Introduction
Economic sanctions are a fixture of American foreign policy. They are deployed to address all manner of disputes in the international system. Though their record of coercing targets to change their behavior is fairly limited, US policymakers value sanctions as a non-kinetic option for signaling resolve and punishing bad behavior in the international arena.Footnote 1
A key development in economic statecraft has been foreign policymakers’ use of “targeted sanctions,” which are imposed on firms, elites, and other entities, rather than on entire economies or economic sectors.Footnote 2 A policy innovation of the late 1990s, targeted sanctions emerged in response to the grim international consensus that comprehensive economic sanctions against Iraq earlier in the decade were responsible for more civilian causalities than the Gulf War itself.
As the “precision-guided munitions of economic statecraft,” targeted sanctions are more humanitarian and better able to limit collateral damage in the sender and target countries than more comprehensive measures.Footnote 3 However, the perceived low cost and ease of implementing targeted sanctions have arguably lowered policymakers’ sanctions inhibitions.Footnote 4 Targeted sanctions are so prevalent today that experts warn of their “re-comprehensivization.”Footnote 5
The US government's entity list, a targeted trade sanctions list managed by the Department of Commerce's Bureau of Industry and Security (BIS), embodies this acceleration in targeted sanctions imposition. Established in 1997, the entity list was originally limited to entities suspected of perpetuating weapons of mass destruction proliferation.Footnote 6 At the end of 2007, there were seventy entities on the list from eight countries.Footnote 7 Then, in 2008, the US government expanded the policy scope of the entity list to include “entities acting contrary to the national security or foreign policy interests of the United States.”Footnote 8 By year-end 2019, there were 1,229 entities on the list from seventy-six international jurisdictions.
Beyond its rapid growth, another striking feature of the entity list is that it increasingly hosts large firms based in interstate rivals. For instance, since 2016, US policymakers have used the list to sanction a variety of multinational Chinese technology companies, such as Zhongxing Telecommunications Equipment (ZTE) Corporation, Huawei Technologies Co. Ltd. (Huawei), Dahua Technology, and Hikvision, among others. Similarly, many of Russia's largest nonfinancial companies, including Gazprom, Lukoil, and Rosneft, are on the list. This qualitative shift in entity list sanctions is consistent with US foreign policymakers’ renewed emphasis on great power competition.Footnote 9
Among their many potential effects, economic sanctions are expected to generate costs for the sender country. Indeed, sanctions work by denying domestic firms’ ability to do business with targets. We know very little about the magnitude of the costs that sanctions generate for sender countries, perhaps because “it is just plain hard to quantify the costs to the sender country.”Footnote 10 Still, there have been efforts to quantify lost or diverted trade in the sender country due to comprehensive or selective sanctions.Footnote 11 A recent study also explores how sanctions affect US firms’ stock price volatility.Footnote 12 Less attention has been paid to the cost of targeted sanctions, which are the instrument of choice for sanctions policymakers worldwide.Footnote 13
Targeted sanctions play out on a firm-to-firm or firm-to-individual level. The domestic effects of targeted sanctions in the sender country are felt most immediately by the firms that have business ties to sanctioned entities. Targeted sanctions undoubtedly incur lower costs for sender countries than more comprehensive measures, but this has arguably led to a sanctioning moral hazard.Footnote 14 The proliferation of targeted economic sanctions compels renewed efforts to quantify the compliance costs these measures incur for sender-country firms. A recent analysis of targeted sanctions calls for more research into this precise issue.Footnote 15 In this article, I answer that call for an important contemporary context by examining how the US government's addition of certain Chinese technology companies to the entity list has affected the sanctioned companies’ American suppliers.
Quantifying the cost of entity list sanctions against Chinese technology companies to US firms is a timely and worthwhile undertaking. These measures have been a salient component of the US government's approach to what political scientists and technology analysts call the US-China technology Cold War.Footnote 16 Economic sanctions generally entail compliance considerations for sender-country firms, but the stakes are high and the potential blowback is significant when sanctions target internationally active firms from large, interstate rivals.
