Corruption level and uncertainty, FDI and domestic investment

Abstract

Based on real options theory and institutional factors, we develop a theoretical framework for investment in the presence of corruption and use a sample of private firms in 13 European countries over 2001–2013 to carry out the first large-scale analysis of the impact of the level of corruption and uncertainty about corruption on post-entry investment of MNE subsidiaries. We employ several waves of managerial surveys (the Business Environment and Enterprise Performance Survey; BEEPS) to construct local- rather than merely country-level measures of corruption level and uncertainty. In combination with a large European firm-level database (Amadeus), we show that corruption uncertainty and corruption level do not have an effect on the investment of MNE subsidiaries. We next carry out the analysis on the sample of domestic firms and find a negative investment effect that is driven primarily by corruption uncertainty rather than corruption level. We also show that investment of domestic firms that are similar (matched) to MNE subsidiaries is unaffected directly by corruption, but is affected by uncertainties related to finances and judiciary. Our results are robust to controlling for various types of uncertainty, and they provide new insights into the effects of corruption on investment.

Résumé

Nous appuyant sur la théorie des options réelles et des facteurs institutionnels, nous développons un cadre théorique lié à l’investissement en présence de corruption, et utilisons un échantillon d'entreprises privées dans treize pays européens sur la période 2001–2013, afin de réaliser la première analyse à grande échelle sur l'impact du degré de corruption et de l'incertitude liée à la corruption sur les investissements post-entrée des filiales des entreprises multinationales (MNEs - Multinational Enterprises). Plutôt que de construire simplement au niveau national les mesures du degré de corruption et de l'incertitude liée à la corruption, nous réalisons plusieurs vagues d'enquêtes managériales (BEEPS) pour développer celles-ci au niveau local. En combinaison avec une grande base de données des entreprises européennes (Amadeus), nous démontrons que l'incertitude liée à la corruption et le degré de corruption n’exercent aucun impact sur l'investissement des filiales des MNEs. Nous effectuons ensuite l'analyse sur l'échantillon d'entreprises domestiques, et observons un impact négatif sur l’investissement, principalement généré par l'incertitude liée à la corruption plutôt que par le degré de corruption. Nous démontrons également que l’investissement des entreprises domestiques, similaires (appariées) aux filiales des MNEs, n’est pas directement influencé par la corruption, mais plutôt par les incertitudes liées aux finances et au système judiciaire. Restant robustes avec le contrôle de divers types d'incertitude, nos résultats apportent de nouveaux renseignements sur les impacts de la corruption sur l'investissement.

Resumen

Con base en la teoría de opciones reales y los factores institucionales, desarrollamos un marco teórico para la inversión en la presencia de corrupción y usamos una muestra de empresas privadas en países europeos entre el 2001 y el 2013 para llevar a cabo el primer análisis a gran escala del impacto del nivel de corrupción y la incertidumbre sobre la corrupción en la inversión posterior a la entrada de las filiales de las empresas multinacionales. Utilizamos varias rondas de encuestas gerenciales (BEEPS) para construir medidas locales, y no sólo a nivel país, de nivel e incertidumbre de corrupción. En combinación con una amplia base datos europea a nivel de empresa (Amadeus), mostramos que la incertidumbre y el nivel de corrupción no influencian la inversión en las filiales de las empresas multinacionales. Después llevamos a cabo el análisis en la muestra de empresas locales y encontramos un efecto negativo en las inversiones está impulsada por la incertidumbre de corrupción y no al nivel de corrupción. Mostramos también que la inversión de las empresas locales que son similares (que equivalen) a las filiales de las empresas multinacionales no se ve afectada directamente por la corrupción, pero sí por las incertidumbres relacionadas con las finanzas y el sistema judicial. Nuestros resultados son robustos controlando por varios tipos de incertidumbre y proporcionan nuevos conocimientos sobre los efectos de la corrupción en la inversión.

