Contingency at the midpoint: the numbers that should be true when the job is half built

Aug 28, 20265 minute readBy Corven & Ashby

There is a point on every GMP job, somewhere around 50 percent complete by value, where the contingency position is knowable and still changeable. Before it, too little has happened to read the trend. After it, the remaining risks are mostly committed and the position is what it is. The midpoint is the one review that can change the ending.

It is also the review that gets skipped, because at the midpoint everyone is busy and nothing has gone visibly wrong. Five numbers, read together on one afternoon, are enough.

One: the contractor's fund against the work done

The first number is the share of contractor contingency drawn against the share of work complete. On a fictional $36 million medical office building, 50 percent of the work billed and 48 percent of the contractor's $1.08 million contingency drawn is on trend. 50 percent of the work and 74 percent of the fund drawn is not. The second case says the fund will run out around month fifteen of twenty, and the owner should know now what the contract says happens then.

The comparison is by value, not by time. Half the calendar with a third of the work billed is a different question, about schedule, and it should not be allowed to disguise the contingency one.

Two: what the draws were for

The second number is the split of draws by cause. If most of the contractor's fund went on buyout shortfalls and coordination, the fund did what it was for and the remaining risks of that kind are smaller now that buyout is largely done. If a meaningful share went on items that read like owner scope or design gaps, the fund has been absorbing changes that should have raised the GMP, and the owner has been paying for them through the savings pool without knowing.

At the midpoint this is still correctable. A conversation with the contractor about reclassifying three or four of those draws as change orders, with the record behind it, is a conversation about maybe $200,000 on the fictional project. At closeout it is an argument.

Three: the owner's fund against the remaining risk

The third number is the owner's contingency balance against what is still unknown. By the midpoint the site is excavated, the structure is up, and the unforeseen conditions that were going to appear mostly have. What remains is finishes, systems and the owner's own late decisions. An owner's fund at 70 percent of its opening balance with those risks ahead is comfortable. One at 30 percent is not, and the board should hear that now, while the late decisions are still decisions.

The allowances as hidden contingency

Allowances are contingency that has already been named. At the midpoint, how many have been reconciled and how did they land? Four allowances that all reconciled over their values are telling the owner something about the two that are still open, and the owner's fund should be read with that expectation in it.

Four: the unclassified changes

The fourth number is the count and value of change orders whose cause is still open. Every one of them is a claim on one fund or the other that has not been decided. Twelve unclassified items worth $410,000 at the midpoint is $410,000 of contingency position the owner does not actually know. The midpoint review should end with a date by which each of them will have a cause.

Five: the projected pool

The fifth number is what the savings pool looks like if the trends hold. Buyout largely done, contractor contingency on its current burn rate, allowances at their current pattern, the unclassified items split as the record suggests. If the answer is a pool, the owner can plan for it. If the answer is a shortfall, the owner has ten months to decide what to do about it, which is the whole point of looking at the midpoint rather than the end.

The afternoon it takes

None of this is analysis. It is five figures read from a ledger that has been kept current, and one page of notes. What makes it hard on most projects is not the arithmetic but the ledger: if the draws have no causes, the allowances have no status and the changes have no classification, the afternoon becomes a fortnight of reconstruction, and the review that could have changed the ending does not happen.

What to do this month

  1. On the first pay application past 50 percent complete by value, read the contractor's contingency drawn as a percentage against the work billed.
  2. Split the draws to date by cause and list any that read as owner scope or design gap.
  3. Put a date against every unclassified change order by which it will have a cause.
  4. Restate the projected savings pool on the current trends and put it in the next report.

Questions on this

Why the midpoint and not every month?

Every month is right for the record. The midpoint is the one review where the trend is readable and the ending is still changeable, which makes it worth an afternoon with all five numbers together.

What if the contractor's contingency is ahead of the work?

Read the causes first. If the draws were contractor risks now largely behind the job, the fund may be fine. If they were changes that should have raised the GMP, that is the conversation to have now.

Is 50 percent by value or by time?

By value billed. Time and value diverge on most projects, and the contingency question is about money against work, not against the calendar.

In the product

Contingency ledger, Allowance register, Anticipated final cost. Free tool: Contingency Runway.

Keep reading

Earlier: Self-performed work inside a GMP: how the owner reads a package the contractor gave itself. Later: Change order pricing review: what the owner checks before the cause is settled. All articles on contingency.

One next step

See the baseline on a contract like yours.

Thirty minutes on a call. We read a fictional GMP amendment into the baseline and show what the basis flags reveal.

Create a free account