Re-assessing the US-China trade war

A compact version of this study can be read here.

In economic history, 2018 will be remembered as the year that the US started a trade war with China. This week the G20 summit will kick-off in Argentina, but the meeting that will catch the eye is the meeting at the side-lines of the G20 between Trump and Xi Jinping on trade. Trump has hinted this month that a deal between both economic giants is in the making, but this was not confirmed by Larry Kudlow, the director of the National Economic Council (see here). Given the reluctance of China to give in to US demands, we are sceptical that a deal could be wrapped up during the G20 and trade tensions will de-escalate. In fact, if the talks fail, there is a chance that early December President Trump will announce a new round of tariffs targeting all Chinese imports.

The question is: what is at stake economically? Many economists and economic institutes have made an economic impact assessment of the current situation and a further escalation of the conflict (see, for example, IMF, 2018a). Rabobank is no exception. We have made some initial attempts to calculate the impact of this trade war mainly by focusing on export market share losses (see, for example, here). With relatively small volumes of trade being targeted at the beginning of the year, a partial analysis was sufficient to get an indication of the economic impact. However, as Trump has raised the stakes in June and September by targeting 250bn USD of Chinese exports to the US (followed by retaliatory measures by China), a more thorough analysis is required.[1]

Therefore, in this Special we have re-assessed the impact of the protectionist measures currently in place and ran a second scenario where the US will target the remaining Chinese exports to the US as well, and China will retaliate. We add two novel aspects to our analysis. First, we use the tariff version of the National Institute Global Econometric Model (NiGEM), which is a macro-econometric world trade model that Rabobank has been using for a decade now to make economic forecasts and run scenario analyses. The upside of using NiGEM is that the model allows us to assess the impact of rising relative prices due to the protectionist measures on key variables in the short to medium term, such as the trade flows, investment and private consumption, something we have not been able to do in the partial analyses on trade wars that we carried out before. A second novel aspect is that we examine dynamic productivity effects. By taking into account these effects, we are able to assess the permanent impact of the current trade war on China’s and US’ growth potential. Many studies indeed have shown that trade is an important conduit for economies to benefit from international knowledge spillovers, which are an important determinant of a country’s ability to accelerate labour productivity growth. Combining these two aspects enables us to assess the economic impact of the US-China trade war much more accurately than we have done so far.

[1] We already ran a more extreme scenario analysis in which the US would start a global trade war by raising tariffs by 20% across the board and all targeted trading partners responding accordingly (see Erken, Every, Giesbergen and Wijffelaars, 2018), but at the current stage such a global scenario seems unlikely.

Before turning to the methodology we use to run our scenario analysis, we have to shed more light on the economic mechanisms of a trade war and expected economic effects. Or to put it more simple: how does protectionism work?

Economic mechanisms behind protectionism

Protectionism can take many forms, ranging from imposing tariffs, raising non-tariff barriers (NTBs[2]) to providing subsidies to exporting companies. In theory, installing protectionist measures directly affects the relative competitiveness of domestic producers vis-à-vis foreign competitors. In the US-China trade war, the aim of the Trump administration by imposing import tariffs on Chinese goods is to foster the relative competitiveness of US trade exposed industries (e.g. manufacturing), which ultimately will result in a higher domestic market share of these US industries and will create jobs for US citizens. An important additional benefit is that import tariffs yield higher government revenues.

The catch of an import tariff is that it directly props up the trading costs for Chinese exporters, who will consequently face two choices (Figure 1). First, Chinese exporters confronted with higher tariffs can pass on these costs in higher retail price of the goods exported to US shores. US consumers and producers will consequently be confronted with higher prices of (intermediate) goods. The economic losses for US consumers and producers will be higher, the lower the price elasticity of demand for the Chinese products. In case of a low price elasticity, US consumers and producers will continue to buy the Chinese (intermediate) products, despite rising price levels, as they do not have suitable alternatives. This is, for instance, the case with certain commodities (tungsten and cadmium) and rare earth metals, which are used in electronic equipment such as smartphones and televisions (see NYT, 2018). In case of high price elasticity, US consumers and producers will shift their demand away from the relatively more expensive Chinese goods and Chinese exporters will lose market share.

A second option for Chinese exporters is to (partly) internalise the higher tariff by lowering their sales price at the expense of their own profits and financial buffers (Figure 1). In this way, the retail price for US consumers will remain equal (i.e. lower sales price + higher tariff = equal consumer price). Chinese exporters will adopt this strategy if they suspect that the tariffs are temporary in nature or in case of a high price elasticity.

[2] There are various forms of non-tariff barriers, such as technical barriers to trade, import quotas, import licenses, safeguards and phytosanitary regulations. For more information and data on NTBs, we refer to the World Trade Organisation.

The problem with protectionism is that the countries that are being targeted by protectionist measures will almost certainly impose retaliatory measures (Figure 1). In the current trade war, China has responded to each imposed US tariff package with a retaliatory package of its own. The result is that the economic pain that the US inflicts on Chinese exports will be reciprocated by China on US exporters in the same manner as described above.

Before discussing the consequences of a trade war, it is important to elaborate why countries start a trade war in the first place (see Erken, Verbruggen and Marey, 2018). Research by Autor (2018) shows that a large group of consumers benefits from international trade, but there are also disadvantages for a small, concentrated and often vulnerable group of employees. Although the rise of China has been positive for global prosperity, it also created losses in some US industries that were vulnerable to import competition from China (e.g. Acemoglu et al., 2016). The most vulnerable sectors were often regionally concentrated, causing the losses to be felt mainly in specific areas. These negative redistributive effects have contributed to more support for protectionist policies in the US, independent of other causal factors such as technological development.

Economic effects

Trade

A trade war has multiple economic effects. First and foremost, protectionist measures and retaliation raise the costs of trade for the participants involved in the trade conflict, in our case the US and China. This will lead to lower export and import volumes on either side. The negative impact on trade volumes is partly mitigated by export substitution effects, as the negative impact is partly offset by redirecting trade to other destinations. To give an example, in the current trade war, China has imposed import tariffs on US soybeans closing out the US which traditionally exported more than half of their soybean exports to China, the world’s largest importer. Consequently, the US will have to ship all of their soybean export to other shores. This divergence in trade limit the economic losses partly, but as our colleagues from Food & Agri Research have argued, if the US does not ship its soybeans to China, it needs to buy almost 100% market share in all other countries in the world by discounting prices well below those in competitor South America. Another mitigating effect might be the response of the exchange rate. China’s exchange rate (CNY) has depreciated by 9% since April, which has partly mitigated the higher soybean prices China has to pay for South American soybeans.

Integrated supply chains

In contrast to the mitigating export substitution effects, the negative impact of the trade war on trade could be aggravated by integrated value chains. Over the past decades, multinational firms have increasingly been exploiting international comparative advantages by relocating parts of their production processes abroad. For example, they benefit from low labour costs in Asia for the assembly of goods, while marketing and R&D are located at the home base. This has enabled them to produce more efficiently and improve their competitiveness. The downside of these so-called sliced up value chains is that multinationals have become more vulnerable to import tariffs on intermediate products or commodities from abroad. For instance, the US tariff package on Chinese imports of USD 50bn implemented this summer applies primarily to intermediate products and capital goods (Bown, Jung and Lu, 2018). Consequently, only about 40% of the tariffs in this USD 50bn package are borne for by Chinese firms while the remaining 60% of the tariffs are absorbed by foreign firms that are active in China (see Lovely and Liang, 2018). The tariffs on computers and electronic products are even borne for 87% by foreign firms. Ultimately, the US firms that rely on intermediates imported from China will become more expensive due to trade barriers (and vice versa), which means that these companies will face either a deterioration of competitiveness due to higher retail prices (and higher export prices and a lower global market share) or an absorption of the higher costs, which will hurt their profitability.

Households and financial markets

Consumers in both countries will feel the pinch as well, as the trade barriers will result in import inflation, which will eat away at real disposable income of households and lead to lower household spending. Trade wars will generally cause financial market turbulence, which affects both consumer and producer sentiment and could ultimately lead to lower private investment and private consumption. We have witnessed the disruptive nature of uncertainty caused by trade wars on financial markets this year. Especially emerging markets (EM’s) have borne the brunt of the US-Chinese tensions over trade, as investor sentiment changed from a risk-on to risk-off modus, which resulted in massive capital outflows and put EM currencies under pressure (see Lawrence, Verbruggen and Erken, 2018). The Chinese CNY lost 7% against the US dollar this fall vis-à-vis the beginning of the year (Figure 2). Although this depreciation certainly has mitigated the impact for Chinese exporters, the losses in terms of trade (i.e. more expensive imports) will be felt by Chinese households and have put additional downward pressure on private consumption. For the US, the opposite is the case: the strengthening of the USD provides US households with windfall gains, as imports will become cheaper, but will be a setback for US exporters, especially the ones shipping their products to Chinese shores.

Relocation of production

Over the last decades, many Western companies have moved parts of their production facilities to China. This was mainly related to a comparative labour cost advantage, which meant that on balance it was cheaper to produce in China than, for example, in their own country. As a result of rising wages, particularly in the coastal areas of China, some of the companies have made a move to other emerging economies in the region (e.g. the Philippines or Vietnam), which are still at an earlier stage of their economic development, therefore profiting from a relatively low wage advantage compared to China. These kinds of shifts are a medium-term strategic story. After all, such reallocations are a time-consuming and costly process, especially when fixed assets are involved. As a result of the increased tensions between China and the US, this process seems to have accelerated. For example, various surveys show that more and more companies are considering a relocation (Forbes, 2018; Bloomberg, 2018). Furthermore, the trade war seems to lead to an increase in foreign direct investment in Southeast Asian countries (Bloomberg, 2018).

