw firsthand how little they knew about our working and living conditions. Attendance was particularly high at “big-ticket” moments, such as when HR introduced their initial salary proposal: a 2 percent wage increase for Ann Arbor graduate workers and 1 percent for Dearborn and Flint. Members derided the offer as “milk numbers” and plastered irreverent posters of milk cartons all over campus. 

Meanwhile, every Monday morning a group of graduate workers would door-knock departments, distributing bargaining newsletters and talking with grads about what had happened at bargaining the previous Friday. This pulled a core group of organizers together and led to dozens of conversations each week, keeping the organizing committee’s finger on the pulse of union membership.

This initial chapter of the campaign laid the foundation for our strike. Members learned that we didn’t have to take what management offered us, and that we could struggle and push for more. We analyzed and exerted our collective power, were willing to fight longer than management was, and ultimately got what we needed. Open bargaining was first and foremost a space where workers could confront management as a collective and develop a shared understanding of who we were up against. In that room, we learned together that good arguments alone would never convince management to do the right thing, and grappled with fear and doubt over whether and how we could win our demands. These experiences served to inoculate us against the trappings of the official bargaining process and foster the militancy required for successful struggle.⁵

Strike Part 1: The University’s Initial Strike-Breaking Attempts

THE RACKHAM PLAN

With little progress made at the table after months of negotiations, particularly on our salary proposal for a living wage, we prepared to present members with a vote to initiate a strike. Days before our strike authorization vote in March, the University announced what they called the “Rackham plan” (named after UM’s Rackham Graduate School) which would provide 12 months of funding to many (but not all) Ann Arbor PhD workers. The Rackham plan was announced by university leaders as a new graduate student policy. Importantly, the higher wages that the Rackham plan would pay graduate workers were presented outside of official collective bargaining, with no guarantee they would not be rescinded or paired with onerous additional work requirements – and there remained many workers left out of the plan (masters workers, Doctor of Musical Arts workers, Dearborn PhDs, and PhDs past their guaranteed funding years), not to mention the fact that it did not address our other demands like parental leave, better trans healthcare, and more. The announcement of the Rackham plan was broadly understood by graduate workers as a strike-breaking measure and led to a reframing of our living-wage demand: that UM put the Rackham Plan’s 12-month funding provision (along with raises) in the contract and expand it to all workers. 

After months at the table and with this demand on the table, members voted overwhelmingly to strike, with 80 percent of active membership casting a vote (approximately 50% of the bargaining unit) and 94 percent voting yes. 

Beginning with a walkout at 10:24am on March 29th (with the slogan “Walk Away from 24k”), our strike started in a fashion similar to previous GEO labor actions: with symbolic demonstrations of our power, including high-participation pickets and a short-lived attempt at hard picketing.⁶ In this initial phase, which lasted from roughly March 29 to April 10, energy was high and most workers were not imagining months-long strike action. While earlier conversations at General Membership Meetings (GMMs) involved our leadership core inoculating workers that, to win our demands, we might need to strike for weeks and months (rather than a day or two as strikes had historically been imagined), this was not yet a shared understanding among graduate workers. Thus in this phase, a significant layer of workers hoping to end the strike quickly on favorable terms invested their attention into factors that might lead to a speedy conclusion: fears of University moves to break the strike on the one hand; and, on the other, hopes that concessions by our bargaining team might “wrap things up.”

INJUNCTION DENIED!

One such fear was grounded in the anti-worker legal and contractual environment in Michigan. Our contract had not expired, meaning we were breaching its “no strike” clause. Moreover, public sector strikes are illegal in Michigan. The illegality of the strike gave the University grounds to seek a court injunction against us, which had broken our most recent strike in 2020. 

The University filed for an injunction a few days after the strike began. The night before the anticipated ruling, the Organizing Assembly – the body of active organizers in GEO – voted overwhelmingly on a plan to hold department meetings to discuss whether we should keep striking in the event of an injunction. After discussions at the department level, the plan was to hold a member-wide vote. We believe that this vote to continue striking would likely have passed, although the length of time we could have sustained a strike in defiance of court order (likely incurring court fines and perhaps other legal repercussions) is unclear. But in a surprising turn of events, a Washtenaw county judge denied the injunction on the grounds that the University’s legal representation had failed to prove that our strike had caused “irreparable harm” to the University. That the University’s injunction was denied was a significant moment in the campaign; both we and the University had embarked on a much longer fight than expected. The boss did not seem to anticipate that they would actually have to fight for their injunction in court,⁷ much less that they might lose the case. Meanwhile, AFT-MI, our parent union, had long urged our leadership to renounce all strike plans to avoid the injunction threat.

