National Statistics (ONS). Several iterations of the CIS survey were performed, each focusing on innovation input and output over a three-year reference period. In this study, we focus on three rounds of the CIS: CIS4, CIS5 and CIS6, which correspond to the innovations from 2002–2004, 2004–2006, and 2006–2008. After merging CIS4, CIS5 and CIS6, we obtain 45,598 observations. We have excluded the 16,000 observations for which the data needed for our analysis (i.e., size, age, human capital characteristics, industry, and region) are missing.
12 To build our sample of LBO companies, we retrieved from Capital IQ all the deals that (1) involved target companies incorporated in the UK, (2) were reported as LBOs, (3) were announced between January 1, 1998 and December 31, 2008, and (4) were either “closed” or “effective”. We identified 4,365 transactions that comply with these criteria. These LBOs were matched to respondents in CIS4, CIS5 and CIS6. 1 A total of 1,535 CIS respondents were involved in an LBO between 1998 and 2008. We distinguish between two types of LBO companies (post-LBO and pre-LBO), depending on when the LBO occurred relative to the CIS observation period. The timing of the LBO for post-LBO and pre-LBO companies is depicted in Figure 1. Post-LBO companies are defined as companies that have been the target of an LBO during the CIS observation period or in the three previous years (i.e., companies in this group are those for which the LBO occurred between 1999 and 2004 for CIS4, between 2001 and 2006 for CIS5, and between 2003 and 2008 for CIS6). For these companies, the innovation variables are measured after the LBO occurred. Pre-LBO companies are defined as companies that have been the target of an LBO in the three years since the end of the relevant CIS observation period (i.e., companies in this group are those for which the LBO occurred between 2005 and 2007 for CIS4, between 2007 and 2009 for CIS5, and between 2008 and 2010 for CIS6). For these companies, the innovation variables are thus observed before the LBO occurred. Figure 1: CIS Surveys and the Timing of LBOs Year t-5 t-4 t-3 t-2 t-1 t t+1 t+2 t+3 CIS observation period Post-LBO company Pre-LBO company 1 To preserve the anonymity of the respondents, the ONS performed the matching for us and we were not able to track their identity.
13 We identify 855 post-LBO companies and 232 pre-LBO companies. We exclude from the sample 448 observations that correspond to LBOs that occurred more than 3 years before the beginning of the corresponding survey. Including non-LBO companies, our final sample comprises 29,150 observations. We report the distribution of our sample across regions, CISs and LBO statuses in Table 1. In our analysis, we distinguish companies based on the technology and knowledge intensity of the sectors in which they operate. Specifically, we define HTKIS and MLT industries using the Eurostat (2009) industry classification. The distribution of HTKIS and MLT is reported in Table 1.
14 Table 1: Distribution of the Sample All HTKIS MLT N % N % N % Region Northeast 1,818 6.2% 682 7.2% 1,136 5.8% Northwest 2,624 9.0% 799 8.5% 1,825 9.2% Yorkshire and the Humber 2,411 8.3% 799 8.4% 1,612 8.2% East Midlands 2,383 8.2% 809 8.5% 1,574 8.0%
- Fall '18