Productivity

Labour productivity growth is the most important pillar of economic growth. There is a vast strand of literature that shows that trade has a significant impact on productivity. First, knowledge developed abroad positively affects domestic productivity, but these spillovers are not automatic or exogenous.[3] Well-known conduits of international knowledge spillovers are human capital mobility (Park, 2003), foreign direct investment (Branstetter, 2006) and trade (Coe and Helpman, 1995; Grossman and Helpman, 1991; Lee, 2005). Trade fosters foreign knowledge spillover effects, as firms can use foreign-produced intermediate inputs (Goldberg et al., 2010; Yu, 2015). Moreover, downstream users can reverse engineer technologies embodied in innovative final imports and use this knowledge in their own production processes. These mechanisms seem much more important for China than the US, as the latter still has a large technology lead over China.

Nevertheless, US productivity is also affected directly by the Chinese-US trade relationship, as openness to foreign trade fosters market competition, which stimulates firms to reduce their X-inefficiencies and increase efforts to innovate.[4] In this sense, import competition will result in more innovative, more efficient firms (the within firm effect). Secondly, there are sector composition effects (the between firm effect): lower trade costs will result in reallocation of labour and capital toward more productive and skill-intensive firms within sectors and toward skill-intensive sectors in all countries (see Burstein and Vogel, 2017). These findings are in line with Bernard, Redding and Schott (2007), who find that within and between-industry reallocations of economic activity during periods of trade liberalisation raised average productivity in all industries, but more so in the comparative advantage industries.

For European firms, Bloom, Draca and Van Reenen (2016) find that Chinese import competition has increased technical change within European firms (within effect) and also caused a shift of employment towards technologically more advanced firms (sector composition effect). Taken together, these effects account for 14% of European technology upgrading in the period 2000-2007. Other studies that find a robust direct positive effect of international trade on productivity are by Edwards (1998) and Alcalá and Ciccone (2004).

As trade has a beneficial impact on labour productivity development, a pullback in trade caused by higher trade costs should have an adverse impact on productivity. We will return to this topic more extensively when we will discuss the productivity models for both the US and China.

[3] Countries also need sufficient ‘absorptive’ research capacity to internalise knowledge developed abroad (Cohen and Levinthal, 1989).

[4] The term X-inefficiencies refers to slack in the production process and higher production costs than necessary, which are the result of a lack of competitiveness pressure in the market.

Trump already referred to China’s unfair trading and economic practices during his election campaign back in 2016, but it took a year before his administration strengthened this anti-China rhetoric by implementation of actual protectionist policies. The delay was probably related to China's strategic role in the conflict between the US and North Korea (see also: Erken and Giesbergen, 2017). Due to this role, it was expected that China would be less willing to cooperate with the US against North Korea in case of bilateral trade tensions, for example by following up on international sanctions against the North Korean regime. We first outline the currently installed and looming protectionist packages on both sides. After that we present an overview of our trade war scenarios.

Installed and looming packages

Figure 3 outlines the most important steps in the US-China trade war. In January 2018, the US first implemented import tariffs on washing machines and solar panels. These tariffs were not directly targeted against China, though they are important producers and exporters of these products. After these first ripples, the US imposed another round of tariffs on steel and aluminium in March 2018. These measures again applied to a wide range of countries, such as Canada and Mexico and European Union member states. A tariff of 25% applied to steel exports to the US, and 10% on aluminium exports. For China, these measures corresponded to an export value of roughly 3bn USD (see Figure 3). This time China decided to introduce countermeasures by an equivalent amount, by implementing tariffs of 15% on American fruit, nuts and wine, and 25% on aluminium and pork. After these first protectionist packages it became increasingly clear that the Trump administration was turning its protectionist focus towards China. And China took the moral high ground vowing to fight back at any cost against any US protectionist policy directed towards them.

Before the US decided to further tighten the screws on China by introducing more tariff packages (Figure 3), several rounds of bilateral negotiations took place. In their statement dating back to May 2018, China committed itself to import more goods from the US, especially agricultural and energy-related products. China would also pay more attention to the protection of intellectual property rights of US companies operating in China. Although this outcome was initially regarded as positive, it soon became clear that China’s pledges were not exactly aligned with the demands by the Trump administration. That is why the US continued to tighten the screws by implementing another round of tariffs, targeting USD 50bn of imports from China, which in turn led to a similar retaliation by China (Figure 3). Meanwhile, no other rounds of negotiations took place between the two sides. In the most recently installed package by the US of USD 200bn introduced in September, three phases are evident (Figure 3). After the 'first' round of 10% tariffs on USD 200bn, the levies will be increased to 25% as of 1 January. If China retaliates, the US is also prepared to impose levies on another USD 264bn worth of Chinese exports, which in effect makes all export from China to US shores subject to a 25% tariff. These are mainly consumer (electronic) products. Next to tariffs, the US is considering non-tariffs barriers (NTBs) as well, as the ZTE example shows..

As we have indicated on several occasions (see for example: Erken, Every and Giesbergen, 2018), China has less options to implement new tariffs on imports from the US than the other way around. This is simply because China imports far less from the US than the other way around. China has taking countermeasures for the USD 200bn package by targeting USD 60bn of US imports, which means that after the previously targeted USD 50bn, 87% of the total US exports to China are now subject to higher tariffs (Figure 3). Because China’s limited ability to match new US tariff packages by similar countermeasures, we foresee that they will use NTBs as well. Finally, China can choose to be less cooperative as a mediating party in the conflict between the US and North Korea. North Korea recently announced that they are unwilling to continue negotiations with the US on denuclearization if the international sanctions against the country remain in place.

Overview trade war scenarios

In order to measure the (potential) trade war impact, we re-assess the impact of the protectionist measures currently in place and run a second scenario where the US will target the remaining Chinese exports to US shores as well, and China will retaliate accordingly. As such, we make a distinction between these different scenarios for our re-assessment of the potential trade war impact, and compare the impact to a no trade war benchmark scenario. The timeline and assumptions of the economic shocks in both trade war scenarios is discussed extensively in Annex A.1 of this report. Figure 3 relates to the timing of the different scenarios.

Benchmark: No trade war

In our benchmark scenario, trade tensions between US and China are absent and no protectionist measures would have been installed on either side. This scenario is based on the forecasts published in Rabobank’s most recent economic quarterly outlook (Van Es and Barendregt, 2018), but we have excluded the effects of the tariff rounds we already took into account in that outlook. This relates to the tariffs imposed up and until August 2018 (Figure 3).

Scenario 1: Currently installed and announced packages

This scenario takes account of protectionist measures that have already been installed, and measures that have been announced. This concerns the tariff packages of 25% on USD 50bn on both sides, as well as a 10% US tariff on USD 200bn worth of export from China to the US, as well as an average 7% tariff by China on an additional USD 60bn imports from the US. In this scenario we have implemented the tariff shock in 2018Q3. In addition, we implement an additional tariff shock in our model in 2019Q1. This relates to the American proposed tariff increase on 1 January 2019 from 10% to 25% on USD 200bn of Chinese goods, and the assumption that China will also raise their tariff rate from 7% to 25% on USD 60bn of American goods (Figure 3). Although there is some level of uncertainty involved, we integrate them in this baseline scenario because of the high probability we attach to these looming tariff increases.

Scenario 2: Full-fledged bilateral trade war

This final scenario assumes a further escalation of the trade war somewhat later in 2019Q1, where both countries decide to levy tariffs on all their bilateral imports. For the US this means a 25% tariff rate on an additional USD 264bn USD, and for China a 25% rate on another USD 50bn of imports from the US (Figure 3). Furthermore, we assume that China will increase non-tariff barriers by an ad valorem equivalent (AVE) to 15% in order to more or less match the protectionist measures introduced by the US. We also assume a heavy depreciation of the Chinese currency (CNY) in this scenario.

Over the last year there has been increased attention on the potential adverse economic impact of a trade war. These scenario studies focus mostly on the impact of the currently introduced and looming packages, as we have described in Section 3, but the scope of these studies vary to a substantial degree and depend on the countries that are involved in the scenario analyses, as well as the assumptions in general. As an example, some analyses also assume that Europe be pulled in the current US-China trade war and not every study covers the same amount of tariffs and related protectionist packages. This of course greatly depends on timing and related assumptions. An overview of the selected studies for comparison, their potential impact assessment and a number of the most important assumptions can be found in Table 1.

In their most recent World Economic Outlook of October 2018 (IMF, 2018a), the IMF draws up the possible impact of the US-China trade war (see scenario Box 1) by using their so-called Global Integrated Monetary and Fiscal Model (GIMF) for a scenario (‘layers’) analysis. It builds on four previously described scenarios in a July 2018 G20 Surveillance Note (IMF, 2018b). Their first trade war scenario includes tariffs that already have been implemented. In a second scenario, the IMF adds the proposed increase of US tariffs from 10% to 25% on USD 200bn USD of Chinese export by the beginning of next year. A difference between our analysis and the IMF, is that the IMF also assumes retaliation by other trading partners (e.g. Europe) next to China, and assesses a scenario where the US decides to implement tariffs on cars. Consequently, the IMF finds a larger impact in the (multilateral) trade war scenario on the US than on China. The tariff shocks in the different scenarios are permanent, and they include also confidence and investment effects by assuming an increase in risk premia for advanced economies (30bps) and emerging markets (60bps) in order to reflect relatively higher financial vulnerabilities. The study does, however, not include any dynamic productivity effects or non-tariff retaliation by China and it is found that the impact for China is bigger than for the US in the short-run, but not in the longer run.