WITHHELD PAYCHECKS

Unable to rely on the courts to end the strike, the University turned to other measures, particularly withholding April paychecks. Throughout the strike, the University sent all workers a weekly attestation form with a simple yes-no question about whether they had worked that week. Based partly on advice from AFT-MI’s lawyers that these attestation forms were an insufficient legal basis for docking pay, workers voted en masse to ignore them, regardless of whether they were on strike. But despite our hopes that the University might be legally and logistically incapable of withholding workers’ paychecks, ultimately at the end of April the University docked pay from every worker who did not attest to working. The University seems to have made no systematic effort beyond these forms to determine whether workers were or were not on strike or for how long they were on strike, which laid the groundwork for numerous individual grievances.⁸ 

In the absence of a strike fund from our parent union, we drew on the experience and methods developed by grad workers at Santa Cruz and Columbia to create our own strike fund and distribute our limited funds based on need. One factor that worked in our favor, and helped us concentrate funds to cover those in the most dire need, was the financial insulation of the University’s aforementioned, strikebreaking Rackham plan. A large subset of workers knew that while we were losing our April paycheck, we would also be receiving roughly $12,000 in expanded summer funding, starting in May. Unfortunately, about a third of all Graduate Student Instructors. were not covered by the provisions of the plan, which was, of course, a major stake of the strike.⁹ 

Having seen the employer harden its position and make moves to retaliate, it was becoming increasingly clear that we would not be able to win our demands without continued struggle. After striking for three weeks with no movement, workers increasingly assessed the situation according to the leverage we had built, and would need to continue to build, over the long haul.

Strike Part 2: Accumulating Power and Tracking Our Leverage

As the morale boost from the injunction’s defeat began to fade, the task ahead came into focus: how to cause enough disruption to force the University’s hand. One takeaway from the injunction proceedings, witnessed by dozens of department organizers, gave some clue: the fact that the University’s witnesses clearly did not have any reasonable understanding of which classrooms were running and which were empty, but that they feared the consequences of being unable to assign grades.

Assessments of the strength of the strike in each department (both quantitative and qualitative), a feature of our organizing since the start of the strike, became increasingly serious. Through department meetings, group chats and other spaces, workers assessed how many in their department were on strike, how courses were being disrupted, the impacts of scabbing, whether and how grades would be entered, and more. Department organizers used strike tracking sheets and weekly strike report-back forms, following a template from striking grads at the University of California. From these assessments, a picture emerged of significant impacts on courses in most LSA (literature, sciences, and arts) departments, which mostly had majority or supermajority strike participation, and little impact in departments (particularly in the College of Engineering) with minimal participation. We held weekly membership meetings where we would vote on whether to continue striking. Stewards would present their department-level strike assessments so that members could make a grounded decision one way or another.¹⁰ Despite a small minority of workers who were adamant that we needed to settle quickly before the end of the term, each week, members voted overwhelmingly to continue the strike. 

We saw graduate workers from all corners of the University taking ownership over their own strike activity in new ways as each department and even each class demanded particular strike acts in order to stop the grades from going into the gradebooks. This involved negotiation with direct supervisors and garnering support from students, faculty advisors, and, where possible, department chairs. One graduate student reported at a membership meeting that they had smuggled the stack of final exams from campus to make sure that no one would put them through the scantron machine. 

With classes ending on April 18, we also looked ahead to grading deadlines. Prior grad worker strikes had demonstrated the significance of final grading deadlines. In the best organized departments, stewards led their colleagues in rigorous mapping of classes and grades, so that they understood their strike power in concrete terms: number of strikers, number of grades withheld, methods of scabbing and their level of effectiveness. Graduate instructors of record, those instructors who serve as the sole or lead instructor for a course, met collectively and recognized the deep impact that could result if they withheld grades. Unlike other grad workers who teach discussion sections under a faculty lead, instructors of record had full control over their courses and their gradebooks and therefore were more insulated from scabbing.

The top leadership of the University put massive pressure on faculty and department chairs to figure out a way to put in the grades, revealing that these grades were in fact valuable to the University. The administration became so desperate to figure out a way to fill in the missing grades that in some cases they pressured department chairs to simply fill in “A” grades where any were missing. The provost directed deans to coerce faculty, misleading several departments into believing that they were the only one without grades submitted. With faculty disorganized, this tactic was effective. We conservatively estimated that at least 15,000 students’ grades were impacted by the strike, what we called “bullshit grades.”  

