Change in unionized and regulated settings

Daniel Osei's plant at Northfield Components is unionized, and its collective agreement was renewed in October after the bargaining the cohort studied in MBA-601. The agreement allows the company to introduce new equipment, but it also contains a technological change clause: the company must give the union notice, discuss how affected employees will be retrained or reassigned, and agree on the classification and pay grade of any new job. The two robotic welding cells will turn eight manual welders into four cell operators and free the other four for work elsewhere in the plant. Daniel's first plan treated the union as a hurdle to clear after the design was complete. Dr. Morin asks him to imagine instead that the union's stewards, who know every welder in the plant, were part of designing the change.

Much writing on change management assumes that managers can redesign work as they see fit, subject only to persuasion. In unionized workplaces, and in industries regulated for safety, privacy or professional standards, that assumption is wrong. Changes to work are governed by collective agreements, labor law, regulators and professional bodies. This lesson explains the rules that shape change in these settings and shows that working within them, and with the people they protect, often produces better changes than working around them.

The collective agreement as a framework for change

A collective agreement is a binding contract between an employer and a union that sets wages, hours, job classifications, seniority rights, grievance procedures and many other terms for the employees it covers. Most agreements contain a management rights clause, which reserves to the employer the right to manage the business, direct the workforce and introduce new methods, subject to the agreement's other terms. In practice, almost any significant change to work touches something in the agreement: job descriptions, classifications, shift patterns, overtime, seniority in selecting new jobs, training rights, discipline or layoffs.

Disagreements about whether a change is consistent with the agreement are usually settled through the grievance procedure, which ends in binding arbitration if the parties cannot resolve them. An employer that implements a change the arbitrator later finds violated the agreement may have to reverse it and compensate employees. Even when the employer wins, a series of grievances consumes time, poisons relationships and slows adoption. The practical lesson is that the agreement should be read at the design stage, not after the launch.

In the United States, private-sector labor relations are governed by the National Labor Relations Act, enforced by the National Labor Relations Board. The Act requires employers and unions to bargain in good faith over wages, hours and other terms and conditions of employment, which are known as mandatory subjects of bargaining. Under the unilateral change doctrine established in NLRB v. Katz (1962), an employer generally may not change a mandatory subject without first bargaining with the union to agreement or impasse. Many changes to work, such as new shift schedules, performance standards, monitoring that can lead to discipline and changes in job duties, are mandatory subjects. Some management decisions, such as closing part of a business for economic reasons, are not themselves subject to bargaining, but under First National Maintenance Corp. v. NLRB (1981) the employer must still bargain over their effects on employees, such as severance, transfers and training. Whether a management rights clause allows a particular change without bargaining depends on its wording and on how the Board interprets such clauses, a question on which the Board's approach has shifted repeatedly as its membership has changed.

Cedarline's Chicago-area distribution center is covered by the National Labor Relations Act. After the warehouse workers voted for the union in June, the company was required to bargain over any change to mandatory subjects, and in practice it could not introduce the new warehouse management system, with scanner-based performance tracking, without bargaining first. The first collective agreement, ratified after the bargaining the cohort studied in MBA-604, now governs: it contains a joint pick-rate committee and a commitment that scanner data will not be used for discipline without review. Those provisions do not prevent the new system, but they mean that the way it measures and uses performance data must be worked out with the union.

From adversarial to joint approaches

Labor relations can be conducted at arm's length, with management designing changes and the union reacting, or jointly, with the parties working on problems together while preserving their distinct roles. Each approach has its uses, and most relationships mix them. But for complex changes to work, joint approaches have important advantages. Union representatives often know the work and the workforce better than managers do, and their involvement gives the change legitimacy with employees who distrust management. Joint processes also tend to surface problems before they become grievances.

