Short answer: Most property companies set spend approval limits once, at a portfolio size and a price level that no longer exist, and never revisit them. The result is one of two failures. Either everything escalates, and approval becomes the bottleneck in every maintenance job. Or nothing escalates, because people found a way around the limit, and the control stopped working without anyone deciding to remove it.

Setting a decision authority matrix is well-covered ground, including in RIOO's guide to scaling a portfolio and team without losing control. This article is about the part nobody schedules: reviewing it afterwards.

Key takeaways

• A threshold is a number in an inflating world. Left alone, it tightens every year without anyone changing it.

• The two failure modes look different. Everything escalating is visible. Nothing escalating is not.

• Invoice splitting is among the clearest signs a limit has become unworkable.

• Authority limits usually sit in the management agreement, so a threshold that no longer works is a contractual problem as well as an operational one.

• Dollar value alone is a blunt instrument. A $2,000 roof repair and a $2,000 landscaping change are not the same decision.

In this guide

• Why approval thresholds go stale

• The two failure modes

• How to tell which one you have

• What the numbers should tell you

• The management agreement problem

• Value alone is not enough

• Emergency work sits outside the matrix

• How to run a threshold review

• Common mistakes

• Frequently asked questions

Why approval thresholds go stale

Short answer: Because a threshold is a fixed number applied to costs that are not fixed. A limit set at $500 several years ago may now approve noticeably less work than it did, without anyone changing the policy. The number stayed still and the world moved past it.

Two forces do this, and they compound.

Costs rose. Analysis of multifamily operating data puts repairs and maintenance costs up more than 28% since 2021, driven by sustained inflation in labour, materials, and insurance, against average operating expenses reaching $8,657 per unit in 2024. Separately, construction cost analysis shows construction input prices up more than 43% since early 2020 according to Bureau of Labor Statistics data. Whatever your threshold covered in 2020, it covers meaningfully less work now.

The portfolio grew. A limit that produced twelve escalations a month at 400 units produces sixty at 2,000, assuming nothing else changed. The threshold did not become wrong. The volume behind it became unmanageable.

There is rarely a moment at which the policy fails visibly. It tends to degrade quietly, and teams adapt around it, which is often why the review never gets scheduled.

The two failure modes

They are opposites, and they need different fixes.

The asymmetry matters. A threshold set too low creates noise, and noise gets fixed because people complain about it. A threshold set too high creates silence, and silence gets fixed only after something goes wrong.

Reviews are more often triggered by the first problem than the second. The second is the one worth going looking for.

How to tell which one you have

Four checks, none of which requires a project.

1. Pull the last three months of approvals and look at the distribution. If a large majority of maintenance spend crossed the threshold, the limit is doing nothing except adding a step. If almost none did, the limit is not being enforced, or work is being structured to stay beneath it.

2. Look for clustering just below the line. A run of invoices at $480, $495, and $499 against a $500 threshold is worth examining. Some vendors genuinely price near a known limit because that is what the work costs. But a consistent pattern across vendors and properties is usually evidence that a limit has become unworkable and the team has adapted.

3. Check for split invoices. Two invoices from the same vendor, same property, same week, each below the threshold, for work that was obviously one job. This is one of the strongest signals available, and it is worth looking for specifically rather than hoping it surfaces.

4. Ask an approver what they actually do. If the honest answer is that they approve almost everything without much review, the threshold is not a control. It is a delay.

What the numbers should tell you

Four metrics, and each answers a different question.

The second one is the most revealing and the least tracked. An approval step where nothing is ever rejected may not be exercising judgment. It can mean the threshold is well calibrated and the work reaching it is genuinely appropriate. More often it means the step is adding delay to work that was going to happen anyway, and the person approving usually knows it.

That does not automatically mean the threshold should rise. Sometimes it means the wrong person is approving, or that the approver lacks the information to make a real decision. Both are worth knowing.

The management agreement problem

This is where an operational irritation becomes a contractual one, and it is the part most reviews miss.

Manager authority limits are usually written into the management agreement. RIOO's guides to property management contract terms and commercial management agreements both cover how those clauses are structured. The operational consequence is that a manager cannot simply raise a limit that has stopped working. The number belongs to the owner.

Three situations follow, and they need different responses.

That third row is worth dwelling on. A management company running properties for eight owners may have eight different authority limits, set at eight different times, each reflecting a different owner's comfort rather than any operational logic. The team has to hold all of them, and in practice they hold whichever they remember.

The practical answer is a per-owner authority record, held against the property rather than in someone's memory, so the system enforces the right limit for the right owner without anyone having to recall which is which.

One caveat worth stating: approval authority, delegation, and what a manager may commit on an owner's behalf are governed by the management agreement and by local law, and both vary. Nothing here is a substitute for reading your own agreement or taking advice on it.

Value alone is not enough

A single dollar threshold treats all spend as equivalent. It is not.

