Making realistic and understandable Infrastructure Commitments through risk based estimating.

Will Willson FRICS AVS Associate Principal Davis Langdon, Philadelphia

Cel 484 467 2524

November 2009

Introduction and Summary of presentation

Range estimating is used in a number of different contexts. Some define range estimating as ‘monte carlo risk analysis’ whilst others suggest it has nothing to do with risk analysis. Some feel range estimating is a completely different process to estimating or risk analysis. The presentation seeks to suggest to the audience range estimating is part of the risk assessment and analysis process and that by addressing the concept that everything in life has a level of uncertainty and in turn through questioning the level (or range) of that uncertainty together with the influence on the projects delivery better informed decisions can be made.

Identifying potential problems before they happen – and therefore helping to make the right choices

So why did that happen Who checked or did not check what Is this going to cause a big problem to the schedule or budget Are we going to look like idiots for not suggesting this might happen Perhaps we should have taken a different alignment

What is a range on something What influences a range The source data is important, recognizing the extremes

and what the ‘base’ reflects

Contingency

Why do we want contingency 1. To ensure we keep our job2. To make the project managers life ‘easier’3. To protect ‘ourselves’ from embarrassing overspends4. To allow for uncertainty

Why do we want to identify and quantify uncertainty To ensure the ‘right’ decision is made on the chosen option To make realistic judgments with respect to mitigationTo ensure public money is well spent To ensure funders investments are protected

One might suggest just applying a ‘historic’ contingency is about as good as a ‘guess’ and is totally useless as far as helping anyone make a rationale decision

Uncertainty in life

Some things in life are more certain than others

Check-in at an airport Changing a ‘flat’ tire Boiling a kettle Pouring concrete (note wind on the Hoover Dam Bridge Mass Pour)Assembling false work / formwork Laying ‘track’Installing ‘piles’ (I-5 undercrossing extraction & drilling)Moving ‘dirt’Permitting

We constantly and sub-consciously make decisions based on our own historic knowledge and experience

Granting Programs

Validation and Due Diligence by potential ‘funders’

Is the ‘applicants’ estimate robustIs the ‘applicants schedule’ realistic Is the money I am putting into this project good value for money Are other projects more ‘worthy’ of the limited funds I have Are the assumptions used for revenue generation well founded (passenger forecasts, expected promotion of economic and social growth) sound

An estimate

Measure quantities from sometimes from preliminary drawings Price quantities using historical data (eg MTO HiCo, Contractor past bid returns, Bottom up estimate, etc)Add escalation depending on length of contract and perceived market Add ‘risk’ / ‘opportunity’ of project / market factor Add a ‘contingency’

The hope everything averages out and the Bids come back within the engineers estimate

The basis of an estimate

Bottom up estimating uses Productivity assumptions (equipment manufacturers data, time / motion study historical data, estimators experience, contractor feedback)Equipment assumptions Gang / Team assumptions

These are influenced by:Site constraints Soil conditions Scheduling phasingInterface constraintsStaging

Its not a perfect ‘science’ – Windmill Quarry

The basis of an estimate

Top Down estimating uses All-in typically averaged rates from Clients, Consultants or Estimators database Will likely include ‘some risk and uncertainty’

These are influenced by:Make up of projects used to arrive at the average Prevailing market conditions

Questions to be addressed Is ‘our’ project similar to the average What was the ‘range’ in the data which the average was sourced from How old is the data

Influencing factors on an estimate

The components of ‘Uncertainty’

Range uncertainty (can include ‘risks’)Risks

Risks – things that may or may not happenNormal ‘expected’ variability in a ‘range’Influencing factors that may or may not happen in ‘discrete’ risks

Either way we are looking for uncertainty and why there is uncertainty and what influences that

The basis of an estimate

Building (accommodation) projects:Buildings, subject to ‘functional space’ chances and a few site specific issues (ground, permitting, ‘iconic special’ features) fall into a fairly predictable range Cost / m2 can be adjusted based on functional mix and site abnormals

Peculiarities with Transit projects:Vary significantly in make-up and influencing factors Component composite rates can also vary significantly

Analysis of many transit projects has shown that to provide a realistic estimate and schedule:

a knowledge of data source (location and project specifics) is requiredproject estimates and schedules must be built up from a component scope based estimate (ie not all-in cost / mile)

Influencing ‘ranges’

