Terror in the City: Local Terrorism and Firm Exports

1 Introduction

Terrorism is a worldwide phenomenon. There have been major attacks since the early 1970s and the number of terrorist attacks has increased significantly since the beginning of the 21st century (Institute for Economics & Peace 2018). While the direct effects of terrorism are evidenced by numerous victims and their suffering, it also has indirect economic effects that may be intended or unintended by the terrorist perpetrators (Sandler and Enders 2008). Empirical studies provide evidence of negative macroeconomic effects of terrorism on investment and growth (Abadie and Gardeazabal 2003; Bandyopadhyay, Sandler and Younas 2014; Chen and Siems 2004; Choi 2014; Enders and Sandler 1996; Tavares 2004).

Macroeconomic research suggests that international trade could be a channel through which terrorism influences economic growth, as an increase in the number and intensity of terrorist attacks can adversely affect international trade. Empirical studies examining the relationship between terrorism and international trade at the macroeconomic level suggest that the empirical relationship between terrorism and trade is either negative or statistically insignificant (e.g. Bandyopadhyay, Sandler, and Younas 2018; Blomberg and Hess 2006; Egger and Gassebner 2015; Nitsch and Schumacher 2004). One explanation for these ambiguous results is that empirical studies in this area focus on macroeconomic relationships. Recently, Gaibulloev and Sandler (2023) have noted that one of the myths associated with terrorism is that terrorism has macroeconomic consequences for the countries that experience terrorist attacks. Gaibulloev and Sandler argue that the harmful effects of terrorist attacks dissipate with increasing distance from the site of the attack and may have little impact on a country’s overall economy. Significant local effects are quite compatible with small or insignificant macroeconomic effects. Thus, the lack of empirical evidence of a link between terrorism and international trade at the macroeconomic level does not necessarily mean that firms’ trade activities are unaffected by terrorism.

While previous macroeconomic research on the relationship between terrorism and international trade has provided valuable insights, our knowledge of the microeconomic links between terrorism and firms’ international trade activities is still limited, and it is still unclear whether the results of macroeconomic research based on aggregate country-level trade data can even be generalized to the microeconomic level, i.e. when using firm-level or region-level data. Empirical analyzes based on country-level data, for example, completely disregard regional differences in the number and intensity of terrorist attacks within countries. Some regions may be particularly affected by terrorism while others are not. Economic activity in provinces affected by terrorism may shift to safer provinces, so that economic losses in one place are offset by economic gains elsewhere (Gaibulloev and Sandler 2023). Indeed, there is empirical evidence that terrorist attacks are not evenly distributed within a country and can vary significantly between regions and over time (Mazhar 2019; Öcal and Yildirim 2010). This could mean that the internationalization activities of firms located in terrorism-ridden regions tend to be more affected by terrorism than firms located in other regions of the country. Of course, this depends on whether the effects of terrorism are localized or whether there are negative interregional spillovers from terrorism. Macroeconomic studies that rely on terrorism at the country level tend to obscure such relationships, which means that the results of these empirical studies do not allow conclusions to be drawn about the microeconomic relationships between terrorism and international trade.

This study attempts to fill this research gap by investigating the relationship between exports and terrorism at the firm- and city-level, thereby testing the microeconomic basis of macroeconomic research in this area. Recent research suggests that firms’ capabilities are determined by local environments, such as urban clusters, and that studies examining firms’ internationalization activities benefit from not only focusing on the country-of-origin, but also zooming in to the level of the city or a subnational region (Côté, Estrin, and Shapiro 2020; Mudambi et al. 2018). Therefore, our disaggregated approach complements macro-level research in two ways. First, this paper follows the call to consider the spatial dimension of terrorism and the corresponding dynamics over time (Gaibulloev and Sandler 2023; LaFree 2010). Our subnational approach allows us to empirically investigate whether firm exports are related to terrorist attacks in the cities where the firms are located and whether firm’ exports are related to terrorist attacks in other cities through interregional spillover effects.

