Export Boosting Policies and Firm Performance: Review of Empirical Evidence Around the World

1.1 Export Promotion Agencies

Given the interest in boosting exports, the number of national export promotion agencies (EPAs) and their export promotion programmes (EPPs) has grown significantly over the past 20 years (Cruz et al. 2018; Lederman et al. 2010; Olarreaga et al. 2020). In the 1990s, approximately 65% of existing EPAs in developing countries were created with the goal of supporting the transition from a heavy reliance on imports to exports (Cruz et al. 2018). Regardless of whether they are in developing or developed countries, many EPAs are public entities that receive a substantial amount of funding from the government, i.e. taxpayers (Van Biesebroeck et al. 2015), although there are already private firms providing similar replacement services (Cruz et al. 2018). It appears that public funding is necessary for existing EPAs and EPPs to function, so it comes as no surprise that some researchers have criticised the efficiency of agencies in developing countries (e.g. Lederman et al. 2010). The efficiency of public money spent on EPPs may be a particular concern when public budgets are tight and policymakers want to prioritise the most efficient interventions.

EPAs assist firms overcome circumstantial difficulties and become successful exporters (Munch and Schaur 2018). If there are private providers of export promotion services and public entities support exporters, the question is why governments intervene in the market with EPAs and what market failure they are trying to address. The main economic rationales for government support are to promote information spillovers and to address trade barriers such as information asymmetries (Aalto and Gustafsson 2020; Copeland 2008). Information asymmetries between the potential exporter and foreign customers and companies, as well as the lack of trust between actors in international business, can be eliminated if the potential exporter incurs the sunk costs of obtaining the needed information. Otherwise, a lack of information can lead to underinvestment, so EPAs aim to share risk with potential exporters and reduce information asymmetry. In terms of information sharing, successful exporting of products can lead to information sharing with other exporting and non-exporting firms about demand conditions for different types of products in a foreign market. Copeland (2008) concludes that information spillovers form the theoretical backbone of government export promotion, but empirical evidence of spillover effects requires further research. Cadot et al. (2013) provide examples of positive spillover effects from exporters to neighbouring firms producing similar products. However, if there are no information spillover effects, Copeland (2008) questions the need for an export promotion scheme.

EPPs can be directed at various promotional activities, such as building a country image through promotion and advocacy or various bundles of support services (Lederman et al. 2010: 257). Export support services may include training, technical regulations, quality standards, capacity building, logistics, customs packaging, pricing, as well as marketing services such as advertising, trade fairs, exhibitions, missions, and follow-up services offered by representatives abroad (Lederman et al. 2010; Munch and Schaur 2018; Volpe Martincus and Carballo 2010a). The range of support services is broad and includes partner search and matchmaking (meetings, recruitment, contact databases, and other forms of market entry assistance) (Munch and Schaur 2018) as well as market research and publications (datasets and information, surveys, publications) (Lederman et al. 2010). In addition, services may include the analysis of political and economic conditions, international law, and business plans (Munch and Schaur 2018; Volpe Martincus and Carballo 2010a).

Several studies evaluate the effects of EPAs and direct subsidies on exporters. Bernard and Jensen (2004) use a panel of U.S. manufacturing firms to examine the factors that increase the probability of entry into exporting. Their study found volatile entry and exit rates of manufacturing firms into the export market, with past exporters having a higher probability of re-entering exporting, while current exporters have a higher probability of remaining exporters in subsequent years. Since entry costs are substantial and the spillover effects of other firms’ export activity are negligible, Bernard and Jensen (2004) conclude that export promotion costs do not have a significant effect on firm-level export prospects. At the aggregate level, Rose (2007) used bilateral gravity model of trade to answer whether presence of the foreign missions is associated with country-level export growth. He finds that each additional consulate is associated with an increase in exports by six to ten percent. Finally, using survey data from 103 developing and developed countries, Lederman et al. (2010) find a statistically significant effect of EPAs and their strategies on country-level exports.

