This is Part 2 of “What Lies Beneath”, a five-part public service “thriller” series following Scott while exploring the concept of administrative truth: what it is, how it works, where it breaks down, and what better-designed systems could look like for people and public servants. Read the intro and Part 1.
Scott’s* Record of Employment (ROE) is filed. Service Canada now knows he was laid off. His Employment Insurance (EI) application is in process. That is one (service-level) administrative truth, being established by one institution, under one piece of federal legislation.
But Scott realized that the gap between his last paycheque and his first EI payment may be longer than he expected. His two teenage boys sure eat a lot! Scott decides to figure out whether he qualifies for provincial income assistance while he waits for his federal EI to come through. While he browses the web, he wonders whether there is retraining support available. Could he start practising a different trade? Construction sure isn’t looking great these days. Could his situation qualify him for any housing assistance?
The inter-webs are overwhelming. Scott’s feeling a bit lost, and gives up his search for the day. Each of his questions belongs to a different institution, a different jurisdiction, a different set of rules. None of those institutions know what Service Canada is already determining about him (ie. “Yes, Scott lost his job”, and “Yes, he is eligible for support”). None of them will ask the other. Each will start from scratch.
*Scott is a fictional character in a speculative fiction story to illustrate the concept of administrative truth. His story returns at the beginning and end of every chapter and serves as a vignette to tell a bigger story. It is set in a not so distant, imperfect future, where (mostly invisible) computers – not humans – handle public service delivery.
* * *
Recap Part 1.
In Part 2 we dive into the institutional and jurisdictional complexities of the country that is now Canada. But before we do, it is worth revisiting some of the concepts introduced in the “Truth Ladder” of Part 1, because it matters for understanding what isn’t working, and what problems new designs would actually address.
Administrative truth is the specific authoritative data that an institution treats as operative for the purposes of making a consequential determination about Scott at a specific moment in time. When Service Canada designates his employment record as the authoritative data for his EI eligibility assessment, that data becomes more than personal information. It becomes the governed, designated source of record for a consequential determination. So the distinction between authoritative data and administrative truth is not in the information itself, but in what the institution does with it, and what flows from that, for example a benefits payment, or access to more services down the road. Or denied access, if you will.
A signal is the moment a change enters the system: unverified, recorded, awaiting confirmation. The administrative truth is what the signal becomes after government has officially verified and recorded it via a service flow. Scott’s Record of Employment was a signal. His confirmed EI eligibility is an administrative truth, and in itself can become a signal or trigger for more service eligibility. Some services have high value and trigger eligibility for many other services, something I call a “service cascade“.
Now let’s look at Canada more closely.
A Country of Layers
In 1867, Canada became a complex federation that unites 10 provinces and 3 territories (Yukon, Northwest Territories, and Nunavut) under a central government that acknowledges two languages (French & English). Federal, provincial, municipal, regional, territorial, and increasingly, First Nations governments exercise their own jurisdiction over their own peoples and lands, amidst the ongoing effects of Canada’s colonial history.
Each layer has its own legislative authority, its own programs and services, its own definitions of eligibility and its own systems for recording and acting on what it knows about the people it serves.
While Scott lives in the Province of British Columbia, his layoff triggers needs that cross several of these layers. Employment Insurance is federal, administered by Service Canada under the Employment Insurance Act. Income assistance is provincial, administered under provincial legislation with its own eligibility criteria. Retraining support may come from both federal and provincial programs (ie. WorkBC), with overlapping and sometimes conflicting eligibility rules. Housing or rental assistance, if it exists, is municipal or provincial depending on the program.
Each institution within each layer will ask Scott to establish his situation, his ground truth, from the beginning: his layoff, his household composition, his income, his address and his assets. Each will apply its own definition of what those facts mean for eligibility for support. Each will make its own determination, independently, without reference to what any other institution has already decided.
Michael Lipsky, writing about street-level bureaucracy in 1980, observed that policy is effectively made at the point of citizen contact, by human frontline workers who interpret rules in practice (Lipsky, 1980). What Lipsky could not have anticipated was a system where the street level is multiplied across jurisdictional layers, and digitised across jurisdictions that do not share a common definition of the person standing in front of them. Scott is one person. But the Canadian administrative state sees several different versions of him (ie. “employed” or “unemployed”), simultaneously, through lenses that were never designed to align.