The urgency of this task is further supported by these instruments’ projected efficacy. Unilateral targeted trade sanctions, such as those under investigation, are expected to be among the least effective forms of sanctions because they are difficult to enforce and incentivize noncompliance.Footnote 17 Moreover, sanctions between great power adversaries are typically ineffectual because rivals expect frequent confrontations, and submitting to pressure is very costly.Footnote 18
Policymakers also claim to care about the cost of sanctions. Though the BIS's first guiding principle makes clear that national security is its top priority, its second guiding principle revolves around “ensuring the health of the US economy and the competitiveness of US industry.”Footnote 19 Considering the potential ramifications of unilateral targeted trade sanctions against Chinese technology companies, their projected inefficacy, and policymakers’ stated objectives, measuring the cost of these sanctions to US firms has important scholarly and policy implications.
Using an event study methodology, I examine how the additions of ZTE, Huawei, and a group of artificial intelligence and surveillance companies to the entity list in 2016 and 2019 affected the firm value of these companies’ US suppliers.Footnote 20 I find that the addition of these firms to the entity list led to a statistically significant 220 basis points reduction in their US suppliers’ risk-adjusted stock returns. In aggregate, the sanctions wiped out about nineteen billion USD in market capitalization.
Complementary cross-sectional analysis of firm-level characteristics shows that US companies that maintained deeper ties with the targeted entities were more negatively affected by the sanctions. This supports the direct connection between entity list sanctions and the firm value of US suppliers. A supplementary finding of the cross-sectional analysis is that, even after accounting for revenue dependence, firms that allocate more resources to research and development are more negatively affected. This suggests that the sanctions are particularly harmful to the United States’ own technology sector.
Collectively, these findings contribute to sanctions scholarship by advancing our understanding of the costs that targeted trade sanctions against internationally active companies incur for sender-country firms. Further, US policymakers, whose stated objectives include ensuring the competitiveness of US businesses, can use these findings to more thoroughly weigh the tradeoffs of targeted sanctions imposition. More broadly, the findings will hopefully spur wider research and theoretical development on the cost of sanctions for firms.
This article also makes a methodological contribution to the study of sanctions. Though data aggregation is challenging in this context and care must be taken to address threats to inference, the event study methodology is generally well-suited for measuring the cost of targeted sanctions to publicly traded companies. Future research could employ a similar framework to study the cost of targeted sanctions in other contexts.
I proceed by laying out the theoretical reasons why targeted trade sanctions are expected to incur costs for sender-country firms. Next, I outline the research design employed in this article, including case selection, data, application of the event study methodology, and the firm-level cross-sectional analysis. I then discuss the article's main findings. Finally, I summarize the key implications of this research.
How targeted trade sanctions affect sender-country firms
The mechanics of sanctions are such that, even though the sender intends to coerce a foreign target to alter its behavior, the specific measures it takes are geared toward economic agents in its own country.Footnote 21 This leads to the “general rule that sanctions entail costs for the sender country.”Footnote 22 Yet quantifying the cost of sanctions to the sender has remained elusive. Nonetheless, a small body of literature has measured sanctions’ domestic economic ramifications in certain contexts.
Hufbauer et al. (Reference Hufbauer, Elliott, Cyrus and Winston1997), for example, estimate that US economic sanctions imposed in 1995 likely reduced exports to target countries by fifteen to nineteen billion USD. They further calculate that lost exports would have, all else equal, resulted in about 200,000 lost jobs. Performing a similar study of US economic sanctions in 2000, DeRosa (Reference DeRosa, Hufbauer, Schott, Elliott and Oegg2009) finds more modest and statistically insignificant lost exports of 2.8 billion USD. Looking at a narrower group of exports, Richardson (Reference Richardson1993) estimates that US export controls on high-technology items deter about 5 percent of such exports.
More recently, Webb (Reference Webb2020) examines how US sanctions affect domestic firms’ stock price volatility. Focusing on eight cases in which the United States imposed or threatened to impose sanctions against seven different countries, he finds that select firms with economic exposure to these countries endured a spike in volatility, while comparable firms without exposure to the sanctioned countries did not. The present article seeks to advance the literature on the cost of sanctions by drilling down on the mechanism behind targeted sanctions, the predominant contemporary form of economic sanctions.
