3. Adoption of Investment Facilitation Measures across Economies

The IFI provides detailed data for 142 WTO Members, including over 90% of the IFDA signatories. It also includes key Members not participating in the IFDA (e.g., India and the USA). Based on the 2024 World Bank classification, the IFI covers 54 high-income, 72 upper- and lower-middle-income, and 16 low-income countries, accounting for 98.2% of global inward FDI stock and 97.6% of FDI flows in 2019 (UNCTAD, 2023).

Figure 1 illustrates the wide variation in the adoption of investment facilitation measures across the 142 economies (grey bars). IFI scores range from 0.22 for the Central African Republic to 1.76 for the Republic of Korea, and the median score is 1.04. On average, economies have adopted 49% of the 101 IFI measures, though there is significant variation: high-income countries have adopted over 62%, while low-income countries have adopted only 29%.Footnote ⁹ Thus, the adoption level of these measures is strongly linked to the country’s level of economic development. The distribution of IFI scores by income level confirms that low-income countries have the lowest average (0.55) and median (0.59) scores, while high-income countries have the highest (1.26 average, 1.29 median). Middle-income countries fall in between, with averages of 0.77 (lower-middle) and 0.98 (upper-middle). However, some low-income countries, like Uganda (1.12), outperform the median upper-middle-income country (Argentina, 0.99) and even some lower-scoring high-income countries such as Guyana (0.40), Barbados (0.61), or Panama (0.75).

Source: Authors for IFI scores, UN for country codes, World Bank for income level.

Figure 1. IFI score per economy and reform gaps relative to the IFDA.

The reform gap relative to the IFDA, shown by the coloured bars in Figure 1, reflects the flipside of current adoption levels. The largest gaps are seen in low-income countries with lower adoption rates (e.g., Djibouti, Central African Republic, Chad, Liberia, Benin, Haiti, Eswatini), while high-income countries with higher adoption levels (e.g., United Kingdom, Netherlands, Republic of Korea, Japan, USA) exhibit the smallest gaps. The reform gap is divided into three categories based on the IFDA text: red bars represent gaps in binding ‘shall’ provisions, orange bars cover conditional binding commitments like ‘shall, to the extent practicable’, and yellow bars indicate gaps in non-binding provisions, such as ‘should’ or ‘may.’ Our assessment shows that even high-income countries have room to enhance their investment facilitation frameworks by adopting more provisions, both within and beyond the IFDA, to reach the maximum IFI score of 2.

The distribution of IFI scores by World Bank regions offers further insights. The lowest median scores are in Sub-Saharan Africa (SSA) and Latin America and the Caribbean, with 0.67 and 0.75, respectively. The Middle East and North Africa (MENA) region scores a median of 0.99, while South Asia reaches 1.09 and East Asia and the Pacific scores a median of 1.13. European and Central Asian countries perform better with a median of 1.31, while North America stands out with an average and median of 1.64, driven by the high scores of Canada and the USA. Sub-regional disparities are also significant. For example, in the Asia-Pacific region, the average IFI score of 1.12 is driven up by 7 high-income countries (e.g., Australia with 1.52, Japan with 1.65, Republic of Korea with 1.76) out of 22 sampled. Even within the EU, with its single market and high economic integration, there is notable variation, from Malta (0.83) to the Netherlands (1.64). The 27 EU countries face adoption gaps in areas such as focal point and review (seven measures unadopted by at least 22 countries), and the application process (7 measures unadopted by at least 19 countries).

4. Overview of the Policy Areas and Underlying Measures

Analysing the adoption of investment facilitation measures by policy area, we find that three areas – responsible business conduct and anti-corruption, electronic governance, and regulatory transparency and predictability – have relatively high adoption levels, with economies achieving on average 60%, 59%, and 56% of the possible maximum score, respectively (Figure 2). In contrast, the other three areas show lower averages: 48% for the application process, 36% for cooperation, and 33% for focal point and review. The largest gaps are in focal point and review and cooperation where 103 and 93 economies, respectively, score below 40% of the possible maximum. In the following, we provide a detailed overview of the six policy areas, highlighting their compounding measures and country-level adoption.

Notes: Whiskers illustrate the min/max values, boxes show first to third quartile, the horizontal bar represents the median, while x indicates the average for each respective group. The differences in policy areas’ averages are statistically significant at the 5% level, except for the following pairs: regulatory transparency vs. electronic governance, regulatory transparency vs. responsible business conduct, electronic governance vs. responsible business conduct, and focal point vs. cooperation. The policy area score on the y axis is normalized to 1 to allow for comparison of different areas. Source: Authors for the IFI data, income groups according to World Bank.

