Skip to main content
Log in

Proximity to hubs of expertise and financial analyst forecast accuracy

  • Original Paper
  • Published:
Eurasian Business Review Aims and scope Submit manuscript

Abstract

This paper investigates whether the geographical proximity of financial analysts to hubs of information and expertise can influence their forecasting accuracy. Recent studies show that the financial analyst forecasting process show a systematic difference in earnings forecast accuracy dependent on the geographical distance of analysts from the companies they follow. The literature argues that local analysts issue more accurate forecasts because they have an informational advantage over analysts who are further away. Industrial centres can constitute important knowledge spillovers by creating formal and informal networks amongst firms and higher education and research institutions. In such a hub, information can easily flow and propagate. Our hypothesis is that physical proximity to these hubs, and not to the companies they follow, is an advantage for financial analysts, leading to the issue of more accurate forecasts. Using a sample of 205 observations related to 33 firms, across seven countries and ten sectors, our results are consistent with the hypothesis.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Notes

  1. Alternative explanations are related to incentive arguments, including compensation and career incentives, and not to information asymmetries.

  2. By financial opening the authors mean the opening of the country to foreign investors.

  3. Petersen and Rajan (2002) and DeYoung et al. (2011) document that advances in information technology facilitating information flow may have eliminated any physical local information advantage.

  4. See also Sect. 3 and the “Appendices A and B”.

  5. Brooks (2008) explains that the Newey–West procedure implies ‘HAC' (heteroscedasticity and autocorrelation consistent) standard errors. This adjustment allows us to deal with the coefficients' standard errors since it produces a variance–covariance estimator which is consistent in the presence of both heteroscedasticity and autocorrelation.

  6. Clement (1998) documented that PMAFE improves the chances of identifying the differences in individual analyst forecast accuracy. Jacob et al. (1999) discussed these benefits in more detail.

  7. As in all the regressions these latter accuracy measures were the best, and consistently with previous literature (see Clement 1998, 1999 and Jacob et al. 1999), we report only the results obtained using this accuracy measure.

  8. See Storper (1997) for a review.

  9. Jacobs and DeMan (1996, p 425) argue that “there is not one correct definition of the cluster concept…different dimensions are of interest”.

  10. A block diagonal matrix can be split up in parts that have no connection with each other. By rearranging sectors appropriately (details of this method are reported in “Appendix A”), the matrix would look like blocks of matrices along the main diagonal.

  11. The intermediate consumption is an economic concept that represents the monetary value of goods and services consumed or used as inputs in production by firms of a sector in a country.

  12. There are several possible algorithms for making the block diagonal matrix by rearranging sectors. “Appendix A” describes an algorithm which does not involve complex computations and is easy to program. An algorithm based on eigenvalues, which has the advantage of ordering clusters according to the strength of their linkages, can be found in Dietzenbacher (1996).

  13. Please see “Appendix A” for a detailed general definition of this table.

  14. In mathematics, the power iteration is an algorithm: given a matrix A, the algorithm will produce a number λ (the eigenvalue) and a non-zero vector v (the eigenvector), such that Av = λv. The algorithm is also known as the Von Mises iteration.

  15. Appendix B” reports further technical details.

  16. We include D05, which is a dummy variable equal to 1 if the observation obtained was in 2005, 0 otherwise; D06 is a dummy variable equal to 1 if the observation obtained was in 2006, 0 otherwise; and D07, which is equal to 1 if the observation obtained was in 2007, 0 otherwise.

  17. It is not feasible to extend our data to more recent years as the available information sources are limited on those selected years.

  18. The sample size issue explains also the low adjusted R-squared of the models. This result is line with previous research that provides regressions in which the adjusted R-squared hits low values as well (approximately 0.15). This evidence claims for more research on this topic in order to shed more light on the issue.

  19. The OECD defines an input–output table as a tool for the presentation of a detailed analysis of the process of production and the use of goods and services (products), and the income generated in that production for any European country.

  20. Hoen (2002) shows that this method brings to same results also selecting other input–output tables.

References

  • Ang, J. S., & Ciccone, S. (2001). International differences in analyst forecast properties. Working paper, University of New Hampshire.