The National Institute of Economic and Social Research (NIESR) looks at the potential trade war effects by continuing earlier research of Liadze (2018), Hantzsche and Liadze (2018), Carreras and Ramina (2017) and Liadze and Hacche (2017). They also incorporate the most recent round of tariffs and run simulations using NiGEM. As such, they use a similar econometric model, but use different assumptions. They shock comparable scenarios exogenously, but these shocks are not regarded as permanent as the shocks are only applied from 2018Q3 till 2020Q4, assuming that prices will adjust after that period. Their results point to slightly stronger negative impact on the US compared to China, but this is due to differences compared to our assumptions and scenarios. Most notably, we include exogenous foreign exchange developments and adverse potential productivity effects, where NIESR abstracts from these effects. Moreover, we assess a further escalation of the bilateral trade war (including NTBs), whereas NIERS only assess the currently imposed protectionist measured.

The trade war scenario analysis by the European Central Bank (ECB) is published as part of their economic bulletin dating back from September 2018. By using the IMF’s GIMF model as well as their own global model (ECB, 2017), the ECB assesses both the trade and confidence channels by which the economy might be impacted by the current trade war. It is, however, important to note that this study does not assess a US-China trade war, but examines a global trade war where the US imposes tariffs on all imports and all trading partners will reciprocate these protectionist US measures. Furthermore, the study assumes that the trade tensions will ease going forward and will only last for two years. Finally, exchange rates and monetary policy are modelled endogenously. The ECB also makes a distinction between direct and indirect trade effects, by additionally taking account of potential confidence and financial market effects, by modelling a tightening of financial conditions assuming an increase in bond premia by 50bps and a stock market decline of two standard deviations in all countries. Still, the impact on trade from dynamic productivity effects is not covered here either. And one major difference with all other studies discussed here is that this analysis takes into account a full retaliation of other trading partners against the US. This explains the relatively higher impact on the US economy compared to China, as the latter can benefit from substitution effects.

The Netherlands Bureau for Economic Policy Analysis (CPB) analysed the potential trade war by using different scenarios (CPB, 2018). WorldScan, a so called computational general equilibrium (CGE) model of the world economy, is used to measure and comprise an analysis on an (inter)national and sectoral level. The CPB takes into account previously installed packages and makes assumptions on looming packages, by using five different scenarios, varying from solely steel and aluminium tariffs, to knock-on escalation scenarios where tariffs from both the EU and China versus the US are installed, and the US even engages in trade wars with all OECD countries. Model specifications include an integration of so-called perfect and imperfect competition mechanisms and the assignment of different elasticity categories. These are based on the assumptions that relatively low elasticity levels should yield larger economic losses and that the adverse impact is larger under a scenario of imperfect competition. On the other hand, the analysis does not show any specific distribution of the impact over time. It only considers what the impact might be of a permanent shock up to and including 2030, with the last-mentioned year as the reference point compared to the baseline.

Although the potential impact discussed is somewhat outdated given the current trade war timeline, Bank of England (BoE) governor Mark Carney highlighted some simulations they had conducted on the potential trade war impact in a July speech. In these simulations, three different impact channels are distinguished: direct trade war effects, tighter financial conditions and greater uncertainty, and permanent tariffs. By using NiGEM they estimate the cumulative impact of a broadly defined 10% tariff on the US, the UK and the Euro Area. As a result, the potential impact on China is not discussed explicitly here, which in turn makes it hard to compare the results with ours. The central bank emphasizes the monetary policy challenges in terms of higher inflationary risks, highlighting that any trade war shock would act as a drag on economic activity, but the initial impact would result in upward inflationary pressure. Carney made notion of lower productivity growth in case of a long-lasting trade war, referring to findings that 20% less trade would incline 5% lower productivity in the long run (Feyrer, 2009). However, it is seems that the BoE has not incorporated any of these productivity effects in their own analysis.

To sum up, a clean-cut comparison between the scenario studies on the trade war is complicated by the fact that each study makes different scenario and modelling assumptions, and uses different methodologies and models (see also Table 1). The upside of our analysis is, first of all, that it is fully up to date. Second, we feel that the focus we have on the two countries currently involved is more realistic than an analysis focusing on a multilateral, global trade war. Third, we feel that China will not only resort to import tariffs as retaliatory measures, but will adopt non-tariff measures in a full escalation scenario as well. Fourth, we use our in-house expertise on FX forecasting to draft accurate exogenous FX paths for the Chinese yuan and the US dollar. Lastly, we also include dynamic productivity effects in our analysis. That is a key difference with all the other studies discussed, which in turn has major implications for the results as well, which are discussed in more detail in Section 9.

There are different models and methodologies to assess a trade shock, be it either a positive one (e.g. a new free-trade agreement) or a negative one (e.g. Brexit, a NAFTA breakup or the China-US trade war). To get a better idea of all the different models and methodologies that are out there, we refer to an excellent overview by Tetlow and Stojanovic (2018) on different studies that have looked at the economic impact of Brexit. As was the case with our economic assessment of Brexit, we again use a two-step approach to examine the economic impact of the US-China trade war. We use the macro-econometric world trade model NiGEM in combination with calculations of two productivity models for the US and China (see next section), developed by RaboResearch.

Step I: NiGEM

To assess the economic impact of the US-China trade war, many institutes use macro-economic models that capture economic effects in case of trade shocks or policy changes. The IMF, for instance, uses the Global Integrated Monetary and Fiscal Model (GIMF) whereas CPB uses Worldscan. In a similar fashion we use National Institute Global Econometric Model (NiGEM). NiGEM is a macro-econometric world trade model, estimated in a ‘New-Keynesian’ framework (see Figure 4). This means agents are forward looking, but rigidities result in a slow adjustment process in case of external events or shocks. Rabobank has been using this econometric model for over a decade now and other institutions, such as the Europe Central Bank and the Bank of England, use the model as well.

Using NiGEM has three main benefits. First, the model allows us to assess the impact of several key variables in the short to medium term, such as exchange rate fluctuations, trade flows, foreign direct investment and the labour market. Second, the US and China have a separate model in the framework, but countries are linked to each other through trade and competition, interaction of financial markets and international asset stocks. NiGEM ensures that all economic variables are viewed within a closed accounting setting and economic shocks, such as the tariffs imposed by US and China, are accounted for via these interdependencies. Third, NiGEM is an error-correction model, which means that short-term deviations of GDP from a country’s growth potential are made up eventually. So in the long-run, growth is driven by structural factors, such as capital formation, structural employment and labour-augmented technological change.

Tariff and extended China model

We do not use the conventional version of NiGEM to assess the trade war. For trade-related topics, the expanded tariff version of NiGEM, which was released in 2017, is more suitable. The tariff version has the benefit that it includes directionality of export prices, meaning that tariffs can be imposed to specific countries. In that sense, higher tariffs result in lower export volumes of the targeted country, higher competitor export prices and import inflation. The latter will affect real wages, real disposable income and private consumption. At the same time, relative competitiveness is hurt by cost-push inflation as intermediates become relatively more expensive, which raises production costs, export prices and lowers competitiveness. A second adaption is that we use the expanded China model, whereas the conventional NiGEM model uses a small model for the Chinese economy. In the expanded China model domestic demand is disaggregated in private consumption, government consumption and total investments, which enables us to better assess the effects on the Chinese economy.

A disadvantage of NiGEM is that productivity effects are more or less fixed, as labour-augmented technological change on the supply side of the model is exogenous. However, the dynamic productivity effects of cross-border trade shock can potentially be very large, and even double the negative economic impact in case of a sudden trade shock (for an overview, see e.g., Tetlow and Stojanovic, 2018). Therefore, we have developed productivity models for the US and China as our second step.

Step II: Productivity models

As technological change in NiGEM is more or less exogenous, we have to calculate the technology shocks in our trade war scenarios separately and impose the calculated technology shock ex post in NiGEM. To calculate these technology shocks, RaboResearch has developed two dynamic productivity models for both the US and China. In this model, we fully endogenise total factor productivity (TFP) and translate these effects to labour-augmented technological change in NiGEM. Admittedly, given data availability and quality the US TFP model is much more comprehensive than the China TFP model. But even the China model is able to capture more than 60% of total factor productivity growth over two decades.

Besides the upside of assessing the impact of trade effects on long-term productivity in China and the US in a more robust fashion, the productivity models also enable us to derive a solid benchmark scenario (which is the no trade war scenario). This is because we are able to fully decompose US and China potential growth, based on assumptions of its underlying drivers.

Limitations

Using NiGEM in combination with a productivity model has its advantages, but there are also some limitations to our methodological approach. First, we merely assess the trade war from a macro economic standpoint, which means we are unable to reflect on the impact in different sectors, firms or product classes. In a similar fashion, we are not able to grasp the impact the trade war has on supply chain integration. The IMF (2018a) poses the same critique: “Global macroeconomic models, such as GIMF […] cannot capture some of the sectoral distortions that the proposed trade restrictions are likely to generate.” Our colleagues from Food & Agri Research (FAR), nevertheless, have conducted extensive research on the impact of the trade war on the Food & Agri sector. Read more on the implications of the trade war on the global soybean industry,global soybean trade, China’s import pattern, the soybean processing industry in China, the US, and elsewhere, as well as on the livestock sector and on US fruits and nuts.