From a theoretical standpoint, targeted sanctions are likely to generate direct and indirect costs for certain firms in the sender country. When a government imposes a targeted sanction against a foreign entity, especially a foreign firm, sender-country firms that have business relationships with the sanctioned entity are most directly affected. Often, these firms must forgo the revenue stream or supply channel associated with the target. The forgone business can be diverted to international competitors or moved in-house at the target, especially in the case of unilateral targeted sanctions. Targeted sanctions can also inflict reputational damage on sender-country firms’ reliability as suppliers and customers.
The imposition of certain forms of targeted sanctions may protect sender-country firms from these costs. Most notably, compared to targeted trade sanctions, targeted financial sanctions are “easier to enforce, harder to evade, and may spur market-reinforcing effects.”Footnote 23 This is especially true for US-led targeted financial sanctions, due to the dominant position of the US financial system.Footnote 24 In effect, international economic agents have few alternatives to dealing with the US financial system and pay a high price for flouting its rules. Targeted trade sanctions, particularly of the unilateral variety, are not expected to shelter sender-country firms from sanction costs. They are difficult to enforce, incentivize trade diversion to a sender's international competitors, and often upset international economic partnerships.
Use of the entity list constitutes a targeted trade sanction, and list members are usually sanctioned unilaterally by the US government. Members of the entity list are typically banned from receiving items subject to the US Export Administration Regulations (EAR), which includes virtually all items exported from the United States and many foreign exports containing American components.Footnote 25 Without a license from the commerce department, US firms that have supply relationships with an entity placed on the entity list must sever ties with that entity, which leads to a loss of revenue.Footnote 26 The sanctioned entity will, in turn, likely seek to establish supplier relationships with the US firms’ international competitors or aim for greater self-sufficiency.
The End-user Review Committee (ERC), the interagency body that determines which foreign entities are added to the entity list, has flexible authority to sanction foreign entities without consulting other relevant stakeholders.Footnote 27 Because entity list sanctions are often sudden and unexpected, US firms that have business ties with sanctioned entities may also suffer reputational damage and business model uncertainty. All else equal, entity list sanctions should have an adverse impact on affected US suppliers’ bottom line and international competitiveness. With these potential consequences in mind, I establish two theoretical expectations:
Expectation 1: Placing entities on the entity list that have US suppliers is likely to reduce those US suppliers’ firm value.
Expectation 2: Firms that have deeper ties with targeted entities are likely to be more negatively affected by entity list sanctions.
Research design
I test my expectations regarding the cost of entity list sanctions for US suppliers by applying the event study methodology, which has been deployed extensively in finance and business research to examine how unanticipated events affect the value of a firm.Footnote 28 As in this article, event studies typically examine the behavior of firms’ stock returns around an event. The granularity of daily stock market returns, their reflection of current and future business prospects, and their quick adjustment to relevant news allow for plausible inference of the effects of specific events on the value of a firm. In the following sections, I discuss case selection, data, application of the event study methodology, and the firm-level, cross-sectional method I use to extend the analysis.
Cases
I examine three entity list cases. These are the ERC's decisions to sanction ZTE in March 2016, Huawei in May 2019, and a group of technology firms in October 2019 that the US government believes is facilitating the Chinese government's monitoring of Muslim minority groups in the Xinjiang Uighur Autonomous Region (XUAR).Footnote 29 Though the October 2019 notice applies to a variety of firms and entities affiliated with the XUAR public security bureau, relevant data for this study are only available for Dahua Technology, Hikvision, IFLYTEK, and Megvii Technology. For ease of reference, I refer to these entities as the “surveillance firms” and the associated event as the “surveillance event.” While the surveillance firms were sanctioned for human rights reasons, ZTE and Huawei were sanctioned for contravening the US government's export and reexport policies related to Iran.
A little over two weeks after ZTE was placed on the entity list, the commerce department issued a temporary general license (TGL), which effectively established the status quo ante licensing policy for ZTE.Footnote 30 In effect, US businesses could carry on doing business with ZTE. The TGL remained in place for the duration of ZTE's time on the entity list. A more limited TGL was also issued for Huawei five calendar days after it was placed on the entity list.Footnote 31 The commerce department consistently extended the Huawei TGL until 17 August 2020.Footnote 32
It is important to emphasize that there is clear separation between the sanctioning of these firms and the issuance of the TGLs, and there is no evidence the TGLs were anticipated at the time of sanctioning. This separation allows me to identify the immediate effect of the sanctions on US suppliers’ market value. The TGLs, among other factors, preclude me from analyzing the long-run effects of the sanctions. In this sense, my findings should be interpreted as the effect on US suppliers in the absence of a subsequent TGL. In any case, event studies are known to be far less reliable for longer-term studies.Footnote 33 As such, most event studies explore very short-horizon effects.
