"From Insufficient Economic data to Economic Big Data – How Trade Data is redefining Reliability in Economic Indicators for Markets" by Tony Nash, Chief Economist and CEO at Complete Intelligence & Mayookh Lad, Head of Analytics at Complete Intelligence
Download as PPTX, PDF0 likes580 views
This document discusses how trade data can provide more reliable economic indicators than traditional economic estimates. It notes that trade data contains a large amount of granular information across countries, products, and time that can be used to generate forecasts. The document describes how a company called Complete Intelligence uses trade data and machine learning to generate consistent economic and financial forecasts with confidence intervals. They aim to provide reliable forecast data to help monitor markets and evaluate strategies.
"From Insufficient Economic data to Economic Big Data – How Trade Data is redefining Reliability in Economic Indicators for Markets" by Tony Nash, Chief Economist and CEO at Complete Intelligence & Mayookh Lad, Head of Analytics at Complete Intelligence
- 1. From Insufficient Economic Data to Economic Big Data How Trade Data is redefining Reliability in Economic Indicators for Markets 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com
- 2. Economic Data or Economic Estimates? • Economic data vs Financial data • How many revisions did major USA economic indicators have, over the last 3 months? • Who uses this data: Traders, Quants, Macro funds, Economists, Media, Corporates and Government institutions • GDP (as a case): • Market sensitivity to GDP figures (first report vs estimate) is quite high • Left or Right side of the equation • Way forward – alternative measures 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 2
- 3. Economic Data or Economic Estimates? 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 3 Source: http://www.tradingeconomics.com/united-states/news
- 4. Trade Data – the internals • Trade Data: More Economic Data and Less Economic Estimate • Speciality of Trade Data - Depth and Granularity • Country x 200 Trading partners x 2 Export/Import x 1400 products x 7 years of monthly data • 25 million data points per country • Quarter million time series per country • Data Structure • Real world functionality • At 50 countries, there are 11 million real world time series • How do we use this? 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 4
- 5. Trade Data – DB sample 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 5 * Sample figures, not actuals
- 6. Trade corridors • Trade corridor as a mathematical function • Correlations: • Equity Indices: • S&P500 (0.96), FTSE100 (0.90), DJIA (0.96), Nikkei225 (0.95), NASDAQ Composite (0.96) • Commodities: • Oil (0.86), Steel (0.84), Aluminium (0.90) • Currencies: • EURUSD (0.85), USDJPY (0.94), USDCNY (0.97) 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 6 * Correlations at HS2 Trade level * Time period over 6 years
- 7. Trade Data - applications • How companies use this data: • Deeper understanding of Supply Chains • Understanding of Trade of products • Using Trade forecasts • Sales Forecast and analytics • Hedging Risk in Treasuries • Forecast of currencies and commodities • Banks, an exception – how they use this information • Global Trade picture • Estimation of Trade Finance 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 7
- 8. Forecasting – a bird’s eye view 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 8 Trade Database Economic forecast Statistical forecast Forecast Trade Database Asset Data (Macro/Currency/ Commodity) Shortlising mechanism (Statistical) Certainty of Forecast Asset Forecast Economic forecast Statistical forecast • Trade Database at HS4, while Forecast Trade DB used as input for shortlisting at HS2 • Calculations PER Asset forecast • Period assumed: since 2010-Jan • Refresh rate = 1 month
- 9. Machine learning • Algorithms used have machine learning capabilities • Previous decisions are stored • Economic and Financial decisions and data over the last 7 years are stored • New data • New data is absorbed as per changes in Economic conditions, Financial price movements as well as new Trade Data at a highly granular level • Data is processed using New Data along with previous decisions • Cycle is repeated once every month, for all Economic indicators, Financial assets and Trade Data • Each forecast asset has not only an accurate forecast, but comes with a confidence level for the prediction • Back-testing indicates lower error rate with higher confidence level 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 9
- 10. CompleteIntel – what we do • Data Analytics • Trade Data • Market Data • Economic and Financial Forecasting • Supply Chain Analysis • Business Structure: • Consulting • B2C Economic Forecasting 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 10
- 11. Who hasn’t used a Google search to get that one piece of data for a presentation? Everyone does it. We all try to locate the exact data we need and get it for free, if possible. Unfortunately, most data are either old, inconsistent or not forecasted. Historical data is pretty easy to find, but it’s tough to get data that is consistent, affordable, regularly updated and comparable across countries. Complete Intelligence uses open data sources and sophisticated big data modelling engines to deliver reliable, frequent, forecast data that can be used to monitor markets, develop your own models, assess economic and industry events, evaluate corporate strategy and more. We have a single global model, which ensures consistency across countries and time series. It also allows us to account for the interdependencies in the global economy. If an economic event happens in Germany, how do you know it’s accounted for in the China forecast? If oil demand surges in India, how do we know it’s accounted for in a global oil price forecast? You can’t unless it’s done with an integrated global model. Like the one we use. 29.Apr.2017 Sat Real forecasts. Precise analytics. For everyone. www.completeintel.com 11 This document contains confidential IP of Complete Intelligence and its client. No part of it may be circulated, quoted, or reproduced for distribution outside the Client without prior written approval from Complete Intelligence; For questions write to: [email protected]