Second, NiGEM is not able to fully capture trade substitution effects. Export volumes of third countries respond to higher export prices increases due to higher tariffs between the US and China. As the relative prices of these countries improve, part of the higher prices in the US and China results in export opportunities for third parties, so we capture some of the expected export substitution effects in a trade war. However, bilateral trade shares are static which implies that import prices in the US and China may be too high in our scenarios, and this might results in a too steep decline of the export market size of third countries. Third and last, NiGEM does not take into account the relocation of firms as a result of the increased tensions between the US and China.

Although we have tried to make separate assessments of export substitution effects (see Erken and Heijmerikx, 2018) and the impact on integrated supply chains (Erken and Tulen, 2018), we cannot integrally take into account these two effects using NiGEM. A model that is capable of capturing the impact on supply chains and export substitution on product level is, for instance, the Global Trade Analysis Project (GTAP). Rabobank has planned to start working with this model in 2019.

What is total factor productivity?

Paul Romer (1990) is the founding father of the so-called endogenous growth theory, for which he will receive the Nobel Prize for Economics (together with William Nordhaus). In contrast to neoclassical growth models in which technological progress is exogenous, in the endogenous growth theory technological progress is explained by investment in human capital and/or Research & Development (R&D) (Romer 1990; Jones 1995). Although endogenous growth models have been tested by means of calibration, it is difficult to empirically estimate endogenous growth models which have been developed from a theoretical perspective. These models are based on knowledge production functions at the global level, which makes them less useful for individual countries.

To examine the long-term impact of the current trade war on productivity development, however, we need to endogenise technological progress for individual countries. Here is where the growth accounting methodology comes in, which is available for individual countries in two important databases: the Total Economy Database (TED) developed by the Conference Board and Penn World Tables (PWT) from the Groningen Growth and Development Center. Both databases enable us to derive total factor productivity series (TFP), which is arguably the purest measure of technological progress, given that it captures the portion of economic growth that cannot be attributed to an increase in labour or capital inputs.

We use both databases as starting point to derive our TFP models for the US and China. There are discrepancies in TFP growth in the US between both databases, as the labour share is slightly higher in TED than in PWT. All in all, these discrepancies are relatively small (Figure 5).

For China, there is much more debate about the accuracy of total factor productivity measurement. In the past, we have discussed the quality of China’s official GDP data and the Conference Board also publishes an alternative set of growth decomposition series in the TED based on the comprehensive work of Wu (2014). The alternative series of TED are much more aligned with the TFP series in PWT (Figure 6). Ultimately, we choose PWT as our main database for China TFP analysis, but will conduct robustness analysis by using the alternative TFP series from TED (Annex A.2).

Drivers of TFP

There is a vast literature on determinants of TFP (see for an overview Erken, Donselaar and Thurik, 2016). Below we give a brief description of these determinants.

Human capital and R&D

Human capital[5] and R&D are the most important driving forces behind technological progress and TFP, not only from a theoretical perspective (see Lucas, 1988; Romer, 1990), but also from an empirical point of perspective: see Griffith et al. (2004), Engelbrecht (1997), Bassanini and Scarpetta (2002), Cameron et al. (2005) and Erken, Donselaar and Thurik (2016).

Domestic R&D expenditure is an important driver of TFP growth, not only to generate domestic innovations, but also to benefit from knowledge developed abroad. A country needs so-called ‘absorptive capacity’ to be able to understand foreign technologies and internalise these in domestic innovation processes (Cohen and Levinthal, 1989). There have been some doubts about the quality of Chinese R&D statistics, as Chinese firms may evade tax by overreporting on R&D (Chen et al., 2018). Nevertheless, König et al. (2018) show that when stripping out overreporting behaviour from the data, R&D investment appears to be as productive in China as in Taiwan.

R&D can be conducted by firms (private R&D) as well as by public knowledge institutes and universities (public R&D). There is much evidence that private R&D generates positive productivity effects (see e.g. Coe et al., 2009; Coe and Helpman, 1995). With respect to public R&D, the evidence is mixed. Guellec and Van Pottelsberghe de la Potterie (2004) find a positive effect of public R&D capital, whereas Khan and Luintel (2006) and Van Elk et al. (2015) have difficulties reproducing these results. In this study, we also examine the impact of public R&D capital on productivity development.

As discussed in Section 2 of this report, knowledge developed abroad is an important source for domestic innovation and productivity development. For China, Liu and Buck (2017) find that learning by exporting and importing significantly improves innovation of the Chinese indigenous high-tech sector. Foreign R&D activities on the other hand affect innovation of domestic firms only when absorptive ability is taken into account, which is in line with the findings of Cohen and Levinthal (1989). Moreover, Filatotchev et al. (2011) and Xiu et al. (2010) show that returnee entrepreneurs mobility are important conduits to benefit from foreign knowledge, whereas Zhang (2017) finds evidence that international knowledge spillovers due to foreign direct investment have a positive impact on the performance of overall research activities. In this study we mainly examine the role of the import quote as a conduit for foreign knowledge spillovers, which is in line with the approach by Coe and Helpman (1995), Lee (2005) and Cameron et al. (2005).

It is pretty evident that at the current junction, China benefits much more from knowledge development by the US as global technological leader than the other way around. As discussed in Section 2, however, the US nevertheless can benefit from trade with China for domestic innovation and productivity purposes as well, as openness to foreign trade fosters market competition, which stimulates firms to reduce their X-inefficiencies and increase efforts to innovate. Bloom, Draca and Van Reenen (2016) indeed find evidence that Chinese import competition has increased technical change within European firms (within effect) and also caused a shift of employment towards technologically more advanced firms (sector composition effect).

Other factors

Entrepreneurship is an important determinant of productivity. Inspired by the limitations of the endogenous growth theory, Braunerhjelm et al. (2010) have developed different models that introduce a filter between knowledge in general and economically-relevant knowledge; they identify entrepreneurship as a mechanism that reduces this so-called ‘knowledge filter’.[6] Bottom-line is that entrepreneurs are needed to valorise knowledge. There is also empirical evidence of a positive relationship between entrepreneurship and productivity (e.g. Carree and Thurik, 2008; Erken, Donselaar and Erken, 2016) or economic growth (e.g. Acs et al., 2018; Audretsch, 2018), although some find that the relationship between entrepreneurship and economic growth rather is U-shaped (e.g. Carree et al. 2002; Prieger et al. 2016) or L-shaped (Carree et al., 2007).

Labour input generates adverse TFP effects (Belorgey et al., 2006; Bourlès and Cette, 2007; Erken, Donselaar and Thurik, 2016). High labour participation is often characterised by increased deployment of less-productive labour, which lowers labour productivity. Working fewer hours may have a positive impact on productivity if less fatigue occurs among workers or if employees work harder during the shorter number of active hours. Another factor that affects total factor productivity are business cycle effects, as labour and capital endowments are not immediately adjusted to business cycle volatility, which makes TFP susceptible to fluctuations of the business cycle. Firm profitability supposedly has a positive impact on TFP. More profits support higher R&D expenditure by firms (Himmelberg and Petersen, 1994). In addition, higher profit expectations can motivate firms to innovate at a higher rate (given a fixed amount of R&D capital). Lastly, higher profits provide firms with financial means to stimulate innovation (given a fixed degree of R&D capital). We use the capital income share as a proxy for the profitability effect. The capital income share is defined as gross capital income as a percentage of the gross value added of businesses. The negative counterpart of profitability is taxation. Taxation could have a negative impact on productivity: a higher rate of taxation implies negative incentives in certain markets and less incentive to innovate, which consequently could result in a less efficient economy. This effect is important to examine as the Trump administration just recently launched the massive new tax scheme the Tax Cuts Act and Jobs Act.

The model: a framework for empirical TFP analysis

In order to model all mechanisms addressed above, we have adopted the framework used by Erken, Donselaar and Thurik (2016). Under the neoclassical conditions of perfect competition in product markets and constant returns to scale in the production factors of capital and labour, the marginal products of capital and labour are equal to the return on capital and the wage rate, respectively. It can be derived that, in that case, the output elasticities of capital and labour are equal to the shares of capital income and labour income in total factor income. The annual growth of TFP can then be calculated as follows:

[5] Human capital is defined as quality improvement of labour due to education and training.

