For all property types--Office, Industrial, Retail, Multifamily, and Hotel--as well as capital markets fundamentals, we update our history and forecasts on a quarterly basis. For example, at the close of Q4 2016, we update our history for Q4 2016, and our forecasts commence from Q1 2017. Histories are typically published two to three weeks after the close of the quarter. Forecasts are typically published in the following three to four weeks.
Yes. Econometric modeling and forecasting, like any science, typically improves with time. We have repeatedly backtested our forecasting models and enhanced them at appropriate intervals. We continue to evaluate the quality of our forecasting models at regular intervals and inform our clients of additional enhancements if we make them.
Yes. The core of our forecasting methodology reflects standard economic theory that relates new capital formation (or investment) to asset prices. We provide clients with documentation of our forecasting methodology.
For all property types, as well as capital markets fundamentals, our data are compiled from internal CBRE sources, including our local market researchers, as well as external sources. For quality assurance, our data undergo an extensive process of validation before they are used for forecasting. Both our historical data and forecasts pass through a final process of quality assurance before they are posted to our data warehouse.
As an organization, CBRE has vast internal data sources. At CBRE Econometric Advisors, one of our primary objectives is to standardize inputs from a variety of data sources, internal and external, resulting in data series that are appropriate for comparisons over time and between markets. CBRE's local research may utilize specific data-collection methodologies appropriate for their local market, and as a result, there may be differences. As an organization, we strive to identify differences and work toward consistent methodologies to resolve them when possible.
An API is an application protocol interface. APIs are typically used to develop web and mobile applications that network different data platforms for end users. We have spent considerable time and effort developing DataDriver, which essentially serves as an API for our data warehouse.