Economic Forecasting Methods

Machine learning beats traditional forecasting methods in multi series forecasting. In one of the latest M forecasting competitions, the aim was to advance what we know about time series forecasting methods and strategies. Competitors had to forecast 40k+ time series representing sales for the largest retail company in the world by revenue: Walmart. These are the main findings: ▶️ Performance of ML Methods: Machine learning (ML) models demonstrate superior accuracy compared to simple statistical methods. Hybrid approaches that combine ML techniques with statistical functionalities often yield effective results. Advanced ML methods, such as LightGBM and deep learning techniques, have shown significant forecasting potential. ▶️ Value of Combining Forecasts: Combining forecasts from various methods enhances accuracy. Even simple, equal-weighted combinations of models can outperform more complex approaches, reaffirming the effectiveness of ensemble strategies. ▶️ Cross-Learning Benefits: Utilizing cross-learning from correlated, hierarchical data improves forecasting accuracy. In short, one model to forecast thousands of time series. This approach allows for more efficient training and reduces computational costs, making it a valuable strategy. ▶️ Differences in Performance: Winning methods often outperform traditional benchmarks significantly. However, many teams may not surpass the performance of simpler methods, indicating that straightforward approaches can still be effective. Impact of External Adjustments: Incorporating external adjustments (ie, data based insight) can enhance forecast accuracy. ▶️ Importance of Cross-Validation Strategies: Effective cross-validation (CV) strategies are crucial for accurately assessing forecasting methods. Many teams fail to select the best forecasts due to inadequate CV methods. Utilizing extensive validation techniques can ensure robustness. ▶️ Role of Exogenous Variables: Including exogenous/explanatory variables significantly improves forecasting accuracy. Additional data such as promotions and price changes can lead to substantial improvements over models that rely solely on historical data. Overall, these findings emphasize the effectiveness of ML methods, the value of combining forecasts, and the importance of incorporating external factors and robust validation strategies in forecasting. If you haven’t already, try using machine learning models to forecast your future challenge 🙂 Read the article 👉 https://buff.ly/3O95gQp

Every FP&A forecasting technique ranked. I've used all of these across P&G, Unilever, and Squarespace. Some are gold. Some are traps. Here's my tier list based on three criteria: → Forecast accuracy over time → How well it supports decision-making → How practical it is to implement and maintain Swipe through to see where each method lands. —-- 💡 Join my free live training: Steal my FP&A Playbook: Get my 6-part framework to become a high-impact FP&A pro https://lnkd.in/e9fEFjmK —-- 📌 E-TIER (Avoid as primary method) • Incremental Approach: Creates bad incentives. Business partners spend every dollar by year-end to protect their baseline. • Market-Based Approach: Sounds sophisticated but you're forecasting the market, not your business. Hides assumptions instead of forcing clarity. 📌 D-TIER (Limited reliability) • Expert Judgment: Experts consider context data can't capture. But optimism bias, recency bias, and groupthink are real. Track accuracy over time to calibrate. 📌 C-TIER (Decent with drawbacks) • Statistical Methods: Fast once set up. But often black-box. When leadership asks "why did we miss?", you can't point to the model. • Time Series Analysis: Good middle ground between judgment and statistics. Gets faster with practice but struggles with newer businesses. 📌 B-TIER (Solid in the right context) • Zero-Based Budgeting: Forces you to challenge status quo. Helps find the 20% of expenses driving 80% of outcomes. But time-intensive and risks short-term bias. • B2B Sales Pipeline: Connects forecast to real opportunities. But reps are often optimistic and it only captures known deals. 📌 A-TIER (Master this) • Driver-Based Forecasting: Identify the 10-15 key drivers that move results. Enables scenario planning. Stays useful all year. This is what I teach at Wharton. 📌 S-TIER (Gold standard) • Statistical Methods + Driver-Based Combined: Statistical models keep experts honest. Expert input catches inflection points models miss. At Squarespace, we tested this and it outperformed either method alone. Which forecasting method does your team rely on most? Drop it below 👇 -Christian Wattig P.S.: Don't miss my next free live training - I cover the same FP&A framework I teach at Wharton Online: https://lnkd.in/e9fEFjmK

