Albert Hirschman gave dissatisfied customers two active responses to decline: exit and voice.

Exit is leaving. Voice is staying and trying to change the thing from within.

Part 1 was about why exit fails in PBM relationships. Employers can leave, but they rarely escape the structure. They can run the RFP, change the logo, renegotiate the guarantee, reset the implementation calendar, and still find themselves inside the same basic architecture: opaque economics, shifting definitions, partial data, layered incentives, and a contract that is easier to celebrate than to verify.

So the natural answer is voice.

If leaving does not work, speak.

Demand transparency. Escalate. Audit. Renegotiate. Ask for the data. Push the account team. Bring in the consultant. Tighten the language. Add the guarantee. Force the issue.

Employers do this all the time.

That is the first thing to understand. The problem is not that employers are silent. The problem is that PBM voice often travels through a system built to absorb it.

Picture the quarterly business review.

The employer comes prepared. Pharmacy trend is worse than expected. Specialty spend is accelerating. A rebate guarantee technically hit, but the net story still feels wrong. A prior authorization change produced patient disruption. The consultant has a list of questions. The benefits leader wants answers before renewal.

The PBM account team is attentive. They are polished, apologetic, and fluent in concern.

“We hear you.”

“We’ll take that back.”

“Let us open a ticket.”

“Clinical will review.”

“Pricing will need to weigh in.”

“Legal may have constraints.”

“Operations says this is working as designed.”

Nobody refuses to listen. Nobody says the employer is wrong to care. Nobody storms out of the meeting.

The voice enters the system.

Then it dissipates.

Not because one person kills it. Not because the account team is personally indifferent. Not because the employer lacks sophistication.

Because the system has many ways to convert a demand into an activity without converting it into a consequence.

That is why voice fails.

Hirschman’s Missing Bridge

Hirschman’s framework is useful because he saw exit and voice not as isolated behaviors, but as a system. When exit is unavailable or ineffective, voice should become more important. In his words, the voice option becomes central when dissatisfied customers or members cannot simply leave. He also warned that exit can undercut voice when the people most capable of complaining are the ones who depart.

That was Part 1’s PBM problem.

Exit is possible but functionally unhelpful. The employer can switch PBMs, but the market does not necessarily learn. The departing employer often moves laterally into the same architecture. The prior PBM loses a demanding account but keeps the system. The new PBM gains a dissatisfied customer and a fresh implementation cycle.

Exit becomes a pressure valve.

Voice is supposed to be what remains.

But Hirschman’s framework depends on loyalty as the bridge between blocked exit and active voice. Loyalty keeps someone engaged long enough to fight for repair. Later EVLN work made this even clearer: Rusbult, Farrell, Rogers, and Mainous added neglect as a fourth response and found that investment and poor alternatives can pull people toward loyalty or neglect rather than active exit.

PBMs break that bridge.

Employers remain, but remaining does not mean loyalty in Hirschman’s stronger sense. It often means technical loyalty: renewal without belief, continuation without trust, staying without repair. The relationship persists, but the employer’s voice has already been engineered down before it reaches anything capable of changing the model.

That is the difference between loyalty and inertia.

One keeps reform alive.

The other keeps the contract alive.

Voice Has Three Requirements

Voice is not just speech.

Voice requires three things.

First, a channel that can carry signal.

Second, an audience that can act.

Third, a consequence if the signal is ignored.

PBM relationships compromise all three.

The employer may speak clearly. The consultant may document the issue. The contract may describe the obligation. The account team may acknowledge the concern.

But if the channel converts complaint into ticket traffic, if the audience is fragmented across silos, and if the consequence is weak because exit is not credible, voice becomes ritual.

A meeting happens.

A deck is revised.

A ticket is opened.

A follow-up is promised.

The system survives.

The Channel Problem

The first failure is the channel.

Employers usually encounter the PBM through relationship infrastructure: account teams, client service leads, quarterly business reviews, escalation calls, issue logs, executive sponsors, and consultant-mediated status meetings.

These channels are not fake. They matter. A good account team can solve real problems, reduce friction, and protect a client from operational chaos.

But relationship management is not the same as decision authority.

The account team often sits between the employer and the machinery that actually determines economics, claims logic, formulary strategy, rebate treatment, specialty routing, utilization management rules, and operational configuration. They can translate the employer’s frustration. They can escalate. They can explain. They can coordinate.

But they may not be able to change the underlying thing.

That creates what I would call the empathy buffer.

The employer voices dissatisfaction. The account team receives it with care. The tone is constructive. The language is partnership-oriented. The issue is validated. The follow-up is real.

But the force of the complaint is softened by the very channel designed to receive it.

The complaint becomes an item.

The item becomes a ticket.

The ticket becomes an internal review.