Typical sensitive ranges in quantity and cost (civils):Tunnels excavation advance rate: machine type, soil type, geotechnical variability, contractor, access, means and methods, interfaces Underground Stations – method (mined / drill and blast / SEM) access, faces available, temporary support, waterproof liners, staging, removal of spoil, interface with tunnel, restrictive working Property acquisitions costs – inability to establish tenants and evaluate businesses, widening constructions land take required, politicalcompromises, zoning changes Earthworks – haulage distance, rock type, soil consistency, re-use / off site disposal, contamination, staging, working hour constraints Large bridges – construction methods, construction ‘restrictions’ (in-water work, fishery protection), plant access, plant type, available capable and qualified contractors Utility relocations Pavement / street repairs and scope creep

All impact the ‘range’ and are not necessarily risks but can be ‘managed’

Influencing ‘ranges’

Typical sensitive ranges in quantity and cost (buildings):Curtain wall – type, fixings, access, finish, available manufacturers Ground conditions / foundation type – temporary support, dewatering, pad footings v’s piled foundations Permits – facings (cladding / siding), lighting (eg to car parks)Functional Space (hospitals, retail, education)Specification (rental market, anchor tenant, changing tastes)Type and size of equipment and systems and controls (industrial / Pharmacutical)

Typical sensitive ranges in quantity and cost (environmental)Noise abatement (sound walls)Wetland mitigation – protection, alternate construction methods Protected endangered species – relocation options, alternative habitats, time and process to implement, ratio of replacements to displacedOdor control (necessary to meet background targets)Discharge quality (purity of water, quantity of water, recharge capacity)

The 10th percentile ‘Concept’

Objective Avoid ‘risk on risk’Identify all uncertainties Remove all contingencies

The principals The lowest ‘sensible cost’ or ‘schedule duration’ that can be done if everything goes right first time and none of the ‘risks happen’

The advantages / OpportunityFlushes out ‘hidden contingency’Quantifies embedded uncertainly Provides a target for ‘least cost construction’Avoids ‘killing’ projects based on historic inappropriate data by challenging source against lowest likely current market costs

Risk analysis (quantification) forces range estimating

Process requires ‘answers’ that are very useful to the Client / Funder:

So which items have a ‘narrow range’ and which have a ‘wide range’ and whyWhich items are sensitive to the overall cost and do they have a wide or narrow range How can we ‘narrow that range’ to become more certain If there is no ‘range’ why do we need contingency

…..what specific risks are there outside of our ‘range’which characterizes our

Data source Means and methods Innovative or known type

Characterizing the ‘range’

Don’t over-thinkAt least set credible high, low and medium (P10, 50 and 90)Establish distribution curve if possible

P90P10 P90P10 P90P10

Estimated Value (P50)

Optimism bias

A ‘range’ and a ‘risk’ outside the ‘range’

Look for ‘spread’ ; acknowledge risks outside range

Appreciate deterministic schedule may not show the true critical path when a range ‘blows’ the float

Recognize that many components make up an infrastructure project and the variability of those components is different. The mix of those components on your project will therefore influencethe level of exposure

Risk analysis (quantification) forces range estimating

Process posses lots of awkward questions to the estimator that are very useful to the Client / Funder:

Where did the quantities come from What assumptions were used What was not included How much ‘bunce’ was built into the base estimate Where did the ‘rate’ come from Was it the ‘highest’ price from a subbie, the mid or the low price – how many ‘check prices did you get and how wide was the feed back you got – if wide why Where did you get the prices and how reliable are they –phone, fax quote, interview, the mother-in-law !

So which items have a ‘narrow range’ and which have a ‘wide range’ and why – how can we ‘narrow that range’

So we know where ….now we need to have a plan

“Big” current Influencing factors on an estimate

The ‘market’ and ‘economy’

New York Manhattan Tunnels

The prevailing ‘market’ clearly has a ‘huge’ impact on the returned bidsReturned bids typically remove ‘range’ and leave ‘risk’ any remaining ‘range’ being transferred to the ‘bidder’However we must not forget civil engineering infrastructure projects are based on a given GBR and one could suggest soil conditions and rock characterization are influenced by the extent of borehole investigations and testing and there remains a ‘range’ of uncertainty which can be separated from ‘risk’ directly proportional to the extent of soil investigations

Contact Information

wwillson@davislangdon.us

davislangdon.com