[6] The idea behind the knowledge filter is that entrepreneurship serves as a conduit for knowledge. R&D by itself is not a growth safeguard, just as fostering entrepreneurship is insufficient for propelling growth. Entrepreneurs have to exploit knowledge (R&D) in order to lead to positive growth. This conclusion is also drawn by Michelacci (2003), who develops an endogenous growth model where innovation requires the matching of an entrepreneur with a successful invention

where Y denotes gross domestic product, K and L denote (physical) capital input and labour (measured in physical units such as hours worked) and is the share of capital income in total factor income, or stated differently, the share of capital income in the gross domestic product. Moreover, Δ denotes mutation in first differences, i stands for country, t is a time index (i.e. year) and log represents the natural log. From (1), we can define our baseline TFP model as:

where α₁ measures the effect of growth of domestic R&D capital (S) on TFP growth. The coefficient α2 picks up the effect of foreign R&D capital (S^f). Foreign R&D is interacted with the import quote, which is defined as the import volume (M) as a ratio of gross domestic product volume (Y). R&D capital is calculated by using the perpetual inventory method on intramural R&D expenditure and taking into account a depreciation of capital of 15% due to the obsolescence of knowledge. This depreciation rate is often used to calculate R&D capital, based on Griliches (2000, p. 54), who refers to this percentage as the ‘‘conventional’ 15 percent figure for the depreciation of R&D-capital’.Nominal R&D expenditure is deflated using an index for the price of R&D, which consists for 50% of a deflator for domestic expenditure and 50% of an wage index. Foreign R&D is measured by a proxy of R&D in 19 other OECD countries (or 20 in case of foreign R&D for China), weighted by bilateral import shares for domestic R&D capital. As we assume that the US is not able to benefit from Chinese R&D capital, we have left out Chinese R&D activity in the foreign R&D capital variable used in the US TFP equation, whereas we have included the US in the foreign R&D capital variable for the Chinese TFP model.

The term α₃ measures the impact of human capital (H). The average years of (tertiary) education are usually used as an indicator to measure the amount of human capital in a country (see Barro and Lee, 2013). This indicator is useful, although a disadvantage is that it does not take into account the quality level of education. Cohen and Soto (2007) claim to produce better data than Barro and Lee, as they use information from surveys based on uniform classification systems of education over time, and an intensified use of information by age groups. Moreover, Cohen and Soto claim that their data is more suitable for models in first differences, which is exactly how our TFP models are estimated.

Lastly, coefficient α₄ is a vector containing different control variables, such as openness of the economy, public R&D capital, entrepreneurship, labour inputs and the business cycle. Two notions are important in this respect. First, to measure openness of the economy, we first use trade exposure[7] as an indicator (see Bassanini et al., 2001), but as a second step adjust trade export data for country size using estimations by Donselaar (2011). Small economies are by definition more exposed to foreign trade, regardless of their trade policy or competitiveness. Second, high-quality entrepreneurship data from the COMPENDIA dataset is only available up to 2009. This forces us to extend the series based on the OECD self-employment series, which goes at the expense of the quality of our entrepreneurship variable. Finally, as we have a much richer dataset for the US than for China, our TFP model for the US contains much more control variables. We will elaborate on each specific TFP model in more detail in the next two subsequent sections. For all technicalities on variable construction, we refer to Erken (2008), Donselaar (2011) and Erken, Donselaar and Thurik (2016).

[7] Trade exposure is defined as: TRADE = X + (1-X) ´ M, where X represents the ratio of exports in relation to GDP, M is the ratio of imports in relation to domestic demand. Domestic demand is calculated by domestic production minus exports plus imports.

Our approach to model the impact of R&D, human capital and foreign knowledge development on productivity within a dynamic setting fits a broad strand of literature (see Park, 1995; Frantzen, 2000; Griffith et al., 2004; Cameron et al., 2005; Buccirossi et al., 2013). Table 2 shows the variables that we use in our US TFP model, as well as their data sources. We use simple OLS to estimate the model and use the Akaike, Schwarz criterion and Hannan-Quinn information criterion to obtain the optimal lag structure. The estimation results are shown in Table 3.

Our baseline estimation is illustrated in column (1) of Table 3 and consists of domestic R&D capital, foreign R&D capital interacted with the import share and human capital. All variables show a statistically significant impact on TFP growth. Domestic private R&D capital (coefficient c₁) shows a coefficient 0.18, which means that a 1% higher growth of R&D conducted by firms on US soil generates a TFP growth of 0.18ppts. As coefficient (c₂) is an interaction variable, the interpretation is somewhat complicated. Our results show that if foreign R&D capital increases by 1%, US domestic TFP growth will increase by 0.06ppts, whereas if the import quote rises by 1%, this adds another 0.12ppts to US TFP growth. Finally, human capital index has a positive effect on TFP growth: the coefficient (c₃) shows that an increase of the human capital index by 1% yields higher US TFP growth by 1.5ppts.

In column (2) we add public domestic R&D capital (c₄), which fails to show a significant impact. This is in line with many studies that have difficulties assessing a significant impact of public R&D (alongside private R&D capital) on economic growth (see for an overview Van Elk et al. (2015)). However, when we add openness of the economy and a business cycle variable in column (3), public R&D shows a statistically significant impact on TFP growth. The coefficient is 0.49, which means that a 1% increase in public domestic R&D capital leads to additional TFP growth of 0.49ppts. Moreover, the two controls (openness (c₅) and the business cycle (c₆)) also have a statistically significant and positive impact on TFP growth and the explanatory power of the model jumps markedly (R² rises from 0.21 to 0.51).

In column (4) we add two labour input variables: participation (c₇) and hours worked per worker (c₈). As expected, the participation variable has a (statistically significant) negative impact on TFP growth. The reason is that a higher growth of labour participation generates a composition effect which serve as a drag on productivity growth, as the additional people that start working are being less productive and skilled than the already participating ones. The amount of hours worked per worker does not have a significant impact on labour productivity growth. This also makes sense, as the US is one of the few OECD countries which is able to combine higher levels of productivity with high participation ratios and many hours worked per worker. In many countries there is a clear trade-off between hours worked and productivity per hour, but the US is clearly an outlier in this respect (see Figure 7, which explains why we do not find significant negative correlation.

In column (5) we add entrepreneurship (c₉) as a control variable, which has a significant impact on TFP growth. This means that if the deviation of the business ownership rate from the equilibrium levels growth, TFP growth increases as well, as there are more people able to valorise knowledge. Reversely, if a country lacks this mechanism, there is a TFP growth penalty.

Finally, in columns (6) and (7), we add two variables related to the effect of firm profitability on productivity, which are the capital income quote (c10) and corporate tax as a % of GDP (c₁₁). The capital income quote has a significant impact and a coefficient of 0.10, which means that an improvement of the capital income quote by 1% results in an increase in TFP growth by 0.10ppts. Corporate tax fails to show a statistically significant effect. We have experimented with a different lag structure, construction of the corporate tax variable and the combination with other variables, but none of these experiments yielded fruitful results.

Although we know that via the route of additional corporate investment, lower corporate taxes (such as the Tax Cuts and Jobs Act) will have a positive impact on productivity growth, we seriously doubt if there is an additional effect of lower corporate taxes on total factor productivity growth. However, if we substitute our corporate tax variable by a more general tax variable, being total tax revenue as a % of GDP, we find slightly better results. Coefficient (c₁₁) now shows a negative sign, which is in line with expectations, and is statistically significant at a 90% confidence level. However, the tax variable eats away at our entrepreneurship variable and does not prop up the explanatory power of the model estimated in column (6). We will examine the role of (corporate) tax on US TFP growth more extensively in a separate research note.

Ultimately, our preferred equation is equation (6), which has a solid fit (see Figure 8). We use this model to construct our benchmark scenario for the US and calculate TFP effects in our two trade war scenarios, which we consequently incorporate in NiGEM.

Table 4 gives a description of the variables used for our China TFP model and its sources. As we want our China model to be comparable with the US TFP model, we draw upon the same data sources, but are not able to include as many control variables as with the US model. Table 5 gives an overview of the estimation results.

Column (1) in Table 5 shows our baseline model for Chinese TFP growth, which contains domestic R&D capital, foreign R&D capital interacted with the import share and human capital. All variables show a statistically significant impact on China’s TFP growth. Domestic private R&D capital (coefficient c₁) shows a coefficient 0.61, which means that a 1% higher growth of R&D conducted by Chinese firms generates a TFP growth of 0.61ppts. The coefficient is much higher than the US model, but in line with research for other countries as well (see Erken et al., 2018). The higher efficiency of domestic R&D could be explained by the fact that Chinese technological progress over the last 20 years emerged from relatively low levels and so-called ‘fishing out’ effects are limited for the Chinese economy and much more prominent in the US. Fishing out means that it becomes harder to generate new ideas, the higher the general level of knowledge and technology (see Jones 1995). The coefficient for international knowledge spillovers (c₂) is also higher than for the US, which again makes sense. The so-called catching-up potential in China is much larger than in the US, which is in line with the technology gap theory. This theory states that countries with low levels of technology can benefit more from knowledge developed abroad than countries operating at the technological frontier (Fagerberg, 1987). Finally, the human capital index (c3) has a positive and large effect on TFP growth: an increase of the index by 1% yield higher Chinese TFP growth by 3.36ppts.

In column (2), (3) and (4), we add the control variables public domestic R&D capital (c₄), the business cycle (c₅), and openness of the economy (c₆). The effects of public R&D and the openness are statistically insignificant and/or show the wrong sign, whereas the impact of the business cycle is also quite weak. The insignificant impact of openness more or less makes sense, as market competition is limited in China due to the major influence the government still holds over the corporate sector. As a consequence, large state-owned enterprises (SOEs) will only have limited incentive to produce more efficiently or prop up innovative efforts in response to higher import competition.

In column (5) we add labour input variables, but also strip out the public R&D capital variable to lower the risk of overfitting the model, as we only have limited degrees of freedom due to the relatively short dataset. The effect of labour participation (c₇) is insignificant and shows the wrong sign. Hours worked per worker (c₈) shows the correct sign, but is insignificant as well. In column (7) we experiment with the impact of corporate tax (as a percentage of GDP), which also show the correct sign but no statistically significant effect.