Stock price forecasting is difficult because prices are driven by many external forces, like macroeconomics, policies, company fundamentals, and investor sentiment, which make the series noisy, unstable, and hard to model. Machine- and Deep-Learning (ML/DL) are widely used for stock price analysis due to their higher predictive accuracy compared to traditional statistical and econometric methods. Although DL can capture nonlinear, high-dimensional market patterns, its effectiveness depends on having large datasets. In practice, daily stock data are limited, especially for IPOs, and standard data-augmentation techniques used in computer vision cannot be applied because they break temporal order. This creates a data-scarcity problem that weakens model performance. Beyond limited data, stock prices also contain intertwined components such as trends, cycles, and randomness. Single- or multi-scale decomposition methods break down a signal such as a time series, stock price data, or a sound wave, etc, into components, each representing the signal's behavior at a different scale or level of detail. These subseries, however remain correlated in practice. Existing models treat them independently and ignore these cross-relationships, losing valuable predictive information. To address both data limitations and structural complexity, the authors of [1] proposes a combined TimeGAN + decomposition learning framework ('TimeGAN + SSA + LSTM'). Multi-view market data (open, high, low, close, volume) are first used to train a TimeGAN model, which generates realistic synthetic sequences to expand the dataset. The closing-price series is then decomposed using SSA (singular spectrum analysis) into smoother subseries, and an LSTM extracts temporal features from each. A self-attention mechanism captures the interactions among correlated subseries, and the fused representation is further enhanced by modelling its dependencies with other market-feature series. A final LSTM produces the closing-price prediction. Experiments were conducted on nearly a decade of data from multiple international stock indices: the U.S. S&P 500 (SP500), China’s CSI 300 (CSI300), Japan’s Nikkei 225 (N225), and the U.K.'s FTSE 100 (FTSE100). The results demonstrate that the proposed 'TimeGAN + SSA + LSTM' integrated approach, combining data augmentation, decomposition, and inter-series attention, achieves superior prediction accuracy (RMSE) and superior Sharpe-Ratio (SR) compared to other advanced baselines (BP, LSTM, VMD-LSTM, N-Beats, SCINet, DLinear, MLSF and MASTER). #QuantFinance The link to the paper is available in the comments.

Inventory planning isn’t just about stock. It’s about balancing demand, supply, operations, and cash flow, at scale. A strong inventory strategy ensures the right products reach the right place at the right time, without locking capital or creating waste. Here’s what a complete inventory planning framework typically covers: 🔹 Why Inventory Planning Matters Drives customer satisfaction, reduces disruptions, improves operational efficiency, and protects margins through smarter stock decisions. 🔹 Inventory Planning Process Starts with historical demand analysis, moves through forecasting, safety stock, reorder points, cross-team collaboration, and continuous monitoring. 🔹 Planning Methods & Models Uses ABC/XYZ classification, FIFO rotation, MOQ, EOQ, and demand-driven planning to match inventory levels with real business needs. 🔹 Role of Data Sales history, stock levels, supplier lead times, demand trends, and forecast accuracy power every planning decision. 🔹 Key Goals Maintain service levels, reduce excess inventory, free working capital, stabilize operations, and support scalable growth. 🔹 Key Inventory KPIs Service level, stock turns, forecast accuracy, working capital, and excess inventory guide performance tracking. 🔹 Tools & Automation Demand forecasting, automated replenishment, exception management, dashboards, and reporting turn planning into an ongoing system. 🔹 Best Practices Accurate master data, ERP integration, continuous model refinement, exception-based management, and strong cross-team alignment. 🔹 Real-World Applications From industrial supplies to electronics, each category applies different planning rules based on demand patterns and lead times. Inventory planning isn’t a back-office function anymore. It’s a strategic capability that connects supply chains to business outcomes. When done right, it transforms uncertainty into predictable growth.

Because with a bad forecast everything else will fail... This infographic contains 7 steps to create and improve a forecast: ✅ Step 1 - Start with Historical Data Collection & Cleaning 👉 gather and clean past sales data (ideally 3 years) 👉 remove outliers, fill in gaps, and ensure data accuracy before analysis ✅ Step 2 - Segment Your Demand 👉 break down your demand into segments to create more granular forecasts 👉 examples: volume, value, product categories, customer types, regions ✅ Step 3 - Generate a Baseline Statistical Forecast 👉 as starting point, generate a baseline forecast using statistical methods like time series analysis ✅ Step 4 - Apply Seasonality and Trend Adjustments 👉 use historical seasonal patterns and emerging trends to fine-tune your forecast for upcoming periods ✅ Step 5 - Collaborate & Fine-tune in S&OP Meetings 👉 collaborate with sales, marketing, finance, and operations to align on one consensus forecast ✅ Step 6 - Adjust for Market Intelligence 👉 incorporate insights from sales teams, marketing campaigns, external research, and product launches to adjust your baseline forecast ✅ Step 7 - Incorporate Forecasts into S&OE (Sales & Operations Execution) 👉 drive actionability in the short term based on this aligned forecast, helping the team respond quickly to deviations 💥 Bonus Step: Build a Continuous Feedback Loop 👉 track forecast accuracy by comparing actual sales to forecasted figures, and regularly update your model based on this feedback Any other steps to consider? #supplychain #salesandoperationsplanning #integratedbusinessplanning #procurement

