Study guide
10.3 Applied Causal Forecasting: Regression, Correlation & Leading Indicators
ISM CPSM Exam 2 — Supply Management Integration Exam
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: Category Management Strategy & Spend Analysis
1.1 Category Strategy Development & Stakeholder Alignment 1.2 Spend Analysis, Data Categorization & Spend Trees 1.3 Supply Market Intelligence & Porter's Five Forces 1.4 Category Segmentation & Kraljic Portfolio Matrix
2 Chapter 2: Strategic Sourcing Waves & Supplier Relationship Management
2.1 7-Step Sourcing Wave Execution & RFx Process 2.2 Total Cost of Ownership (TCO) & Landed Cost Modeling 2.3 Supplier Evaluation, Weighted Scoring & Award Strategy 2.4 Supplier Relationship Management (SRM) & KPI Scorecards
3 Chapter 3: Supply Chain Network Strategy & Resilience
3.1 Supply Chain Network Design & Facility Location 3.2 Make-or-Buy & Insourcing vs Outsourcing Analysis 3.3 Supply Chain Risk Management, Resilience & Continuity 3.4 Sustainable, Ethical & Diverse Supply Chain Integration 3.5 Global Sourcing Footprint, Regionalization & Trade-Policy Strategy
4 Chapter 4: Supply Chain Strategy & Digital Transformation
4.1 Functional Alignment, SCOR Model & Value Chain Strategy 4.2 Agile vs Lean Supply Chains & Postponement Strategies 4.3 Digital Supply Chain Systems: ERP, WMS, TMS & AI
5 Chapter 5: Logistics & Transportation Management
5.1 Transportation Modes, Selection Analysis & Trade-offs 5.2 Freight Rating, Carrier Contracting & 3PL/4PL Management 5.3 Incoterms® 2020 Rules, Customs & Global Trade Compliance 5.4 Logistics Fundamentals, Customer Service & Total Logistics Cost 5.5 International Freight Operations: Ocean, Air, Intermodal & Forwarders 5.6 Export Controls, Trade Security Programs & Shipping Documentation
6 Chapter 6: Warehousing, Distribution & Reverse Logistics
6.1 Warehouse Operations, Layout & Slotting Optimization 6.2 Distribution Network Models, Cross-Docking & Hubs 6.3 Reverse Logistics, Remanufacturing & Circular Supply Chains 6.4 Materials Handling, Packaging & Unitization 6.5 Warehouse & Logistics Technology: WMS, TMS, Auto-ID & Automation 6.6 Logistics Performance Measurement, KPIs & Carrier Scorecards
7 Chapter 7: Demand Forecasting & Sales and Operations Planning (S&OP)
7.1 Demand Forecasting Principles & Quantitative Models 7.2 Sales and Operations Planning (S&OP) & CPFR 7.3 Demand Management, Volatility & Bullwhip Effect Mitigation 7.4 Aggregate Planning, Capacity Balancing & Master Scheduling
8 Chapter 8: Inventory Management & Control Policy
8.1 Inventory Types, Carrying Costs & Working Capital 8.2 Inventory Control Models: EOQ, Reorder Point & Safety Stock 8.3 ABC Analysis, VMI, Consignment Inventory & JIT Systems 8.4 MRP, Bills of Material & Distribution Requirements Planning 8.5 Inventory Positioning, Risk Pooling & Multi-Echelon Policy 8.6 Inventory Record Accuracy, Cycle Counting & Excess Disposition
9 Chapter 9: Project Management Integration & Governance
9.1 Project Charter, Scope Management & Work Breakdown Structure 9.2 Project Scheduling, Critical Path Method (CPM) & PERT 9.3 Project Risk Management, Earned Value Management (EVM) & RACI
10 Chapter 10: Forecasting Methods, Models & Accuracy