  • Aroche-Reyes, F. (2001). The question of identifying industrial complexes revisited: a quality perspective. In M. L. Lahr & E. Dietzenbacher (Eds.), Input-Output Analysis: Frontiers and Extensions. Houndsmills: Palgrave.

    Google Scholar 

  • Bacmann, J. F., & Bolliger G. (2001). Who are the best? Local versus foreign analysts on the Latin American stock markets. Working paper, University of Neuchâtel, Neuchâtel, Switzerland.

  • Bae, K., Stulz, R., & Tan, H. (2008). Do local analysts know more? a cross-country study of the performance of local analysts and foreign analysts. Journal of Financial Economics, 88(3), 581–606.

    Article  Google Scholar 

  • Barker, R. (1998). The market for information: evidence from finance director, analysts and fund managers. Accounting and Business Research, 29(1), 3–20.

    Article  Google Scholar 

  • Bolliger, G. (2004). The characteristics of individual analysts’ forecasts in Europe. Journal of Banking and Finance, 28(9), 2283–2309.

    Article  Google Scholar 

  • Breton, G., & Tafler, R. J. (2001). Accounting information and analyst stock recommendation decisions: a content analysis approach. Accounting and Business Research, 31(2), 91–102.

    Article  Google Scholar 

  • Brooks, C. (2008). Introductory econometrics for finance, 2nd edn. Cambridge, UK: Cambridge University Press.

  • Brown, L. (1993). Earnings forecasting research: its implications for capital markets research. International Journal of Forecasting, 9(3), 295–320.

    Article  Google Scholar 

  • Brown, P. G., Foster, G., & Noreen, E. (1985). Security analyst multiyear earnings forecasts and the capital market. Studies in Accounting Research, 21.

  • Brown, L., Call, A., Clement, M., & Sharp, N., (2013). Inside the ‘Black Box’ of Sell-Side Financial Analysts. Working Paper, University of Texas.

  • Cetorelli, N., & Peristiani, S. (2009). Prestigious stock exchanges: a network analysis of international financial centers. Staff Report, Federal Reserve Bank of New York.

  • Chang, C. (2010). Information footholds: isolating local presence as a factor in analyst performance and trading. Journal of International Money and Finance, 29(6), 1094–1107.

    Article  Google Scholar 

  • Chang, J. J. Khanna, T., & Palepu, K. (2000). Analyst activity around the world. Working Paper, Harvard Business School, Strategy Working Paper Series.

  • Clement, M. B. (1998). Some considerations in measuring analysts’ forecasting performance. Working Paper, University of Texas, Austin.

  • Clement, M. B. (1999). Analyst forecast accuracy: do ability, resources, and portfolio complexity matter? Journal of Accounting and Economics, 27(3), 285–303.

    Article  Google Scholar 

  • Conroy, R., Fukuda, Y., & Harris, R. S. (1997). Securities houses and earnings forecasts in Japan: what makes for an accurate prediction? Financial Analysts Journal, 53, 29–40.

  • Davis, E. P. (1990). International financial centres: an industrial analysis. Economics Division, Bank of England, discussion paper 51.

  • DeYoung, R., Frame, W. S., Glennon, D., & Nigro, P. (2011). The information revolution and small business lending: the missing evidence. Journal of Financial Services Research, 39(1), 19–33.

    Article  Google Scholar 

  • Dietzenbacher, E. (1996). An algorithm for finding block-triangular forms. Applied Mathematics and Computation, 76(2–3), 161–171.

    Article  Google Scholar 

  • Doeringer, P., & Terkla, D. (1995). Business Strategy and cross-country clusters. Economic Development Quarterly, 9(3), 225–237.

    Article  Google Scholar 

  • Du, Q., Yu, F., & Yu, X. (2013). Cultural Proximity and the Processing of Financial Information. Working Paper, Indiana University.

  • Eurostat (2008). Eurostat Manual of Supply, Use and Input-Output Tables. Eurostat Methodologies and Working Papers, European Communities, Luxemburg.

  • Forni, M., & Paba, S. (2001). Knowledge Spillovers and the Growth of Local Industries. CEPR Discussion Paper, no. 2934.

  • Hoen, A. R. (2002). Identifying linkages with cluster based methodology. Economic System Research, 14(2), 131–146.

