World GDP 2026F3.0%World CPI 2026F4.7%Brent$88/bblFed funds3.50-3.75%ECB deposit2.25%Data centre load485 TWhReadings as at 1 Aug 2026
QuantiaQuantitative economics
Pedestrians crossing before high-rise buildings
The practice

How we work

Quantia is an independent quantitative economics practice. We take on engagements where the principal difficulty is measurement rather than presentation, and where the client must be able to defend the resulting figure after the engagement closes.

Approach

Three requirements on every deliverable

A figure is only useful if the person relying on it can identify where it came from, what it excludes and how wrong it could be. Most published analysis addresses the first two and omits the third.

01

Attribution

Every figure names the publication it derives from and links to it. Each report closes with a source register listing the series, the issuing body and the vintage.

02

Stated uncertainty

Point estimates are published with an interval or a sensitivity. Where a series is modelled rather than measured, the exhibit says so.

03

Reproducibility

Method, parameters and assumptions are documented in full, so a third party can rebuild the result without reference to us.

Methodology

Four analytical disciplines

We hold no attachment to a particular technique. The requirement is that any estimate can be explained in a single paragraph, including the conditions under which it would change.

Index construction

Chained Laspeyres and Fisher aggregation, hedonic quality adjustment where the basket changes faster than the weights, and explicit treatment of rebasing.

Disagreement about whether a price has risen is usually disagreement about weights.

Passthrough and incidence

Distributed lag models on panel data, with the lag structure selected by information criterion, and incidence apportioned between producer margin, distributor margin and final consumer.

The apportionment is generally where the commercial question sits.

Scenario design

Scenarios constructed around the two or three variables carrying most of the variance, identified by decomposition rather than by workshop.

A scenario that varies eight parameters simultaneously does not support inference.

Forecast evaluation

Vintages retained and scored against outturn as data arrives, using mean absolute error and directional accuracy.

The assessment is provided to the client irrespective of the result.

Engagement models

Three ways to work with the practice

Scope, duration and deliverables are fixed at the outset. Larger programmes combine formats in sequence, beginning with measurement.

Measurement buildSix to ten weeks

For decisions that depend on a series which does not yet exist in published form.

  • Documented data series to specification
  • Refresh code and maintenance notes
  • Source register and definitional appendix
  • Handover session with the analytical team
Typical output: one to three series
Estimation studyFour to eight weeks

For questions about magnitude: how much of a cost reaches a price, how demand responds, who bears an incidence.

  • Estimated model with stated identification
  • Elasticities with confidence intervals
  • Sensitivity analysis across key parameters
  • Written argument for the specification
Most requested engagement
Standing retainerOngoing, quarterly cycle

For teams that maintain a house view and need it marked against outturn on a regular cadence.

  • Maintenance of core client series
  • Quarterly scoring of internal forecasts
  • Priority access for ad hoc questions
  • Annual methodology review
Minimum four quarters
Analytical review at a working desk
Source policy

Desk research, fully attributed

Our research draws on the statistical and forecasting output of institutions that collect the underlying data, peer-reviewed literature, and where it represents the best available estimate, published sell-side and consultancy analysis. We do not conduct household surveys or customs audits and do not represent otherwise.

Where a figure reaches us through a compiler rather than the issuing body, the note names the compiler. Where an estimate originates with a bank or consultancy rather than an official statistical agency, the note identifies it as such.

Principal sources

  • International Monetary Fund
  • OECD
  • International Energy Agency
  • World Trade Organization
  • Food and Agriculture Organization
  • UN-Habitat
  • USDA Economic Research Service
  • US Energy Information Administration
  • Federal Reserve
  • European Central Bank
Standards

What this commits us to

These constraints apply to client deliverables and to everything published on this site.

Units, base year and frequency stated on every chart
No point forecast without an indication of dispersion
Definitional differences disclosed on cross-country comparison
Revisions to published projections recorded visibly
Every figure attributable to a source the reader can verify
Derived figures labelled as ours, with inputs specified

Who we work with

  • Corporate strategy
  • Pricing and commercial
  • Asset allocation
  • Industry bodies
  • Public agencies
  • Investor relations

We do not provide expert testimony, and we do not accept engagements where the conclusion is specified in the brief.

Begin with the decision, not the figure

The most useful initial brief describes what changes depending on the answer. Given that, we can usually confirm within a day whether the question is answerable from available data.

contact@quantiaconsulting.online