Peter Hingley

The extent to which national research and development expenditures affect first patent filings in contracting states of the European Patent Convention

The objective of this paper is to investigate the relationship between R&D expenditures and the numbers of patent applications coming to contracting states of the European Patent Convention (EPC). Annualised data on patent application filings and R&D expenditures from 1963 to 1993 were collected for each state. Regression models were constructed to relate domestic first filings (DF) in contracting states of the European Patent Convention (EPC) to R&D expenditure in those states. The models gave an indication of the time-lag effect of R&D on DF, as well as on the nature of the effect (data for Germany were restricted to the former West Germany). Two types of regression models were used. Time-series analysis transfer functions were successfully fitted to 10 of the 13 states, with an average time lag estimated as 3.6 years. Linear regression analysis also gave acceptable models when fitted to the whole data sets, but only in six of the 13 states. However, if linear regression was restricted to DF data from 1980 onwards, acceptable models were found for 11 of the 13 states, with an average lag estimated as 2.8 years. When the most appropriate model was selected for each state, the average delay was estimated as 3.7 years. Comparison of the forecasts with actual results suggests that the method is reasonably accurate.

Forecasting Patent Filings at the European Patent Office (EPO) with a Dynamic Log Linear Regression Model: Applications and Extensions

We describe the application of a structural econometric model to forecast annual and quarterly total filings of patents at the EPO, amid cyclical shocks. Such models are applicable for national, regional, and international patent offices for budgetary planning purposes. We also provide out-of-sample forecasts out to 2019. The exercise predicts strong further growth in EPO filings from China. The forecasts appear largely optimistic out to 2019. Other findings are that quarterly data do not exhibit as much sensitivity to cyclical influences as do annual data, consistent with the view that innovation decisions are made based upon longer term factors. Furthermore, filings respond more directly to market fluctuations rather than indirectly to deviations in research and development (R&D) spending from trend. The estimates also vary by technological field. Further possibilities for an extension of the approach are discussed, including the usage of country-to-country correlations of forecasts to unearth communalities in patenting behaviour between countries.