Resumo

Com base na teoria de opções reais e fatores institucionais, desenvolvemos um modelo teórico para investimento na presença de corrupção e usamos uma amostra de empresas privadas em treze países europeus durante 2001-2013 para realizar a primeira análise em grande escala do impacto do nível de corrupção e incerteza sobre corrupção em investimentos pós-entrada de subsidiárias de MNE. Empregamos várias ondas de pesquisas gerenciais (BEEPS) para construir medidas locais, em vez de meramente nacionais, do nível de corrupção e incerteza. Em combinação com um grande banco de dados de empresas europeias (Amadeus), mostramos que a incerteza da corrupção e o nível de corrupção não afetam o investimento de subsidiárias de multinacionais. Em seguida, realizamos a análise na amostra de empresas domésticas e encontramos um efeito negativo de investimento que é impulsionado primariamente pela incerteza da corrupção, e não pelo nível de corrupção. Também mostramos que o investimento de empresas domésticas que são semelhantes (pareadas) a subsidiárias de MNE não é afetado diretamente pela corrupção, mas é afetado por incertezas relacionadas a finanças e sistema judiciário. Nossos resultados são robustos por controlar vários tipos de incerteza e eles fornecem novos insights sobre os efeitos da corrupção sobre investimento.

摘要

基于实物期权理论和制度因素, 我们提出了一个有腐败存在情况下的投资理论框架, 并用2001–2013年间13个欧洲国家的私营企业的样本对腐败水平和腐败不确定性对跨国公司 (MNE) 子公司市场投资后的影响进行了首次大规模的分析。 我们采用了几波管理调查 (BEEPS) 来构建本地的而不仅仅是国家层面的腐败水平和不确定性的量表。 结合大型的欧洲公司级数据库 (Amadeus), 我们显示, 腐败的不确定性和腐败水平对MNE子公司的投资没有影响。 我们接下来对国内公司的样本进行了分析, 发现负面的投资效应主要是由腐败不确定性而不是由腐败水平所引起。 我们还显示, 与MNE子公司相似 (匹配) 的国内公司的投资不受腐败的直接影响, 但受与财务和司法有关的不确定性的影响。 我们的研究结果对控制各类不确定性是可靠的, 并对腐败对投资的影响提供了新见解。