The review becomes a status update.

The status update becomes a future meeting.

Voice doesn’t disappear.

It dissipates.

This is not censorship. It is conversion.

The demand is converted from a challenge to the model into a manageable workflow object inside the model.

You can hear it in the hallway language.

“We escalated it.”

“Pricing is reviewing.”

“They said the guarantee was met.”

“Clinical says the criteria are appropriate.”

“They’re going to get back to us.”

“They opened a case.”

The employer has not been ignored. That is what makes the channel so durable. Ignoring a sophisticated buyer would create conflict. Absorbing the buyer creates process.

Process feels like movement.

Sometimes it is.

Often it is containment.

The Audience Problem

The second failure is the audience.

Voice needs someone who can act on what is being said. In PBMs, that audience is often fragmented across incentive partitions.

A rebate issue belongs to one group.

A formulary decision belongs to another.

A specialty pharmacy concern belongs elsewhere.

A utilization management complaint is routed to clinical.

A network issue goes to network.

A pricing question goes to pricing.

A data request goes to reporting.

A contract interpretation goes to legal.

A patient disruption issue goes to operations.

Each group can explain its own piece. Each group can be technically correct inside its own frame. But the employer is rarely frustrated by one isolated piece. The employer is frustrated by the combined effect.

That is the audience problem.

The buyer experiences the PBM as one system.

The PBM answers as many departments.

Everyone has an explanation.

No one has ownership.

This is not merely bureaucratic inconvenience. It is structural power.

Fragmented accountability makes voice harder to aim. If no single audience owns the full economic and operational effect, then no single audience has to respond to the full complaint. The employer is left arguing with shadows.

The rebate team can say the rebate terms were honored.

Clinical can say the prior authorization criteria were appropriate.

Operations can say the turnaround time guarantee was met.

Network can say access standards were satisfied.

Reporting can say the requested field is unavailable.

Legal can say the contract governs.

Each answer may be defensible.

Together, they may still describe a failing relationship.

This is how voice loses force. Not because the employer cannot speak, but because the system has divided itself into compartments where no one has to hear the whole sentence.

The employer asks: “Are we actually better off?”

The PBM replies: “Which exhibit are you referring to?”

The Consequence Problem

The third failure is consequence.

Voice matters when ignoring it costs something.

In a normal buyer-seller relationship, a dissatisfied customer can credibly threaten to leave. The threat changes the conversation. It gives voice weight. Even if the customer stays, the possibility of exit disciplines the supplier.

But PBMs operate inside quasi-lock-in.

The employer can threaten an RFP. The PBM knows that. The employer can switch. The PBM knows that too.

The PBM also knows what switching costs.

It knows the employer will have to manage member disruption, data migration, prior authorization resets, new ID cards, new specialty pharmacy pathways, communications, implementation risk, consultant coordination, and internal political exposure if the transition goes badly.

It knows the alternatives are not clean.

It knows the employer may dislike the current PBM but still fear the switch.

So voice loses its threat.

The PBM does not need to win the argument.

It needs to survive the meeting.

This is why so many escalations feel oddly weightless. The employer is angry, but not necessarily mobile. The consultant is pressing, but still operating inside the same vendor set. The account team is concerned, but the contract is still renewable. The PBM can make concessions at the margin without changing the core architecture.

A reporting concession here.

A guarantee adjustment there.

A one-time credit.

A tighter meeting cadence.

A “strategic roadmap.”

A promise to revisit after renewal.

The consequence is not absent. It is diluted.

And diluted consequence produces ritualized voice: the performance of accountability without the pressure that makes accountability real.

Why Consultants Don’t Solve It

Consultants can improve the buyer’s position. They can benchmark pricing, run RFPs, identify contract weaknesses, pressure vendors, manage implementation, and translate PBM language into employer language.

But consultants operate inside the same structural constraints.

They can help an employer move from one PBM to another. They can help negotiate stronger language. They can help identify better terms. They can sometimes surface problems the employer would not have found alone.

But if the available choices are still clustered inside the same opaque architecture, consulting can become lateral movement with professional choreography.

The employer feels action.

The market sees motion.

The underlying structure remains intact.

This is the diversive energy problem. Hirschman used the idea of diversion to describe energy that looks active while pulling attention away from deeper repair. In the PBM market, the diversive energy is the RFP treadmill: finalist meetings, pricing grids, disruption analyses, implementation plans, legal review, guarantee modeling, member communications, post-go-live cleanup.

A huge amount of work.

A limited amount of structural change.

The consultant may help the employer choose better. But “better” is still constrained by what can be measured, compared, enforced, and verified.

If the contract improves but the employer still cannot observe the operational truth beneath it, voice has not become power.

It has become procurement theater.

