The implications of IFRS adoption on foreign direct investment in poor countries

Globalisation has contributed to the acceleration of international capital transactions and has increased investors’ need to access homogeneous, reliable and comparable financial reports. The objective of the study is to investigate the impact of International Financial Reporting Standards adoption on foreign direct investment flows in poor countries. In order to achieve this objective, the propensity score matching method was applied on a sample of 38 poor countries between 2008 and 2014. Results indicate that International Financial Reporting Standards adoption has a positive impact on foreign direct investment flows in poor countries.


Introduction
Globalisation has contributed to the acceleration of cross-border capital transactions.In this context, foreign direct investment (FDI) has become a tool used by countries in pursuit of economic development.FDI entails a number of important benefits for poor countries, such as the introduction of new production processes, creating connections between different business sectors and enabling domestic companies to access international capital markets (Agrawal, 2013).The increasing degree of interconnection among global capital markets has generated a need for investors to access homogeneous, reliable and comparable financial information.Thus, it became essential to create a common financial language (Rakes and Shilpa, 2013).
The relationship between International Financial Reporting Standards (IFRS) adoption and FDI has been extensively investigated.Numerous studies indicate that the transition to the international accounting framework has led to increased FDI flows, particularly in developing countries (Marquez-Ramos, 2011; Gordon, Loeb and Zhu, 2012; Chen Ding and Xu, 2014).However, little attention has been paid to poor countries.Although, according to the literature, poor countries are part of the developing countries group, they may exhibit certain economic and social particularities relevant to the process of accounting harmonization (Perera, 1989; Irvine and Lucas, 2006).The present study contributes to the literature by examining the relationship between IFRS adoption and FDI growth in poor countries.
The research paper is structured as follows: the first section presents a review of the literature with reference to relevant studies addressing the relationship between IFRS adoption and FDI flows.The second section describes the research methodology and within the third section results are presented and discussed.The last section concludes the study.

Literature review
A number of studies suggest that FDI contributes to economic development in poor countries (Acaravci and Ozturk, 2012;Adeniyi et al., 2012;Rakes and Shilpa, 2013).Investments tend to be concentrated in less developed countries where higher economic growth rates can be achieved (Rakes and Shilpa, 2013).Scarcity of financial resources has prompted many poor countries to consider FDI as a key tool designed to facilitate the transfer of new production technologies (Hossein and Yazdan, 2013).
The acceleration of cross-border financial flows, between developed countries as well as between developed and developing countries, has contributed to the internationalization of trade, businesses and capital markets (Trabelsi, 2015).In this context, it became vital to develop a single set of accounting standards with the aim to achieve uniformity in financial reporting worldwide (Zeghal and Mhedhbi, 2006).The literature points to three relationships that can occur between IFRS adoption and FDI growth (Zeghal and Mhedhbi, 2006;Lasmin, 2011;Marquez-Ramos, 2011).The first one is the unidirectional relationship running from IFRS adoption to FDI growth.This relationship implies that the adoption and implementation of the international accounting framework in a country contributes to an increase in FDI flows.The second relationship is the unidirectional one running from FDI growth to IFRS adoption.Under this hypothesis countries are pressured to adopt the international accounting framework as they gradually integrate into the global economy.The third relationship identified is the bidirectional relationship between IFRS adoption and FDI growth.According to this relationship, the two variables are mutually dependent.Efobi and Nnadi (2015) argue that the use of a single set of global accounting standards reduces information barriers across capital markets.The authors invoke this argument to explain the relationship running from IFRS adoption to FDI growth.Differences between accounting standards may hinder the dynamics of cross-border capital transactions.Most frequently, foreign investors have less informational advantages compared to domestic investors.Consequently, transaction costs are higher for foreign investors.This contributes to a decrease in FDI flows (Efobi and Nnadi, 2015).The role of using a single set of financial reporting standards is to reduce information asymmetries in the investment decision-making process (Chen, Ding and Xu, 2014).
The transition towards the international accounting framework enables poor countries to access external capital sources (Yu and Wahid, 2014).This in turn should help increase liquidity and stimulate the financing of worthwhile projects (DeFond et al., 2011).
Gordon, Loeb and Zhu (2012) argue that countries FDI inflows can increase if financial statements prepared in accordance with IFRS exhibit a higher level of quality than those prepared in accordance with domestic standards.Accounting practices in poor countries are underdeveloped.Therefore, there is a higher probability for poor countries to experience a more significant increase in FDI inflows as a result of IFRS adoption in comparison to developed countries.The previous studies provide empirical evidence that IFRS adoption and implementation contribute to FDI growth.However, some authors argue that countries seeking FDI growth are prone to IFRS adoption (for instance Judge, Li and Pinkster, 2010; Lasmin, 2011).Lasmin (2011) uses the neo-institutional theory developed by DiMaggio and Powell (1983) to predict the likelihood of IFRS adoption by countries due to mimetic institutional pressures caused by FDI growth.The research has revealed that an increase in FDI flows generates the likelihood of IFRS adoption within a country.The observed relationship, however, is not statistically significant.
Guler, Guillen and Macpherson (2002) also use the neoinstitutional theory and empirically prove that the degree of network cohesion contributes to national adoption of international standards.According to the authors, FDI growth increases cohesion between countries involved in a transaction.Consistent with this line of reasoning, Judge, Li and Pinkster (2010) provide empirical evidence to support the theoretical relationship between mimetic pressures generated by FDI growth and IFRS adoption.
The study conducted by Marquez-Ramos (2011) emphasises the importance of examining the bidirectional relationship that might occur between IFRS adoption and FDI growth.The author also argues that there may be a number of factors affecting both variables simultaneously.While analysing the relationship between the adoption of the international accounting framework and FDI growth it is important to control the effects of factors that could affect both variables simultaneously, as this could lead to different results.
In light of the literature that investigates the economic effects of IFRS adoption, the general hypothesis of this study is developed: H1: IFRS adoption contributes to FDI growth in poor countries.In the present study, the research hypothesis is tested by means of the propensity matching method (Rosenbaum and Rubin, 1983).This methodological approach allows comparisons between countries with similar characteristics.