1.2 Export Boosting Policies

While EPP are the most commonly used policy, they are not the only policy to directly boost exports. Some authors (e.g. Broocks and Van Biesebroeck 2017) include financial subsidies under the umbrella of EPPs, but there are also other policies such as subsidised export loans (i.e. export discount credit programmes) and public export credit guarantees. For example, Defever et al. (2020a) provide evidence for subsidised export loans, Agarwal et al. (2018) for public export credit guarantees, while Defever and Riaño (2017) using a two-country model of trade with heterogeneous firms, provide a quantitative assessment of the effect that subsidies with an export share requirement have on exports and conclude that this type of subsidy increases exports more than an equivalent unconditional subsidy accessible to each exporter. All of these policies aim to boost exports and should therefore be reviewed together. Therefore, together with EPPs we defined these policies as export boosting policies (EBP):

Export boosting policies are defined as a group of public policies that aim to increase firm exports in the short-run. These include export-oriented capacity building, attending trade fairs, exhibitions, missions, partner search matchmaking, pricing, advertising, analysis of the context abroad (i.e. export promotion policy), export credit guarantees, subsidized export loans, or export-oriented public grants.

EBP could be particularly effective for the smaller firms (Atkin et al. 2017; Broocks and Van Biesebroeck 2017; Munch and Schaur 2018), primarily because they face higher information asymmetry. Credit export guarantees may be particularly relevant for guaranteeing larger shipments of smaller firms, while public grants and subsidized export loans could be particularly relevant for smaller firms with greater capital constraints (see McKenzie 2017; Srhoj et al. 2021).

1.3 The Position of Export Boosting Policies in a Wider Set of Public Policies

EBPs potentially contribute to firm exports, however EBPs are not the only public policies that contribute to exports. To understand this point, two dimensions are useful for placing EBPs within a broader set of public policies: the time horizon and the policy objective. Many public policies may have the indirect objective of increasing exports in the medium or long term^[3] (see also Coad et al. 2022; Bloom et al. 2019), however, EBPs are short-term in nature, have the immediate goal of increasing exports of target firms in the short-run, and typically include the words “exports” or “foreign market” in their title and description. International trade theories (e.g. those of Vernon and Posner) emphasise the importance of “research and development” (R&D) processes for innovation and thus for firm exports. Altomonte et al. (2013) suggested that EBPs should be considered “under the same umbrella” as innovation policies, yet they are usually analysed separately.

In particular, for small open economies, R&D grants and tax incentives can be considered EBPs, but they do fall outside our definition.^[4] To explain why, we provide Figure 1, which depicts the flowchart of the R&D process in two stages and a commercialization stage. Stage 0 includes basic research, which is usually considered “research” in “R&D” and is dominated by universities and research institutes (Aghion et al. 2021), while few private firms are actively involved in this stage. Stage 1 is applied research, which is also referred to as “development” in “R&D”, but is often referred to as R&D in the analysis of private firms, R&D investment, grants, and tax incentives (for review, see Dimos et al. 2022; Vanino et al. 2019; Zúñiga-Vicente et al. 2014). There is a time gap between conducting additional R&D activities, developing new patents, innovative products or services, and selling the new product in domestic or foreign markets (e.g. Vanino et al. 2019). EBPs focus on boosting exports in the short term during the commercialization stage, which can come after stages 0 and 1, but can also be independent of the R&D process and focused on completly standardized or imitative products.

Figure 1:

Flowchart of the R&D process. Source: Aghion et al. (2021: 207).

Other forms of state aid, such as public grants outside R&D that are not explicitly aimed at boosting exports are not considered in our analysis. The effects of public grants on SME performance in the EU has been reviewed by Dvouletý et al. (2021), while Kersten et al. (2017) provide an overview of other forms of SME financing in developing countries. We also do not focus on the impact of entrepreneurship support institutions, although these institutions may implement EBPs or provide information on EBPs to firms. For example, cluster policies allow for a finer division of labour and greater proximity between firms, lead to greater trust between firms, and facilitate trade credit between firms, but tend to affect firms’ export performance in the medium to long term (Figal Garone et al. 2015; Long and Zhang 2011; Schmitz 1999). Hallen et al. (2020) discuss business incubators and accelerators that can implement EBPs or provide information on EBPs. However, because business incubators and accelerators provide many activities to a small number of both domestic and foreign market-focused firms, among the firm participants of business incubators and accelerators it is difficult to dissentangle the impact of EBP on firm performance. Various types of entrepreneurship promotion infrastructure, such as entrepreneurship zones, export processing zones, or business parks, provide quick access to business infrastructure under favourable conditions. These zones can attract foreign direct investment and have positive effects on the productivity and exports of participating and nearby firms.^[5] However, since they do not always have to be export-oriented or require considerable time from the idea to the implementation of the policy measures and finally to the impact at the firm level, we do not deal with them in our study.