I focus on targeted trade sanctions against Chinese technology companies because these sanctions are increasingly probable and highly consequential. In other words, these are important cases. One limitation of these cases is they are not representative of typical entities placed on the entity list. Adding multinational Chinese technology companies to the entity list is likely to have more significant ramifications for US suppliers than adding other probable targets. For this reason, my findings should be thought of as an upper bound on the effects of adding US firms’ business partners to the entity list. Data availability also contributes to my case selection. Supply chain data that are integral to this study are difficult to aggregate because there are few requirements internationally to disclose such information. The presently investigated firms have comparatively robust data availability.
Data
I examine the stock return behavior of publicly traded US firms that had supply relationships with ZTE, Huawei, and the surveillance firms around the date of their addition to the entity list. I obtain data on the Chinese companies’ US suppliers through Bloomberg's supply chain data.Footnote 34 Because I focus on stock returns, I can only use US firms that have actively traded stocks throughout the estimation period and event window that I utilize. My sample comprises 103 unique US firms and 154 supply relationships. Some companies are suppliers to more than one of the Chinese companies. I obtain data on these US firms’ stock returns from the Center for Research in Security Prices (CRSP).Footnote 35 Table 1 summarizes the entity list events I investigate.
Table 1. Entity List Events
Event study methodology
I implement the event study in roughly three phases. First, I establish the event window and estimation period. Second, I derive and aggregate abnormal returns. Third, I test whether aggregate abnormal returns in the event window deviate significantly from 0. I discuss these phases in the following sections.
Event windowFootnote 36 and estimation period
Though the three events have different calendar dates, they all share a common event timeline. I designate t = 0, the event date, as the first full day of trading following the entity list announcements. For the ZTE and Huawei events, t = 0 also corresponds to the effective date of the sanctions. For the surveillance event, t = 0 is the day in between the announcement and the effective date.
For the Huawei and surveillance events, setting t = 0 as the first full day of trading following the announcements is clearly most appropriate. The announcements for these sanctions were made late in the day, and the financial news summarizing stock market developments discusses the effects of the sanctions on t = 0, not t = −1, the announcement date. For ZTE, setting t = 0 as the first full day of trading after the announcement is less straightforward. Stories addressing the forthcoming ZTE sanctions were published early enough on the announcement day to affect trading. Nevertheless, I believe the first full day of trading following the announcement is most appropriate. For all three events, the preponderance of news covering the events occurred on t = 0. As illustrated in the Event Window Appendix, there was five to six times more news coverage for the events on t = 0 than on t = −1.
I use an event window of eleven days, which corresponds to +/ − 5 days surrounding t = 0. The event window is excluded from the estimation period, which is used to estimate expected returns. I establish an estimation period of 253 days prior to the beginning of the event window, which corresponds to the average number of trading days in a year.
As documented in the Event Window Appendix, I conduct a search to determine if there are other events in the event window, particularly on t = 0, that could have abnormally affected the US suppliers’ returns. Though I find some evidence of other events on t = 0 that could have abnormally affected the firms’ returns, I believe these events pose minimal threat to assuming the principal drivers of abnormal returns among these sets of US firms on t = 0 were the entity list actions.Footnote 37 The reason for this is the other events’ connections to the US suppliers under investigation are more indirect than the entity list sanctions.
To substantiate my assumption, I conduct a magnitude check that examines the US firms’ stock price reactions to events that are very similar to the alternative events on t = 0. The effects of these alternative events, which are reported after the main results, are statistically insignificant and substantially smaller than the effects I find in the main results.
I also find evidence of economic and political events that could have abnormally affected the US firms’ returns at other days in the event window. These include the release of Chinese economic data and developments in the US-China trade dispute. Additionally, the Huawei TGL was issued during the event window, toward the end of the day at t = 2, with effects felt at t = 3. The presence of these other events during the event window contributes, among other reasons, to my focus on the effects of the entity list sanctions on t = 0.