Contract Language as Failed Voice

The contract is often where voice tries to become durable.

The employer pushes for definitions. Audit rights. Pass-through language. Rebate guarantees. Specialty terms. Network requirements. MAC appeal provisions. Reporting obligations. Performance guarantees. Termination rights. Transparency riders.

This is voice in written form.

But the contract is a map.

It is not the territory.

A map matters. Bad maps are dangerous. Better maps can improve navigation. But the map does not tell you whether the terrain shifted after you signed it. It does not tell you whether a definition is being optimized around. It does not tell you whether an economic guarantee is meaningful once the category underneath it has been engineered.

Enforceability requires measurement.

Without measurement, voice is just speech.

That is the contract problem in PBMs. The employer may negotiate a right, but rights are only useful when violations can be detected, attributed, and acted upon. If the employer cannot see the underlying data, cannot compare definitions across time, cannot monitor policy drift, cannot connect access friction to abandonment, and cannot reconcile financial promises to operational behavior, then contract language becomes symbolic power.

It says what should happen.

It does not prove what happened.

This is where PBM voice so often fails. It gets written down, but not instrumented.

The employer wins a clause.

The system keeps the interface.

The Exhaustion Cycle

Voice costs energy.

This is easy to miss when talking about market discipline in the abstract. Voice is not a free signal. It requires staff time, expertise, memory, data, patience, and political capital.

Someone has to notice the issue.

Someone has to document it.

Someone has to escalate.

Someone has to sit through the meeting.

Someone has to interpret the response.

Someone has to decide whether the answer is acceptable.

Someone has to keep the thread alive after three other priorities arrive.

PBMs have institutional memory. They have playbooks, account histories, legal templates, pricing teams, reporting systems, and specialized staff. Employers often have thinner continuity. Benefits leaders change. Consultants rotate. Internal priorities shift. Renewal calendars compress everything. The employer’s memory leaks faster than the seller’s playbook.

That asymmetry matters.

Voice resets.

The same questions return under new labels. The same concerns are rediscovered by new teams. The same ambiguity survives because the buyer-side record is fragmented, while the seller-side system is built to persist.

Eventually the employer learns the emotional logic of quasi-lock-in.

“We raised this last year.”

“They gave us the same answer.”

“We don’t have time to re-open it.”

“We’ll deal with it at renewal.”

“We just need to get through open enrollment.”

This is how neglect forms.

Neglect is not apathy.

It is the outcome of repeated non-response.

It is what happens when voice becomes expensive, inconclusive, and hard to accumulate. It is what happens when the buyer keeps speaking but the system keeps converting speech into process.

The employer does not stop caring.

The employer stops expecting the conversation to change anything.

The Same Equilibrium

Part 1 ended with neglect as the resting state of quasi-lock-in.

Part 2 explains how the system gets there.

Exit fails because switching is possible but often unhelpful.

Voice fails because the channel absorbs, the audience fragments, and the consequence dilutes.

Loyalty fails because it becomes technical continuation rather than engaged repair.

What remains is neglect.

Not dramatic neglect. Not negligent neglect. Not a benefits leader asleep at the wheel.

Operational neglect.

The quiet normalization of opacity.

The quarterly review that becomes theater.

The audit right nobody has time to use.

The guarantee nobody fully trusts.

The access policy nobody can track across versions.

The renewal that happens because switching would be worse.

PBMs call this retention.

A more honest term is compliance-by-fatigue.

The Instrumentation Question

The answer is not “complain harder.”

Voice cannot become power unless the system can carry signal. And in PBMs, signal depends on instrumentation.

Can the rulebook be turned into data?

Can coverage criteria be tracked over time?

Can policy drift be detected?

Can financial guarantees be reconciled to claims reality?

Can rebate narratives be tested against formulary behavior?

Can access friction be linked to abandonment?

Can exit leave a receipt?

This is where the next era of pharmacy benefit strategy has to move. Not toward louder escalation, but toward better legibility.

Hirschman saw that exit and voice can fail in relation to each other. Exit can drain voice. Voice can be too weak without exit. Loyalty can preserve reform, or decay into passivity. Later EVLN work gave us the missing endpoint: neglect.

The PBM market shows the whole arc.

Exit becomes lateral.

Voice becomes ritual.

Loyalty becomes inertia.

Neglect becomes retention.

The counter-move is not another meeting.

It is receipts.

The rules can be found.

The rules can be compared.

The rules can be tracked.

The contract can be tested against the territory.

And when the system becomes visible enough, voice changes character. It is no longer a complaint waiting to be absorbed. It becomes a signal attached to evidence, history, and consequence.

That is what PBMs fear less than anger and more than churn.

A buyer with memory.

A market with receipts.

A voice that lands.