Deciding on large scale investments

ABSTRACT

This study analyses the importance of investment factors across investment sizes and the frequency of large investment decisions within the firms. We use data from 1442 investment decisions made by 226 Brazilian firms between 1997 and 2010. The results indicate that the influence of investment factors is different for investments of different sizes. The results also indicate that the increase in the frequency of large-scale investment decisions made by the firm influences especially the role of cash flow, diminishing its importance as an investment factor. This result allows one to argue that, probably, firms more experienced in large-scale decisions could be bolder in their decision-making process, relying less on the accumulation of funds through cash flow.

KEYWORDS:

• Investment decisions
• financial constraints
• investment size
• investment frequency

JEL CLASSIFICATION:

• G3
• G30
• G31

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Since 2001, Brazilian corporate governance rules are similar to those applied in the US.

² There was a decrease in 2008, as a result of the international crisis.

³ This publication is equivalent to the Fortune 500. The daily economic newspaper Gazeta Mercantil collets the data, which is analyzed by the Department of Accounting of the University of São Paulo (http://www.fipecafi.org/mm/). We thank the Department of Accounting of USP for allowing us the access to the database.

⁴ BM&F Bovespa Stock and Exchange http://www.bmfbovespa.com.br/capitalizacao-bursatil/ResumoBursatilHistorico.aspx?idioma=pt-br

⁵ http://www.ibge.gov.br/home/estatistica/economia/cadastroempresa/2012/default_xls_empresa.shtm

⁶ For more information on the Brazilian economic scenario, see ApexBrasil – Brazilian Trade and Investment Promotion Agency (2014). Investment Guide to Brasil 2014. http://arq.apexbrasil.com.br/legado/investmentguidetobrasil2014.pdf, visited 07/24/2015, and Deloitte (2015), Doing Business in Brazil. http://www.deloitte.dbbrazil.com.br/show.aspx?idCanal=171nnV02FxuWEQm1suf4EA==, visited 07/24/2015.

⁷ A discussion of how the grouping of investments by size was established will be presented in the following section.

⁸ A likelihood test to compare the fixed and the mixed models resulted in χ2 = 259.63, Prob > χ2 = 0.00, confirming that the mixed model is the most adequate for our study.

⁹ As robustness checks, we have used other variables to represent investment size, such as Investment/Total Equity and Investment/Total Assets. The main conclusions do not change.

¹⁰ We have introduced sectoral dummies, but they came out nonsignificant, and therefore were dropped. This is not unexpected, since our definition of investment size already takes into account not only sectoral specificities, but even firm specificities.

¹¹ We have also tested other cut-off points, such as the median, the terciles, and investment rates exceeding 20% and or 2.5 times the median (following Nilsen and Schiantarelli 2003). In all cases, the main conclusions remain.

¹² Regressions including the variable ‘frequency’ are estimated for the period 2000–2010, to allow 3 years of potential experience with large scale investment decisions for every firm.

Additional information

Funding

This work was supported by FAPESP grant (Fundação de Amparo à Pesquisa do Estado de São Paulo, SP- Brasil) [grant number 2011/11562-4].