With each passing day, our labor was scabbed bit by bit with missing grades slowly making their way into gradebooks. A group of faculty concentrated in the History department issued a public letter saying that they would not interfere in the strike by grading and that they too would not be issuing grades until the University settled the strike, but such solidarity was a minority tendency. As scabbing continued, some workers argued that we should undertake disruptive protest actions, such as blocking the bus depot and classroom entrances during final exams, fearing that our strike power was waning and that we needed to augment it with additional types of disruption. Others argued for greater concessions in bargaining, believing that we needed to settle quickly. By the final grading deadline, the only missing grades were from those classes that were taught by graduate workers who were instructors of record, and by mid-May our efforts to continue withholding grades dwindled as more and more departments entered cheapened or fabricated grades behind the backs of strikers.

The University’s maneuvering around the winter grading strike exceeded our expectations of what they would be willing to withstand, dashing our hopes of a strike-settling offer. Even with the likelihood of this outcome diminishing with each day, the majority of workers were nevertheless not ready to concede at the bargaining table. The winter semester concluded without a strike settlement and with little movement at the bargaining table. A new set of disagreements and anxieties settled in among core organizers regarding whether we had been mistaken about our leverage and whether and how to hold through the summer and prepare to continue the strike in the fall. 

Strike Part 3: The Summertime Holding Pattern: Uncertainties and Expectations

In the early summer period (roughly the month of May), the focus of the strike concentrated on two organizing projects: convening the roughly 100 spring/summer GSIs,¹¹ and attempting to foster the scandal around winter semester’s “bullshit” grades as much as possible. Without previous preparation, spring GSIs voted overwhelmingly not to begin their semester on strike, and organizing meetings did not culminate in new labor action. We did see the University take action on multiple fronts to try to manage a potential summer strike, including attestation forms, “no-strike” pledges, and forcing workers to enter grades regularly in Canvas or other easily trackable systems. These moves would reappear in the leadup to a potential fall strike.

After grades were submitted, we made various attempts to drum up scrutiny and public outrage around the University’s shocking choice to falsify student grades. Once we filed an accreditation complaint with the Higher Learning Commission, we entered into a period of uncertainty, compounded by the general demobilization typical of academic workplaces in the summer.

Our efforts to create a “PR crisis” for the University caused no meaningful developments at the bargaining table. Meanwhile, the American Association of University Professors (AAUP) decided not to take on our case about the fake grades, and the Higher Learning Commission did not find the University at fault. Without concrete evidence that our strike had worked in the form of better proposals from management, some organizers took the position that we had “lost” our power after grades were submitted. Others continued to organize members through direct actions, including demonstrations highlighting the PhD students left out of the Rackham Plan. Ultimately, as the summer wore on, organizers became increasingly unsure about whether continuing to struggle against the University would be worth the consequences and whether, collectively, we wanted to continue the strike into the fall semester. 

At the same time, there were signs that UM was straining in its preparations for a fall semester with a potentially huge gap in teaching labor: departments were tasked with hiring more lecturers to replace as much GSI labor as possible (for instance replacing some GSI instructors of record), attempting to hire only non-striking GSIs (most notably by making hiring conditional on “no-strike” pledges), and increasing class sizes. These plans meant extra labor for faculty instructors, on top of the overwork they had experienced in the winter when many of them scabbed GSI labor. While there were not very serious collective efforts to organize opposition to scabbing, it appeared that the prospect of another semester of overwork did take its toll.¹²

Strike Part 4: Settling and Its Discontents

As the fall semester approached, GEO was on unstable ground. The employer’s apparent capacity to absorb the blow of our winter strike had undermined the consensus around the theory of accumulating power that many of us had articulated. With grad workers dispersed and atomized over the summer, we were unable to facilitate the serious mass democratic deliberation necessary to build a new consensus. 