Your SKU-level forecast has 42% error, but the exec dashboard shows only 15%. How’s that possible? This isn’t a data error. It’s a statistical truth—and a hidden power of demand planning most teams underestimate: Forecast accuracy improves with aggregation. Here’s why At granular levels like SKU-Store-Week, your forecast is vulnerable to noise—local promotions, competitor actions, supply delays, even weather. But as you move up—SKU to Product, Store to Region, Week to Month—the randomness starts to cancel out. What you get is a clearer, more stable signal. In one of my projects, here’s what I observed:  SKU-Store-Week: 42% MAPE  SKU-Region-Month: 31% MAPE  Product-Region-Month: 22% MAPE  Product-National-Month: 15% MAPE This trend isn’t accidental. It’s due to the law of large numbers in statistics, where larger sample sizes smooth out volatility - often referred to as noise cancellation. But here’s the catch: While higher-level forecasts are more accurate, they lack the granularity necessary for operational decisions, such as replenishment. The key is to build your forecasting hierarchy smartly: - Aggregate where stability matters (e.g., financial planning) - Disaggregate where action is needed (e.g., order fulfilment) Always balance visibility with accuracy. Not everything needs to be forecasted at the lowest level. How do you design your aggregation hierarchy in demand planning? Would love to hear your approach. #DemandPlanning #SupplyChain #Forecasting

This is a demand planning knowledge base built on >50 demand planning projects and many more conversations and readings (the link is in the comments). It should be a good starting point for beginners but also covers more advanced topics, and it includes an overview of software providers as well (no. 10 in the advanced section). The list of topics covered in the demand planning fundamentals: 1. Demand Planning basics 2. Data for Demand Planning 3. Forecast accuracy, bias and forecast value add 4. The link between demand planning and sales forecasting 5. Artificial Intelligence (AI) in demand planning And in the advanced section: 1. Demand segmentation & classification 2. Advanced data sources, leading indicators, outside-in planning and causal forecasting 3. Product lifecycle planning/Portfolio management 4. Forecasting methods for baseline generation 5. Understanding forecasting hierarchy and levels: aggregation, disaggregation, and manual adjustments 6 & 7. Order/forecast consumption & demand sensing 8. Company specifics to take into account in demand planning 9. Demand planning in the organization: which department should own it? 10. Finding the right demand planning tool & overview of software providers This knowledge base often links to strong articles/videos written by other people and companies - some important ones I like to mention: Lora Cecere, Nicolas Vandeput, Ivan Svetunkov, Institute of Business Forecasting & Planning Arkieva, John Galt Solutions, Slimstock, Logility, o9 Solutions, Inc., Kinaxis, OMP If you have any feedback, suggestions, or mistakes that you found, I’ll be happy to hear about them! Other parts of supply chain planning will be included soon. #supplychain #planning #knowledgebase

I see Finance teams spending days and weeks building Excel forecasts that break the moment business patterns shift. There's a better way. I just published a walkthrough showing how to implement 𝗠𝗟-𝗯𝗮𝘀𝗲𝗱 𝗳𝗼𝗿𝗲𝗰𝗮𝘀𝘁𝗶𝗻𝗴 𝗶𝗻 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗙𝗮𝗯𝗿𝗶𝗰 - achieving >95% accuracy with a setup that takes hours, not the days/weeks Excel requires. Once configured in Fabric notebooks, forecasts refresh automatically. No more monthly Excel gymnastics. CFOs get conservative/baseline/stretch scenarios from the same model. And it adapts to trend changes without manual recalibration. The approach works beyond AR (Accounts Receivable) - I've used similar frameworks for sales forecasting, inventory planning, and capacity projections across Telco, Oil & Gas, and Pharma clients. 𝗪𝗵𝗮𝘁 𝘁𝗵𝗲 𝘁𝘂𝘁𝗼𝗿𝗶𝗮𝗹 𝗰𝗼𝘃𝗲𝗿𝘀:  • Prophet framework for automatic seasonality detection  • 12-month cash flow predictions with confidence intervals for scenario planning  • Lakehouse integration for automatic Power BI refresh  • Cross-validation workflow that tunes parameters automatically 𝗥𝗲𝗮𝗹 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆 𝗺𝗲𝘁𝗿𝗶𝗰𝘀: With the sample data I was able to achieve 3% MAPE (Mean Absolute Percentage Error) - that's $50K average variance on $1.5M monthly collections. Industry target is under 5%. 𝗧𝘂𝘁𝗼𝗿𝗶𝗮𝗹 𝗮𝗻𝗱 𝗻𝗼𝘁𝗲𝗯𝗼𝗼𝗸 𝗹𝗶𝗻𝗸 𝗶𝗻 𝗰𝗼𝗺𝗺𝗲𝗻𝘁𝘀 🎥👇 ____ #MicrosoftFabric #PowerBI #MachineLearning #DataAnalytics #Forecasting 