Figure 2. Distribution of IFI scores by policy area.

Regulatory transparency and predictability represents the most important policy area according to our expert survey, carrying a weight of 23%. It incorporates 23 individual measures, with economies achieving between 7% (Central African Republic) and 93% (Korea and the USA) of the possible maximum score of 0.46 in this area.

Investment facilitation emphasizes enhancing transparency and stakeholder involvement. Regulatory transparency can be achieved through measures such as online publication of all investment-related information, establishing special enquiry points to assist investors, and providing advance notice of proposed changes to laws, regulations, and fees. Regarding involvement of interested parties, the concept envisages an opportunity to comment on drafts of investment regulations and acts before implementation. Our findings reveal that over 96% of included economies provide online access to applicable legislation, regulations, and procedures, and more than 92% offer this information in one of the WTO’s official languages. Additionally, 95% of economies maintain an enquiry point (often via investment promotion websites), and 92% publish information on competent authorities. However, fewer economies disclose the purpose and rationale of laws (57%), publish application processing timeframes (46%), or make judicial decisions on investment matters available (8%).

Electronic governance covers 14 measures related to the use of information and communications technology (ICT) and single-window mechanisms. Based on the expert survey, this policy area is the second most important in the IFI, with a weight of 18.7% and a maximum possible score of 0.37. While Djibouti and Central African Republic have not adopted any measures in this area, 11 economies have achieved the maximum score: USA, Chinese Taipei, Finland, India, Israel, Luxembourg, Mauritius, Mexico, Oman, Sri Lanka, and Thailand.

Our results suggest that over 96% of economies in the sample have a national investment website for information purposes, but only 74% provide a minimal set of relevant details, such as on licensing requirements, fees, screening, and approval. Other positive findings include that competent authorities in 94% of economies use email or electronic tools to communicate with investors, 86% recognize electronic signature as legally valid and over 73% accept authenticated copies of documents. However, ICT-based single-window or one-stop shop measures are less common. Specifically, 44% of economies have a single entry point for submitting investment-related applications, but only 27% offer a fully operational single-window service with electronic processing. Furthermore, 60% provide phone or online contacts for complaints related to mandatory registrations through the single window, 58% allow receipt of business registration certificates online, and just 38% ensure regular updates on the single window portal. While simultaneous submission of mandatory registrations (e.g., business registry, tax ID, social security) is possible in 36% of economies, only 28% offer this service through an online platform. Finally, only 36% of economies allow all fees for mandatory registrations to be paid through the single window.

The policy area focal point and review includes 23 measures related to alternative dispute resolution, appeal procedures, cooperation, and information-sharing, as well as ombudsperson-type mechanisms. This area has an expert weight of 18% and a possible maximum score of 0.36. Analysed economies achieve between zero (Chad) and 89% (Republic of Korea) of the possible maximum score. Our data demonstrate that measures related to judicial review and appeal are widely adopted among economies. Of the analysed economies, 95% provide independent or higher-level administrative and/or judicial appeal procedures, 90% allow investors to support or defend their position in judicial reviews, while the decision of such reviews is based on submitted evidence and arguments in 89% of covered economies. Additionally, nearly 72% have set timeframes for providing appeal or review decisions, and 70% ensure sufficient time to study a contested decision and prepare an appeal. In contrast, dispute-prevention mechanisms are available only in 8% of economies. Similarly, most WTO Members lack a fully functioning focal point or ombudsperson-type mechanism to support investors. Only 28% of the economies have a focal point to guide investors on legislation, agencies, and processes, 24% allow a focal point to coordinate and handle complaints, and 30% enable online feedback channels to the focal point. Around 22% of analysed economies have focal points that assist investors in obtaining information from governmental agencies and respond to enquiries of different stakeholders. The least adopted focal point functions include suggesting corrective recommendations on illegal or unfair administrative measures (6%), holding regular meetings with foreign-invested companies and government officials to mitigate conflicts and facilitate their resolution (4%), operating a single window (4%), and enforcing solutions for investment complaints (3%).

The application process policy area deals with formalities (documents, automation, and procedures) as well as fees and charges. It includes 25 measures and has an expert weight of 17.6%. Economies in our sample achieve between 12% (Central African Republic, Guyana, and Niger) and 92% (Republic of Korea) of the possible maximum of 0.35.