10.1 Forecast Horizons, Hierarchy & the Aggregation Principle 10.2 Seasonality, Decomposition & Seasonal Index Forecasting 10.3 Applied Causal Forecasting: Regression, Correlation & Leading Indicators 10.4 Managing Forecast Accuracy: Bias, Weighted Error & Forecast Value Added 10.5 Demand Sensing, Downstream Data & Machine-Learning Forecasting 10.6 New-Product Forecasting: Analogs, Attributes & Diffusion Models 10.7 Supply-Side Forecasting: Lead Time, Commodity Price & Capacity
11 Chapter 11: Product & Service Development in S&OP
11.1 New Product & Service Development and the Stage-Gate Process 11.2 Early Supplier Involvement, Co-Design & Supplier-Led Innovation 11.3 Specifications, Statements of Work & Standardization 11.4 Value Analysis, Value Engineering & Design for Manufacturability 11.5 Product Life Cycle, Phase-In/Phase-Out & Obsolescence Management
12 Chapter 12: Quality Management & Continuous Improvement
12.1 Quality Fundamentals, Cost of Quality & the Quality Philosophies 12.2 Quality Management Systems, ISO Standards & Registration 12.3 TQM, Six Sigma, Lean & Continuous Improvement Methods 12.4 Statistical Process Control, Process Capability & Acceptance Sampling 12.5 Supplier Quality Management, Certification Levels & the Quality Tools
10.3 Applied Causal Forecasting: Regression, Correlation & Leading Indicators
Key Takeaways Explain This Topic AI
The correlation coefficient r ranges from -1 to +1 and measures direction and strength; the coefficient of determination R-squared is r squared and states the share of variation in demand explained by the model.
Correlation is not causation, and a high R-squared on historical data guarantees nothing about forecast accuracy outside the observed range.
A leading indicator must move before demand and be published in time to act on — an indicator that moves simultaneously with demand has no forecasting value.
Multiple regression adds independent variables but risks multicollinearity when those variables are correlated with each other, which destabilizes the coefficients.
Causal models earn their keep at medium and long horizons and at turning points, where time-series extrapolation is structurally blind.
Last updated: August 2026
Applied Causal Forecasting: Regression, Correlation & Leading Indicators
Time-series methods answer "what has this series been doing?" Causal (associative) models answer "what drives this series, and what are those drivers about to do?" That difference matters most exactly where time series fail — at turning points and over long horizons . Exam 2 tests both the arithmetic and the judgment.
Simple Linear Regression — Fully Worked
The model fits a straight line Y = a + b X Y = a + bX Y = a + b X , where Y Y Y is demand (the dependent variable) and X X X is the driver (the independent variable), by minimizing the sum of squared vertical distances between the observations and the line — hence least squares . b = n ∑ X Y − ∑ X ∑ Y n ∑ X 2 − ( ∑ X ) 2 a = ∑ Y − b ∑ X n = Y ˉ − b X ˉ b = \frac{n\sum XY - \sum X \sum Y}{n\sum X^2 - (\sum X)^2} \qquad a = \frac{\sum Y - b\sum X}{n} = \bar{Y} - b\bar{X} b = n ∑ X 2 − ( ∑ X ) 2 n ∑ X Y − ∑ X ∑ Y a = n ∑ Y − b ∑ X = Y ˉ − b X ˉ
Scenario: A supplier of commercial HVAC filters believes demand tracks regional non-residential construction permits issued two quarters earlier.