Ultimately, we arrive at column (7) as our most comprehensive TFP model for the Chinese economy. We use this model to derive a potential growth path for China and use the model to calculate dynamic TFP effects in our two scenarios. In Annex A.3 we conduct several robustness analyses.

In this section we present the results of our scenario analyses. Our two scenarios (scenario 1: current & announced protectionist measures, scenario 2: escalation to a full-fledged trade war) are compared to our benchmark scenario, which assumes a trade war would not have occurred. We will first briefly describe our benchmark scenario to give a broad idea what the economic growth potential of both the US and China are in absence of a trade war, before elaborating on the effects in our two trade war scenarios for the US and China, respectively.

Benchmark scenario: no trade war

For the US economy, we more or less assume business as usual in our benchmark scenario, which results in the potential growth path up to 2030 illustrated in Figure 9. US growth potential is more or less stable at 2%, which is remarkable given the fact that the TFP growth projection is fully endogenised based on assumptions about eleven underlying drivers.

For the Chinese economy, we have to take into account that the government has altered its policy course considerably. In 2015 Premier Li Keqiang presented the so-called Made in China 2025 strategy which focusses on transforming China’s export-led investment growth model to a more domestic-oriented consumption-driven growth model. This plan focuses on more technology-driven manufacturing, where the most advanced sectors of the economy play a crucial role. It is part of a broader strategy to be a leading manufacturing power by 2049, when the country has its 100th Anniversary of the founding of the People's Republic of China. An important goal in the Made in China 2025 plan is to increase Chinese-produced content of core material to 50% by 2020 and 70% by 2025. Moreover, the plan puts heavy emphasis in fostering high-tech industries, such as the aerospace industry, ICT, robotics and clean-energy cars. From this perspective, we assume a pickup in imports in the short-run, as China still needs foreign imports for upgrading their economy, but a gradual decline in openness to foreign trade in the longer run. Furthermore, we assume that growth of private R&D activity will continue at a relentlessly high pace. Ultimately, given our assumptions, we are able to construct the potential growth path for the Chinese economy depicted in Figure 10.

For the upcoming years, China is able to counterbalance the negative impact of lower employment growth by a combination of continuing investment and TFP growth in our benchmark scenario, but ultimately the lower connectivity with knowledge developed abroad bears the risk that China’s TFP growth will trend lower and lower. Given the current efficiency of domestic R&D activity, it is questionable if China is able to replace productivity-enhancing knowledge spillovers from abroad by domestic innovative activity. Keep in mind that the Chinese education system is less supportive to generate creativity and problem-solving skills than OECD countries (see Zhao, 2015). Moreover, in China the strong state-led intervention in the economy causes capital misallocations and disrupts price competition. Consequently, private fast-growing innovative firms have difficulties competing on the internal market against inefficient state-owned enterprises (SOEs) and bank funding is also largely calibrated towards these SOEs (see, for example: Huang and Du (2017) and Wei, Xie and Zhang (2017)). That is probably one of the reasons why regulators recently announced targeted credit measures for privately owned enterprises (POEs). This goes at the expense of a favourable entrepreneurial environment and Schumpeterian creative destruction, which is necessary to foster TFP gains. Ultimately, we find that China in the long-run is capable to grow at a pace slightly below 4% annually.

Scenarios: impact on US economy

Figure 11 shows the impact on the US economy in our two trade war scenarios. In scenario 1, the US economy would miss out on 0.9ppts of growth in 2030 compared to our benchmark scenario. In case of a further escalation, the US would miss out on 1.6ppts of economic growth in the long-term, which is still relatively mild. As we adopt a different set of assumptions, it is difficult to compare our results to other studies, but in general these effects are in line with both the IMF and the CPB findings (see Section 4).

In absolute terms, the calculated effects imply that each US citizen in 2030 will miss out on USD 600 of wealth in total (Figure 12). In case of a further escalation (i.e. scenario 2), this price tag could be almost as high as USD 1,100 per capita in 2030 (Figure 12). As US GDP per capita currently is as high as 55,000 per capita, the impact is still relatively mild. Then again, we also have to bear in mind that these are average effects, whereas we know that the impact is distributed unevenly. A banker in Manhattan will feel the pinch of the trade war to a far lesser extent than a cranberry or soybean farmer in the Mid-West.

Explanation

In Figure 13 and 14, the adverse impact on various expenditure components is illustrated. Export growth is especially taking a beating in scenario 2 on the back of higher export prices due to the tariffs and NTBs implemented by the Chinese on US goods and the appreciation of the US dollar. The impact on private consumption due to higher inflation (Figure 15) caused by the US tariffs is mitigated by the strengthening of the US dollar. A stronger dollar improves the terms of trade of US consumers, which alleviates the negative impact of higher import prices on their purchasing power. The impact on private investment growth is also relatively mild. Ultimately, the development of private consumption and investment growth is much more important than growth of the external sector, so this provides some ground as to why the US economy does not experience a large adverse shock in both our trade war scenarios.

Roughly one-fifth of the calculated negative GDP effects in the US are due to lower dynamic productivity effects. The damage of the trade war on the supply side of the economy is limited, as US firms do not heavily rely on Chinese technologies. Of course, lower import competition from China will alleviate the competitive pressure on US firms somewhat, which leads to less incentives for these firms to innovate and increase operational efficiency, but also these negative productivity effects are marginal at most.

Scenarios: impact on China

For China, the trade war comes with a much heavier price tag than for the US. Our calculations (Figure 16) show that under the current circumstances (scenario 1), China would lose out on 1.5ppts of economic growth in 2030 compared to a situation without a trade war (benchmark scenario). In case of a further escalation (scenario 2), i.e. Trump would decide to target all of China’s export to US shores, GDP losses could end up being as high as 5% in the long term.

In absolute terms, the calculated effects imply that each Chinese citizen (1.4 billion) in 2030 will miss out on USD 400 of wealth due to the current set of implemented protectionist measures. In case of a further escalation, this price tag could be almost as high as USD 1,500 per capita (Figure 17). Given these numbers, one also has to keep in mind that GDP per capita in China (currently USD 15,000) is 3.5 times as low as in the US. This underlines that the economic pain from the trade war weighs much heavier on the average Chinese than on the average American.

Explanation

The impact on the Chinese economy in our first scenario is more or less in accordance with findings in other studies, but in case of a further escalation (scenario 2) we find much more prominent effects. This is due to a number of reasons. First, we use expert knowledge from our Asia FX strategist to include an exogenous path for the Chinese currency (Chinese Yuan: CNY), which is - as far as we can tell - not done in other studies. In this exogenous path, we expect the CNY to depreciate quite profoundly if Trump would announce another round of tariffs on all Chinese exports, as it will raise the pressure further on the CNY due to an expected liquidity impulse by the central bank and more pressure from capital outflows. This will deteriorate the terms of trade of Chinese households and weigh additionally on their purchasing power, besides higher consumer prices caused by tariffs and non-tariff barriers (see Figure 18). The inflationary impact has a large adverse impact on Chinese private consumption (Figure 19), which cumulatively is almost 10ppts lower in the second scenario compared to a trade war-free world.

Secondly, due to lower trade with the US (Figure 20), the Chinese economy would be seriously restricted in its capacity to benefit from technological knowledge developed in the US. This would consequently weigh on productivity growth in China going forward, which would have been much higher in case a trade war would not have occurred.

One important condition we have to mention here is that we did not incorporate any fiscal response from China in case of a further trade war escalation. This is a likely response from the Chinese government. However, the room to manoeuvre is becoming more and more restricted, as we have emphasized before (see, for example, Giesbergen and Erken, 2018). Especially in the area of monetary policy, the Chinese government has to weigh its options carefully, given the already high and increasing debt levels. There is some more room on the fiscal side, but this stimulus needs to be focused on productive (non-SOE) investments.

Impact on F&A

As we have discussed in Section 5, a limitation of the current study is that we mainly assess the economic impact of the trade war from a macro standpoint. Nevertheless, as Rabobank has a key focus on Food & Agri, our colleagues from Food & Agri Research (FAR) have conducted many in-depth studies to assess the impact on various industries in the F&A sector. Read more on the implications of the trade war on the global soybean industry, global soybean trade, China’s import pattern, the soybean processing industry in China, the US, and elsewhere, as well as on the livestock sector and on US fruits and nuts. In Box 1, we briefly discuss the far-reaching impact of the current trade war on the global soybean market.

As we use an econometric world trade model, it is possible to assess the second-order impact on third countries spilling over from the direct confrontation between the two largest economies in the world. Before discussing these results, however, we have to address the caveats here. First of all, we have mainly focussed on the US-China relationship in our scenario analysis and have formulated detailed assumptions to calculate the impact of the trade war to the best of our ability. We have not included specific detailed assumptions for other countries. For instance, an announcement of another round of protectionist measures by president Trump (i.e. our second scenario) could easily result in large market volatility and safe haven investment in the Japanese yen, which would go at the expense of the Japanese export position and result in lower exports. Therefore, the results below have to be taken with a grain of salt. Furthermore (as stated earlier), one of the limitations of NiGEM is that it does not take into account export substitution perfectly and the impact on integrated global supply chains.