Second, beyond the macroeconomic relationship between terrorism and international trade, this is the first study, to the authors’ knowledge, to use firm-level panel data to empirically examine the relationship between terrorism and exports. A negative correlation between terrorist attacks and exports is found by Osgood and Simonelli (2020), who examine the relationship between terrorism and vertical foreign investment by US firms. However, their study focuses on the very specific case of intra-firm exports, i.e. production abroad for sale at home. In contrast, our study examines firm exports as a whole, and we focus on Pakistan, which has been consistently ranked as one of the countries most affected by terrorism since 2002 according to the Global Terrorism Index (Institute for Economics & Peace 2015; 20–22). Since the economic impact of terrorist attacks is particularly high in countries that are severely affected by terrorism (Mirza and Verdier 2008), it is to be expected that the export activities of firms in such economies could also be severely affected by terrorism. Rather than conducting a cross-country analysis, we use a single country approach, which also implies that all firms in our sample tend to face the same institutional environment, e.g. trade regulations and legal frameworks. At the same time, we take into account sub-national differences in terms of the location of firms and terrorist incidents.

Our empirical strategy allows us to control for a substantial part of unobserved heterogeneity that may influence the results of empirical studies at the country-level. We use a panel data estimation with firm-specific and year-specific fixed effects to eliminate variation due to unobserved time-invariant factors (stable firm and city characteristics) as well as unobserved year-specific factors (systematic factors affecting all firms and cities in a given year) that may confound the relationship between a firm’s export ratio and terrorism. Our results may therefore be less affected by problems of aggregation and omitted variables than the results of previous studies. However, like previous empirical studies in this area, our study is based on observational data. Without a natural experiment or valid instrumental variables, it is not possible to identify causal effects. Therefore, we do not claim to be able to identify causal effects of terrorism and interpret our results with due caution, pointing to correlations or empirical relationships rather than (causal) effects.

Our empirical analysis is based on a unique panel dataset containing the State Bank of Pakistan’s financial data for 301 firms listed at the Pakistan Stock Exchange (PSX) from 1999 to 2014. To account for the regional dimension of terrorism, we have compiled detailed terrorism data from the South Asian Terrorism Portal (SATP) for different regions (cities) of Pakistan over fifteen years. Based on the number of terrorist incidents, the number of fatalities, and the number of injuries we constructed a City Terrorism Index (CTI) for each city following the Global Terrorism Index (Institute for Economics & Peace 2018). In addition, we also consider the impact of terrorism outside the city in which a firm is located by constructing a distance-weighted indicator for terrorism outside a city, assuming that firm exports tend to be stronger disrupted by terrorist attacks in geographical proximity than by terrorism in more distant regions. Our empirical panel estimations show that there is a negative and statistically significant correlation between a firm’s export ratio and the level of terrorism in the cities where the firms are located. The empirical correlations are strongest in the medium term, and slowly decrease over time. Although we also find some empirical evidence of a correlation between firm exports and terrorism in other regions, suggesting interregional spillovers of terrorism, the strongest correlation is found between firms’ export ratios and terrorist incidents in the same city.

2 Terrorism, Firm Exports, and Interregional Spillovers

Terrorism can be defined as “the threatened or actual use of illegal force and violence by a non-state actor to attain a political, economic, religious, or social goal through fear, coercion, or intimidation” (Institute for Economics & Peace 2022, p. 6).^[1] Individual terrorist attacks are quite random and unpredictable, creating a broader atmosphere of fear in addition to physical destruction and human suffering. Terrorism can affect the business operations of firms – even if this is not necessarily the main intention of the terrorists.

Terrorism can directly affect the capability of firms to innovate, produce and sell their products by disrupting R&D, organizational and production processes. For example, terror-induced uncertainty over returns from R&D projects reduces significantly firm’s R&D investments as firms opt to defer their investments (Li et al. 2022). In addition, employees may not show up for work or some investment plans may not be implemented due to uncertainty, which may negatively affect the scale of production, productivity or quality of products. The negative effects of terrorism on companies’ capabilities can also affect their exports (Bandyopadhyay, Sandler, and Younas 2018).