A $2,000 emergency plumbing repair and a $2,000 discretionary landscaping upgrade carry the same number and entirely different risk. One is unavoidable and time-critical. The other is a choice that could wait a week for an owner conversation.

Most mature approval structures add a second dimension:

The last row is a useful gate that does not require a threshold change at all. Work should not be assignable to a vendor whose certificate of insurance has lapsed, regardless of value, and that check belongs at work order assignment rather than at invoice approval. Catching it at approval means the work is already done.

A workable structure often looks like a higher limit for budgeted routine work in an approved category, a lower one for discretionary spend, and a separate path for anything statutory.

Emergency work sits outside the matrix

Most authority matrices have no defined emergency path, which means the emergency path is whatever the manager decides in the moment.

That is not necessarily wrong. Someone has to be able to stop a flood at 2am without waiting for an owner. But it needs to be a policy rather than an improvisation, because the alternative is that the exception becomes the routine.

Four things a defined emergency provision needs:

• What qualifies. Habitability, safety, further damage, or statutory. Not "urgent" as a judgment call.

• A ceiling, even if a high one. Unlimited emergency authority is not authority, it is an absence of one.

• A notification requirement. The owner is told within a defined window, before the invoice arrives.

• A documentation standard. Why it qualified, what was done, what it cost. The record is what makes the exception defensible.

Without those, the emergency threshold is effectively undefined, which is difficult to explain after the fact.

How to run a threshold review

Seven steps. Most of the work is in the first three.

1. Establish what the agreements actually say. Per owner, per property. Not what the system does, what the contract states. These diverge more often than teams expect.

2. Pull the data. Three to six months of approvals, with values, dates, categories, and time to approval.

3. Find the clustering and the splits. Look specifically. They will not present themselves.

4. Decide what the threshold is for. Cost control, owner visibility, fraud prevention, or budget discipline. These want different numbers, and a threshold trying to do all four usually does none well.

5. Propose per-category limits rather than one number. Higher for budgeted routine work, lower for discretionary, separate for statutory and emergency.

6. Take it to owners as a contract conversation. With the data. "Our current limit means 84% of maintenance jobs need your sign-off, averaging three days" is a better argument than a request.

7. Set a review cadence. Annual as a minimum, and off-cycle whenever the portfolio grows materially, a new owner is onboarded, or costs move sharply.

That last step is what stops the whole problem recurring.

Common mistakes

Frequently asked questions

1. How much should a property manager be able to approve without owner sign-off?
It depends on the management agreement rather than on a standard. Common structures give managers independent authority for routine maintenance up to a set figure, a middle band requiring internal approval, and owner approval above that. The right number depends on portfolio size, spend volume, and the owner's preference, and it should be reviewed as costs and portfolio change.

2. Why do spend approval thresholds stop working?
Usually because they are fixed numbers applied to rising costs and growing portfolios. Repairs and maintenance costs have risen substantially since 2021, so a limit set several years ago approves less work than it did, without anyone changing the policy.

3. What are the signs an approval threshold is too low?
A high proportion of spend crossing the threshold, invoices clustering just beneath it, split invoices from the same vendor for one job, long approval queues, and vendors complaining about delays.

4. What are the signs an approval threshold is too high?
Very little spend requiring approval, an approval step that never rejects anything, and owners querying costs they did not know about. This failure is quieter than the opposite one and usually surfaces only after an incident.

5. What is invoice splitting and why does it matter?
Breaking one job into multiple invoices, each below the approval limit, so that no approval is required. It is a signal that the threshold has become unworkable rather than that anyone is acting badly, and it means the control has stopped functioning without anyone deciding to remove it.

6. Can a property manager raise their own approval limit?
Usually not. Authority limits are typically written into the management agreement, so raising one is a contract amendment requiring owner agreement, not an internal policy change.

7. Should approval thresholds vary by spend category?
Generally yes. A single dollar figure treats an emergency repair and a discretionary improvement as the same decision. Most mature structures set different limits by category, urgency, and budget status.

8. How should emergency spend be handled in an approval matrix?
With a defined provision rather than an improvisation: what qualifies as emergency, a ceiling even if a high one, a notification requirement to the owner within a set window, and a documentation standard covering why it qualified and what was spent.

9. How often should approval thresholds be reviewed?
Annually as a minimum, and off-cycle when the portfolio grows materially, a new owner is onboarded, or costs move sharply. Thresholds that cause problems were often reasonable when they were first set.

10. How do you manage different approval limits across multiple owners?
Hold the authority limit as a per-owner record against each property, enforced by the system rather than remembered by the team. A management company running for several owners will otherwise apply whichever limit someone recalls, which is either over-restrictive or unauthorised.

Approval thresholds are one of the few operational controls that degrade purely through the passage of time. Nobody changes them and they stop working anyway, because the number stands still while costs and portfolio do not.

The question worth asking is not whether your limit is the right number. It is when it was last set, what has changed since, and whether anyone has looked at what the team does when it gets in the way.