Trade policy uncertainty decimates investment

We have shown how heightened firm-level trade policy uncertainty during the trade war leads to reductions in firm-level investment, R&D expenditures and profits for Chinese listed firms. However, our results may be capturing instead how rising U.S. tariffs and Chinese retaliatory tariffs have negatively affected Chinese firms directly, by reducing demand for Chinese exports or by increasing the cost of imported inputs. For this reason, it is important to validate that the trade policy uncertainty channel effects on firm operational outcomes is present when the direct impacts of tariffs is included. To this end, we augment our baseline equations (4) and (5) with U.S. and Chinese tariff exposure measures as follows:

<<<TEXT CONDENSED, NONE OMITTED>>>

where β1 measures the trade policy uncertainty effect while the coefficients β2 and β3 capture the direct impacts of tariffs. Tables 11 and 12 report the estimation results. Reassuringly, the sign and significance of the coefficients of the trade policy uncertainty effect on investment and profits remain unchanged, even after controlling for the direct impact of tariffs. Notably, the direct impacts of U.S. and Chinese tariff exposure measures on both investment and profit are 33 statistically insignificant.

Discussion: There may be several factors that could explain the absence of direct tariff impacts in these new results. First, the average ratio of exports to the U.S. to total sales for Chinese listed firms in the sample is about 1.7 percent in 2016 according to Chinese customs data and COMPUSTAT Global data.32 The low ratio suggests that the output loss resulting from rising U.S. tariffs would be quite limited even if U.S. tariffs in a firm’s industry have risen substantially. Correspondingly, reactions in investment and profits would be small for these firms.33 32The calculation is based on the COMPUSTAT sample restricted to the manufacturing sector. The average ratio of exports to the U.S. is computed as U.S. exports relative to total sales that include domestic sales. As not all manufacturing firms are exporters, we set exports of the non-exporting firms as zero in the above calculation.

33Firm shares of imports from the U.S. are also small, possibly leading to the null effect of rising retaliatory tariffs. In Appendix C, we further investigate it by adding interaction terms. The exercise seeks to explore whether the direct impact of U.S. and Chinese tariff exposures are stronger for firms that export to and/or import from the U.S. more. Results are reported in Appendix Tables C.1 and C.2. The coefficient ofthe interaction of U.S. tariffs and U.S. export shares is insignificant, while the interaction between Chinese tariffs and U.S. import shares is significantly negative. Though the overall direct impact of Chinese retaliatory tariffs is insignificant, Chinese firms that concentrate their imports in the U.S. suffer from the trade war, suggesting that import sourcing is probably less flexible for critical parts embedded in the ongoing relationship between Chinese firms and U.S. exporters. 34

A second explanation is that a third country may have been chosen by Chinese exporters to re-route exports. Liu and Shi (2019) find that trade re-routing has been used by Chinese firms in the past to avoid antidumping duties.34 Though we cannot formally test firm re-routing behaviors because of data limitations, the recent news report by Chau and Boudreau (2019) suggests that there is a high possibility that some re-routing took place during the U.S.-China trade war. According to the report, exports from Vietnam to the U.S. have grown strongly in 2019 and many products such as plywood are produced in China might be shipped to the U.S. with ‘Made in Vietnam’ labels.

Third, trade diversification can help mitigate the negative impacts of tariffs. For instance, if Chinese firms easily switch buyers, then the direct negative impacts of tariffs on firm-level investment and profits would be reduced.35 In fact, the Chinese government has implemented policies to help affected Chinese producers to switch to other partners.36 34Trade re-routing means firms send their products to a third country where U.S. tariffs are not applicable. After that, goods are reissued certificates of origin and sent to the final destination country without being subject to the U.S. tariffs. 35Our previous analysis has uncovered a real hedging channel where Chinese exporters that are more diversified in terms of destination markets see a lower increase in trade policy uncertainty.

36According to the report by CNBC, the Chinese government has taken mainly four ways to bolster business during the trade war, which includes increasing government support, opening channels to other international markets through programs such as free trade zones and the Belt and Road Initiative, improving the environment for state-owned and foreign enterprises and implementing policies such as tax and fee cuts. See https://www.cnbc.com/2019/08/26/ 35

Lastly, as firms in the sample are large by nature and may engage in a variety of business activities other than trade-oriented ones, firms may have adapted necessary operation change in response to future tariff increase (e.g., switching from the trade-oriented business to non-trade-oriented ones), leading to a limited direct impact of tariffs. This is in line with the finding that even tariff-induced TPU change does not significantly affect a firm’s total revenue that may consist of sales generated by non-trade-oriented business activities.