Market opportunities versus lower global growth

There are two effects in third countries that work against each other. First, relative export price levels in these countries improve vis-à-vis the US and China, which enables exporters in third countries to increase their global market share. This consequently would result in a positive impact on net exports in these countries due to substitution away from Chinese and US products in favour of products manufactured in these countries. Second, what is weighing on exports is less global demand in general and lower global economic growth, which ends up 0.7ppts lower in scenario 1 and 2.0ppts lower in scenario 2 (Figure 21). This second effect is dominating, which leads to the conclusion that virtually every third country ends up with lower growth than in a situation where a trade war would not have occurred.

Another effect that is weighing on growth in third countries is the fact that firms that use intermediate goods from the US or China will be faced with more expensive products, which will be partly incorporated in their own prices. So higher inflation in the US and China will partly feed into higher inflation and lower private consumption in third countries as well.

Euro Area and the Netherlands

In the Euro Area, the US-China trade war would shave off 0.25-0.5ppts of growth in total up to 2030. This suggests that as long as the Euro Area won’t be involved in the trade war, the impact on the European economy will be very limited. However, some European countries are affected more than others. The Netherlands, for instance, is a very open economy and also has strong trade ties with the United States. Therefore, the current string of events weighs more heavily on Dutch economic growth (see Figure 22) compared to countries that are relatively less dependent on foreign trade, such as France.

Cumulatively, the Dutch will lose out on 0.4ppts of economic growth at the peak of the trade war in 2021 in scenario 1 against our benchmark scenario, with a long-term effect of -0.3ppts in 2030. In scenario 2, however, the total adverse effect in 2021 is -0.9ppts with a long-term effect of -0.7ppts. In absolute terms, the price (in terms of missed economic growth) each Dutch resident pays at the peak of the trade war in 2021 is approximately 170 euros in scenario 1, whereas a further escalation would more than double that price (380 euros).

In this study we made a re-assessment of the economic impact of US-China trade war. In contrast to earlier Rabobank studies on this topic we adopt more advanced methodologies in this study. Besides examining the direct impact on economies using a large econometric trade model, we also assess the impact on labour productivity development, given the fact that trade is an important conduit to benefit from knowledge developed abroad.

We examine two scenarios. Our first trade war scenario is the most realistic one and includes all protectionist measures currently in place, or announced. In the second scenario we assume that the trade war is taken a step further and the Trump administration will target the remaining Chinese exports to the US of USD 264bn (and China will retaliate). Both trade war scenarios are compared with a benchmark scenario in which a trade war would not have occurred.

Our findings show that trade wars in general will only lead to economic losses, most of which end up in the countries directly involved (the US and China), but also negatively affects third parties. In our first trade war scenario, global economic growth will be 0.7ppts lower in 2030 than in a trade war-free world. In our escalation scenario, global economic growth will be 2.0ppts lower in 2030.

Given these losses, however, our scenario analysis also shows that China disproportionately is bearing the brunt of the US-China trade war. China would miss out on 1.6ppts of economic growth in 2030 in our first scenario, and 5.7ppts in our second scenario against the benchmark scenario of no trade war. The US will have to absorb economic losses of 0.9ppts under the first scenario, whereas in the escalation scenario these effects will be 1.6ppts. In absolute terms, the calculated effects imply that each Chinese citizen in 2030 will miss out on USD 400 of wealth due to the current set of implemented protectionist measures, whereas each US citizen misses out on USD 600. In case of a further escalation, this price tag for China could be almost as high: USD 1,500 per capita, and in the US the price ends up being USD 1,100 per capita. Given these numbers, one also has to keep in mind that GDP per capita in China (currently USD 15,000) is 3.5 times as low as in the US. This underlines that the economic pain from the trade war weighs much heavier on the average Chinese than on the average American.

Our results for the US are largely in line with finding in other studies, but we find a much larger impact on China in case of a further escalation (scenario 2). This is for two reasons. First, we expect that the Chinese currency will depreciate quite profoundly if Trump would announce another round of tariffs on all Chinese exports. Secondly, due to lower trade with the US, the Chinese economy would be restricted in its capacity to benefit from technological knowledge developed in the US. This would consequently weigh on productivity growth in China going forward, which would have been much higher in case a trade war would not have occurred.

We have shown in this report that the stakes are high. In the short-run, any trade deal could ease tensions somewhat, but our expectation is that the tensions between both countries will continue for over a longer period of time. This study provides a clear picture of what the detrimental potential economic effects would be in that case.

Acemoglu, D., Autor, D., Dorn, D., Hanson, G. H., & Price, B. (2016). Import competition and the great US employment sag of the 2000s. Journal of Labor Economics, 34(S1), S141-S198.

Acs, Z.J., Estrin, S., Mickiewicz, T. and L. Szerb (2018). Entrepreneurship, institutional economics, and economic growth: an ecosystem perspective. Small Business Economics, 1-14.

Audretsch, D.B. (2018). Entrepreneurship, economic growth, and geography. Oxford Review of Economic Policy, 34(4), 637-651.

Autor, D.H. (2018), Trade and labor markets: Lessons from China’s rise. IZA World of Labor.

Barro, R.J. and J.W. Lee (2013). A new data set of educational attainment in the world, 1950–2010. Journal of Development Economics, 104, 184-198.

Bassanini, A. and S. Scarpetta (2002). Does human capital matter for growth in OECD countries? A pooled mean-group approach. Economics Letters, 74(3), 399–407.

Bassanini, A., Scarpetta, S. and P. Hemmings (2001). Economic growth: the role of policies and institutions. Panel data evidence from OECD countries. OECD Economics Department Working Papers, no. 283.

Belorgey, N., Lecat, R. and T.P. Maury (2006). Determinants of productivity per employee: An empirical estimation using panel data. Economics Letters, 91(2), 153–157.

Bernard, A.B., Redding, S.J. and P.K. Schott (2007). Comparative advantage and heterogeneous firms. Review of Economic Studies, 74(1), 31-66.

Bloom, N., Draca, M. and J. Van Reenen (2016). Trade induced technical change? The impact of Chinese imports on innovation, IT and productivity. The Review of Economic Studies, 83(1), 87-117.

Bourlès, R. and G. Cette (2007). Trends in “structural” productivity levels in the major industrialized countries. Economics Letters, 95(1), 151–156.

Bown, C., Jung, E. and Z. Lu (2018). Trump, China, and tariffs: From soybeans to semiconductors. Vox.

Branstetter, L. (2006). Is foreign direct investment a channel of knowledge spillovers? Evidence from Japan's FDI in the United States. Journal of International economics, 68(2), 325-344.

Braunerhjelm, P., Acs, Z.J., Audretsch, D.B. and B. Carlsson (2010). The missing link: knowledge diffusion and entrepreneurship in endogenous growth. Small Business Economics, 34(2), 105-125.

Buccirossi, P., Ciari, L., Duso, T., Spagnolo, G. and C. Vitale (2013). Competition policy and productivity growth: An empirical assessment. Review of Economics and Statistics, 95(4), 1324-1336.

Burstein, A. and J. Vogel (2017). International trade, technology, and the skill premium. Journal of Political Economy, 125(5), 1356-1412.

Cameron, G., Proudman, J. and S. Redding (2005). Technological convergence, R&D, trade and productivity growth. European Economic Review, 49(3), 775-807.

Carree, M.A., Van Stel, A.J., Thurik, A.R. and A.R.M. Wennekers (2002). Economic development and business ownership: An analysis using data of 23 OECD countries in the period 1976–1996. Small Business Economics, 19(3), 271–290.

Carree, M.A., A.J. Van Stel, A.R. Thurik and A.R.M. Wennekers (2007), The relationship between economic development and business ownership revisited, Entrepreneurship and Regional Development, 19(3), pp. 281-291.

Carree, M.A. and A.R. Thurik (2008). The lag structure of the impact of business ownership on economic performance in OECD countries. Small Business Economics, 30(1), 101–110.

Chen, Z., Liu, Z., Suárez Serrato, J.C. and D.Y. Xu (2018). Notching R&D investment with corporate income tax cuts in China. National Bureau of Economic Research, NBER Working Paper no. w24749.

Coe, D.T., & Helpman, E. (1995). International R&D spillovers. European Economic Review, 39(5), 859–887.

Coe, D.T., Helpman, E. and A.W. Hoffmaister, A. W. (2009). International R&D spillovers and institutions. European Economic Review, 53(7), 723–741.

Cohen, W.M. and D.A. Levinthal (1989). Innovation and learning: the two faces of R&D. The Economic Journal, 99(397), 569–596.

Cohen, D. and M. Soto (2007). Growth and human capital: good data, good results. Journal of Economic Growth, 12(1), 51-76.

CPB (2018). Trade Wars: Economic impacts of US tariff increases and retaliations. An international perspective. CPB Background Document, The Hague.

Donselaar. P. (2011). Innovatie en productiviteit. Het Solow-residu ontrafeld (in Dutch). Erasmus University Rotterdam.

ECB (2017). ECB-Global: introducing ECB's global macroeconomic model for spillover analysis. ECB Working Paper Series, no. 2045, Frankfurt.

ECB (2018). Macroeconomic implications of increasing protectionism. ECB Economic Bulletin 6/2018, Frankfurt.

Edwards, S. (1998). Openness, productivity and growth: what do we really know?. The Economic Journal, 108(447), 383-398.

Engelbrecht, H.J. (1997). International R&D spillovers, human capital and productivity in OECD economies: An empirical investigation. European Economic Review, 41(8), 1479–1488.

Erken (2008). R&D, Productivity and Entrepreneurship. ERIM, Erasmus University Rotterdam.