    Article  Google Scholar 

  • Hope, O.-K. (2003). Disclosure practices, enforcement of accounting standards, and analysts’ forecast accuracy: an international study. Journal of Accounting Research, 41(2), 235–272.

  • Jacob, J., Lys, T., & Neale, M. (1999). Expertise in forecasting performance of security analysts. Journal of Accounting and Economics, 28(1), 51–82.

    Article  Google Scholar 

  • Jacobs, D., & Demann, A. (1996). Cluster, industrial policy and firm strategy: a menu approach. Technology Analysis and Strategic Management, 8(4), 425–437.

    Article  Google Scholar 

  • Katz, L. (1953). A new status index derived from sociometric analysis. Psychometrika, 18, 39–43.

  • Krugman, P. (1991). Geography and trade. Cambridge: MIT Press.

    Google Scholar 

  • Lawson, C., & Lorenz, C. (1999). Collective learning, tacit knowledge and regional innovation capacity. Regional Studies, 33(4), 305–317.

    Article  Google Scholar 

  • Lorenz, M., & Maskell, P. (2004). The cluster as a nexus of knowledge creation. In P. Cooke & A. Piccaluga (Eds.), Regional Economies as Knowledge Laboratories (pp. 77–92). London: Edward Elgar.

    Google Scholar 

  • Malberg, A., & Maskell, P. (2002). The elusive concept of localization economies: towards a knowledge based theory of special clustering. Environment and Planning A, 34(3), 429–449.

    Article  Google Scholar 

  • Malloy, C. J. (2005). The geography of equity analysis. Journal of Finance, 60(2), 719–755.

    Article  Google Scholar 

  • Marshall, A. (1920). Principles of economics (8th ed.). London: McMillan.

    Google Scholar 

  • Orpurt, S. (2004). Local analyst earnings forecast advantages in Europe. Working Paper, University of Chicago.

  • Petersen, M., & Rajan, R. (2002). Does distance still matter? The information revolution in small business lending. Journal of Finance, 57(6), 2533–2570.

    Article  Google Scholar 

  • Porter, M. (1990). The competitive advantage of nations. London: MacMillan Press.

    Google Scholar 

  • Previts, G., Bricker, R., Robinson, T., & Young, S. (1994). A content analysis of sell-side financial analyst company reports. Accounting Horizons, 8(2), 55–70.

    Google Scholar 

  • Ricardo, D. (1963). The principles of political economy and taxation. Homewood: Irwin.

    Google Scholar 

  • Rosenfeld, S. (1996). Industrial Strength Strategies: Regional Business Clusters and Public Policy, Aspen Institute Rural Economic Policy Program. Best Practice Series, Industrial Strength Strategies.

  • Rosenfeld, S. (1997). Bringing business clusters into the mainstream of economic development. European Planning Studies, 5(1), 3–23.

    Article  Google Scholar 

  • Storper, M. (1997). The regional world. New York: Guildford Press.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Elisa Cavezzali.

Appendices

Appendix A

This appendix illustrates how we identify and define the hubs of expertise following a cluster-based approach.

The key point is the definition of an index allowing us to rank the European sectors by knowledge intensity. The methodology, based on the literature about cluster theory, is for the identification of two different proxies of hub of expertise.

Firstly, in order to identify the boundaries and size of hubs of expertise, we apply the Hoen’s algorithm (Hoen 2002), based on symmetric input–output tables.Footnote 19 Hoen asserts that any sector needs linkages with other sectors in order to develop its own business. The symmetric input–output tables of different countries should represent these relations. Therefore, by analysing the input–output tables, we identify the strongest linkages between sectors and thus the clusters, i.e. aggregations of sectors within a national economy.

The symmetric input–output tables are analytical tables derived from the supply and use system. An input–output table is a quantitative economic tool which represents the interdependencies between different branches of the national economy or different, even competing, economies. The transformation procedure converts the product-by-industry system of the supply and use tables into a product-by-product system or industry-by-industry system (Table 4). Input–output tables are used to identify economically-related industry clusters and also so-called ‘key’ or ‘target’ industries of a specified economy.

Input–output tables often contain an enormous amount of detailed data. In order to deal with these data, it is necessary to aggregate the data. One possibility is to search for clusters of sectors with strong linkages. The clusters then denote how the sectors may be aggregated (Aroche-Reyes 2001).