It is also important to evaluate whether the events were anticipated. To examine this, I conduct an extensive search in Nexis Uni for all English language news reports dealing with the entity list actions. These searches are documented in the Event Window Appendix. Though there is evidence of limited references to the sanctions and related events prior to the announcement dates, the events were not anticipated in such a way that the effects of the sanctions would have been reflected in the US firms’ returns in the days leading up to their announcement.
Abnormal returns
Next, I find abnormal stock returns for the individual firms. Abnormal returns are residual returns from a model of expected returns. They are used to measure the effect of the event on a firm's returns. To generate expected returns, I employ the market model, which measures risk-adjusted returns using the correlation of firms’ returns with a market index.Footnote 38 Using the market model, I estimate the relationship between the US firms’ returns and returns to the Nasdaq Composite IndexFootnote 39 over the estimation period:
Where R it are returns for an individual firm i at time t, R mt are returns to the Nasdaq Composite Index at time t, and ε it is a disturbance term. The Nasdaq Composite Index is more appropriate for generating expected returns than the S&P 500 index because the Nasdaq, like my sample, is more heavily weighted toward technology and communications stocks than the S&P 500 Index. Using the Nasdaq also helps extract intra-industry cross-correlation that could pose a threat to inference.
Using the estimated intercept and slope coefficient from equation [1], I estimate abnormal returns (AR) for firm i at time t as:
In this article, I analyze the average abnormal return (AAR) for all firms during the event window, focusing on the AAR at t = 0. For a given day, the AAR is:
Alternatively, the cumulative AAR (CAAR) enables analysis of the effect of events over the entire event window, which is useful for testing market efficiency and when t = 0 is not precisely known.Footnote 40 I do not investigate CAARs because I believe t = 0 is well identified, I am less interested in testing market efficiency, and there are other events in the event window that could have abnormally affected returns.
Tests of significance
Finally, I conduct statistical tests to determine whether average abnormal returns differ significantly from 0. The range of parametric and non-parametric tests proposed in the event study literature is extensive.Footnote 41 The most appropriate tests for my research design are the parametric and non-parametric tests developed by Kolari and Pynnonen (Reference Kolari and Pynnonen2010, Reference Kolari and Pynnonen2011). These tests are robust to event-induced volatility and cross-sectional correlation due to event-date clustering, the two most serious threats to inference in my study.Footnote 42
Both tests make use of the sample average cross-sectional correlation $( {\bar{r}} )$
, which is about 0.09 in my study. Kolari and Pynnonen (Reference Kolari and Pynnonen2010) have shown that even small amounts of cross-sectional correlation can lead to substantial over-rejection of the Null hypothesis of no price effects. To generate $\bar{r}$
, I calculate pairwise correlations within event blocs and over the estimation period. The $\bar{r}$
for the full sample is the average of all firms after figuring pairwise correlations within each event bloc.
Kolari and Pynnonen's parametric test uses $\bar{r}$
to adjust the parametric standardized cross-sectional test proposed by Boehmer, Musumeci, and Poulsen (Reference Boehmer, Musumeci and Poulsen1991) (“BMP test”).Footnote 43 The BMP test is a hybrid test that performs a cross-sectional t-test on standardized ARs. In discussing the results, I refer to Kolari and Pynnonen's parametric test as the “Adjusted BMP” test. The non-parametric test, referred to as the Generalized Rank (GRANK) test, performs a Rank test on generalized standardized ARs, which are adjusted for $\bar{r}$
.Footnote 44
Because the GRANK test is non-parametric, an added benefit is that it is robust to non-normality of stock returns. Simulations have shown that, consistent with the Central Limit Theorem, averages of abnormal returns are approximately normally distributed with a sample size of about fifty.Footnote 45 Though my study eclipses this threshold, there may be some threat to the application of the Central Limit Theorem due to the non-independence of firm selection.