In summary, our contribution is to review EBPs and firm performance.

1.4 Research Statement

There is a vast array of empirical studies on the topic of EBPs with heterogeneous empirical rigour. Over the years, the sheer accumulation of such studies has led to some confusion among policymakers and practitioners about how to support and boost exports. The purpose of this paper is to systematically review robust empirical evidence on the effectiveness of EBPs. In our review, we included only studies with high methodological rigour, this way supporting the so-called credibility revolution in applied economics (Angrist and Pischke 2010). The importance of policy evaluation studies with high methodological rigour has been emphasized by many international organisations (e.g. World Bank, OECD, European Commission), many scientific studies (e.g. Khandker et al. 2010; Gertler et al. 2016; Storey 2017) and used as a search filter in systematic literature reviews of policy evaluation studies (e.g. Dvouletý et al. 2021). Since selection into an EBP is not random, various experimental and quasi-experimental approaches have been used, most of which falls on the matching techniques, two-step estimation methods (instrumental variables or Heckman), fuzzy regression discontinuity design (RDD), fixed effects methods and randomised control trials (RCT). We consider a range of EBPs and provide a structured review of 33 studies from 26 countries worldwide. Our study provides insights into the effectiveness of EBPs in terms of extensive and intensive export margins as well as firms’ production function and output. We show heterogeneity of effects with respect to different firm characteristics and discuss new evidence on spillover effects of export policies. We move from micro to macro effects and provide an overview of back-of-the-envelope calculations that shed light on macroeconomic effects. Finally, we discuss the future research agenda and provide recommendations for policymakers.

2.1 Methodology and Code

Systematic reviews are very convenient for surveying scientific research using quantitative methods (e.g. randomized controlled trials, quasi-experimental designs, etc.) (Tranfield et al. 2003). In order to provide a homogeneous analysis of a policy, this paper provides a systematic overview of microeconometric effects focusing specifically on EBPs and firm performance. To ensure a scientifically rigorous literature review we followed the guidelines of Xiao and Watson (2019) in defining the purpose and intended objectives of the search, clarifying a detailed and consistent protocol, and developing and refining the screening process and quality assessment. The first step of the research process was to minimize human error and bias using a data extraction form (Tranfield et al. 2003). To systematically find and select articles, we developed a search code based on the original code by Dvouletý et al. (2021: 17). The Thomson Reuters Web of Science Core Collection database (Clarivate Analytics 2020) was selected because it provides the most valuable collection of high-impact articles (Dvouletý et al. 2021). The next step was to form a code review panel composed of five senior scholars respected in the field of international economics, international trade, and/or firm-level public policy evaluation. We received feedback from the review panel in January 2020. After the experts in methodology and theory approved the code, we applied the following search code in the Web of Science database from February 15–20, 2020, to find relevant articles:

• TS ^[6]  = (export subsid OR export-promotion service OR export promotion OR export grant OR export credit guarantee OR export credit insurance facilit OR export rediscount credit OR subsid export loans OR trade missions)

• AND

• TS = (business OR firm OR enterprise OR compan)

• AND

• TS = (firm performance OR productivity OR profit OR employment OR sales OR revenue OR turnover OR value added OR export status OR export intensity OR import intensity OR capital OR raw material OR intermediate input OR wage bill OR destination countr OR number of product OR age OR location)

• AND

• TS = (experiment OR field experiment OR randomized evaluation OR random OR impact evaluation OR impact assessment OR counterfactual evaluation OR propensity score OR regression discontinuity OR diff-in-diff OR difference-in differences OR difference in differences OR OLS OR fixed-effect OR instrumental variable OR identification strategy OR impact on OR impact of OR causal

Structurally, our code consists of four parts reflecting the policy instruments used to boost exports, the firm as the unit of observation, the outcome variables of interest and the methods used to evaluate counterfactual impacts. At the end of this step, our initial dataset from the Web of Science database encompassed 228 research articles.