Hoen (2002) developed an algorithm based on these symmetric input–output tables. His algorithm aggregates sectors into clusters according to the following rule: two sectors compose a cluster if their relations, the so-called linkages, are large according to a certain restriction.

Table 4 A simplified input–output table

This algorithm is based on the matrices of intermediate consumption across industries. Then, to identify a cluster empirically, the author uses the block diagonal matrix method.Footnote 20

Hoen’s algorithm allows us to identify hubs by the intermediate consumption flow in a national perspective.

Appendix B

This appendix illustrates how we identify and define the hubs of expertise following a network-based approach.

We base our methodology on Cetorelli and Peristiani (2009) approach, which adopted network analysis to deal with a comparison of stock exchanges in a global perspective. Following their procedure, we determine the so-called prestige index which allows us to compare hubs of expertise from different countries. We regard each country as a network and the sectors of the country as nodes of the network. The production patterns are ties between nodes.

As in the approaches used previously, we start from an input–output matrix and we build a network matrix. Each element of the matrix is considered as a bidirectional flow.

Table 5 describes a typical network matrix, used in our framework. The row entries represent the origin of the flow, while the column entries present the destination of it. In this way, the main diagonal accounts for flows due to the sector activity (measured by the sum of intermediate consumption and value added of each sector) and off-diagonal entries represent interactions between different nodes. For instance, I.C. 11  + V.A. 11 indicates the flow produced and accumulated by industry 1 itself, I.C. 12 indicates the flow of intermediate consumption from industry 1 (origin) to industry 2 (destination), while I.C. 21 is the flow of intermediate consumption from industry 2 to industry 1.

Table 5 A network matrix example

By analysing the matrix by row, we can identify the intensity of the interaction of each unit towards other destination nodes. This indicator is called the out-degree index and is calculated as the row sum, excluding the main diagonal entry. In examining the matrix by column, it is possible to compute the so-called in-degree index, which represents the ability to influence the origin of flows. Neither index offers details about where flows are coming from.

In order to consider the out-degree and in-degree indices simultaneously, Cetorelli and Peristiani (2009) suggested using the betweenness index, which exploits network ties and captures the uniqueness of a given node in a network. Let m jk (n i ) be the maximum flow between nodes (nj,nk) which goes through node n i . We obtain the overall betweenness of node n i aggregating across all possible pairs of nodes in the network, other than n i . In order to allow for comparison over time, normalisation is recommended, so that the betweenness index of node n i is:

$$ P(n_{i} ) = \sum {\sum {\frac{{m_{jk} (n_{i} )}}{{m_{jk} }}} } $$
(B1)

Therefore, the prestige index of node n i is:

$$ \Pr (n_{i} ) = x_{1i} P(n_{1} ) + x_{2i} P(n_{2} ) + \ldots + x_{Ni} P(n_{N} ) $$
(B2)

where the weights are represented by the flows from each of the nodes onto n i . We have N equations in N unknowns for each network.

This sophisticated and standardised index allows for the judgement of the importance of each node in a network, fully exploiting the information contained in the entire network structure.

This metric allows us to normalise the data from symmetric input–output tables and identify an international ranking for hubs of expertise. This index is a proxy for the knowledge level of every industry in each country. The greater values in this index are associated with the greater influence of the sector in the production of goods and services for the whole economy.

In simple terms, we dispose the flows of intermediate consumption between every pair of sectors on the off-diagonal entries, while the main diagonal includes the sum of the intermediate consumption flows within every sector with the sector value added. This matrix represents all of the data on a country’s production. We divide every column of the matrix by the column sum and apply the power method in order to calculate the eigenvector associated with the largest eigen value of the matrix. This eigenvector contains the index of prestige of any sector. Lastly, we identify the main sector of each firm and assign to each analyst covering that firm the index of prestige associated with that analyst’s country location.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Cavezzali, E., Crepaldi, J. & Rigoni, U. Proximity to hubs of expertise and financial analyst forecast accuracy. Eurasian Bus Rev 4, 157–179 (2014). https://doi.org/10.1007/s40821-014-0007-8

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s40821-014-0007-8

Keywords

Navigation