However, management provided us an opportunity to reorient when they issued their first serious offer on August 2 (just a few weeks before the first day of fall classes), thus confirming the impact of our strike and the threat we posed in the fall. Notably, this offer included something the University had been adamantly arguing was outside the scope of bargaining since the start of the strike: summer funding for Ann Arbor PhD students under the Rackham plan. At the same time, the lack of movement on other key issues was disappointing for many. One of the most problematic aspects of the offer was that the summer funding proposal left out some of the most financially precarious constituencies, including masters students, doctors of musical arts (DMAs), workers on the Dearborn campus, and all PhDs beyond their funding packages. Additionally, the wage article broke pay parity between UM’s three campuses (Ann Arbor, Dearborn, and Flint).¹³ The University also did not improve its offer on childcare benefits or disabled workers’ rights. And, perhaps most importantly, the offer came with a 48-hour deadline, after which the university threatened to revert to its previous, entirely inadequate offer. This psychological warfare was made more potent by the boss’s enlisting of state and local labor leaders to deliver and reinforce their message: settle now or lose everything.¹⁴

GEO members received the “exploding” offer with a mix of relief and anxiety. On the one hand, the fact that the University had made a big move revealed that we still had leverage. On the other hand, members were understandably concerned with the substance of the offer and divided by the offer’s deadline. One group, who organized a petition signed by 82 members, advocated for a swift response, arguing that making a counter within the employer’s timeline would allow the University to save face while also using our strong bargaining position to get a better offer.

Another group argued that more time was necessary to proceed collectively. This group argued that the exploding character of the offer was best ignored in exchange for continuing negotiations at the pace that best served us rather than being rushed into making decisions out of panic. This perspective was informed by the reality that the university made the offer to avoid a fall strike, and that their 48-hour deadline was unlikely to be a principle so important that they would endure a strike rather than enforce it.

Shortly after receiving the exploding offer, hundreds of GEO members gathered at an online mass meeting to decide how to proceed. As leaders, our goal was to create space for members to make this decision deliberately and democratically. During the winter, we had learned that it was only possible for members to make an informed decision of such magnitude by engaging in decentralized group discussions among rank-and-file workers, chiefly at the department level. Therefore, we proposed to members that we initiate a “Week of Discussion,” in which as many departments and interest groups as possible would meet to discuss the offer and potential responses. Following this, workers would gather to decide what our counter would look like. The proposed week of discussion implied, of course, that we would run beyond management’s imposed deadline. The alternatives were to somehow pull a counter together in 48 hours, or else accept the exploding offer right away. Members ultimately agreed, nearly unanimously, that we should enter a “week of discussion.”

The week of meetings at multiple levels of the union was one of the most robustly democratic moments of the campaign. Leadership did not make a recommendation on whether or not to accept the offer, and there was no contingent pushing hard in any one direction. Instead, the deliberations led to a shared assessment of our position and a counteroffer composed by nearly 300 people. This was seriously difficult work. There was both a sense of deep disappointment as we started to grapple with the limits of our power at the same time that we were celebrating a spectacular victory. 

Above and beyond the substance of any counteroffer, the real question was whether to settle or strike. This was a decision that necessarily involved department-level discussions about the labor power of GSIs in their specific departments and deliberation about the presence or absence of collective will to keep striking and withstand continued retaliation. As a membership, there was a broad consensus that the current contract on offer from the employer was unacceptable. There were over a dozen outstanding issues of great significance to various segments of the membership. However, there were also misgivings about continuing the strike, especially with the specter of another lost paycheck on the horizon. In the disorienting long summer, organizers had been unable to collectively envision and prepare for a potential fall strike, limiting the space of possibility for those workers most vocal about winning more. 

As the week of discussion concluded, members met again to decide whether to pass the counter we had developed together during the week of discussion, or else agree to the terms of management’s now-”exploded” offer. We voted overwhelmingly to pass the counter, which matched management’s last salary proposal, though including parity for the Dearborn and Flint campuses, and adding significant additional provisions that would benefit parents, disabled and immunocompromised graduate workers, and trans graduate workers. These proposals, while meaningful, cost little in the way of resources. We hoped that management would look foolish to refuse them and risk a strike on that basis. 

After we passed back a counteroffer, the University insisted that their offer had, in fact, exploded, but then continued to bargain with us on the basis of its terms. After two more rounds of counters and with less than a week to go before the first day of classes, the University made an offer with considerable additional gains. In addition to the summer funding side letter, they agreed to a side letter committing to a transitional funding program that would apply to all graduate workers rather than just those in our bargaining unit (GSIs).¹⁵ They also agreed to extend paid parental leave by six weeks, implement 12-month funding for Dearborn PhD workers by 2025, offer a “signing bonus” of $1,000 to all fall GSIs (later extended to graduate researchers and grads on fellowship), and establish an informal agreement that President Santa Ono would make a public statement voicing support for the implementation of an unarmed, non-police emergency response team at the University.¹⁶

GEO members overwhelmingly voted to settle two days after receiving the University’s offer. The offer was in many ways disappointing, but the decision to settle was based on a collective assessment of our ability to continue struggling against the University, and the majority of people agreed that, based on our power, this was the correct moment to reach a truce even if that meant signing a contract that did not meet our needs.