Terrorism can also impact access to markets, as it may detract firms’ business activities through a decrease in customer demand as well as additional costs of doing business and changes in government regulations and policies (Czinkota et al. 2010; Suder 2004). In particular, terrorism can impede firms’ access to world markets in at least three ways. First, terrorism can be conceptualized as an increase in trade costs, such as a hidden tax or tariff on trade, whereby transportation costs associated with bilateral trade increase when terrorism increases in trading countries (Egger and Gassebner 2015; Lenain, Bonturi, and Koen 2002; Mirza and Verdier 2008; Nitsch and Schumacher 2004; Walkenhorst and Dihel 2006). Second, terrorism increases uncertainty and creates an atmosphere of fear and mistrust among business partners. Consequently, firms might be unwilling to enter into business contacts with actors operating in an insecure and risky business environment. Business plans become less reliable, if not obsolete, and travel to terrorism-prone countries becomes more dangerous and more expensive (e.g. higher insurance premiums). Terrorism can deter international business contacts, as international customers are not only less familiar with local risks but are also more likely to choose alternative business opportunities and locations than local firms (cf. Becker and Rubinstein 2011; Greenbaum, Dugan and LaFree 2007). Even if there are alternative solutions for communicating with customers in foreign markets, the lack of face-to-face interaction can diminish the likelihood of future business deals. Tan and Colleagues (2024) show that suppliers located near terrorist attacks are 2.9 percentage points more likely to terminate business relationships with their main customers within two years following the attacks. Increased perceptions of risk in the supply chain by major customers is the main reason for terminating the relationship. Similarly, terrorism affects acquisitions as top managers of acquiring firms are often reluctant to invest in terrorism-afflicted areas, in part because they are less willing to travel to the headquarters of target firms near terrorist attack sites (Nguyen, Petmezas, and Karampatsas 2023). Third, trade is indirectly affected by legislative and organizational counterterrorism measures introduced by the government to reduce the likelihood of future terrorist attacks, which also raises the cost of doing business (Gaibulloev and Sandler 2019). For example, the goods and supply chain can be affected by delays and temporary disruptions due to longer inspections, enhanced border security and stricter safety regulations (Walkenhorst and Dihel 2006), which increases the cost of exporting (Lenain, Bonturi, and Koen 2002).

While previous research has not specifically examined the link between terrorism and firm exports, there are several empirical studies that examine the effects of terrorism on trade at the macroeconomic level, and their results suggest either that terrorism has a negative effect on trade or that the effects are statistically insignificant (e.g. Bandyopadhyay, Sandler, and Younas 2018; Blomberg and Hess 2006; Egger and Gassebner 2015; Nitsch and Schumacher 2004). These ambiguous or insignificant effects may be due to the fact that the focus of previous studies has been on the macroeconomic consequences of terrorism, which may have obscured the true consequences of terrorism, which are only identifiable at the micro level. Gaibulloev and Sandler (2023, p. 294) point out that “significant localized economic impacts are entirely consistent with little overall economy-wide macroeconomic impacts.” With regard to the relationship between international trade and terrorism, Egger and Gassebner (2015, p. 60) conclude: “The average terror event is relatively small and, for countries and country-pairs, infrequent. Certainly, that does not mean that terror does not matter. It says, though, that trade might be the wrong domain to look for big effects.” Furthermore, Nitsch and Schumacher (2004, p. 424 ff.) state: “In fact, it is possible that terrorism has almost no measurable effect on trade since the overwhelming majority of terrorist actions are operations with only local implications.” Given the relevance of the local dimension of terrorism for economic activities in general, Greenbaum and Hultquist (2006, p. 128) suggest that “future research should use the disaggregated nature of city-level data further to more carefully examine the spatial relationships among cities and to better understand the implications of a terrorist attack on the regional economy.”