In addition, we thoroughly searched the Scopus database via the search engine using the search term “TITLE-ABS-KEY” and found no article that was not already included in the results of the Web of Science Core Collection database. We also searched for studies via Google Scholar search and ResearchGate search to include relevant working papers and press articles. To determine whether we may have missed important studies, we began by searching for cited references, i.e. we searched the reference lists of the selected studies. We also went through the tables of contents of the journals in which most of the 228 studies were published. Finally, we invited a panel of experts to suggest possible studies that might have been overlooked in the search. Based on the citation search and suggestions from the leading scholars in the field, we were able to identify and include 16 additional informative studies. Ultimately, a total of 244 articles met the research criteria and were included in the primary analysis.

2.2 Selection of Articles

In the next step, we downloaded the abstracts of all 244 articles that met the search criteria. Each of the three authors conducted a parallel independent evaluation, scoring each article from 5 (most relevant) to 1 (least relevant) based on the policy instrument, unit of observation, outcome variable, and method used. The mean and standard deviation were calculated from the three authors’ scores, after which all articles with a mean score of less than 3.67 were discarded, allowing us to finally include 64 articles in the structured table. In creating the structured table we followed the work of Dvouletý et al. (2021) and focused on several key variables: Country and policy programme, sample and time period, type of policy and target, dependent variables, methods, and key outcomes. In our review, we included only studies with high methodological rigour, supporting the so-called credibility revolution in empirical economics (Angrist and Pischke 2008, 2010).^[7]

In a further step, we obtained the full texts of the 64 selected articles and conducted an in-depth analysis of the relevance of the articles to our review. The main reasons for excluding an article were i) lack of empirical rigour, e.g. studies based on unidentified ordinary least squares or survey data using structural equation modelling^[8] ii) analyses conducted at the country or regional level but not at the firm level, and iii) a policy instrument that did not focus on promoting exports, i.e. we excluded impact evaluations of R&D grants. Based on the relevance of the articles, we have selected the final 33 studies that were included in Table 1 (see Appendix).^[9]

After preparing the first draft manuscript, its structure, and the draft Table 1, we conducted an in-depth analysis of the bibliography list of the 33 core studies during October 1–20, 2020, to identify articles that may not have been selected by our search code (Table 1). We found no additional robust studies and therefore produced the final list of 33 studies in Table 1. The selected articles were published in journals including the following: Quarterly Journal of Economics, European Economic Review, Journal of International Economics, Journal of Development Economics, American Economic Journal: Economic Policy, Review of World Economics, World Economy, Structural Change and Economic Dynamics, Canadian Journal of Economics, Economics Letters, Economics Bulletin, Applied Economics, Economic Policy, and Economia Politica.

3.1 Export Boosting Policies: Heterogeneous Design

As can be observed from the 33 EBP studies in Table 1, there may be underlying heterogeneity of effects because EBPs are not homogeneous. We begin our in-depth analysis by dividing the policies into demand-side and supply-side export boosting policies. This division is followed by a subdivision of supply-side types and a review of the supply-side policy bundle.

Research Design: Most empirical studies can be classified as observational studies that use data from customs, official statistical surveys or export promotion agencies to draw conclusions about the effects of EBPs by comparing the performance of firms that benefit from an export boosting policy measure (the treatment group of firms) with firms that do not take part in an export boosting programme (the control group of firms). Rare exceptions are studies that use an experimental approach. Here some firms are randomly selected to receive an export promotion measure and others are randomly allocated to the control group of non-treated firms (Atkin et al. 2017; Breinlich et al. 2017).