"One-week-ahead electricity price forecasting using weather forecasts, and its application to arbitrage in the forward market: an empirical study of the Japan Electric Power Exchange " by Takuji Matsumoto Misao Endo 2021 in Risk.net Abstract: Although forecasting one-week-ahead average electricity prices is necessary for decision-making such as evaluating forward contracts, its modeling has not been sufficiently studied. Therefore, to find a suitable forecasting approach, this study constructs and compares multiple parsimonious models using widely published weekly weather forecasts and then applies them to arbitrage trading in the forward market. In particular, we clarify the following empirical results using the data from Japan Electric Power Exchange. First, instead of using forecasted temperature directly as an explanatory variable, the two-step forecasting method using measured temperature as an intermediate variable is more likely to reduce forecast errors. Second, quantile regression has better density forecast accuracy than the generalized autoregressive conditional heteroscedasticity model. Third, the logarithmic conversion for prices tends to improve forecast accuracy. Fourth, one-week-ahead weather forecasts can significantly improve both the price forecast accuracy and the arbitrage profit. The proposed arbitrage strategy can be used by many participants because it can be flexibly changed according to the player’s risk tolerance. In addition, our forecasting/trading method, based on published weather forecasts, has wide applicability in that it can be constructed even in markets where system information is not sufficiently disclosed.

Sales and Stock Forecasting: In FMCG In the context of Fast-Moving Consumer Goods (FMCG), SSF usually stands for Sales and Stock Forecasting. This involves predicting future sales and managing stock levels to meet customer demand efficiently. Here’s how SSF plays a critical role in FMCG: 1. Sales Forecasting - Demand Prediction: Estimating future customer demand based on historical sales data, market trends, seasonal variations, and promotional activities. - Promotional Impact: Forecasting the impact of upcoming promotions, discounts, and marketing campaigns on sales volumes. - Sales Channel Performance: Predicting sales across different channels (e.g., retail, online, wholesale) to allocate resources and stock effectively. 2. Stock Forecasting - Inventory Management: Ensuring that the right amount of stock is available to meet predicted sales without overstocking, which can lead to high holding costs, or understocking, which can lead to stockouts and lost sales. - Supply Chain Coordination: Aligning stock levels with supply chain activities to ensure timely replenishment, considering lead times, supplier reliability, and logistics. - Shelf Space Optimization: Managing stock levels to optimize shelf space in retail environments, ensuring high-demand items are always available and minimizing space for low-turnover products. 3. Key Factors in SSF for FMCG - Data Accuracy: High-quality, up-to-date data is essential for accurate forecasting. This includes sales data, market trends, consumer behavior, and external factors like economic conditions. - Technology Integration: Utilizing advanced forecasting software and data analytics tools to enhance the accuracy and efficiency of SSF processes. - Collaboration: Cross-functional collaboration between sales, marketing, supply chain, and finance teams to ensure alignment on forecasts and strategies. - Agility and Flexibility: The ability to quickly adjust forecasts and stock levels in response to changing market conditions, unexpected events, or shifts in consumer behavior. 4. Benefits of Effective SSF in FMCG - Improved Customer Satisfaction: Ensuring that products are available when customers want them, reducing the risk of stockouts. - Cost Efficiency: Minimizing overstocking and associated costs while avoiding lost sales due to understocking. - Enhanced Decision-Making: Providing data-driven insights that inform sales strategies, marketing campaigns, and supply chain planning. - Competitive Advantage: Being able to react swiftly to market changes and consumer trends, giving the company an edge over competitors. In FMCG, where products move quickly and customer preferences can change rapidly, effective SSF is crucial for maintaining smooth operations and achieving sales targets.