Against this background, we argue that the local dimension of terrorism should be considered when examining the relationship between exports and terrorism, and that the most promising way to identify potential effects of terrorism on exports is to focus on countries that are heavily affected by terrorism and to analyze the link between terrorism and exports at the micro level (cf. Mazhar 2019). In particular, firms operating in terrorism-prone countries suffer from frequent and severe terrorist attacks, which can hamper their exports. But even within a country threatened by terrorism, not all regions and cities are affected by terrorism in the same way, and it is to be expected that the export activities of firms located in cities and regions with many and severe terrorist attacks will be particularly negatively affected by terrorism.

However, not only terrorism in the same city or region can affect firm exports, but also terrorism in other regions due to interregional spillover effects of terrorism, and these spillover effects can be positive or negative. On the one hand, terrorism in other regions can have a negative impact on firm’s access to world markets, for example through rising export costs. In addition, intense media coverage of terrorist attacks can trigger widespread uncertainty and fear among economic actors far beyond the city or region attacked (Nellis and Savage 2012). On the other hand, terrorism in other regions can also facilitate firms’ access to world markets because international customers can shift their imports from regions threatened by terrorism to other regions.

Due to the aggregation of data, these micro-level effects of terrorism may be blurred in empirical analyses that focus on the macro level, which could explain the ambiguous results of macroeconomic studies. Therefore, to shed light on the relationship between terrorism and firm exports at the micro level, we use firm-level data and city-level terrorism indicators in our empirical analysis.

3 Data

3.3 Terrorism in Pakistan: Differences Between Cities and Over Time

The intensity of terrorist attacks varies remarkably across cities (regions) in Pakistan and over time, as shown in Figure 1 for the City Terrorism Index (CTI) in five provincial capitals. On the one hand, there are cities where the intensity of terrorism is high, e.g. Peshawar, Karachi and Quetta. On the other hand, there are cities where the intensity of terrorism is low, e.g. Faisalabad. Although terrorism has generally increased since 2006, it is worth noting that not all cities are affected in the same ways. The CTI peaks differ not only in intensity (e.g. 1,112 for Karachi and 2,165 for Peshawar), but also over time (e.g. for Karachi in 2013 and for Peshawar in 2009). As can be seen in Figure 1, our measure of terrorism at the city-level (CTI) does not follow a clear or systematic pattern: The CTI may rise sharply in one year, followed by a sharp decline in the next. Moreover, each terrorist attack is quite random and unpredictable, which means that terrorist attacks can be seen as exogenous shocks from firms’ perspective.

Figure 1:

Terrorism in cities across Pakistan (CTI index). Source: South Asian Terrorism Portal (2018).

Figure 2 depicts the intensity of terrorism for all target cities between 2000 and 2007. Each pie chart represents a city and contains the CTI intensity for up to four two-year periods. Pie charts with a single shaded pattern indicate that terrorist attacks only occurred within one specific two-year period. Pie charts with multiple patterns mean that terrorism occurred in more than just two years. The larger the relative CTI value within a time period, the larger the shaded area in each pie. For 2000–2001, very few cities are affected by terrorist attacks. Thereafter, more and more cities are attacked, especially in the northwestern border provinces of Balochistan and Khyber Pakhtunkhwa (KPk), which share a border with Afghanistan.

Figure 2:

City Terrorism Index (CTI) in Pakistani cities for the 2000 to 2007 period. Source: South Asian Terrorism Portal (2018). Each circle represents the CTI intensity for a given city or region in the case of FATA, Gilgit-Baltistan, and Azad Kashmir. Patterns within the pie show the accumulated value of CTI for two years (e.g. 2000–2001). Dot in the circle represents the geographic location of the labeled city in the map of Pakistan.