On the contrary, it is less controversial that the rising TPU makes firms less certain of what to plan for the future, leading them to postpone investment (Handley et al. (2020)), and it is exactly as what we find in the data.37 Nonetheless, the stylized facts highlighted by this paper contribute to the understanding on how firms respond to rising trade policy uncertainty, which can be used later as models of dynamic endogenous effects of TPU are developed.

5.5 Robustness Checks

To the extent that some Chinese firms are focused on serving the domestic market, while others are more heavily involved in exporting, we explore if our main results change according to firm orientation. For this purpose, we focus on subsets of firms based on indicator variables 1(Persistent Exporter) (1(Persistent Importer)) that equal one if the firm exported (imported) in all years between 2013 and 2016. Alternatively, orientation could be defined via the measure of exports-to-revenue or imports-to-cost ratio. Table D.1 reports the result under both specifications. The impact of tariffs on firm TPU remains similar to that in Table 3 among both groups. In particular, we observe that the increase in firm TPU is more pronounced among exporters. We further include those firm orientation variables in regressions investigating the role of trade diversification, reporting the results in Table D.2. We continue to find that firms exporting to a larger number of destinations experienced smaller increases in TPU.

As suggested by Table 6, firms with a more diversified set of trade relationships experienced smaller increases in TPU. However, this outcome might also be shaped by differences in firm financial positions. If it possible that more diversified firms may have better access to funds and that the access to funds, rather than the diversification, may help these firms to adapt to the changing environment. To address this concern, we follow trade-war-what-it-means-for-china-firms-as-trumps-calls-us-firms-to-go.html for details. In addition to the above discussion, Cavallo et al. (2019) also find that imports of U.S. retailers increased after the initial announcement of possible tariffs, but before the full implementation of tariffs.

Therefore, by completing sales/purchases beforehand, the impact of tariffs on firm sales for Chinese firms can be very limited overall. 37By modeling firm’s dynamic sourcing decision, Handley et al. (2020) show that firms tend to postpone making investment decision in the presence of trade policy uncertainty.36 Manova and Yu (2016) and construct firm-level financial measures as follows:

We interact the financial terms with tariffs in our new regression specification. This allows us to determine whether firms that were in a better financial position experienced a lower increases in TPU in response to increases in tariffs. In Appendix Table D.3, both interaction terms are statistically insignificant. Hence, the results confirm that firm liquidity and leverage do not mediate the impact of tariffs on TPU during the trade war.

In this paper, we explore the sources and consequences of trade-policy uncertainty during the ongoing U.S.-China trade war. Our analysis is based on a novel measure of firm level TPU constructed from a textual analysis of firm-level statements, and firm-specific measures of exposure to trade war tariffs based on customs data and tariff lines. The firm-level TPU measure accurately tracks existing aggregate indices of TPU in China, and reveals a dramatic TPU spike during the trade war.

Our first contribution is to open the TPU black box. While it has been generally acknowledged that TPU must have played a role in the trade war (IMF, 2018), there is little understanding of how the process works, the magnitude of this channel, and which firms are most exposed to it. We move one step in this direction and take advantage of the variation across Chinese firms in their exposure to the trade war. We show that firm-level increases in TPU experienced during this period are systematically associated with firmlevel exposure to both U.S. tariffs (which lower U.S. demand for Chinese exports) and Chinese tariffs (which raise the cost of imported inputs for Chinese firms). We further show that the impact of tariffs on TPU is heterogeneous across firm characteristics, as the tariff effect on TPU is largest for smaller and less diversified firms.

The second contribution of our paper is to document the negative consequences of the TPU spike on firm investment, R&D expenditures and profits. Our work allows us to identify the timing of TPU effects, including the changes over the short- to medium-run. The effect exists, and is quantitatively important, after controlling for the direct impact of tariffs on all these outcomes. We find that a one standard deviation increase in TPU leads to a reduction in investment, R&D expenditures, and profits by 1.4, 2.7, and 8.9%, respectively.

Overall, our work highlights the importance of the TPU channel during the ongoing U.S.-China trade war; the paper also illustrates the benefits of new measures of firm37 level uncertainty based on textual analysis of firm statements (Hassan et al., 2020, 2019; Caldara et al., 2019). The documented stylized facts contribute to understanding how firms respond to rising trade policy uncertainty, which can be used later as models of dynamic endogenous effects of TPU are developed.