Firm-level cross-sectional analysis
If entity list sanctions directly affect US suppliers by shutting off revenue streams with sanctioned entities, then we would expect firms that have deeper ties with targets to be more negatively affected by the sanctions. I test this expectation by exploring how supplier dependence affects abnormal returns on t = 0. I measure dependence as the share of revenue a given US supplier derives from the sanctioned Chinese entities. I obtain the revenue dependence data from Bloomberg.Footnote 46
Analytically, I regress AR i0 on revenue dependence and a series of firm-specific control variables that account for size, profitability, leverage, research and development (R&D) spending, and growth. To measure these characteristics, I use market capitalization, return on assets (ROA), the debt to equity ratio, R&D spending as a percent of sales, and the price to book ratio, respectively. I obtain the market capitalization data from CRSP and the rest of the firm-level data from S&P Global Market Intelligence's Compustat database.Footnote 47 To account for cross-sectional correlation of returns, I cluster standard errors by event-date.
Results
Figure 1 depicts the AAR(t) for the sample of 154 US suppliers over the 11-day event window (t[ − 5, 5]). Table 2 contains the precise AAR(t) over the same period, along with the statistical significance of the Adjusted BMP and GRANK tests.
Figure 1. How Sanctions Against Chinese Technology Companies Affect Their US Suppliers.
Table 2. Average Abnormal Return Test Results
Note: *p < 0.1; **p < 0.05; ***p < 0.01
The Adjusted BMP and GRANK tests are developed in Kolari and Pynnonen (Reference Kolari and Pynnonen2010) and (Reference Kolari and Pynnonen2011).
At −2.2 percent, AAR(0) is by far the largest move over the event window and is strongly statistically significant according to both the Adjusted BMP and GRANK tests. In absolute terms, it is 2.8 times the next highest move at AAR(4), which is the only other day in the event window that displays some statistical significance, according to the GRANK test. In the days leading up to the event, there are much more moderate fluctuations. The mean absolute AAR over t[ − 5, − 1] is 0.30 percent, while mean AAR is −0.10 percent over the same period. The mean absolute AAR and mean AAR over t[1, 5] are 0.52 percent and −0.33 percent, respectively.Footnote 48 Using AR i0 and market capitalization data for each firm, I derive the aggregate reduction in market value from the entity list sanctions as 18.6 billion USD. The results show strong support for my expectation that entity list sanctions against Chinese technology companies have an adverse impact on their US suppliers.
To further examine the robustness of the results, I conduct a placebo test, in which I replicate the event study using alternative entity list sanctions that would, from a theoretical standpoint, have no adverse effect on these US suppliers. Specifically, I swap the event dates to correspond to the closest entity list action, temporally, to each event that does not target a Chinese entity. For the ZTE firms, I use the entity list action on 23 February 2016, in which eight entities from the United Arab Emirates were added to the list.Footnote 49 For the Huawei firms, I use the entity list action on 26 September 2018, in which fourteen entities from Belarus, Iran, Russia, and Singapore were added to the list.Footnote 50 For the surveillance firms, I use the entity list action on 13 November 2019, in which twenty-two entities from thirteen different countries were added to the list.Footnote 51
Additionally, to benchmark the magnitude of AAR(0) in table 2, I conduct a magnitude check, in which I replicate the event study using alternative events that are very similar to the alternative events that occurred on t = 0 of the entity list event study and could be expected to negatively affect the sample of US firms.Footnote 52 For the ZTE firms, I set the magnitude check date to 4 January 2016, on which, according to the financial media, US equity markets fell due to the release of weak Chinese economic data. For the Huawei and surveillance firms, I set the magnitude check dates to 6 May 2019 and 23 August 2019, respectively. Both of these dates correspond to escalations in the US-China trade dispute. The financial media reported both escalations as hitting technology and industrial stocks with Chinese exposure particularly hard.Footnote 53
Table 3 displays the results of the placebo test and magnitude check. AAR(0) is negative in both tests but statistically insignificant and small in magnitude. None of the days in either event window differs significantly from zero. In terms of magnitude, AAR(0) in the magnitude check is only about 1 percent of the size of AAR(0) in the entity list event study. These results reinforce that the −2.2 percent effect of the entity list sanctions on US suppliers is both economically and statistically significant. The results also reassure me that the entity list actions are the primary driver of abnormal returns at t = 0 in the entity list event study.