Similarly, Figure 3 shows the intensity of terrorism for terrorism-affected cities between 2008 and 2015, showing a rapid expansion of terrorism across the country, with terrorist activities shifting from Balochistan and KPk to new cities in the eastern provinces of Punjab and Sind. Overall, the descriptive findings show that terrorist incidents and their intensity can vary remarkably between cities (regions) and over time, supporting the notion that terrorist incidents in Pakistan are rather unpredictable and random. Consequently, cross-regional and country-specific studies, that look at the country as a whole, ignore any subnational variation and thus do not take into account the heterogeneity of terrorism between cities and over time. Therefore, in our empirical analysis, we focus on the geographical and intertemporal dispersion of terrorism in Pakistan.

Figure 3:

City Terrorism Index (CTI) in Pakistani cities for the 2008 to 2015 period. Source: South Asian Terrorism Portal (2018). Each circle represents the CTI intensity for a given city or region in the case of FATA, Gilgit-Baltistan, and Azad Kashmir. Patterns within the pie show the accumulated value of CTI for two years (e.g. 2008–2009). Dot in the circle represents the geographic location of the labeled city in the map of Pakistan.

4 Econometric Specification

Our panel data allow us to control for unobserved heterogeneity in several ways. We use a two-way fixed-effects model in which we include firm-specific fixed effects that control for all unobserved variables that do not change over time, and time-specific fixed effects that control for all unobserved variables that affect all firms in a given year in the same way. Our econometric specification is as follows:

where ER _ijt is the export ratio of firm i (percentage of export sales in total sales) located in city j in year t, γ _i is a firm-specific fixed effect, γ _t is a time-specific fixed effect, and ∈ _it is an error term. The independent variables of interest are the lagged values of the Total Terrorism Index (TTI) where the maximum lag is five (T = 5). We also include three control variables, namely the logarithm of financial leverage (FL), the logarithm of sales (LS) and the logarithm of total assets (LA). In addition, all explanatory variables are lagged by one year for two reasons: First, terrorist attacks that occur at the end of a year, for example, cannot affect export ratios in the preceding months (also known as time aggregation bias) and by using lagged values we can rule out such cases (Egger and Gassebner 2015). Second, lagged values further reduce potential endogeneity problems. Finally, we take into account that the error terms of firms located in the same city may not be independent. Therefore, we make use of cluster-robust standard errors, i.e. robust errors are clustered at the city level.

Our two-way fixed effects model allows us to eliminate a substantial amount of variation that may include time-invariant and year-specific confounding factors that would otherwise lead to biased parameter estimates. It should be noted that the inclusion of firm-specific fixed effects eliminates the between-firm variation and our parameter estimates are therefore based on the within-firm variation over time. The inclusion of year-specific fixed effects ensures that our parameter estimates are based on city-specific variation in terrorism levels over time. Including only lagged values of the terrorism variables in our regression analyzes further reduces potential endogeneity problems. However, we acknowledge that controlling for firm-specific and time-specific fixed effects and including only lagged terrorism variables cannot completely rule out potential endogeneity biases due to unobserved variables.

More precisely, our panel estimator allows us to control for unobserved variables that are constant over time (time-invariant characteristics of a firm and the city in which it is located) and common year-specific effects, but not for unobserved city-specific variables that change over time. Therefore, endogeneity problems could arise due to omitted variables, which would mean that our estimated coefficients of the terrorism variables would be biased and could not be interpreted as causal effects. However, unobserved city-specific variables that change over time only lead to biased results if they are also correlated with changes in our terrorism measures, such as the City Terrorism Index (CTI). For example, one such unobserved variable could be changes in demand factors in export destination countries, which could affect changes in firms’ exports. However, we find it unlikely that such demand-side factors are city-specific and thus correlated with the year-to-year change in terrorism in a given city. While lower demand for Pakistani imports in destination markets may be caused by factors such as exchange rate changes, recessions, changes in tastes/preferences, or simply trade diversion to other markets, these factors are likely to affect exports by firms in all cities, but are unlikely to correlate with city-specific annual changes in terrorism. Therefore, we argue that such unobserved demand-side factors are likely to be captured by time-specific fixed effects, as they affect all cities in a similar way. Another unobserved variable could be annual changes in local socioeconomic conditions in the cities where the firms are located. Unobserved local socio-economic conditions that do not change over time are already captured by the firm-specific fixed effects in our analysis. However, our econometric specification does not allow us to account for unobserved city-specific variables that change over time.