None of the analysed economies provides a time period of more than 20 days between the publication of new or amended fees and their entry into force, making this the least adopted measure in the entire dataset. Only five countries – Albania, Austria, Denmark, Mexico, and Myanmar – adopt the ‘silent yes’ approach, where an application is considered approved if no response is received within a specified timeframe. Other measures in this area show better results: 88% of economies do not charge for answering enquiries or providing forms, 51% provide information on all fees and charges, 43% limit fees to the cost of rendered services, and 38% conduct periodic fee reviews, though only 25% adapt fees to changing circumstances. Furthermore, 80% inform applicants about their decision, and 73% accept applications year-round, but just 25% allow applicants to submit additional information to complete their application, and only 20% permit resubmission of previously rejected applications.

Another set of frequently adopted measures focuses on facilitating the entry and sojourn of investment personnel. Data suggest that 91% of economies publish their requirements for temporary entry of business visitors and 88% issue visas with multiple entries. On average, it takes 10 days, requires eight documents, and costs $97 to obtain a business visa. Moreover, 51% of covered WTO Members accept and process visa applications electronically, and 48% allow for visa renewal or extension.

The next policy area is cooperation, encompassing 11 measures related to cooperation among investment-related authorities at national and international levels. According to the expert survey, this area contributes 10.5% to the overall IFI score. While the EU countries score quite similarly (between 0.14 and 0.19, except for Malta and Belgium with 0.12), other countries achieve between zero (Djibouti, Sierra Leone, and Solomon Islands) and 95% (United Kingdom) of the possible maximum of 0.21.

In our sample, 91% of economies cooperate with neighbouring and third countries through multilateral or regional agreements containing investment promotion and facilitation provisions. Examples include the ASEAN Comprehensive Investment Agreement (ACIA), the USA–Mexico–Canada Agreement (USMCA), the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP), or the Regional Comprehensive Economic Partnership (RCEP). Over 70% of economies regularly organize business–government networking events with partner countries, and 58% support cooperation of investment agencies. However, only 13% have a specific inter-agency coordination body, 19% exchange information about domestic investors or investment opportunities with other countries, and 21% establish best practice sharing programmes. Unfortunately, only 20% of economies provide a comprehensive domestic supplier database that is online in one of the WTO languages and searchable by sector, product, location, and other criteria, along with local production capacity and corporate social responsibility.

The last policy area, responsible business conduct and anti-corruption, includes five measures based on fundamental international conventions and has an expert weight of 12.2%. This area shows the highest average adoption level at 60% with economies ranging from 20% (Brunei, Chinese Taipei, and Tanzania) to 100% (22 economies in our sample) of potential maximum score of 0.24.

According to our results, 11% of analysed economies have ratified at least three and another 87% at least seven of the eight fundamental International Labour Organization conventions on freedom of association, forced labour, discrimination, and child labour. Brunei and the USA are the only countries with no more than two ratified conventions. Moreover, 98% of economies adopt measures in accordance with the United Nations Convention against Corruption, while only 31% have measures to combat bribery of foreign public officials in international transactions. Around 58% of considered WTO Members adopt double taxation measures similar to the OECD multilateral convention to implement tax treaty related measures to prevent base erosion and profit shifting, and only 21% have a specific national action plan to implement the UN guiding principles on business and human rights.

5. Robustness of the Methodological Approach

We undertake a factor, principal component, and uncertainty analysis to address two key questions about our composite index: (i) does the correlation of measures support the allocation of individual measures into policy areas; and (ii) does our specific set of weights create a specific (potentially biased) picture of economies’ performance that changes considerably with alternative weights?

5.2 Correlations of Policy Areas and Principal Components

Despite the diversity of IFI policy areas, the data are well-suited for a composite index of the policy areas since their scores are relatively strongly correlated. While most correlations between the policy areas exceed 0.5, the average interim correlation is 0.62, and Cronbach’s Alpha amounts to 0.91. A Principal Component Analysis (Greenacre et al., Reference Greenacre, Groenen, Hastie, d’Enza, Markos and Tuzhilina2022) shows how much of the overall variation of the individual policy area scores can be explained by their common variation. It reveals that the first component (the greatest common variation of the six policy areas) explains 68% of their overall variation. Together with the second principal component, 82% is explained. Figure 3 shows that all policy areas load quite similarly on the first component, showing that all policy area scores exhibit substantial common variation. For the second component, it is a bit more differentiated: while two policy areas do not load on that component at all, two others do positively and two negatively. These results support using a composite index to capture the concept of investment facilitation. However, the correlations between the policy areas are not perfect, implying that distinguishing between different policy areas still provides additional insights.