- Period Permits (X, hundreds) Filter demand (Y, thousands) XY X² Y²
- 1 4 26 104 16 676
- 2 6 32 192 36 1,024
- 3 5 30 150 25 900
- 4 9 43 387 81 1,849
- 5 8 39 312 64 1,521
- 6 10 46 460 100 2,116
- Σ (n = 6) 42 216 1,605 322 8,086
Step 1 — Slope. b = 6 ( 1,605 ) − ( 42 ) ( 216 ) 6 ( 322 ) − ( 42 ) 2 = 9,630 − 9,072 1,932 − 1,764 = 558 168 = 3.321 b = \frac{6(1{,}605) - (42)(216)}{6(322) - (42)^2} = \frac{9{,}630 - 9{,}072}{1{,}932 - 1{,}764} = \frac{558}{168} = \mathbf{3.321} b = 6 ( 322 ) − ( 42 ) 2 6 ( 1 , 605 ) − ( 42 ) ( 216 ) = 1 , 932 − 1 , 764 9 , 630 − 9 , 072 = 168 558 = 3.321
Step 2 — Intercept. X ˉ = 42 / 6 = 7.0 Y ˉ = 216 / 6 = 36.0 \bar{X} = 42/6 = 7.0 \qquad \bar{Y} = 216/6 = 36.0 X ˉ = 42/6 = 7.0 Y ˉ = 216/6 = 36.0 a = 36.0 − ( 3.321 ) ( 7.0 ) = 36.0 − 23.25 = 12.75 a = 36.0 - (3.321)(7.0) = 36.0 - 23.25 = \mathbf{12.75} a = 36.0 − ( 3.321 ) ( 7.0 ) = 36.0 − 23.25 = 12.75
Step 3 — The forecasting equation. Y ^ = 12.75 + 3.321 X \hat{Y} = 12.75 + 3.321X Y ^ = 12.75 + 3.321 X
Step 4 — Forecast. If permits two quarters ago were 12 (hundred): Y ^ = 12.75 + 3.321 ( 12 ) = 12.75 + 39.85 = 52.6 thousand filters \hat{Y} = 12.75 + 3.321(12) = 12.75 + 39.85 = \mathbf{52.6\ \text{thousand filters}} Y ^ = 12.75 + 3.321 ( 12 ) = 12.75 + 39.85 = 52.6 thousand filters
Step 5 — Interpret the coefficients in business language.
Slope b = 3.321 b = 3.321 b = 3.321 : each additional hundred permits is associated with roughly 3,321 additional filters . This is the number a category manager uses to size a capacity commitment.
Intercept a = 12.75 a = 12.75 a = 12.75 : about 12,750 filters of demand exist independent of new construction — replacement and maintenance demand. Note the caution: the intercept is only meaningful if X = 0 X = 0 X = 0 falls within a plausible range of the data.
Correlation and the Coefficient of Determination r = n ∑ X Y − ∑ X ∑ Y [ n ∑ X 2 − ( ∑ X ) 2 ] [ n ∑ Y 2 − ( ∑ Y ) 2 ] r = \frac{n\sum XY - \sum X \sum Y}{\sqrt{\left[n\sum X^2 - (\sum X)^2\right]\left[n\sum Y^2 - (\sum Y)^2\right]}} r = [ n ∑ X 2 − ( ∑ X ) 2 ] [ n ∑ Y 2 − ( ∑ Y ) 2 ] n ∑ X Y − ∑ X ∑ Y
Continuing the example:
- Numerator = 558 558 558 (computed above)
- n ∑ X 2 − ( ∑ X ) 2 = 168 n\sum X^2 - (\sum X)^2 = 168 n ∑ X 2 − ( ∑ X ) 2 = 168
n ∑ Y 2 − ( ∑ Y ) 2 = 6 ( 8,086 ) − ( 216 ) 2 = 48,516 − 46,656 = 1,860 n\sum Y^2 - (\sum Y)^2 = 6(8{,}086) - (216)^2 = 48{,}516 - 46{,}656 = 1{,}860 n ∑ Y 2 − ( ∑ Y ) 2 = 6 ( 8 , 086 ) − ( 216 ) 2 = 48 , 516 − 46 , 656 = 1 , 860
r = 558 168 × 1,860 = 558 312,480 = 558 559.0 = 0.998 r = \frac{558}{\sqrt{168 \times 1{,}860}} = \frac{558}{\sqrt{312{,}480}} = \frac{558}{559.0} = \mathbf{0.998} r = 168 × 1 , 860 558 = 312 , 480 558 = 559.0 558 = 0.998
R 2 = ( 0.998 ) 2 = 0.996 R^2 = (0.998)^2 = \mathbf{0.996} R 2 = ( 0.998 ) 2 = 0.996
Reading these numbers:
Statistic Range Meaning
r −1 to +1 Direction and strength of the linear relationship. r = 0 r = 0 r = 0 means no linear relationship — a strong curved relationship can still produce r r r near zero.
