Our platform is the product of blending big data concepts, disparate statistical techniques, and actual factual firm-level credit performance data. Our monthly data set starts from December 2009 and we have year-end snapshots from 2005 onwards. We refresh the firm-level data every month based on the performance of nearly 28 million US businesses.
The platform is comprised of three components — firmographic structure of US businesses, monthly computation of credit performance metrics, and weekly forecasts of creditworthiness. This platform powers the high-frequency, high-resolution indicators we publish concerning consumer & small business credit.
The firmographic structure is constructed using several Census Bureau data sets. It is based on geography at county resolution; industry at 4-digit NAICS industry group resolution; and four business size classes based on headcount. This results in over 2 million segments that track nearly 28 million US firms. Each segment is assigned a weight that reflects its economic impact and is suitable for rolling-up a dimension such as county to state.
Credit performance metrics. Ideally, we would have liked to conduct a statistically valid survey of businesses every month and gauge their credit performance. That is cost prohibitive, if at all possible. As a pragmatic approach, we source monthly data concerning the actual performance of all businesses nationwide from our partners. From this data we derive the incidence and severity of business defaults for each firmographic segment. Defaults arise whenever there is non-performance of a financial obligation across the spectrum of delinquencies, charge-offs, and bankruptcies. Firm-level default events are pooled within segments using weights designed to reflect default severity. A segment’s default rate is computed as the product of the segment default incidence and severity. When required, firmographic segments are aggregated using the weighted harmonic mean.
A default index is a time series of aggregated segment default rates quoted as a seasonally adjusted annualized rate (SAAR). By including select segments, various indices can be constructed such as state-level or business size-based indices. We publish an aggregate and business size-based sub-indices.
A segment’s health summarizes current business credit performance and forecasted trend. The forecast is based on relevant macroeconomic factors including labor market conditions, industrial production, sales, housing market conditions, and financial conditions at both national and sub-national levels depending upon data availability.
A health score provides a point in time ranking of related segments such as states or industry sectors. A health score is obtained by transforming individual segment health values onto a scale that is located at 100 with a scale parameter of 15.