Interest-based bargaining, developed from Fisher and Ury's principled negotiation and studied in MBA-604, applies these ideas to labor negotiations: parties share interests, generate options jointly and use objective criteria. Joint labor-management committees apply them between negotiations, for health and safety, training, job evaluation or the introduction of new technology. Thomas Kochan and colleagues at MIT have studied labor-management partnerships for decades. Well-known examples include the partnership between General Motors and the United Auto Workers at the Saturn plant, founded in 1985, which involved the union in decisions from product design to supplier selection before it was wound back in the 2000s; and the Labor Management Partnership at Kaiser Permanente, the American health care provider, established in 1997 with a coalition of unions, which has been studied as one of the largest and longest-lasting partnerships of its kind. Researchers have found such partnerships associated with improvements in quality and service, but also fragile: they depend on trust between particular leaders and can collapse when leaders change, when business pressures intensify or when either side feels the other has used the partnership against it.

A joint approach does not mean that the union co-manages the business or that managers give up their responsibilities. It means agreeing on how a change will be designed and introduced, who will be involved, what information will be shared and how disagreements will be handled, while each party remains accountable to its own constituency. It also does not mean bypassing the union by consulting employees directly about matters the union represents them on, which can breach labor law and will certainly damage trust.

Daniel’s robotic cells

Daniel revises his plan. He meets the union's plant chair before any design is final and proposes a joint working group of two stewards, two welders, a supervisor, an engineer from the equipment supplier and himself. The group's mandate is to design the cell operator job, propose a classification using the point factor job evaluation plan the parties agreed in the new collective agreement, plan retraining for the welders who will move to other work, and agree on how the cells' output data will be used. The stewards raise three issues Daniel had not considered: the most senior welders, who have seniority rights to the new jobs, are also the ones closest to retirement and least interested in learning robot programming; the cells' safety guarding must be reviewed by the joint health and safety committee; and the welders want assurance that cell data will not be used to compare individuals.

The group's solution is to post the cell operator jobs under the seniority provisions but to offer retirement-transition roles training new operators to the senior welders who prefer them, to schedule the safety review before the cells arrive and to agree that cell data will be reported by shift, not by person, for the first year. The classification lands one grade above the current welder rate, which the union accepts because operators will need new skills. The change takes six weeks longer to design than Daniel's original plan. But there are no grievances, the first cell reaches its target output in eleven weeks rather than the sixteen the supplier predicted, and two of the senior welders turn out to be excellent trainers.

Regulated settings

Unions are not the only source of rules about change. Many industries are regulated for safety, privacy, quality or professional standards, and regulators often require that changes be assessed, documented and approved before they are made. These requirements can feel like obstacles. They are better understood as formalized versions of good change practice.

Process safety is the clearest example. In industries that handle hazardous materials, such as gas processing, refining and chemicals, a formal management of change procedure requires that any change to equipment, chemicals, technology or procedures be reviewed for its safety implications, documented, approved and communicated to affected workers before it is implemented. In the United States, the Occupational Safety and Health Administration's process safety management standard, 29 CFR 1910.119, requires covered facilities to have written management of change procedures. Prairie Ridge's gas plants already operate under such a procedure; its merger team must decide how the procedure applies to renewables sites and to organizational changes, such as merging maintenance crews, that can affect safety as much as a new valve.

The phrase is apt: process safety engineers carry out management of change in a literal, disciplined form. Their procedures ask what could go wrong, who must be told, what training is needed and who must approve, which are the questions of this course. Investigations of major industrial accidents have repeatedly found that changes made without such review, including changes to staffing and organization, contributed to disaster. Several safety bodies now recommend applying management of change reviews to organizational changes as well as technical ones.

Health care adds professional and privacy regulation. Lakeview's documentation system records conversations between physicians and patients, so the hospital had to confirm that its use complied with the HIPAA privacy and security rules, sign a business associate agreement with the vendor, obtain patients' consent under its own policies and state law, and agree where recordings are stored and how long they are kept. Physicians remain professionally responsible for the content of their notes, whatever software drafted them, which is one reason the cardiologists' concern about errors was so serious. Nurses at Lakeview are unionized, and the hospital must consider whether changes to their documentation duties require bargaining with the union.