Research methodology
The use of the propensity score matching method in the analysis of the relationship between IFRS adoption and FDI growth has two methodological advantages compared to the classic linear regression model.First, the propensity score matching method is nonparametric.Therefore, it is not necessary to specify a parametric relationship between the dependent variable and independent variables included in the analysis.
Second, this method reduces the number of untreated observations (in the case IFRS non-adopters) to a subsample of treated observations with similar characteristics (IFRS adopters) (Tucker, 2011).These properties of the propensity score matching method prevent errors in estimating the average treatment effect of the treated observations (Balsmeier and Vanhaverbeke, 2016).
Testing the research hypothesis involves identifying the extent to which IFRS adoption contributes to FDI growth.
In the analysis based on propensity score matching the main indicator is the average treatment effect defined by Rosembaum and Rubin (1983) as: where r 1 i represents the reaction of unit i if it has been treated (T=1) and r 0 i represents the reaction of unit i in case the treatment (T=0) has not been applied.
By adapting equation ( 1) in accordance with the research question of the present study, the following equation is obtained: where FDI 1 i and FDI 0 i represent FDI of country i in case it has adopted IFRS (T=1) or if it uses domestic standards for financial reporting purposes (T=0).
When estimating the causal effect of IFRS adoption, FDI 1 i and FDI 0 i cannot be observed for the same country simultaneously.According to Rosembaum and Rubin (1983) in this case, equation ( 1) can be re-written as follows: where P is the probability to notice observation i for which T=1 within the statistical sample.
Similarly, in the case of the causal inference between IFRS adoption and changes in FDI flows, equation (3) becomes: Propensity scores are estimated through a probit model where the dependent variable is the treatment variable and the independent variables are those on which treated and control observations will be matched (Tucker, 2011).Subsequently to the estimation of the scores that capture similarities between countries, each treated observation is matched with the most similar control observation.Matchings are performed through stratification matching (Becker and Ichino, 2002;Tucker, 2011).
Stratification matching is based on dividing the observations in blocks.The average propensity score is computed for each block of observations.Subsequently, the difference between the average values of the dependent variable is computed for each block of observations based on the following equation: where: I(q)-observations from block q; Y T i -value of the dependent variable for the treated observation i; Y C j -value of the dependent variable for the control observation j; N T q -number of treated observations in block q; N C q -number of control observations in block q.
Within the stratification matching method, the average treatment effect (ATE) is the weighted average of all observed effects in each block and it is computed based on the following equation: where T i is the binary variable that indicates if observation i is treated (T=1) or not (T=0) (Becker and Ichino, 2002).
The main advantage of using the propensity score matching method in solving causal inferences is that it allows improving the conditions of a random experiment in order to estimate the causal effect as in a controlled one (Rosenbaum and Rubin, 1983).This method yields to reliable results only if the conditional independence assumption is satisfied.In this instance, the conditional independence implies that the decision of using a certain set of financial reporting standards is random and uncorrelated with FDI flows, once the vector of exogenous variables effects has been controlled for.The inferences performed through the propensity score matching method are valid only for those values of the propensity score for which matchings between treated and control observations are possible.The interval where these values can be found is called common support region (Tucker, 2011).

Research sample
The sample on which the research hypothesis was tested consists of low income and lower middle income countries (according to the World Bank classification), with active capital markets.The period of time between 2008 and 2014 was analysed in order to control for the effects of the business cycle and due to data availability.
The rationale for selecting this sample is represented by the international consensus concerning the low income levels of these countries.Furthermore, the international organizations have undertaken actions for the development of these countries and put pressure on them to adopt an international accounting perspective in order to facilitate the monitoring of implemented programs.

Variables
The variables used in this study are presented in Table no.2. The selection of variables was performed using the model developed by Gordon, Loeb and Zhu (2012).
According to the authors, this model includes the factors that are most frequently used in the literature to examine the evolution of FDI flows.Thus, 12 variables, for which it was possible to collect a complete data set were included in the analysis.