CONCLUSION

Grounding ourselves in sober assessments of our strike power and making important decisions collectively and frequently provided a basis for forging ahead in our strike month after month with little external support and much external criticism. The vantage point that emerged from department-level assessments of the strike led us to make a much different set of calculations at the table than the way we were advised to make them. Week after week, management would pass back our proposals with every line crossed out, reverting to their original proposals. We followed suit. Week after week, we decided collectively to keep our proposals more or less the same. Instead of horse trading between different planks of the platform or asking for less and less of a wage increase each time management said we were asking for too much, we instead told them what we actually wanted and waited for a reason to ask for something different or less based on what we thought we could win. It wasn’t until we had a sign that management was actually serious about reaching a deal – when they made us the “exploding offer” – that we came to the collective conclusion that it made sense to start making moves to reach a deal.  

Had we engaged in bargaining as it conventionally happens, by cutting back critical proposals to show our willingness to arrive at a deal and to demonstrate our reasonableness, we would have won significantly less in the end. We were under tremendous pressure from all sides to follow these conventions. Indeed, some GEO members as well as allies among faculty and the local labor movement literally pleaded with us to back off of our demand to secure a living wage for ourselves. Faculty were adamant that it was simply not possible, and many suggested that the proposal was borderline ridiculous. Local labor leaders, for their part, urged us to bargain “correctly” by narrowing our demands to arbitrary bottom lines that could fit on an 8.5 by 11 piece of paper.

Our experience showed us that when given the space, the collective body of members made rigorously strategic decisions. We were told by both management and some of our labor allies that unless the bargaining team had the flexibility to make strategic decisions when the moment called for it, negotiations would certainly break down. We found the opposite to be the case. It was because the bargaining team was not empowered to sidebar or to revise proposals or to prioritize some demands over others that we arrived at the decision to hold our demand for a living wage steady until we had accumulated enough leverage to win it. 

Collective deliberation and decision making was important beyond the bargaining table as well. In our experience, rigid or pre-determined understandings of what our strike should look like led to inflexibility and difficulty in adapting to changing circumstances. In order to counter unexpected administrative attacks, we found departmental organizing to be a potent antidote. Where grads engaged in grounded, collective analysis to make strategic and tactical decisions with their department coworkers, they were able to maintain a united core of strikers and continue onward. This required a shift in “ownership” of the strike from officers to the rank and file. By the end of the strike, many departments were making moves collectively and thinking rigorously about where their strike leverage lay – something we wish we had fostered from the beginning of the campaign. In general, a common understanding at the strike’s conclusion was that its significance and success lay in the extent to which rank-and-file workers led the strike.¹⁷

We learned that it is impossible, particularly for a small group of organizers, to predict what the boss will do and what effect a strike will have. One cannot simply “copy and paste” the experience of previous grad strikes and expect management to respond in the same manner. Our errors were partly in making overly bold predictions based on other higher ed strikes (for instance that we would be able to successfully negotiate back pay for striking workers, and that withheld grades would lead to immediate moves at the bargaining table).¹⁸ The strike also pushed forward our understanding of graduate instructors’ strike leverage and strategy, yielding insights about grade withholding and about timing.¹⁹

Beyond yielding a new contract, the 2023 strike ushered in a qualitatively new situation for academic organizing at UM. On the one hand, the strike demonstrated the success of a militant and workplace-grounded approach to building power as a path available to workers; on the other, the University has attempted to quash this approach as best it can, particularly through various forms of retaliation and workplace restructuring.²⁰ As the waves of repression against the pro-Palestine movement on US campuses (in many ways manifesting in our workplaces) have shown, there are major forces arrayed against the student and worker movements, against which the power of the academic labor movement may be currently quite limited. In this challenging context, the approach developed throughout the strike – developing workplace power at the department level, fostering a culture of workplace fights, and reasoning about risk and accumulating leverage in situations of significant uncertainty – is more relevant than ever. 

₁ Graduate Student Research Assistants (GSRAs) – who until recently were barred from unionizing under Michigan law – are not in the bargaining unit. Regardless, many GSRAs were invested in the contract campaign and many of them benefited from the gains that we made.

