National culture and R&D investments

ABSTRACT

By using time-collapsed data for 12,362 firms from 40 countries, I document that national culture is significantly associated with R&D investments, controlling for other determinants including a country’s financial development. In a more individualistic, less masculine, and more indulgent culture, firms tend to make more R&D investments. These findings are robust to alternative proxy variables for financial development and remain unchanged regardless of whether tested by instrumented two-stage least squares or fixed effects panel regressions. Culture’s impact is different depending on firm size and age. Further, its limited impact on equity issues suggests that it relates to R&D directly rather than being mediated by other financial market channels.

KEYWORDS:

• National culture
• R&D investment
• equity issuance
• financial development

JEL CLASSIFICATIONS:

• G15
• G30
• G32

Disclosure statement

No potential conflict of interest was reported by the author.

ORCID

Kyeong-Seop (KS) Choi http://orcid.org/0000-0001-6413-9229

Notes

1 A few prior studies on national culture and R&D are from the field of human resources. Hoppe (1993) discusses Hofstede’s national culture and the management of highly educated R&D professionals. Kedia, Keller, and Jullan (1992) compare four countries – Austria, Belgium, Finland, and Sweden – based on Hofstede’s national culture and their R&D units’ productivity. Targeting the Chinese context only, Wang et al. (2010) emphasize corporate culture’s importance in breeding an innovative atmosphere. To my knowledge, Varsakelis (2001) and Shao, Kwok, and Zhang (2013) are the extant literature on national culture and corporate R&D on the global scale of 50 and 44 countries, respectively. Lorca and de Andrés (2018) provide statistically biased results using only 13 European countries as their sample.

2 The sixth culture variable Indulgence has missing data entries on Chile and Israel. Thus, all the subsequent regressions testing six culture variables simultaneously are actually losing these two countries. Hence, for all such regressions, I conducted additional tests excluding Indulgence. The statistical results remain qualitatively the same (unreported).

3 Corporate investments have been associated with product demand and internal funds. With corporate investments decomposed into ordinary and R&D investments, it is documented that ordinary investments depend on product demand proxied by sales and sales growth, whereas R&D investments are determined by internal funds whose proxy is cash flow (Hall 1992; Himmelberg and Petersen 1994). Therefore, all regressions in this study control for Sales, Sales growth, and Cash flow.

4 I also tested other substitutes for Hofstede’s culture variables. Taras, Steel, and Kirkman (2012) recognize the sample limitation as well as time invariance in Hofstede’s national culture construction. They collected all the extant empirical studies that utilize Hofstede’s culture variables and meta-analytically constructed decade-based longitudinal indices for Power Distance, Individualism, Masculinity, and Uncertainty Avoidance. I tested Taras’ indices through models (1) to (10) and obtained similar results. Individualism and Masculinity relate to R&D investments significantly. As Taras’ meta-indices are not widely received, I do not report the results here. Instead, they are provided upon request.

5 This is true when debt is considered as a uniform integral variable. Recent research decomposes debt into bank debt (used credit lines, term loans, etc.) and public debt and examines their link with firm performance and R&D investments. For European firms, bank debt can be an efficient channel to fund R&D, because emphasis on firm−bank relationship in Europe enhances the monitoring effect on firm R&D activities financed by bank debts (Campello et al. 2012; Guney, Karpuz, and Ozkan 2017). For U.S. firms, public debt and equity are the more effective financial resources because they are arm’s length financing that provide greater flexibility and tolerance for risky R&D entrepreneurship (Atanassov 2016). In my study, I follow the general line that, worldwide, equity is better suited for intangible risky R&D asset investments.

6 To check consistency in statistical results, I used three other measures of external dependence that are eventually a median value in each 2-digit U.S. industry. I calculated the value, first, as the sum of fixed investment minus cash (Compustat code: CH) over the sample period, scaled by the sum of fixed investment plus R&D expenditure; second, the sum of fixed investment minus cash and short-term investments (CHE), scaled as previously; third, the sum of fixed investment minus income before extraordinary items (IBC), scaled as previously. In unreported tests, all the three measures, as well as their interactions, have much diminished (insignificant) coefficients than the measure I use in my main study. Therefore, I obtain the rationale for my measure of external dependence.

7 Following Rajan and Zingales (1998), I tested manufacturing firms (SIC 2000−3999) and obtained the results consistent with their findings for the entire period as well as for both subperiods. The interaction between external dependence and financial development was strongly significant and positive with Accounting standards and Country equity issues as the proxy variables for financial development. With MCAP/GDP, the interaction term shows far weaker coefficients. With Stocks traded/GDP, the interaction term turns insignificant. Also with Credit/GDP, the interaction term is significant and positive. This implies that credit market development has a positive impact on R&D investments for manufacturing firms. The interaction term with Country mean leverage (a) shows strongly significant negative coefficients. This indicates that consistent with prior studies, greater countrywide corporate debt hinders R&D investments in the manufacturing industries. These test results are provided upon request. Following Brown, Martinsson, and Petersen (2013), testing all firms except for financial and utilities firms, I do not find my results strongly congruent with the results from their study.

8 The data are downloadable from: http://www.worldvaluessurvey.org

9 I consulted the report titled ‘International Statistics on Crime and Justice’ (Helsinki, 2010) produced by collaboration between United Nations Office on Drug and Crime (UNODC) and the European Institute for Crime Prevention and Control (HEUNI). The United Nations collected sparse data on crime and criminal justice systems for their member countries, whereas the HEUNI collected detailed data only for Europe and North America. The report I consulted is a product of their collaboration containing global-scale data in far greater detail.

10 Recent research alludes to a possible mediation of cash holdings in the relationship between culture and R&D investments. Chen et al. (2015) document a negative association between Individualism and corporate cash holdings around the world. He and Wintoki (2016) assert that the increase of cash holdings from 1980 to 2012 in U.S. firms is explained by enhanced R&D investments. This evidence, in turn, supports the extant finding that U.S. firms use their cash reserves to smooth R&D expenditures (Brown and Petersen 2011). Here, the financial channel between culture and R&D seems very likely to be cash holdings. I reserve a detailed study on this for a later paper.