Macroeconomic data reveals trends shaping global economies

Why this topic matters for IFRS 9 practitioners

Macroeconomic data are the primary drivers of forward‑looking probability of default (PD) and loss given default (LGD) adjustments under IFRS 9. Firms that ignore, mis-specify, or inadequately document macro scenarios expose themselves to audit findings, regulatory pushback and unexpected earnings volatility. Macroeconomic variables link portfolio credit behaviour to the broader economy — see how Economic risks & ECL influence model outcomes in stress periods. Using robust macroeconomic inputs ensures that your ECL outputs are credible, auditable and aligned with the expected credit loss objective of IFRS 9.

Regulatory and audit expectations

Supervisors and external auditors expect a documented rationale for the chosen macro indicators, source quality, scenario design, and traceability to model outputs — from staging decisions to lifetime ECL. This is particularly relevant where the Three‑Stage Classification triggers movement to Stage 2 (significant increase in credit risk) and Stage 3 (credit impaired), both of which demand forward-looking evidence.

Core concept: what “Macroeconomic data” are and how they enter ECL models

Macroeconomic data are national and regional time series (GDP growth, unemployment, inflation, house prices, FX rates, commodity prices, interest rates) and derived indicators (credit growth, corporate profit indexes) that are linked to borrower behaviour. They can be used in three ways in ECL models:

1. Direct explanatory variables in statistical PD and LGD models (e.g., unemployment increases PD for retail portfolios by x% per 1pp).
2. Scenario weighting: applying different macro scenarios (base, upside, downside) and probability weights to generate point-in-time expected losses.
3. Top‑down shock adjustments: when granular models are not feasible, apply macro-driven overlays using mapping factors.

Components and examples

Key components to capture:

• Primary indicators: Real GDP growth, unemployment rate, policy interest rate, CPI inflation.
• Sectoral indicators: residential/commercial property prices, industrial production, retail sales.
• Market variables: equity indices, corporate bond spreads, FX rates.
• Derived indices: credit-to-GDP gap, household debt-service ratio.

Example mapping: for a retail mortgage portfolio, a 1 percentage point fall in residential property prices observed historically increased PD by ~0.12% and LGD by ~0.3 percentage points in the subsequent 12 months. Use Historical Data and Calibration to estimate these elasticities and document confidence intervals.

From macro to ECL: a simple formula

At a high level, lifetime ECL = Σ_t (PD_t * LGD_t * EAD_t * DF_t). PD_t and LGD_t can be expressed as functions of macro variables M_t:

PD_t = baseline_PD * f(M_t); LGD_t = baseline_LGD * g(M_t). Build f() and g() through regression, survival analysis or expert overlays, and validate them using Model Validation techniques.

Practical use cases and scenarios for finance, risk and model teams

Below are recurring situations where robust macroeconomic data practices materially affect ECL outputs and business decisions.

1. Quarterly ECL reporting with three macro scenarios

Process: select base, upside and downside scenarios; map GDP, unemployment and house prices to PD/LGD; run scenario ECLs and apply probability weights (e.g., 60/25/15). Deliverable: documented scenario impacts by portfolio and movement between Three‑Stage Classification buckets.

2. Stress testing and capital planning

When performing IRC or ICAAP, you need to link severe macro paths into the ECL engine to quantify capital adequacy. Use long-run historical shocks (e.g., 2008/2020 analogues) and bespoke severe paths for sensitivity testing.

3. Crisis response and re‑calibration

In crises, models based on pre-crisis data understate risk. Implement rapid model review: increase weight on recent adverse data, expand scenario breadth, and enhance qualitative overlays. See guidance on ECL during crises for operational templates.

4. Data-limited portfolios

Smaller lenders with limited historical default data can adopt top‑down approaches: apply sectoral macro elasticities, conservative overlays and increased documentation. For source identification, consult the inventory in ECL data sources.

Impact on decisions, performance and accounting outcomes

Macroeconomic inputs directly influence reported provisions, earnings volatility and capital metrics — collectively the firm’s financial stability. For regulators and boards, linking macro assumptions to material outcomes clarifies risk appetite and strategic choices. Explore the link between macro inputs and systemic outcomes in ECL & global economy.