Table 3. Placebo Test and Magnitude Check Test Results
Note: *p < 0.1; **p < 0.05; ***p < 0.01
Turning to the firm-level cross-sectional analysis, table 4 captures the modeling results, in which I regress AR i0 on firm dependence and a series of controls.Footnote 54 Consistent with my expectations, the results show that there is a strong negative association between abnormal returns on t = 0 and the share of revenue a given US firm derives from the Chinese firms. Stated differently, US suppliers that are more dependent on the sanctioned entities are more negatively affected by the sanctions. This lends support to the argument that entity list sanctions directly affect US suppliers by cutting off revenue streams with sanctioned entities. The results do not speak to potentially powerful indirect effects of sanctions, such as reputational damage and additional business model uncertainty.
Table 4. Firm-Level Cross-Sectional Modeling Results
Note: *p < 0.1; **p < 0.05; ***p < 0.01
An interesting supplementary result captured in table 4 is the strong negative association between R&D spending as a percent of sales and entity list sanctions. After accounting for dependence, entity list sanctions still inflict greater harm on firms that spend more on R&D. This suggests that sanctions against multinational technology companies are particularly damaging to the United States’ own tech sector.
Conclusion
American firms bear notable costs from the sanctioning of Chinese technology companies. On average, I find that the US government's addition of select Chinese technology companies to the entity list reduced their American suppliers’ market value by 2.2 percent on a risk-adjusted basis. The reduction likely occurs because the sanctions force US firms to sever ties with the sanctioned entity, thereby relinquishing a revenue stream to international competitors. This line of reasoning is supported by complementary firm-level cross-sectional analysis, which demonstrates that US firms that are more dependent on the sanctioned entities are more negatively affected by the sanctions. The US suppliers also likely suffer reputational damage for being unreliable, along with additional business model uncertainty.
Why does the US government impose sanctions that inflict steep costs on US companies? Many have found that US sanctions policymakers prioritize security issues over domestic business interests.Footnote 55 Mastanduno (Reference Mastanduno1992) suggests that, in this issue area, relative neglect for domestic business interests owes to a powerful executive and the “absence of effective countervailing power from the US private sector.”Footnote 56 This characterization certainly has relevance for entity list sanctions. The US government is typically required to seek formal public comment on regulations that affect US businesses. For national security reasons, the executive branch imposes entity list sanctions without seeking public comments, which makes it more difficult for the private sector to weigh in.
However, I also contend that the dearth of studies quantifying the domestic economic cost of sanctions prevents policymakers from thoroughly evaluating their tradeoffs. The shortage of such studies owes, in part, to the difficulty of measuring the cost of sanctions. The event study methodology I employ in this article is well-suited for capturing the effects of targeted trade sanctions on the public firms in the sender country that have business relationships with the sanctioned entities. The granularity of daily stock returns, their quick adjustment to relevant news, and their reflection of current and future business prospects, allow for plausible inference of the effects of sanctions.
The nature of the results presented in this article points to areas of potential future inquiry, particularly related to the longer run impact of sanctions. To this end, future research could explore whether the effects of sanctions persist over time and, relatedly, whether sanctions generate redistribution effects among groups of competing firms. Likewise, it would be valuable to examine whether sanctions prompt strategic or organizational changes for affected firms. Future research could also perform event studies akin to the one presented in this article to measure the effects of sanctions on firms in the sender country in other contexts.
Targeted sanctions are a valuable instrument of economic statecraft because, compared to more comprehensive sanctions and the use of force, they allow governments to signal resolve at a relatively low cost. As demonstrated, however, they are certainly not cost-free. This article expands our understanding of the cost of targeted sanctions, which is important for framing the domestic economic ramifications of these measures for relevant stakeholders. Sanctions policymakers, whose stated objectives involve ensuring the competitiveness of US businesses, can leverage the findings presented in this article to better calibrate sanctions policy.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/bap.2020.21.
Acknowledgements
I thank Charles Doran, David Steinberg, Dennis Yang, Pavithra Suryanarayan, Christopher Sands, Timothy Peterson, and the anonymous reviewers for their excellent comments.
Data Availability
The replication data for the analysis presented in this article are available at: Allen, Jeffrey. 2020. “Replication Data for: Do Targeted Trade Sanctions Against Chinese Technology Companies Affect US Firms? Evidence from an Event Study.” https://doi.org/10.7910/DVN/NET3BA, Harvard Dataverse, V1.