Note: The graph illustrates the loadings of the six policy areas on the first and second principal component of a 142 (countries) x 6 (policy area score variables) data matrix, which explain 68% and 14% of the overall variation of the policy area scores, respectively. Source: Authors.

Figure 3. Loading plot for the first two components of the Principal Component Analysis.

5.3 Uncertainty Analysis

To assess the robustness of the overall IFI scores regarding the chosen weighting scheme, we conduct an uncertainty analysis (Saisana et al., Reference Saisana, Saltelli and Tarantola2005). It examines how uncertainty in the selection of input factors (i.e., policy area weights) affects the values of a composite index. We evaluate $s \in \left\{ {1,\,\, \ldots ,\,\,S} \right\}$ Monte Carlo simulations with different assumptions for weights. For each simulation $s$, weights are sampled independently from each other using a quasi-random scheme (Sobol, Reference Sobol’1967), based on a discrete uniform distribution from our expert survey responses. This process is repeated 14,366 times, which is the total number S of conducted Monte Carlo simulations.Footnote ¹⁰ After sampling, these weights are normalized to a unit sum and employed to evaluate the propagated uncertainty using two output variables illustrated by equations (1) and (2):

(1)\begin{equation}{Y_c} = \mathop \sum \limits_{q = 1}^Q P{A_{q,c}}{w_q}\end{equation}

The first output variable, ${Y_c}$, represents the total score of a given economy $c \in \left\{ {1,\,\, \ldots ,\,\,M} \right\}$. It is calculated as a sum over the product of each policy area score $P{A_{q,c}}$ and the respective weight ${w_q}$, where $q \in \left\{ {1,\,\, \ldots ,\,\,Q = 6} \right\}$ indicates the respective policy area.

Figure 3 summarizes the results of all 14,366 simulated scores by displaying for each economy the distribution of resulting ranks, ${\text{ }}rank\left( {{Y_{c,s}}} \right)$, with the fifth and 95th percentiles as bounds. Black bars indicate the median rank of each economy, while the red bars indicate the rank based on the original IFI score using expert weights, $ran{k_{ref}}\left( {{Y_c}} \right)$. For all economies, the original IFI ranking is very close or even equal to the median from the simulations. This implies that the original expert weighting provides an unbiased picture of the economies’ investment facilitation frameworks. The largest difference between the original rank and simulated median is 3 for Cuba and Lao PDR. The broadest range between the fifth and 95th percentiles is found for Lithuania and Taiwan, which indicates that those economies’ IFI scores are most sensitive to the choice of weights.

The second output variable defines the average shift in economies’ ranks as a function of ${Y_c}$:

(2)\begin{equation}\overline {{R_s}} = \frac{1}{M}\mathop \sum \nolimits_{c = 1}^M \left| {ran{k_{ref}}\left( {{Y_c}} \right) - rank\left( {{Y_{c,s}}} \right)} \right|\end{equation}

The statistic $\overline {{R_s}} $ is calculated as an average over M=142 economies of the absolute difference between economies’ reference rank based on the original IFI, $ran{k_{ref}}\left( {{Y_c}} \right)$, and the ranking obtained during a simulation ${\text{ }}s$, $rank\left( {{Y_{c,s}}} \right)$. This provides an average rank shift across economies for every simulation, allowing us to describe the empirical probability density function resulting from all 14,366 values. The median average rank shift $\overline {{R_s}} $ across all simulations is 3.37, meaning that 50% of the simulations show an average rank shift of 3.37 or less.

Overall, we conclude that our initial choice of weights provides a robust picture of economies’ IFI performance. Figure 4 indicates that economies with very high or low scores are least sensitive to the selection of policy area weights, as their scores are quite high or low across all areas. The highest sensitivity is found for economies that perform well in some areas but poorly in others, leading to greater rank shifts when weights vary.

Note: Red bars indicate the original IFI rank, black bars represent the median rank resulting from the 14366 simulations. If no black bar is indicated, the original and median value are equal.Source: Authors, UN for country codes.

Figure 4. Distribution of simulated ranks per economy.