
When a company controls the schedule, software, quotas, and discipline of delivery work but outsources the employment relationship, accountability doesn’t disappear; it migrates. New York City’s proposed “Delivery Protection Act” (Intro 518) is an attempt to close that gap by regulating last‑mile delivery facilities and, in its original form, by compelling direct employment of drivers who today move millions of packages for the city’s dominant e‑commerce operator through subcontractors.
At a Glance
- Intro 518 targets the structure of last‑mile delivery itself: licensing facilities, enforcing safety and training standards, and—at least in early drafts—requiring direct employment of drivers rather than outsourcing to small “Delivery Service Partners.”
- The bill would give the Department of Consumer and Worker Protection a formal lever to investigate complaints tied to licensed last‑mile sites, creating a clearer line of accountability.
- Amazon’s public counter‑campaign centers on costs and jobs, warning of up to $5 per order, threats to more than 40 small contractors and 5,000 workers, and potential relocation of operations outside the city.
- The broader policy fight is not Amazon‑specific: it tests whether a city can regulate a category of facilities where a lead firm exerts de facto control while avoiding employer liability through subcontracting.
What the legislation is trying to fix
Last‑mile delivery has been reorganized around a simple idea: keep operational control in the lead firm’s hands—routing, package flow, timetable pressure, handheld devices, performance dashboards—while placing the employment relationship with small contractors. Amazon’s Delivery Service Partner (DSP) program is the canonical model. Hundreds of DSPs, many of them single‑client businesses, hire drivers who wear Amazon uniforms, use Amazon’s logistics apps, and deliver exclusively Amazon packages on Amazon‑assigned routes. The policy problem the Council is targeting is not whether this model moves parcels efficiently; it’s whether workers and the city have a responsible counterparty when things go wrong—injuries, sudden mass terminations when a contract ends, or unsafe performance incentives embedded in algorithms. New York’s comptroller has tied modern subcontracting in logistics to a long lineage of “sweating” systems that separated control from liability; the bill’s licensing and direct‑employment thrusts are designed to reunify them.
Mechanically, coverage described for Intro 518 has two pillars. First, a licensing regime for last‑mile facilities, administered by the Department of Consumer and Worker Protection (DCWP), which would formalize oversight and create an investigative channel for worker complaints tied to specific, regulated sites. Second, in earlier reported iterations, a direct‑employment mandate requiring covered operators to put drivers on their payroll rather than rely on DSPs. The licensing half is the backbone—regulatory visibility, complaint intake, and enforceable standards—while the employment question is the lever that most directly challenges the current structure of control.
How we got here: control, liability, and the last mile
The last decade turned the curb into a production line. Barcode scans and dispatch software translate consumer promises—“arriving today by 10 p.m.”—into human workloads that are tracked to the minute. In this environment, accountability follows control. Federal and state proceedings have repeatedly probed whether Amazon is a joint employer over DSP drivers because it sets the onboarding standards, assigns routes, controls the handheld systems that measure pace, and can remove drivers from the platform—actions that go beyond an arm’s‑length vendor relationship. Recent labor cases and agency findings have moved in that direction, at least in specific facilities, indicating that the legal system increasingly recognizes substantive control despite formal outsourcing.
For a city, joint‑employer doctrine is not the only path. Municipal police powers—licensing premises that generate traffic, noise, and worker risk; setting safety and training baselines; auditing recordkeeping—offer a complementary route. If a facility concentrates economic activity that produces externalities on city streets and in neighborhoods, the city can license that node, condition its operation on meeting standards, and open a channel for adjudicating complaints. That is the core structure Intro 518 aims to build for last‑mile hubs.
The competing claims: costs, jobs, and who bears the adjustment
Amazon has mounted a sophisticated public‑facing campaign against the bill, warning New Yorkers directly in the app and on the website that orders could cost up to $5 more if the measure passes. The company ties that increase to two asserted consequences: losing the ability to use about 40 local DSP firms that collectively employ more than 5,000 workers, and possibly relocating last‑mile facilities outside the five boroughs—moves it says would slow delivery and raise costs. These claims are specific in tone and vivid in their implications for households and small businesses. They are also, on the present record, unverified projections by an interested party; there is no independent cost study in the public domain that cleanly separates regulatory compliance costs from corporate pricing choices or strategic relocation.