Appendix: Definitions of variables and sources of data

Variable	Definition
Corruption mean and corruption uncertainty measured at a cluster level
Corruption mean	Mean of the normalized answers [=(x−1)/5] to the question “It is common for firms in my line of business to have to pay some irregular 'additional payments or gifts' to get things done with regard to customs, taxes, licenses, regulations, services, etc.” at the cluster level Source: BEEPS
Corruption uncertainty	Standard deviation of the normalized answers [=(x−1)/5] to the question “It is common for firms in my line of business to have to pay some irregular 'additional payments or gifts' to get things done with regard to customs, taxes, licenses, regulations, services, etc.” at the cluster level Source: BEEPS
Definition of clusters and the level of details: (1) high detail/low aggregation (default variable), (2) medium detail (medium aggregation), (3) low detail (high aggregation)	(1) High detail/low aggregation. This is the default variable. Cluster is formed by country, industry (2-digit ISIC rev. 3.1), firm size (micro-, small, and medium–large firms) and urban location (capital, city with more than 1 million inhabitants, city with less than 1 million inhabitants) in the corresponding BEEPS wave (2000–2002, 2003–2005, 2006–2009, and 2010–2013). (2) Medium detail/medium aggregation. Cluster is formed by country, industry alphabet (letter level, NACE2 codes) and firm size (micro-, small, and medium–large firms) in the corresponding BEEPS wave (2000–2002, 2003–2005, 2006–2009, and 2010–2013). (3) Low detail/high aggregation. Cluster is formed by country and industry (industry alphabet (letter level, NACE2 codes) in the corresponding BEEPS wave (2000–2002, 2003–2005, 2006–2009, and 2010–2013)
Other uncertainties and restrictions measured on country level
Political risk ratings: government cohesion and government stability	Two main components of ICRG Risk Ratings are Government Cohesion and Government Stability; created from monthly level index constituents as an end of year measure. Source: ICRG Risk Ratings, https://epub.prsgroup.com/available-data
Political stability and absence of violence	Political stability and absence of violence/terrorism measure perceptions of the likelihood of political instability and/or politically motivated violence, including terrorism. Source: Worldwide Governance Indicators, World Bank, http://info.worldbank.org/governance/wgi/Home/Documents
Election uncertainty	Election Uncertainty (political-type uncertainty) steaming from election year; a dummy variable equal to 1 if there is a local(municipal/state) election and equal to 0 otherwise. Source: hand-collected
Investment freedom index	Starting score of 100 on the investment freedom component of the Economic Freedom Index. The index evaluates a variety of regulatory restrictions that typically are imposed on investment. On page 498 of the handbook, there is a detailed list of deductions from an ideal freedom, i.e., 100 points. See: https://www.heritage.org/index/pdf/2020/book/index_2020.pdf. Source: Heritage Foundation’s Annual Index of Economic Freedom
Economic uncertainty	Monthly stock market volatility. Sources: Czech (PX): https://www.pse.cz/; Slovak (SAX): http://www.bsse.sk/; Latvia (OMXR) and Estonia (OMXT): Nasdaq; Slovenia (SBITOP), Bosnia and Herzegovina (BIFXX), Croatia (CROBEX), Bulgaria (SOFIX), Hungary (BUX), Poland (WIG), Romania (BET), Serbia (BELEXLIN), and Ukraine (UX): Bloomberg
Inflation uncertainty	Inflation is the annual change in Consumer Price Index. Inflation uncertainty is the squared residual from the AR (2) forecasting model of inflation (see Ghosal & Ye, 2015) Source: World Development Indicators, World Bank
Inversed CPI (corruption perception index)	Composite index drawing on corruption-related data from expert and business surveys carried out by a variety of independent institutions. In our estimations, we use the inverted and scaled index, so that the higher index indicates a higher level of corruption. It ranges from 0 to 10 with 10 indicating the highest corruption Source: Transparency International
Judiciary uncertainty	A Judiciary uncertainty measure is constructed as a standard deviation of normalized answers [=(x−1)/3], at the cluster level, to the question “Can you tell me how problematic [is the functioning of the judiciary] for the operation and growth of your business?”. This variable measures the predictability of the judicial system Source: BEEPS, World Bank.
Financing uncertainty	A Financing uncertainty measure is constructed as a standard deviation of normalized answers [=(x−1)/3], at the cluster level, to the question “Can you tell me how problematic [is access to financing (e.g., collateral required) or financing not available from banks] for the operation and growth of your business?”. This variable may also partially capture the effect of financing constraints faced by firms Source: BEEPS, World Bank.
Firm-level control variables
Cash flow	Profits/loss plus depreciation (CF) scaled by total assets (TOAS) Source: Amadeus database provided by the Bureau van Dijk
Ln (employees)	Natural logarithm of the number of employees (EMPL). Source: Amadeus database provided by the Bureau van Dijk
Ln (total assets)	Natural logarithm of total assets (TOAS) in million USD. Source: Amadeus database provided by the Bureau van Dijk
Sales growth	Sales (TURN)t minus lagged sales (TURN)t–1 scaled by lagged sales (TURN)t−1 Source: Amadeus database provided by the Bureau van Dijk
Leverage	Long-term debt (LTDB) plus current liabilities (CULI)) scaled by total assets (TOAS) Source: Amadeus database provided by the Bureau van Dijk
Ln (age)	Firm age, since the (local) incorporation. Computed as year minus year of incorporation plus 1 Source: Amadeus database provided by the Bureau van Dijk
Missing age	If age is missing, then missing age is equal to 1, otherwise 0
Country-level macroeconomic variables
Private credit/GDP	Private credit scaled by GDP. Private credit is the deposit by money banks and other financial institutions. Source: WDI (World Bank)
Market cap/GDP	Total value of all listed shares on the national stock exchange as a percentage of GDP Source: WDI (World Bank)
GDP growth	The annual percentage nominal growth rate of GDP denominated in the local currency Source: WDI (World Bank)
GDP per capita	Real GDP per capita in 2010 USD (a proxy for country income) Source: WDI (World Bank)
GDP	Real GDP in 2010 USD (a proxy for country size) Source: WDI (World Bank)
Control variables groupings
Macrocontrol variables	Consist of private credit to GDP, stock market capitalization to GDP, GDP growth, GDP in constant USD, and GDP per capita (constant USD)
Political uncertainty variables	Include the inversed CPI, inflation uncertainty, economic uncertainty, election uncertainty, investment freedom index, absence of violence, government cohesion, and government stability
Partial firm controls	Include Ln(total assets) and its square
Full firm controls	Include Ln(total assets) and its square, cash flow, Ln(employees), sales growth, leverage, Ln(age) and missing age