Methods: Observational studies must address a problem familiar from evaluations of any type of policy. A simple comparison of the average performance of firms with and without export promotion (treated vs. non-treated firms) cannot reveal a causal effect of export boosting policy because firms in the two groups may differ in several characteristics (observable and unobservable) that are relevant to performance. While this problem can be addressed by randomly assigning firms to the two groups, observational studies that compare firms from the two groups after this assignment is nonrandom (by self-selecting firms into the treatment group or by agency selection based on criteria not fully known to the researcher) must control for these differences between treated and nontreated firms before treatment. Most of the studies reviewed here do so by using variants of a matching approach, often combined with difference-in-differences (DiD). Some studies use IV methods (2SLS), fixed effects regressions, fuzzy RDD, or RCT.^[10]

The heterogeneity of the research design used, the methods employed, and the data examined makes it impossible to perform a formal meta-analysis of the quantitative estimates provided in the studies, so we must provide a more qualitative summary and review of the results.

3.2 From Export Boosting Policy to Impact

While the previous section found positive effects for all EBP types and country-level of development, this section examines the relationship between EBP and individual firm outcomes in more detail. As we discussed in the introduction, one of the distinguishing characteristics of EBPs is their ability to quickly generate positive effects on firm performance. This feature of EBPs is also of interest to policymakers who want to achieve results during their political mandate and can therefore combine EBPs with other medium-and long run policies and reforms to increase exports. Our Table 1 shows that ten articles^[19] examine the time to positive effect, and eight articles find that the effect occurs in the first year after the introduction of the EBP.

Since the goal of EBP is to increase exports, the first-order outcomes are the export-related outcome variables. For first-order outcomes, we report not only the direction of the effects (i.e. positive, negative, or nonsignificant), but also the magnitudes from the original studies (i.e. point estimates and standard errors). We divide first-order outcomes into five main types: first, the start of exporting by nonexporters, second, and third, the export-intensive margin variables: total exports and export intensity, followed by two extensive margin variables: the number of exported goods and the number of exported markets.

3.3 Heterogeneity of Export Boosting Policy Effects

The empirically identified heterogeneity of effects can help researchers understand the mechanisms by which export boosting policies operate and can help policymakers design more effective policies. It should be emphasised that heterogeneity was found along several dimensions and much remains to be learned. Most evidence in the structured review is found for firm size, with some evidence for sectors, and export experience.

3.4 Spillovers and Cost-Benefit Analysis

As shown in Table 1, all 33 articles examine direct effects on intensive or extensive export margin or firm performance. Of the 33 articles, six^[36] examine possible spillover effects on firms that do not directly benefit from EBP. Examining spillover effects is important for at least two reasons, one economic and one methodological. From an economic perspective, a true cost-benefit analysis (CBA) should calculate not only direct spillover effects but also indirect spillover effects. In particular, if the spillover effects are large, they should be included in the CBA calculation. In this sense, one of the economic justifications for government intervention with EPPs is to address market failures in the form of information spillovers, and therefore the study of spillovers is particularly important. Evidence from France (e.g. Koenig et al. 2010) suggests that when a firm is active in a foreign market, the probability of exporting increases for firms close to the exporter. In terms of methodological rationale, note that 28 articles identify the effects of EBPs using matching algorithms (alone or with DiD), two using the IV approach, two using RCT, and two using fixed effects estimation (see Table 1). Matching, IV, and RCT are all counterfactual methods with the same basic assumption, the stable unit treatment value assumption (SUTVA), which essentially assumes that there are no spillover effects from treated firms to control firms. When the SUTVA is violated, the estimates are biased: On the one hand, the effect would be overestimated if there are negative spillover effects on the control group such that the use of EBP leads to a decrease in the control group, while on the other hand, the effect would be underestimated if the treatment leads to an increase in the control group.

Finding 1

This article presents an arsenal of EBPs divided into two levels: the demand side and the supply side. We show that five types of supply-side EBPs have a positive effect. While we cannot draw conclusions about which of the supply-side policies yields the largest benefits per public euro invested. Two studies suggest that providing subsidies or soft loans for technological improvements and machinery is more cost-effective than subsidising working capital and marketing activities. Of the EBPs, few studies examine the effects of export credit guarantees and subsidised export loans. We call for further research on these two policies.

Finding 2

Combination of multiple EBPs has a more beneficial effect than a single intervention (Table 2). Current evidence focuses on export promotion activities, which are sometimes combined with public grants, but further evidence is needed on the bundling of different types of EBPs (i.e. with loans and guarantees). EBPs support firms in the final stage of commercialization, which may be after the R&D process. However, EBPs and R&D policies are usually analysed separately, although they should be “under the same umbrella” (Altomonte et al. 2013). Future research could address the heterogeneous effects of EBPs as a function of firms’ R&D intensity, public R&D subsidies, and tax incentives.