In each of these settings, the practical advice is the same. Identify early which rules apply and who must be involved. Treat the people who enforce the rules, whether union stewards, safety committees, privacy officers or professional leaders, as part of the design team rather than as reviewers at the end. Document decisions and reasons. And remember that a change that violates the rules can be reversed by someone outside the organization, which is the most expensive kind of failure.

Weighing the joint approach

Joint approaches take longer to design. Are they worth it? Dr. Morin asks the cohort to compare two ways of introducing Cedarline's warehouse management system. In the unilateral approach, management designs the system's performance tracking and announces it, relying on the management rights clause. In the joint approach, the system is designed with the joint pick-rate committee, which adds about three months. The figures are estimates, but they make the trade-off explicit.

# Introducing the warehouse system: unilateral launch versus joint design (thousands of dollars)
annual_benefit = 1800      # staffing and error savings at full adoption
years = 4
rate = 0.09

def pv_benefits(start_month, ramp_months, ceiling):
    total = 0.0
    for m in range(1, years * 12 + 1):
        if m <= start_month:
            continue
        share = min(1.0, (m - start_month) / ramp_months) * ceiling
        total += annual_benefit / 12 * share / (1 + rate) ** (m / 12)
    return total

# Unilateral: launch in month 3, but there is a chance of a grievance that the arbitrator upholds,
# forcing a redesign; slower, lower adoption because of distrust
p_grievance_upheld = 0.35
unilateral_ok = pv_benefits(3, 9, 0.85)
unilateral_redo = pv_benefits(12, 9, 0.80) - 250     # redesign cost and remedy
unilateral = (1 - p_grievance_upheld) * unilateral_ok + p_grievance_upheld * unilateral_redo

# Joint: three more months of design, faster ramp and higher ceiling, small grievance risk
p_joint_grievance = 0.05
joint_ok = pv_benefits(6, 6, 0.95) - 120             # cost of committee time and training
joint_redo = pv_benefits(12, 6, 0.95) - 370
joint = (1 - p_joint_grievance) * joint_ok + p_joint_grievance * joint_redo

print(f"unilateral: if no grievance {unilateral_ok:6.0f}, if upheld {unilateral_redo:6.0f}, expected {unilateral:6.0f}")
print(f"joint:      if no grievance {joint_ok:6.0f}, if upheld {joint_redo:6.0f}, expected {joint:6.0f}")
# How likely would an upheld grievance have to be for the unilateral route to win?
for p in (0.0, 0.1, 0.2, 0.3):
    ev = (1 - p) * unilateral_ok + p * unilateral_redo
    print(f"   grievance probability {p:.1f}: unilateral expected {ev:6.0f} vs joint {joint:6.0f}")

unilateral: if no grievance   4294, if upheld   2800, expected   3771
joint:      if no grievance   4475, if upheld   3441, expected   4424
   grievance probability 0.0: unilateral expected   4294 vs joint   4424
   grievance probability 0.1: unilateral expected   4144 vs joint   4424
   grievance probability 0.2: unilateral expected   3995 vs joint   4424
   grievance probability 0.3: unilateral expected   3846 vs joint   4424

With these estimates the joint approach has the higher expected value, even though it starts three months later and costs more to design. Two assumptions drive the result: adoption is faster and reaches a higher level when workers have helped design how their performance is measured, and the unilateral route carries a real risk of a grievance that forces a redesign. The sensitivity table shows that even if the chance of an upheld grievance were zero, the joint approach would still do slightly better, because of the faster, fuller adoption. Dr. Morin cautions that the numbers are judgments, and that in some cases speed matters more. But in a workplace with a new union, low engagement and a recent history of disputes about pick rates, the case for the joint approach is strong.