Data analisys and results
Descriptive statistics are presented in  Consistent with the propensity score matching algorithm, the probability of getting the treatment (in this case IFRS adoption) was estimated by means of a probit model.Within the model, the dependent variable is the treatment variable (IFRS), and the exogenous variables are those on the basis of which matching between treated and control observations will be made (Tucker, 2011).The probit model has the following form: where β i are the coefficients of the probit regression and ε i are the residuals.We can be more than 95% confident that these phenomena were not random and will be reflected by the population.
The chi 2 probability ratio is 181.95 and has an associated probability of 0.0000 showing that the research model is valid.
Propensity scores were computed using the probit model.Based on these scores, matchings between the two categories of observations were made.
The common support region was determined using the functions implemented in Stata 12.0.Thus, for the data set collected, the common support region is given by the interval [0.0351, 0.9999].Observations with propensity scores outside this range were dropped from analysis because it was not possible to find them a match.The final sample size is 197 observations.
In order to apply the stratification matching method, blocks of observations were created using the modules implemented in Stata 12.0.Within each block the average propensity scores of treated and controls must not be significantly different.Thus, six blocks of observations were obtained.

Conclusions
The objective of this study was to examine the relationship between IFRS adoption and FDI growth in poor countries.The general research hypothesis was tested on the sample of 38 poor countries between 2008 and 2014 using the propensity score matching method.
Results are statistically significant at 5% significance level and suggest that the transition to international accounting framework contributes to FDI growth in poor countries.
The relevance of the research results is subject to limits.First, the initial data collection was possible for only 38 countries.Based on this data a panel data set of 266 observations was obtained.This sample was subsequently reduced to 197 observations which allowed matching.Second, the model estimates the average effect of IFRS adoption on FDI growth once the effects of the vector of exogenous variables have been controlled for.Limited data availability allows us to operationalize ten exogenous variables.According to Tucker (2011), the existence of other exogenous factors omitted from the analysis may affect the validity of the results.Including other factors such as the exchange rate, the interest rate and the cost of labour into the analisys can open new research avenues.
The authors examine a panel data set of 124 countries between 1996 and 2009.Results indicate that IFRS adoption has led to an increase in FDI inflows.

2.1. Propensity score matching
Gassen and Sellhorn (2006)ear regression on a panel data set consisting of 208 developed and developing countries between 1996 and 2009.To examine the hypothesis regarding the distinction between the two categories of countries in terms of inward FDI, the authors use a difference-in-difference model.Lasmin (2012)employs the Cobb-Douglas production function to represent the relationship between physical capital, labour and efficiency.These parameters are used as control variables in estimating the effects of IFRS adoption on trade and investment activities.Chen, Ding and Xu (2014) apply a gravity model on a sample of 20 OECD countries between 2000 and 2005.The analysis revealed that FDI flows are positively associated with the degree of IFRS compliance of financial statements published by companies.Over the past years, the propensity score matching method has been frequently applied in accounting and finance research(Tucker, 2011).For instance,Gassen and Sellhorn (2006)use propensity score matching to investigate the determinants and effects of IFRS adoption on a sample of German companies.DeFond et al. (2014) use the propensity score matching method to examine the extent to which IFRS adoption affects the frequency of negative returns reported by publicly traded companies.

Table no
Consistent with prior literature, according to which IFRS is relevant only for countries with active capital market (Amiraslani, Iatridis and Pope, 2013) the sample includes 38 poor countries with active capital markets ( . 1).Out of these, 10 require or permit listed companies to use IFRS, 6 adopt IFRS between 2008 and 2014 and 22 apply domestic standards for financial reporting purposes of listed companies.Liberia, Leshoto, Pakistan, Cambodia, Lao, Sierra Leone, Somalia and Syria were not included in the sample because these did not have active capital market during 2008 and 2014.Zimbabwe, Republic of Congo, Ivory Coast, Egypt, Honduras, Kenya, Myanmar, Nicaragua, Nigeria, Papua New Guinea, Uzbekistan and Zambia were not included in the analysis due to data unavailability.

Table no . 3 .
(Stock and Watson, 2003)the mean, median, standard deviation, skewness and kurtosis indicate the absence of symmetry of the data series(Stock and Watson, 2003).

Table no .
5shows the results obtained after running the model on the data set.

Table no . 6. Descriptive statistics of the propensity scores Blocks of observations IFRS Number of observations Mean t-statistic p-value
IFRS adoption) effect was computed as the weighted average of all observed effects in each block of observations (Table no.7).average treatment effect (ATE) suggests that IFRS adoption has generated on average a 0.4410 units growth in FDI flows in poor countries.The result is statistically significant at 5% significance level and validates the general hypothesis of this study, according to which IFRS adoption contributes to FDI growth in poor countries.This result is consistent with those obtained by previous studies (Marquez-Ramos, 2011; Gordon, Loeb and Zhu, 2012; Chen Ding and Xu 2014).

Table no . 7. Average treatment effect Method Number of treated observations Number of control observations ATE Standard deviation t-statistic
* significant at 5% significance level.