Serving a system you did not design

Writing about transforming systems tends to assume a reader with a blank sheet. In practice the reader is a mid-level civil servant, a hospital manager, a bank's credit officer, a head of department — inside an institution built by people who left long ago, operating rules they cannot change, accountable for outcomes they only partly control.

That is not a lesser version of the problem. It is the problem.

Three things that are actually movable

Formal authority is the least of what a person inside a system controls. Three other things move more, and none requires permission.

What gets measured and reported. Most institutional harm persists because nobody counts it. The complaint that is resolved but not recorded. The category of applicant who is quietly declined. Someone who begins counting a thing, and putting the count where others see it, has changed the institution's behaviour without changing a single rule.

Where discretion sits. Almost every rule has slack in it. A person who decides that, in ambiguous cases, the benefit goes to the party with least power has changed how the system lands for thousands of people, entirely within the rules.

What is normal to say. Institutional culture is largely a record of what people have watched each other say without consequence. The first person to name something in a meeting bears nearly all the cost; the fourth bears almost none. Being first is a real contribution and it rarely feels like one.

Three that are not, and pretending otherwise wastes years

The institution's fundamental purpose. A regulator will not become a development agency. Trying to convert it consumes a career and produces nothing.

Anything requiring a coalition you do not have. Enthusiasm is not a coalition. Nor is agreement in private from people who will not say it in public.

The pace. Institutions move at the speed of their slowest necessary approval. A person who cannot tolerate that either leaves, or burns out attempting to make it move faster.

The distinctive temptation

For Christians in public institutions the specific temptation is not corruption. It is significance — the pull towards visible, symbolic wins that demonstrate you were there and change nothing, over the invisible procedural fix that materially improves how the institution treats people.

Rewriting a form so it stops asking a question that excludes people is dull. Speaking at a conference about values is not. The first does more good, and no one will ever know you did it. That asymmetry is exactly why the first is worth teaching.

Staying, and knowing when not to

Most of the value a person adds to an institution accrues after year five, when they understand where the slack actually is and have enough standing to use it. Which means the most common way to fail is to leave at year two out of frustration.

But staying is not always right, and "influence from within" can become the story someone tells while their standards quietly move. Three signals that it has:

• You have stopped noticing the thing that appalled you when you arrived.
• You would not describe your work honestly to someone outside the institution.
• You are defending decisions you would not have made, from inside the assumption that there was no alternative.

The first is the most dangerous, because it feels like maturity.

What this means for our councils

This is why our industry councils are formation and research bodies rather than advocacy fronts. The people who improve institutions are usually working inside them, cannot say so publicly, and need somewhere to think honestly with others facing the same constraints.

That work does not produce announcements. It produces people who stay, notice, and use the discretion they have — which is most of what institutional change has ever actually consisted of.

Voice of Generations is non-partisan. We do not endorse political parties or candidates, we do not campaign, and nothing published here should be read as the organisation taking a position on a specific policy or piece of legislation.