This can result in quite strange situations, with distorted truths.
Different Rules for the Same Reality
There is no single Canadian administrative truth in the ontological sense of what it means to be unemployed and in need of support.
Federal EI defines eligibility based on insurable hours worked and the regional unemployment rate. British Columbia, where Scott lives, income assistance program has its own means test, its own asset limits, its own household composition rules. Retraining programs have their own eligibility criteria, sometimes requiring that a person be receiving EI to qualify, sometimes requiring that they have exhausted EI first. A person can qualify for one program and not another. They can be navigating three programs simultaneously, each at a different stage, each requiring different documentation, each operating on a different timeline.
For Scott, the divorce adds a layer the systems are not designed to handle well. BC’s income assistance will assess his household. But which household? The one recorded in provincial systems — still reflecting a marriage that ended a year ago — or the one he actually lives in now, as a single parent with shared custody of two teenagers? The answer matters enormously for the income test. The system itself has no mechanism to prompt him to clarify, because it does not know there is anything to clarify. The administrative untruth about his marital status sits silently in the system, lurking, waiting to govern a decision he has not yet triggered.
This reflects something deeper. There is no shared human ontology (what it means to be human, and under which conditions a human would deserve support) across Canadian jurisdictions. No common framework for what a person is, what conditions matter, what facts are real, and what they entitle someone to. Each jurisdiction built its own administrative and ontological model of the people it serves, for its own purposes, under its own legislative authority. Those models were never designed to be compatible, let alone digitally compatible, to share things across country layers.
Let’s take a look at another example. Somebody is experiencing a life-long disability. The Federal government could provide a person with a disability determination based on its understanding of what disability is, making them eligible for benefits (“Yes, this person has a life-long disability”), while a Provincial government would not (“No, this person does not have a life-long disability”), because each applies different meaning of the term, and different thresholds. Which is then administratively most true? And more importantly, which truth is more powerful: that of the person experiencing an event in their lives, or the truth of the State?
A Power Asymmetry
Canadian administrative systems are built, like many others, by default and design on the assumption that citizens can not be trusted, and must prove their reality to institutions before being eligible for support. For example, the institution’s record is treated as reliable until a citizen demonstrates otherwise. The citizen’s account of their own life is treated as a claim requiring expert verification, ie. from a doctor or an employer.
Scott knows he was laid off. He knows he is divorced. He knows his two teenagers eat a lot and live with him half the time. He has the documentation for all of it. What he is being asked to do, across multiple institutions simultaneously, is to perform each of these facts into each system and jurisdiction separately. And not only that, it has to be exactly in the format each system recognises, according to each system’s evidence standards, on each system’s timeline.
As mentioned in Part 1, the burden of closing the gap between what is true about his life and what each system believes to be true falls entirely on Scott. At exactly the moment when his capacity to carry that “administrative burden” (Pamela Herd & Donald P. Moynihan, 2018) is most reduced, for example during a stressful life event. Laverne Jacobs (2024), writes about the Canadian benefit system for people from intersecting backgrounds, and refers to the term “administrative violence” coined by Dean Spade (2015), pointing to the accumulated friction of systems designed around institutional risk management rather than human dignity, producing harm that is diffuse enough to be deniable, and specific enough to be felt by every person who navigates it. Within it, there is a new, complexifying asymmetry of knowledge in systems: institutions (“Big Brother”) hold data about people in digital infrastructures (“Big Other”) that those people cannot see, contest, or correct (Soshana Zuboff, 2019).
This all runs quite deep, and the societal backlash (or whiplash) is, amongst others, a decline in trust; or maybe that is my working assumption. Can we design our way out, or through?
A Note on First Nations Jurisdiction
The fragmentation described in this piece is experienced by all Canadians navigating complex life situations. But it lands differently and more severely for Indigenous (First Nations, Inuit and Metis) peoples, whose relationship to administrative truth carries the additional weight of colonial history. This short 5-part series does not intend to leave this aspect untouched, yet can not claim to do it justice either. So here goes.