Of course, we have thought hard about how to solve the problem of unobserved time-varying city-specific variables. One possible solution would be to use city-specific instrumental variables that change over time. Unfortunately, reliable city-level information is not available for Pakistani cities due to lack of data. However, even if such data were available, it would be difficult to find variables that are sufficiently correlated with year-to-year changes in the City Terrorism Indices (CTIs) while being exogenous and satisfying the exclusion condition. Our descriptive analyzes of the variation in terrorism in Pakistan across cities and time show that CTIs often rise sharply in one year and fall sharply in the next. Since individual terrorist attacks tend to be random and unpredictable, annual CTIs tend to exhibit an erratic pattern. However, this may also suggest that unobserved time-varying city-specific variables do not pose too much of a problem for our analysis, as the generally more uniform changes in city-specific socioeconomic conditions are unlikely to follow the same temporal pattern as CTIs and are therefore unlikely to be strongly correlated with them. Nevertheless, we interpret our results with due caution, as our coefficient estimates are correlations or empirical associations that do not necessarily reflect causal effects.

5 Descriptive Statistics and Estimation Results

6 Discussion and Conclusion

Using a unique dataset that merges information from 301 firms listed on the Pakistan Stock Exchange with measures of terrorism in 105 cities/regions in Pakistan between 1999 and 2014, this study provides insights into the relationship between firm exports and terrorism at the micro level. While Pakistan has been one of the most terrorism-plagued countries in the world over the last two decades, our data shows that the level of terrorism varies significantly across Pakistani cities and also over time.

Overall, our empirical results point to a negative correlation between firms’ export ratio and terrorism. The negative correlation is quite stable in the years following a terrorist attack, which may suggest that exports might be impaired by terrorism for longer than expected. Moreover, we find robust empirical evidence for a correlation between firm exports and local terrorist attacks that occur in the same city where a firm is located. We also find somewhat weaker empirical evidence for an association between export performance and terrorism in other cities, suggesting possible interregional spillovers of terrorism in Pakistan.

While previous studies have focused on the empirical analysis of the relationship between terrorism and international trade at the macro level and have found ambiguous empirical evidence for such a relationship, we go beyond the aggregate country-level analysis and examine the underlying microeconomic relationship between terrorism and exports using firm-level data, taking into account the geographical variation of terrorism within a country and over time. In doing so, we show that firms’ exports are negatively correlated with local terrorism and thereby providing empirical evidence at the micro level for a relationship between terrorism and exports found at the macro level (e.g. Egger and Gassebner 2015; Gassebner, Keck, and Teh 2005; Nitsch and Schumacher 2004).

Moreover, we find that the correlation between firm exports and local terrorism is negative and statistically significant, while domestic sales appear to be unrelated to local terrorism. The latter result may seem surprising at first glance, as it can be assumed that domestic sales are also negatively associated with local terrorism. Empirical evidence suggests, for instance, that successful terrorist attacks in a region lead to fear and reduce consumer confidence in the respective region (Brodeur 2018), and the study by Gaibulloev, Oyun, and Younas (2019) suggests that there is a negative link between terrorism at the district level and subjective financial well-being at the individual level in Pakistan. However, our measure of domestic sales at the company level includes not only sales to consumers in the same region where the company is located, but also sales to other companies in the same region and sales to customers in other regions of Pakistan.

