Show the union your labour-cost model before the first offer
Employers usually bargain by revealing as little as possible about their costs. When inflation has outrun a wage scale, I think the better course is to share the cost model with the bargaining committee early, within clear limits, and to accept the loss of a tactical advantage.
In most collective bargaining I have taken part in since 1999, the employer's first move was to say as little as possible about its costs. The union asks for more than it expects, the employer offers less than it can afford, and the two sides spend weeks discovering through proposals what each already knew. I do not think that pattern is always wrong. My position is narrower: when inflation has outrun the wage scale and both sides know a settlement must be large, an employer should show the bargaining committee its labour-cost model before making a first offer, within limits it states in advance.
I came to this view in the winter of 2022, preparing to renew our agreements at the Drummondville distribution centre and the feed mill nearby. Consumer prices in Canada had risen by close to seven per cent that year. The union asked for 19% over three years; our initial mandate was 10% over four. The gap was wide, and everyone involved knew that the eventual number would sit somewhere between.
Why we opened the model
Three things made me think the usual approach would fail. The first was the calendar. The distribution centre supplies all of our Quebec and New Brunswick stores, and our spring season carries close to a third of the year's feed and seed sales in those stores. Quebec's labour code restricts an employer's use of replacement workers during a strike, so we could not have kept the warehouse running through a stoppage. A long negotiation that ran into April was a genuine risk to the business, not a tactic.
The second was trust. The bargaining committee had been told in earlier rounds that the company could not afford more, and then watched it open new stores. Whether or not that was fair, it meant that any number we presented without its basis would be treated as a bluff.
The third was that our model was not, in fact, very secret. The committee's members worked in the building. They knew the volumes, the overtime and the turnover better than our finance team did. What they did not know was how we converted those into cost per case shipped, how our wages compared with other warehouses in the region and what a stoppage would cost us. Those were the numbers that would decide whether a settlement was affordable.
What we showed and what we kept back
In February 2023, before any offer, we presented the committee with four things: cost per case shipped for the previous three years, a wage survey of six regional warehouses, the volume plan for the next two years and our estimate of the cost of a six-week stoppage in spring. Committee members signed a confidentiality undertaking, and we told them in writing what we would not show: store-level results, individual salaries and anything about our other agreements.
The committee challenged the wage survey. Two of the six warehouses we had chosen paid less than the regional market because of their location, and the union's representative showed us why. We replaced them. That correction moved our view of the market rate by about forty cents an hour, and it did more for the credibility of the rest of the model than anything I said.
The agreements that followed run for five years, with increases of 6.5%, 4% and then 3% a year, and a scheduling clause that posts weekly schedules fourteen days ahead and caps mandatory overtime outside the peak. They were ratified by 81% at the distribution centre and 77% at the mill. There was no stoppage, and turnover at the distribution centre fell from 34% to 18% by 2025. I cannot prove that sharing the model produced that result. I can say that the committee spent its time arguing about the right numbers rather than whether our numbers existed.
The argument against, taken seriously
Our external labour counsel opposed the approach, and his objections deserve a fair statement. Once a model is disclosed, he said, it becomes the floor of every future round, and the employer loses the ability to argue later that circumstances have changed. Disclosure also creates a risk that commercially sensitive figures reach competitors. And a model shared with a committee may be misunderstood by members who hear about it second-hand.
The first objection is the strongest, and I accept part of it. We will open the 2027 round with the same model, and the union will expect us to. That is a cost. I think it is smaller than the cost of a negotiation in which neither side believes the other's figures, but it is real, and an employer whose costs are about to rise sharply for reasons it cannot yet disclose should think carefully before setting the precedent.
The second objection was answered by limits. We disclosed distribution-centre costs, not store margins, and nothing we showed would have helped a competitor price against us.
The third objection proved right. Because only the committee saw the figures, rumours about secret numbers circulated among members during the ratification vote. Information sessions for all members, with the committee presenting the model, would have cost us little and prevented most of that. I would do it next time.
There is also a case in which I would not share the model. If an employer is genuinely in financial difficulty and needs concessions, disclosure becomes a different kind of decision, closer to a restructuring negotiation, with obligations and advisers of its own. My argument concerns the ordinary renewal in a year of high inflation, where both sides know the settlement will be large and the only real question is how large.
Sharing the model also improved our own preparation. Because we knew the committee would test every figure, our finance and operations colleagues checked the volume plan and the overtime history more carefully than in any earlier round, and found an error in how maintenance labour had been allocated between the warehouse and the mill. A model prepared for scrutiny is a better model, whoever ends up reading it.
I would ask a board approving a bargaining mandate to ask management one question: what does the union's committee already know about our costs, and what would it change if they knew the rest? If the honest answer is that they know most of it already, the case for keeping the remainder hidden is weaker than it looks.