The Canadian state has, for most of its existence, claimed the authority to define who Indigenous peoples are, where they belong, and what they are entitled to, and continues to do so to this day. Administrative truths such as “Status Indian registration”, band membership rules, the boundaries of reserves are imposed by the federal government, often in direct contradiction to the self-knowledge and governance structures of Indigenous peoples themselves.
First Nations governments are increasingly exercising their own jurisdiction over their own peoples, including the authority to establish their own services, their own administrative truths within their own treaties.
For example, the First Nations Information Governance Centre’s OCAP principles (Ownership, Control, Access, and Possession) articulate a framework for Indigenous data sovereignty that directly challenges the assumption that the Canadian state’s administrative truth about Indigenous peoples is legitimate or authoritative (First Nations Information Governance Centre, 2014). The BC government’s definition of Indigenous Data Sovereignty — the ability of an Indigenous government to access, collect, control, and utilize data about its citizens and territories — reflects this growing recognition. Another example is Jordan’s Principle, a legal framework to ensure First Nations children can access the services they need. It exists because the governance of administrative truth across jurisdictions failed Indigenous children so utterly and completely that a specific legal remedy had to be created.
The governance of administrative truth in Canada cannot be thought through seriously without engaging with Indigenous data sovereignty at every level.
To conclude
In Canada, governing is fragmented across so many layers that Scott’s administrative truth is not one thing held in one place. It is a set of partial, inconsistent, jurisdiction-specific records, each held by a different institution, each acting as if its version is complete, and thus true.
Yet, as this part 2 set out to demonstrate, it does not only part Scott into pieces in terms of data, the same tendency happens at the level of ontological meaning (what is “unemployed”, and what it means in terms of support is different across jurisdictions in Canada), also intersecting with state power depending on your status (immigrant, resident, Canadian, First Nations, indigenous). The question of who holds the authority to establish administrative truth about a person is a technical, an administrative question and in the Canadian context, it is also a political, cultural and decolonizing question.
Part 3 looks at what happens when the humans (bureaucrats! public servants!) who might once have exercised in situ human judgment within these systems are replaced by automated and digitized processes, as per the direction of many jurisdictions enabling data-interoperability and proactive service delivery. What happens when the machines start talking to each other, and when the administrative untruth about Scott’s household composition is already in the system when they do?
Back to Scott
Six weeks after his layoff, Scott is navigating three separate processes simultaneously. His EI claim is being assessed. His income assistance application is pending provincially. He has not yet started the retraining application because he cannot figure out which program he qualifies for without first knowing the outcome of the other two.
He also can’t choose if he is going to go for the digital route with WorkBC, or whether he’s going to go into an office to meet a person. Apparently he can’t do both and the two processes don’t talk to each other, is what his friend told him? Something he does not understand.
That Friday afternoon, he admits to the friend that he yelled at a government service worker on the phone. He didn’t mean to, he said. The poor person didn’t deserve it, but he felt so unconsidered and invisible, that it felt justified?
The friend said: “Bro, the system was designed around different jurisdictions, programs and around different legislative authority and institutional mandates, not around you. You were the last thing considered, if considered at all … So don’t yell at those poor people in public service roles.”
* * *
References
First Nations Information Governance Centre. (2014). Ownership, control, access and possession (OCAP™): The path to First Nations information governance.
Herd, P., & Moynihan, D. P. (2018). Administrative burden: Policymaking by other means. Russell Sage Foundation.
Jacobs, L. (2024). The Canada Disability Benefit Act and women with disabilities: Pursuing disability equality and reducing administrative violence. Osgoode Hall Law Journal, 61
Lipsky, M. (1980). Street-level bureaucracy: Dilemmas of the individual in public services.
Zuboff, S. (2019). The age of surveillance capitalism: The fight for a human future at the new frontier of power.
What’s next
Here’s what’s coming next.
Part 3 looks at the dark side. What happens as administrative decisions move out of human hands and into automated systems? What happens when an administrative untruth is already in the system when the machines start talking? Who is accountable when no one makes the decision?
Part 4 looks at what better design could make possible, such as a “receipt” of administrative truth, informed navigation within a jurisdiction and portable truths across jurisdictions. It wouldn’t necessarily require digitizing the entire service. It would require governing the records that already exist, and where applicable, digitise those well in consistent ways.
Part 5 makes the governance argument. Because the real problem is about who holds power over administrative truth, and what obligations come with that power.