Presentation on theme: "CBI Regional Trends Survey: Innovation Question Analysis Ciaran Driver, Tanaka Business School Imperial College, University of London Christine Oughton,"— Presentation transcript:
CBI Regional Trends Survey: Innovation Question Analysis Ciaran Driver, Tanaka Business School Imperial College, University of London Christine Oughton, Management Department, Birkbeck College, University of London
CBI Regional Trends Survey Analysis: Objectives Are there significant differences between regions in the way in which innovation responds to its determinants? Are the effects of one regions activities on another observable, and do these linkages vary across regions?
CBI Regional Trends Survey: Innovation Question – Data Source The data source is the CBI-Experian Regional Trends Survey. Quarterly data on innovation and other variables, with the innovation data dating from Data allows us to observe changes in innovation over time and across regions.
CBI Regional Trends Survey: Regional Data Individual responses are aggregated to the regional level using a regional weighting matrix of value added weights for six industrial groups, and three size groups, with the simplification that the size distribution is assumed similar across the UK.
Innovation question (17a): Do you intend to authorise more or less expenditure in the next twelve months than was authorised in the past twelve months on product and process innovation? These balance replies are interpretable as rates of growth of the underlying variable ( Driver and Urga 2004) CBI Regional Trends Survey: Innovation Question
Ino for South_East, Scotland and East-Midlands
Product and process research + Innovation through Market Research and Design Thus it is a broad innovation indicator What does the Innovation Question Capture?
0.75 EM 0.74 EE 0.55 NE 0.69 NW 0.72 SC 0.87 SE 0.79 SW 0.80 WA 0.86 WM 0.78 YH Regional correlation between: Investment and Innovation intentions
Granger causation from investment intentions to innovation intentions at 10% significance for six out of ten regions Regional correlation between: Investment and Innovation Intentions
Optimism Regressor Are you more, or less, optimistic than you were four months ago about the general business situation in your industry? We define the variable opt as the balance statistic (% responding more minus % responding less) Explaining Variation in Innovation Responses
Effects on Ino of regional and national training for each region
Sub Panel Results for Extended Specification
Dominance of national-level training In regions that have high levels of general qualifications (A-levels or degrees), own-region training is not significant Panel Results: Interpretation
National training variable may not be stationary ADF tests and autocorrelation function suggest non-stationarity Caveat 1
Significance for Own-Region Training when National Training Excluded
Caveat 2 Basic equation potentially mis-specified in that innovation experiences a downward step break around 1999 Introducing the regional training variable removes the need for a break dummy But is regional training truly exogenous?
CBI Regional Trends Survey Analysis: Objectives - Recall Are there significant differences between regions in the way in which innovation responds to its determinants? Are the effects of one regions activities on another observable, and do these linkages vary across regions?
Effects of regional innovation on unit cost A regional model of unit cost Is national or regional innovation more important?
Does innovation affect unit cost with a lag? We use the survey data to form a dependent variable of change in unit cost Change in unit cost is then regressed on: Lagged innovation intentions Survey-based capacity utilisation A regional model of unit cost: innovation effects
Average Cost Specification No regional innovation effects found
Significance for National Innovation on Regional Unit Cost AVC (SURE)
Strong external effects? Errors in variables? Why no Regional Innovation Effects?
Confirmation of National Innovation Effect in Panel Estimation of AVC
Sub Panels: Effects on Average Cost Note that for Higher General Education regions, both variables are jointly significant and significant when entered singly
Innovation in the CBI survey mainly seems to be associated with capital investment A reasonable first-cut model is obtained with a lagged dependent variable and the index of business optimism There is similarity across regions to the extent that the optimism elasticities can not be shown to be different across the 10 main regions Conclusions_1
National training affects innovation strongly but the variable may be non-stationary When national training is excluded, regional training is significant (at least at 10%) in all but one region Conclusions_2
Innovation intentions at national level affects perceived change in unit cost in each region There is little support for regional innovation affecting unit cost An exception is for the panel of regions with higher general education where both regional and national innovation variables are jointly significant or significant when entered singly Conclusions_3