Anyone who has graduated with a business degree, worked in finance, or held a senior management position has general knowledge of regression analysis forecasting techniques, weighting components, and scenario testing.
Few are aware of leading indicator inputs. Even fewer are aware of how to conduct timing analyses using the leading indicator inputs. And fewer still use business cycle theory.
But if ITR Economics' forecasting accuracy is unparalleled in the industry thanks to this methodology, why is it not adopted by more firms?
Because it is unique to ITR Economics. The system was developed in the 1940s, has been refined over the years, and provides unparalleled results. Just because it isn’t mainstream or commonly utilized doesn't make it inferior. In fact, the data shows us that this type of methodology provides superior forecasts.
ITR’s Forecasting Explained by Our CEO…
ITR Economics’ success in forecasting is a result of our proprietary forecast modeling that emphasizes the value of separating the timing aspect of trend reversals from the quantitative aspects of current and future trends. “Subject series” as used below refers to the data stream being forecasted. It could be a company, an industry or market, or an entire economy.
Timing trend reversals is inherent to all of our forecasts. The forecast solution of the timing analysis is derived from a working construct of business cycle change derived from ITR’s business cycle theories, a statistical analysis of the subject data series’ endogenous trend characteristics (cyclical and seasonal), and lead times of selected statistically significant leading indicators to the subject series. The endogenous trend characteristic input is derived from ITR’s analysis of the subject series’ monthly data, 3-month moving total/average (3MMT/A), 12-month moving total/average (12MMT/A), and all associated rates-of-change of the data.
Regarding leading indicators, ITR uses the relationship an indicator will have to the subject series as a weighting input into resulting timing analysis, but this is not the exclusive criteria.
Timing analysis also encompasses studying the relationship between non-leading indicators (such as industry and/or market series) and the subject data series. ITR will use well defined relationships and our existing forecasts of the relevant industries and markets to further define the timing of probable trend reversals for the subject series.
The timing of trend reversals is absolutely imperative to accurate forecasting because economic trends are not linear. Indeed, change in the rate of growth/decline is a constant, and we excel at defining, measuring, and projecting that change.
It is our understanding of macroeconomic and microeconomic business cycles, and then defining, measuring, and correlating the inherent cyclicality of the subject series, which enables us to have superior results compared to normal econometric modeling.
Quantitative forecasting encompasses the timing analysis summarized above and adds quantitative value expectations by quarter through the forecast horizon. The statistical technique for the quantitative forecast relies on a proprietary matrix created by ITR Economics.
This matrix enables us to combine the monthly, quarterly, seasonal, and cyclical measures of the subject series with a subjective weighting applied each quarter based on: statistic reliability and either causal or business cycle related precedent(s) for both current and future trends.
Business cycle related precedents are determined using ITR theory input and/or derived statistical relationships and analyst experience. We also define near-term quantitative trend probabilities using ITR’s Universality of Cycles Concept™. ITR’s quantitative forecasts of related extant variables will also be factored into the forecast to further narrow the range of trend probabilities.
For each quarter, the myriad of forecast inputs (the exact number of inputs varies from one subject series to the next) is compiled into a resulting quantitative median input. The median input is frequently (but not exclusively) used as a quarterly focal point for the forecast.
Exceptions to the use of the median input can occur because of unexplained anomalous activity in the subject data series, one-off events driven by business decisions within the subject series if it is a company and explained abrupt external changes that impact the subject series.
Determining causal and otherwise logically related exogenous inputs into the forecast comes from knowing the subject data series (i.e., the company and its industry/markets, with a correlation analysis run against the industry/market variables).
Because the quantitative forecast is on a timing analysis platform and because ITR has developed business cycle expectations relative to the subject series, our forecast results may very well differ from standard econometric modeling efforts.
ITR’s forecasts are an experienced blend of statistical analysis and the art of business cycle theory within the context of applied economics.