₂ This figure reflected the MIT Living Wage Calculator estimate of a living wage for a single person with no dependents living in Ann Arbor, where most UM graduate student instructors are employed; a smaller number work at the Dearborn and Flint campuses.

₃ Unfortunately, masters and Doctor of Musical Arts (DMA) students were excluded from this side letter and continue to piece together an income from semesterly contracts.

₄ This essay was collaboratively written by several organizers who held union office for the duration of the campaign, articulating our particular shared perspective.

₅ A minority of members were more concerned with the legal and procedural aspects of bargaining. Throughout the campaign, this group proposed slightly pared down proposals or stronger arguments as the way to get closer to winning our demands. Their alternative approach was a source of tension among members throughout the campaign.

₆ “Hard picketing” involved appealing to students and faculty not to cross the picket line or hold class. After two days of attempting the “hard picket” and students breezing right through the picket line, we gave up on this practice.

₇ One witness called to testify to the scope of grade impacts admitted to sloppy spreadsheet calculations, while another showed up with statistics scribbled on a piece of paper.

₈ AFT-MI would later refuse to provide legal support for those grievances, arguing that they would lose.

₉ One of the ways we inoculated members around lost paychecks was by emphasizing that we could negotiate back wages as part of a strike settlement. In retrospect, we spoke about the possibility of winning this with too much confidence. We did win a ratification bonus in the final settlement, but some workers did lose pay that was never made up. The sense that officers had over-promised, resulting from shortcutting a more collective and bottom-up risk assessment, was a source of tension.

₁₀ For an idea of the information that was presented and the different metrics we were tracking, see slides 19-22 in this slide deck from one membership meeting. Presentations like this were standard for Organizing Assembly meetings and membership meetings throughout the strike.

₁₁ UM subdivides the summer into two semesters called “spring” and “summer.” Fall and winter semesters make up the regular academic year.

₁₂ This likely contributed to the president of the UM lecturers’ union (LEO) ultimately stepping in to try and “get a deal done” with the regents – if not out of solidarity, then to prevent a greater labor burden on lecturers and to achieve labor peace before LEO began their own collective bargaining in fall.

₁₃ The University notoriously underfunds the Dearborn and Flint campuses, and correspondingly pays their workers insulting wages whenever it can. Pay equity across campuses was the central achievement of the UM lecturer union contract campaign in 2021, since lost in the most recent contract. Before the latest GEO contract, Dearborn GSIs (of whom there were roughly 60 in 2023) had the same hourly pay rate as Ann Arbor GSIs – albeit without guaranteed 12-month funding, and frequently involving overwork – while Flint GSIs (typically between zero and several in a given semester) had a lower pay rate.

₁₄ These leaders, who had always characterized some of graduate workers’ demands as unrealistic and even ignorant, presented management’s significantly improved offer as the result of their own heroic efforts to persuade the UM regents, and therefore as the ultimate limit of negotiations.

₁₅ Transitional funding programs extend temporary funding for workers exiting toxic advising or PI relationships, mitigating the pressures that keep workers in a hostile work environment. Recently UC workers have demanded similar programs to enable research workers to exit military-funded labs.

₁₆ President Ono made this statement one month later on September 21st at the first Board of Regents meeting of the academic year. However, like the city-wide model envisioned by local organizers that our demand was inspired by, there has been no actual implementation as of summer 2025 and UM’s use of militarized police against (particularly pro-Palestine) protesters has only expanded.

₁₇ A reflection by dozens of department organizers post-strike suggested a high degree of commonality in the contract campaign’s perceived successes (open bargaining, participatory structures that enabled workers to hold the line for a long time and achieve wins) and weaknesses (incorrect predictions from officers in the early stages, lack of solidarity from and strong organizing relationships with faculty and other campus workers).

₁₈ Not all officer predictions were overly bold (we fully expected to be enjoined; a host of other considered strike risks, for instance to international workers, ended up as non-issues). For obvious reasons, overly bold predictions are more remembered than overly conservative ones, but both point to the need for a wide, workplace-grounded approach to strike preparation.

₁₉ We now believe, for instance, that a strike beginning in fall, passing through grade deadlines and potentially into the winter term, is a more strategic timeline than beginning in winter. The long summer is conducive to demobilization and meanwhile presents a greater opportunity for the administration to prepare for the upcoming term.

₂₀ For instance, in the aftermath of our strike, UM mandated secondary instructors of record for GSIs to diminish the impact of future strikes, and made cuts to the number of GSI positions at UM.