R² 0 to 1 Share of variation in Y Y Y explained by the model. Here 99.6% of filter-demand variation is explained by permit activity.
Exam trap: a high R 2 R^2 R 2 describes fit to history , not forecast accuracy. Adding variables always raises R 2 R^2 R 2 mechanically, which is why adjusted R 2 R^2 R 2 and out-of-sample validation exist. A model with R 2 = 0.98 R^2 = 0.98 R 2 = 0.98 built on six observations and eight variables has fitted noise.
Leading, Coincident and Lagging Indicators
A causal model is only useful if the driver is known before the demand it predicts .
Type Timing relative to demand Forecasting value Examples
Leading Moves before High — this is what you want Building permits, new orders indices, consumer confidence, order backlog, quote and bid activity
Coincident Moves at the same time None for forecasting; useful for confirming the current state Industrial production, employment, retail sales
Lagging Moves after None for forecasting; useful for confirming a turning point occurred Unemployment duration, inventory-to-sales ratio, average prime rate
Validating a candidate indicator — the practical checklist:
Is there a plausible causal mechanism? Without one, the correlation is probably spurious.
Is the lead time stable? A driver that leads demand by two quarters sometimes and five quarters at other times is not usable.
Is it published in time? An indicator with a three-month reporting delay and a two-month lead provides no actionable warning.
Is it revised? Many economic series are revised substantially after first release; a model built on first prints behaves differently in production.
Does it hold out of sample? Fit on one period and test on a period the model never saw.
For supply management specifically, the ISM Report On Business and its Purchasing Managers' Index are the classic leading indicators of industrial activity, with the new-orders sub-index leading production. Sector-specific leaders include housing starts and permits for building products, vehicle production schedules for automotive components, and rig counts for oilfield supply.
Multiple Regression and Multicollinearity
Multiple regression extends the model to several drivers: Y ^ = a + b 1 X 1 + b 2 X 2 + ⋯ + b k X k \hat{Y} = a + b_1X_1 + b_2X_2 + \cdots + b_kX_k Y ^ = a + b 1 X 1 + b 2 X 2 + ⋯ + b k X k
Example: filter demand as a function of construction permits, average regional temperature, and a regulatory air-quality index.
Multicollinearity occurs when the independent variables are strongly correlated with each other — for instance, including both GDP and industrial production. The model may still fit well overall, but the individual coefficients become unstable and can even take the wrong sign, so any interpretation of an individual driver's effect becomes unreliable. Remedies: drop one of the correlated variables, combine them into a single index, or collect more data.
Other standard cautions:
Overfitting — too many variables relative to observations. A workable rule of thumb is at least ten observations per independent variable.
Extrapolation beyond the observed range — the fitted relationship has no evidence outside the data. Forecasting demand at X = 40 X = 40 X = 40 from data spanning X = 4 X = 4 X = 4 to X = 10 X = 10 X = 10 is extrapolation, not prediction.
Structural breaks — a new competitor, a regulation, or a pandemic changes the relationship itself. Historical coefficients then describe a world that no longer exists.
Autocorrelation — successive residuals correlate in time-series regressions, which understates the true standard errors and makes the model look more reliable than it is.
When to Use Causal Models in Supply Management
- Situation Use causal? Why
- Weekly replenishment of a mature SKU No Time series is cheaper, faster, and adequate
- Sizing a three-year supplier capacity commitment Yes The horizon exceeds what extrapolation can support
Anticipating a demand turning point in a cyclical industry Yes Time series is structurally blind to turning points
Forecasting a commodity input price Yes Price is driven by identifiable supply and demand fundamentals
Forecasting a brand-new product with no history No No historical relationship exists to fit; use analogy and judgment
Justifying a network redesign Yes Long-horizon structural decision requiring scenario analysis
The strongest practical answer on Exam 2 is usually a blend : a statistical baseline for the short horizon, a causal overlay for the medium horizon, and structured judgment applied through the S&OP consensus process — with each adjustment documented so that forecast value added can be measured afterward.