Accounting impact on profitability

Changes to scenario weighting or key macro assumptions can swing quarterly profit before tax significantly. For example, a downgrade in the base GDP forecast by -1.5pp might increase ECL by 10–15% for a retail portfolio, compressing pre-tax profit margins. For a bank with 1% ROE and 30bp increase in ECL ratio, net income impact can be material. See analyses of Impact of ECL on financial statements for concrete templates to quantify this effect.

Operational and strategic decisions

Accurate macro inputs support pricing adjustments, provisioning policies, collection strategies and capital allocation. They also feed into credit limits, product design and liquidity planning, improving decision quality and reducing surprise outcomes tied to economic cycles.

Macro-data quality and financial stability

Incomplete or biased macro inputs can create procyclical provisioning. Strong governance over input sourcing, scenario governance and stress calibration contributes to broader Financial stability & ECL.

Common mistakes in using macroeconomic data and how to avoid them

1. Over‑reliance on a single indicator: Focusing only on GDP ignores sectoral dynamics. Remedy: construct a small basket of indicators relevant to each portfolio (e.g., unemployment + house prices for mortgages).
2. Poor source traceability: Using proprietary forecasts without version control. Remedy: log source, vintage, assumptions and probability weights in a data catalog.
3. Ignoring timing and lags: Macros affect PD/LGD with different lags (e.g., unemployment reacts quickly; property prices slower). Remedy: test lag structures during calibration and validation.
4. No sensitivity testing: Running a single scenario hides uncertainty. Remedy: perform sensitivity testing across +/- shock bands and include results in governance packs.
5. Insufficient documentation for Model Validation: Regulators expect clear linkages from macro inputs to model outputs. Remedy: produce validation scripts, model performance metrics and a judgement log.

Practical, actionable tips and a checklist

Below is an operational checklist and recommended steps for embedding reliable macroeconomic data into your ECL framework.

Step-by-step implementation checklist

1. Define portfolio‑specific macro baskets (e.g., mortgages: GDP, unemployment, house prices).
2. Choose high‑quality sources and retain vintages; see recommended sources in ECL data.
3. Estimate elasticities using Historical Data and Calibration: regress PD/LGD on macro lags and include confidence bands.
4. Design at least three plausible scenarios and assign transparent weights; document scenario governance.
5. Perform Sensitivity Testing: shock individual variables +/- 1/2/3 std dev and analyse ECL delta by portfolio.
6. Integrate outputs into accounting processes and disclose key assumptions in financial statements.
7. Schedule periodic Model Validation reviews and post-implementation monitoring (PIP) to capture performance drift.

Practical tips for data and modelling teams

• Store macro vintages — this lets you reproduce historical ECL runs for audits.
• Keep model code and macro mapping transparent; prefer parametrised transformations for easy updates.
• When data are sparse, combine top‑down mapping with conservative overlays rather than forcing low-confidence statistical fits.
• Document expert judgement clearly: why a certain elasticity or overlay was chosen, with alternative outcomes.

Example: sensitivity test template

Run three sensitivity cases for each portfolio: baseline, GDP -1.5pp, GDP -3.0pp. Report:

• Change in 12‑month PD (%)
• Change in lifetime ECL (absolute and %)
• Impact on Tier 1 capital (bps)

KPIs / success metrics for macroeconomic data in ECL

• Backtesting accuracy: % of observed defaults explained by model (e.g., within 95% confidence).
• Model calibration error: root mean squared error (RMSE) of PD predictions vs observed defaults over rolling 12‑24 months.
• Scenario sensitivity: change in ECL per 1pp GDP shock (bps of portfolio exposure).
• Documentation completeness: % of models with full source traceability and vintage storage.
• Governance cycle time: time from scenario approval to ECL reporting (target ≤ 5 working days).
• Proportion of portfolios using at least 3 macro indicators in final model (target 100% for material portfolios).

Frequently asked questions

Reference pillar article

This article is part of a content cluster on macro and data-driven ECL practices. For a broader view on data governance, sources and the central role of data in ECL modelling, see the pillar piece: The Ultimate Guide: The importance of data in calculating expected credit losses – why data is central to ECL models and its role in forecasting risk and complying with IFRS 9.

For further reading on how macro inputs affect broader ECL themes, consult: Importance of ECL, ECL data and thematic pieces on ECL & global economy.