Finding 3

Access to foreign markets had a positive effect on firm performance, with the mechanism being learning-by-exporting – increased product quality, technical efficiency and productivity of firms that were randomly allocated an export opportunity (Atkin et al. 2017). This finding opens a debate on the use of public procurement as a means to build capabilities (Stojčić et al. 2020). For example, Srhoj and Dragojević (2023) estimate a positive impact of public contracts on employment in construction firms, with a mean public cost per job created at €58,600. However, due to learning-by-exporting mechanism, benefits might be larger for international procurement, and initial evidence in 19 low-income Sub-Saharan African countries show how cross-border procurement auctions can provide growth opportunities for firms (Hoekman and Sanfilippo 2020).

Finding 4

Export promotion agencies are frequently mentioned in the context of developing countries, which might have lead to an opinion of EBPs sole focus on less developed countries, but we show evidence of positive EBP impacts also in developed countries. Both less and more developed countries show favourable impacts of EBPS, and it is not clear in which country context EBPs are more effective. A country’s level of development is a potential source of heterogeneity in policy effects, for example, due to country-of-origin effects (Verlegh and Steenkamp 1999) or a country’s political risk (Moser et al. 2008). We therefore call for an international micro-level study with standardized dependent variables.

Finding 5

Most studies find positive impact on first-order outcomes at intensive and extensive margin. In particular, current evidence shows positive EBPs impact on whether non-exporters start exporting, on total exports and export intensity (intensive margin). Current evidence also shows positive impacts on the number of export products, and number of export markets (extensive margin), but it is not yet clear whether additional export products are new and innovative, nor whether firms manage to sell their products in more developed countries for a higher price.

Finding 6

For second-order outcomes, results are somewhat mixed. Current evidence show positive EBP impacts on firm sales, but mixed evidence for productivity, value added, profits and employment, with insufficient evidence for capital.

Finding 7

EBPs have heterogeneous impacts. EBPs are more effective for smaller firms. Most evidence for positive impacts in accumulated among manufacturing firms. EBPs seem to be effective for both, firms with export experience, and non-exporters with export potential.

Finding 8

One of the EBP indirect targets are spillovers to other firms. On the one hand, three out of five studies examining the spillover effects show positive results. On the other hand, there are also studies showing no or negative effects, which begs for more evidence to support the argument of positive spillover effects. Future research on this topic is critical, as spillover effects could significantly change the cost-benefit analysis (CBA) of EBPs.

Finding 9

Back-of-the-envelope calculations of CBA show that the value added or exports generated by EBP are much higher than the direct costs of EBP. Few studies use additional tax revenues instead of value added and exports to calculate benefits and show mixed results and lower cost-effectiveness of EBPs. Calculations tend to focus on causally identified benefits and direct costs, ignoring the costs of public officials involved in design, implementation and monitoring, or spillover effects. Future studies should include these other costs and potential benefits in the estimation.

There is still much to be learned on the EBPs impact. Luckily, there is considerable research interest in this exciting field. To highlight this point, we briefly discuss three interesting studies published in the late 2022. Manaresi et al. (2022) examine a novel EBP aimed at SMEs in Italy, which subsidizes contracting a temporary export manager for the period of six months. Study finds positive effects of this policy on revenues, profits, employment and productivity. Revenues increased substantially due to exports outside EU, while effects were particularly positive for smaller firms and firms with lower productivity. Buus et al. (2022) provide a detailed examination of mechanisms though which export promotion has an impact on firms. Study does not find a positive impact on prices, markups, quality, and marginal costs, but finds a positive impact on exports within markets. Finally, Cusolito et al. (2022) conduct a six-country experiment in Western Balkans, where treatment was a combination of training and consulting on topics such as search engine optimization and Facebook content to increase firms’ digital presence and better reach foreign customers. Study finds positive impact on the number of customers and exports.

As our review shows, export boosting policies are successful public policies implemented around the world, in developing and developed countries, and will remain an interesting field of academic research.