Business Group Analysts

AbstractWe exploit the rich data of business groups in China and identify sell-side analysts following multiple listed firms within a business group (BG analysts). For a group firm, we find that BG analysts issue more accurate forecasts than non-BG analysts. Such an effect is more pronounced when the focal firm shares stronger economic links with, and when its covering analysts have greater information demand for, its group peers. Further analyses suggest evidence of intragroup information flows. Around group peers’ annual earnings announcements, BG analysts are more likely than non-BG analysts to revise their forecasts for the focal firm along the same direction as group peers’ earnings surprises. Confined to analysts’ coverage initiations for a group firm, those who have covered the focal firm’s group peers prior to the coverage initiation show superior forecasting performance during the initiation year. By adopting the unique perspective of business groups, i.e., a prevalent organizational structure, our study shows that information commonalities within a business group shape analyst behaviour.Keywords: Business groupsSell-side analystsForecast accuracyInformation commonality AcknowledgementsWe thank Jeffrey Ng (Editor) and two anonymous reviewers for their valuable comments and suggestions in improving the paper. Additional materials are available in an online Supplement at the journal’s Taylor and Francis website. All errors are our own.Disclosure statementNo potential conflict of interest was reported by the author(s).Supplemental Research MaterialsSupplemental data for this article can be accessed on the Taylor & Francis website, doi:10.1080/09638180.2023.2238787.Online Appendix 1: Correlations among Empirical VariablesOnline Appendix 2: Non-linear EffectsOnline Appendix 3: Intragroup Economic Links and BG Analysts’ Forecast AccuracyOnline Appendix 4. BG Analysts’ Prior Experience in Covering The Business GroupOnline Appendix 5. Determinants of BG Analyst Choice and Main Effect RobustnessOnline Appendix 6. Spinoffs of Group Affiliations and BG Analysts’ Forecasting PerformanceOnline Appendix 7. Addressing Other Economic LinksOnline Appendix 8. The Importance of Group-affiliated Firms to BG Analysts’ PortfoliosOnline Appendix 9. BG Analysts’ Forecasts for Non-group FirmsNotes1 Our definition of business group is consistent with that of Larrain et al. (Citation2019) and Fang et al. (Citation2017). Larrain et al. (Citation2019) describe business groups as ‘sets of firms with a common controlling shareholder’ (p.3,036), and Fang et al. (Citation2017) view a business group as ‘a structure in which at least two legally independent firms are controlled by the same ultimate owner’ (p. 40).2 Business groups can also be formed by informal ties, which relate to linkages by relations of interpersonal trust, on the basis of similar personal, ethnic or commercial background (Granovetter, Citation2005). Considering informal ties will broaden the definition of a business group. However, compared with formal ties, the information of informal ties is less observable to both researchers and sell-side analysts.3 More precisely, a BG analyst is defined as an analyst following the focal firm who also covers at least one of the focal firm’s group peers during the same year.4 Cheng et al. (Citation2022) examine the common ownership between brokerage houses and firms covered by their analysts. The common ownership explored in our study is different as it refers to firms covered by one analyst.5 Using the ultimate controlling shareholder to define the boundary of a business group is consistent with the existing literature (e.g., Buchuk et al., Citation2014; Fang et al., Citation2017). Ultimate controlling shareholders exert control over subsidiaries significantly in excess of their cash flow rights, primarily through the use of pyramids and participation in management (La Porta et al., Citation1999). Corporate policies of group affiliates are significantly influenced by the ultimate controlling shareholder’s incentives (e.g., Gopalan et al., Citation2014; Fang et al., Citation2017), thus creating intragroup economic links to be explored in our study.6 Other government agencies ultimately controlling SOEs include local bureaus of state asset management, the ministry of finance, etc.7 Listed firms in China have Dec 31 as the fiscal year end date and calendar years serve as their fiscal years.8 Online Appendix 1 reports correlations among empirical variables. Forecast accuracy is positively correlated with the indicator BG_Analyst, providing preliminary evidence that BG analysts, compared with non-BG analysts, issue more accurate forecasts for group-affiliated firms.9 In untabulated analyses, we perform the t-tests using raw values of empirical variables. We find consistent evidence – that absolute forecast errors (AFE) are significantly lower for BG analysts than for non-BG analysts.10 To assess the robustness of our finding, we re-perform our main regression by restricting the sample to firm-years including both BG analysts and non-BG analysts. The restricted sample has a smaller sample size (22,493 versus 42,138 of our primary sample). We consistently find positive and significant coefficients on BG_Analyst. In addition, we examine whether the marginal benefit of covering an additional group peer gradually weakens. We construct BG_Analyst_Npeers, the number of covered group firms (including the focal firm) for each analyst-group-year, and its squared term. The results in Online Appendix 2 show that covering an additional group peer initially enhances the informational benefit, reflected by positive and significant coefficients on BG_Analyst_Npeers. However, the coefficient on the squared term is negative and significant, indicating a decreasing marginal effect.11 In Online Appendix 3, we construct the focal firm’s economic links with group peers covered by the BG analyst and decompose BG_Analyst into two indicators: BG_High_Link (BG_Low_Link) which equals 1 when a focal firm has higher (lower) than median economic links with its group peers also covered by the BG analyst, and 0 otherwise. Economic links are proxied by RPT and Corr Strategy. We replace the BG_Analyst in Equation (2) with the two indicators and re-estimate the regression. We find consistent evidence that when a focal firm has stronger economic links with group peers that are also covered by the BG analyst, the effect of BG analyst status on forecast accuracy is more pronounced.12 In untabulated analyses, we find that the absolute correlations of firm performance (ROA and EPS) for non-state-owned groups are on average higher than those for state-owned groups, supporting the assumption of stronger economic links within the former, than the latter.13 A counterforce exists as BG analysts’ ability to collect soft information may dampen their forecast revisions in response to public disclosures, i.e., group peers’ earnings announcement in our context. We thank an anonymous reviewer for pointing out this possibility.14 Empirical evidence here does not rule out the possibility that information can also flow from the newly-covered group firm to the early-covered group firm. Upon coverage initiation, the BG analyst acquires and processes additional information relevant to the business group, which may facilitate the analyst’s forecasting performance for early-covered group firms. Our documented information flow depends critically on the research design – analyst forecasts issued upon coverage initiations. We thank an anonymous reviewer for pointing out this issue.15 In Online Appendix 4, we examine whether the information advantage of covering multiple firms within a business group varies with the analyst’s prior experience in covering the business group by decomposing BG_Analyst into two indicators: BG_High_Exp (BG_Low_Exp) which equals one when a BG analyst has higher (lower) than median experience in covering the business group (group experience hereafter), and zero otherwise. An analyst’s group experience is measured as the number of years since the analyst has covered any of the firms within the business group. We replace the BG_Analyst in Equation (2) with the two indicators and estimate the regression. The coefficient on BG_High_Exp is approximately two times that on BG_Low_Exp, with their difference being statistically significant. Therefore, the information advantage of BG analysts for group firms is more pronounced when a BG analyst has more experience in covering the business group.Additional informationFundingAuthors acknowledge financial support from Ministry of Education in China (20YJC630106, 21YJC790094), the National Natural Science Foundation of China (71772110, 71902036, 72172037), the National Social Science Fund of China (22BJY078), and the MOE Project of Key Research Institute of Humanities and Social Science in University (22JJD790094).