On the other side, the Council’s rationale is concrete but under‑instrumented in the public record. Supporters cite new safety, training, and labor standards and the need for a direct line of responsibility when drivers are injured, facilities generate neighborhood harms, or contractors collapse. Yet, while the city comptroller has documented high injury counts and safety concerns in last‑mile facilities broadly, we do not have a causally sharp, New York‑specific evaluation showing that direct employment, as opposed to licensing and targeted standards alone, would deliver lower injury rates, better training outcomes, or faster complaint resolution. That evidentiary gap doesn’t negate the accountability logic; it does caution against treating the direct‑hire lever as a proven remedy rather than a policy bet.
Is this an “Amazon law,” or a category rule?
Although Amazon is the primary affected operator, reporting describes Intro 518 as facially broader—aimed at companies that operate covered last‑mile facilities regardless of logo. In practice, the city’s market structure means Amazon is the main entity immediately touched, with potential spill‑in for other integrators and regional couriers. Category design matters legally and administratively: it reduces claims of singling out a company and clarifies that the trigger is a facility’s function and footprint, not the brand of the parcels. That is the right frame for a municipal tool.
Two structural uncertainties remain. First, the exact contours of the final bill—especially the scope and timing of any direct‑employment requirement—were reported as fluid, with the Speaker signaling potential adjustments that double down on licensing while revisiting the employment mandate. Second, litigation risk is real: small‑business advocates have threatened suit over the city’s authority to condition or prohibit certain contracting structures. Neither uncertainty should paralyze policy, but both argue for drafting precision and a phased implementation plan that preserves service continuity while building the enforcement muscle of DCWP.
What credible oversight would look like
If the city proceeds, three design choices will determine whether this becomes meaningful regulation or a paper shield. First, facility licensing has to be coupled with auditable standards—induction training hours, maximum route metrics calibrated to street reality, heat and vehicle safety protocols—and data rights that let DCWP see what the handhelds and dispatch systems already record. Second, complaint investigation must be resourced and time‑bound; a hotline without investigators is a cul‑de‑sac. Third, transition planning for the current DSP workforce is not optional. If the final law keeps a direct‑hire component, the city should require continuity offers for incumbent drivers, portability of tenure for benefits calculation, and an orderly handoff that avoids the abrupt mass‑termination cycles that have scarred this sector. Those measures operationalize the promise of accountability rather than assuming it.
Cost concerns are not illegitimate; they are, however, testable. A credible, independent fiscal and logistics analysis—using actual city order volumes, route mileage, facility capacity, wage and benefit assumptions, and phased compliance timelines—would bring rigor to a debate currently dominated by dueling talking points. Until such a study is public, dollar‑precise claims of price hikes or savings deserve skepticism proportional to the claimant’s incentives.
The bottom line
Intro 518 is best understood as a governance reform for a sector whose power center has shifted from the storefront to the sortation floor and the dispatch algorithm. The licensing architecture is the durable core—aligning authority with impact at the facility level and giving workers and neighborhoods a real venue for redress. The direct‑employment lever, while defensible on accountability grounds, is where the empirical case needs to catch up to the moral intuition; it is also where legal and operational friction will be highest. Amazon’s warnings about costs, jobs, and relocation are politically potent but analytically unproven at the magnitudes advertised. The city’s promise to improve safety and fairness is principled but will require disciplined implementation to avoid unintended harm to small contractors and service reliability. Good policy here will not come from who shouts loudest; it will come from writing a rule that matches where control actually sits—and then building the capacity to enforce it.
Sources:
afp.com, yahoo.com, syracuse.com, foxbusiness.com, techspot.com, nadler.house.gov, amsterdamnews.com, wunc.org