Theoretically, this raises the question of how the observed difference in the empirical relationships between local terrorism and domestic sales and local terrorism and export sales can be explained. It can be expected that the negative effects of local terrorism on (production) capabilities inhibit domestic sales and exports in a similar way. Hence, it is more likely that the differences in the relationships between local terrorism and firm exports and local terrorism and domestic sales are due to differences in the effects of local terrorism on access to domestic and foreign markets. We have argued that firms operating in highly terrorism-prone cities suffer from frequent and severe terrorist attacks because customers may switch to business partners in other regions of Pakistan or even other countries because they expect less insecurity and lower trade costs. International business customers in particular may be deterred by local terrorism because they are less familiar with local conditions than domestic business customers and may therefore avoid doing business with firms operating in terrorism-plagued regions and cities altogether. In contrast, domestic customers located in the same or a different region of Pakistan are likely to be more familiar with local conditions and may be less deterred by local terrorism. Although the relationship between firms’ domestic sales and local terrorism is not the focus of our study, this additional finding is interesting in its own right and should be investigated further in future research.

Our results suggest that terrorism may be a major obstacle for the exports of firms located in terrorism-ridden regions, despite trade and market liberalization policies that allow firms in emerging economies to more easily discover and exploit foreign market opportunities through exports (Aulakh, Rotate and Teegen 2000; Johanson and Vahlne 1977). As the number of terrorist attacks has increased dramatically since September 11, 2001, especially in emerging markets (Institute for Economy & Peace 2015; 2022), the export performance of firms in terrorism-affected regions in countries such as Pakistan can be severely affected by local terrorism.

Although we believe that our study contributes to a better understanding of the links between local terrorism and firm export activities, our study is not without limitations. First, as mentioned in the introduction, we do not claim to speak about causal effects of terrorism. Since, like previous empirical studies on the relationship between terrorism and international trade, we do not have a true natural experiment to identify causal effects, we cannot rule out the possibility of omitted variables. Nonetheless, our empirical analysis provides empirical associations that provide an interesting starting point for future research in this area. Second, due to the limited availability of reliable firm-level data in Pakistan, we focus on a specific type of firms, namely those listed on the Pakistan Stock Exchange. The companies in our sample may differ from other listed companies and non-listed companies and therefore may not be representative of all companies in Pakistan, and they may have multiple production locations. We implicitly assume that a company’s headquarters and manufacturing location are the same. While most manufacturing firms in Pakistan have their main production site at their headquarters, we cannot exclude the possibility that some firms in our sample have more than one location in Pakistan. Therefore, terrorism in cities where firms have subsidiaries could also have a negative impact on the business activities of firms. Admittedly, our City Terrorism Index (CTI) does not fully reflect the impact of terrorist attacks in locations other than headquarters, but we do take into account the impact of terrorism in other regions, at least through the Interregional Terrorism Index (IRTI). If firms have subsidiaries in other cities, the impact we estimate may even underestimate the real impact of terrorism on firm exports. In addition, the location of a firm’s headquarters may be particularly important for a firm’s access to global markets, as foreign customers are particularly familiar with a firm’s headquarters and tend to visit it on their business trips. Finally, our study is a single-country study focusing on Pakistan. This approach has its merits, as Pakistan can be representative of a developing country plagued by terrorism. However, it could also be criticized that the results may not hold for companies in other countries.

We suggest several avenues for future research: First, future studies could examine the relationship between firm exports and local terrorism for other countries affected by terrorism to see if our findings can be generalized. For example, a comparison between Pakistan, Nigeria, India, and Colombia, all of which have been affected by major terrorist attacks but have very different political, economic, and social situations, could be instructive. Moreover, it can be expected that firms will have varying degrees of willingness and ability to adopt counter-terrorism strategies and plans, which in turn may lead to heterogeneity in the impact of local terrorism on firm exports. Some firms may implement formalized programs and contingency plans to respond to terrorist attacks, while others may not. To better understand how terrorism affects exports and how firms deal with the threat of terrorism, future empirical research could further investigate at how firm exports are affected by terrorism and what measures firms take to counter terrorism.