ACCA APM study materials
ACCA APM study resources

Variance Analysis on a Fixed-Price Contract

ACCA APM · Put your knowledge to work in context.

All case studies

Variance Analysis on a Fixed-Price Contract

Calder Engineering won a fixed-price contract to build a bridge deck for $24m. At the end of period 4 of an expected 10 periods, the following is reported: planned cost to date $8.0m; actual cost to date $9.2m; planned work completed 40%; actual work completed 34%. The project manager reports 'cost overrun of $1.2m, being 15% over budget - we must cut costs'. Materials prices have risen 9% since tender due to a supplier failure; the design was changed twice at the client's request without a variation order being raised; and a key subcontractor was mobilised a period late because of a permit delay. The finance director wants a clear view of whether the project will make a profit and what to do about it.

1. Using earned value, calculate cost and schedule performance and interpret what the variances actually mean.

Read the worked answer

Earned value build-up:

  • Budget at completion (BAC) = $24m
  • Planned value (PV) = 40% × $24m = $9.6m
  • Earned value (EV) = 34% × $24m = $8.16m
  • Actual cost (AC) = $9.2m

Variances:

  • Cost variance (CV) = EV − AC = 8.16 − 9.2 = −$1.04m (over cost)
  • Schedule variance (SV) = EV − PV = 8.16 − 9.6 = −$1.44m (behind schedule)
  • Cost performance index (CPI) = 8.16/9.2 = 0.887
  • Schedule performance index (SPI) = 8.16/9.6 = 0.85

Interpretation - the PM's reading is wrong. The reported '$1.2m over budget' compares actual cost with planned cost, which conflates cost and schedule. We have spent $9.2m to buy only $8.16m of work. The true cost overrun on work done is $1.04m; the remaining $0.16m of the apparent gap is simply work not yet performed.

Forecast: EAC = BAC/CPI = 24/0.887 = $27.06m, implying a loss of $3.06m if the CPI persists. On a time-based view the project will also finish late (1/0.85 ≈ 18% over schedule).

Crucial caveat: the CPI is contaminated. Part of the overrun is price inflation and client-driven design changes - recoverable - and part is genuine inefficiency. The variances must be decomposed before any forecast is relied upon.

2. Advise the finance director on the causes of the variances and the actions to take, distinguishing controllable from non-controllable elements.

Read the worked answer

Decompose the $1.04m cost variance:

  • Price variance (non-controllable in the short term): materials up 9%. If materials are, say, 45% of cost, that is roughly 9% × 45% × $9.2m ≈ $0.37m. This is a market/supplier event and should be claimed or mitigated through re-sourcing and hedging, not squeezed out of site productivity.
  • Scope/design changes (recoverable): two client-requested changes with no variation order. These are not a cost overrun at all in economic terms - they are unbilled additional work. Raising variation orders is urgent and is a commercial control failure, not a site failure.
  • Schedule-driven cost (partly controllable): the late subcontractor mobilisation from the permit delay causes standing costs, idle labour and preliminaries. Some is excusable (permit), some is the contractor's own mobilisation planning.
  • Residual = genuine efficiency variance: what remains after the above is the only element the PM truly controls.

Actions:

  1. Immediately raise variation orders for both design changes and quantify them; restate the contract value and hence BAC.
  2. Submit an extension-of-time and cost claim for the permit delay and supplier failure per contract terms.
  3. Re-forecast EAC on the adjusted BAC and a CPI stripped of recoverable items, so the board sees the true expected margin.
  4. Introduce change control: no design change executed without a signed variation.
  5. Address the SPI separately - recovery plan and resource re-sequencing - because schedule slip drives cost through preliminaries.

Warning: a blanket 'cut costs' instruction would attack the controllable element while leaving the recoverable money unclaimed, and would likely worsen the schedule variance.