Erken, H.P.G., Donselaar, P., and A.R. Thurik (2016). Total factor productivity and the role of entrepreneurship. Journal of Technology Transfer, 1-29.

Erken, H.P.G., Every, M., Giesbergen, B. and M. Wijffelaars (2018). Is a US trade war imminent? Rabobank.

Erken, H.P.G., Every, M. and B. Giesbergen (2018). US-China trade war: Back to the future. Rabobank.

Erken, H.P.G. and M. Heijmerikx (2018). Mexico/Canada: are free-trade agreements with the EU beneficial? Rabobank.

Erken, H.P.G. and M. Tulen (2017). US global value chain integration: a major impediment for Trump’s protectionist trade agenda. Rabobank.

Erken, H.P.G., Hayat, R., Prins, C., Heijmerikx, M. and I. de Vreede (2018). Measuring The Permanent Costs of Brexit. National Institute Economic Review, 244(1), R46-R55.

Erken, H.P.G. and B. Giesbergen (2017). North Korea’s pressure point: the economy. Rabobank.

Erken, H.P.G., Verbruggen, K. and P. Marey (2018). Trade wars then and now: Smoot-Hawley all over again. Rabobank.

Fagerberg, J. (1987). A technology gap approach to why growth rates differ. Research Policy, 16(2–4), 87–99.

Feyrer, J. (2009). Trade and income. Exploiting time series in geography. National Bureau of Economic Research, NBER Working Paper no. w14910.

Filatotchev, I., Liu, X., Lu, L. and M. Wright (2011). Knowledge spillovers through human mobility across national borders: Evidence from Zhongguancun Science Park in China. Research Policy, 40(3), 453-462.

Frantzen, D. (2000). R&D, human capital and international technology spillovers: A cross-country analysis. Scandinavian Journal of Economics, 102(1), 57–75.

Goldberg, P.K., Khandelwal, A.K., Pavcnik, N., & Topalova, P. (2010). Imported intermediate inputs and domestic product growth: Evidence from India. Quarterly Journal of Economics, 125(4), 1727-1767.

Griffith, R., Redding, S. and J. Van Reenen (2004). Mapping the two faces of R&D. Productivity growth in a panel of OECD industries. Review of Economics and Statistics, 86(4), 883–895.

Griliches, Z. (2000). R&D, education, and productivity. A retrospective. Cambridge (MA): Harvard University Press.

Grossman, G.M., & Helpman, E. (1991). Innovation and growth in the global economy. Cambridge, MA: MIT Press.

Guellec, D., & Van Pottelsberghe de la Potterie, B. P. (2004). From R&D to productivity growth: Do the institutional settings and the source of funds of R&D matter? Oxford Bulletin of Economics and Statistics, 66(3), 353–378.

Himmelberg, C.P. and B.C. Petersen (1994). R & D and internal finance: A panel study of small firms in high-tech industries. Review of Economics and Statistics, 38-51.

Huang, Z. and X. Du (2017). Government intervention and land misallocation: Evidence from China. Cities, 60, 323-332.

IMF (2018a). World Economic Outlook. Challenges to steady growth, Washington.

IMF (2018b). G-20 Surveillence Note. Washington.

Jones, C. I. (1995). Time series tests of endogenous growth models. Quarterly Journal of Economics, 110(2), 495-525.

Khan, M., and K.B. Luintel (2006). Sources of knowledge and productivity: How robust is the relationship? OECD, STI/Working Paper 2006/6.

König, M., Song, Z., Storesletten, K. and F. Zilibotti (2017). From imitation to innovation? Where is all that Chinese R&D going? Working Paper in progress, Yale University.

Lawrence, C., K. Verbruggen and H.P.G. Erken (2018). Emerging Markets Vulnerability Heatmap. Rabobank.

Lee, G. (2005). International R&D spillovers revisited. Open Economies Review, 16(3), 249-262.

Liadze, I. (2018). Trade war saga continues. NiGEM Observations, No 13.The National Institute of Economic and Social Research (NIESR).

Liu, X. and T. Buck (2007). Innovation performance and channels for international technology spillovers: Evidence from Chinese high-tech industries. Research Policy, 36(3), 355-366.

Liu, X., Lu, J., Filatotchev, I., Buck, T. and M. Wright (2010). Returnee entrepreneurs, knowledge spillovers and innovation in high-tech firms in emerging economies. Journal of International Business Studies, 41(7), 1183-1197.

Lovely. M.E. and Y. Liang (2018). Revised tariffs against China hit non-Chinese supply chains even harder. Peterson Institute for International Economics.

Lucas, R.E. (1988). On the mechanisms of economic development. Journal of Monetary Economics, 22(1), 3–42.

Michelacci, C. (2003). Low returns in R&D due to the lack of entrepreneurial skills. The Economic Journal, 113(484), 207–225.

Park, J. (2004). International student flows and R&D spillovers. Economics Letters, 82(3), 315-320.

Prieger, J.E., Bampoky, C., Blanco, LR. and A. Liu (2016). Economic growth and the optimal level of entrepreneurship. World Development, 82, 95–109.

Romer, P.M. (1990). Endogenous technological change. Journal of Political Economy, 98(5), S71–S102.

Tetlow, G. and A. Stojanovic (2018). Understanding the economic impact of Brexit. Institure for Government, London.

Van Es, F. and E. Barendregt (2018). Global economic outlook: same picture, but far from boring. Economic Quarterly Report, Rabobank.

Van Elk, R., Verspagen, B., Ter Weel, B., Van der Wiel, K. and B. Wouterse (2015). A macroeconomic analysis of the returns to public R&D investments. CPB Netherlands Bureau for Economic Policy Analysis, 313.

Vogel, S. (2018). Changing the soybean trade rhythm. Rabobank.

Wei, S. J., Xie, Z. and X. Zhang (2017). From" Made in China" to" Innovated in China": Necessity, prospect, and challenges. Journal of Economic Perspectives, 31(1), 49-70.

Wu, H.X. (2014). China’s growth and productivity performance debate revisited. The Conference Board Economics Working Papers (No. 14-01).

Yu, M. (2014). Processing trade, tariff reductions and firm productivity: evidence from Chinese firms. The Economic Journal, 125(585), 943-988.

Zhang, L. (2017). The knowledge spillover effects of FDI on the productivity and efficiency of research activities in China. China Economic Review, 42, 1-14.

Zhao, Y. (2014). Who's afraid of the big bad dragon?: Why China has the best (and worst) education system in the world. John Wiley & Sons.

Benchmark: no trade war

In our benchmark scenario, we assume that the US and China are not engaged in a trade war and no protectionist packages are installed on either side. Starting point for this scenario are the forecasts published in Rabobank’s most recent economic quarterly outlook (Van Es and Barendregt, 2018). As some tariffs had already been imposed when that outlook was being made, we have to strip out these effects from that outlook. This is done by adjusting the figures for China and the US with the negative effects on export that we have calculated earlier in this Special. NiGEM consequently adjusts the effects on other variables, such as inflation, private consumption and the import volume. This results is a 0.1ppts higher GDP growth for US in 2018, and a 0.1ppts higher GDP growth for China in both 2018 and 2019 vis-à-vis our latest outlook.

By using this benchmark scenario we are able to compare the damage of the trade war by making comparisons with the two other scenarios.

Trade war scenario assumptions

Below we discuss all assumptions of our trade war scenarios for a set of key factors determining the economic impact (Table A.1).

Tariffs and timing

Our scenarios differ both in terms of assumptions and timing. With regard to scenario 1, we include the measures that have already been implemented and that have already been announced. The shocks we implement in NiGEM in the third quarter of 2018 (for packages that have already been installed), and in the first quarter of 2019 (for the announced tariff increase from 10 percent to 25 percent in January 2019). In the two scenarios we increase prices of both commodity and non-commodity exports for both countries. We first calculated the commodity/non-commodity share for each introduced or announced protectionist package. In all cases, we have applied the actual and announced tariffs as a base, but given the 5-10% range on China’s most recent USD 60bn retaliatory package, we apply an average 7% base tariff on that package. In our second scenario we assume that China will introduce non-tariff barriers as an additional retaliatory policy. Therefore, we introduce an add-on of 15 percent on China’s average weighted import tariffs. This is based on China’s current ad valorem equivalent (AVE) of NTBs and its country tariff profile derived from the World Bank’s World Integrated Trade Solution.

Exchange rates

In our scenarios we have included the expert opinion of our Asia FX strategist on the projections for the Chinese currency (the Chinese yuan: CNY) in order to construct an exogenous path (Table A.2). In our first scenario we use an inverse relation for the USD compared to CNY. In the second scenario, we assume a partial (21%) USD pass-through of so-called safe haven effects. Given that the US also plays an important role in the trade war, we assume a lower safe haven USD effect than during other geopolitical events. We shock both currencies on a permanent basis in 2019Q3.

Risk premia

In our second scenario we assume that investors are less willing to buy Chinese government bonds due to the full escalation of the trade war. This not only deteriorates China’s export position, but also (long-term) investments in China. We additionally take into account an increase of the Chinese term premium by 67 bps, which increases the user costs of capital, and thus negatively affects China’s total investment. Our rationale is related to the overall rise in spreads since the trade tensions took off in May 2018, which was relatively more substantial than in the previous months. In Table A.3, we illustrate that the increase in the spread was 67bps between the Chinese 3m and 10y government bond between May and September 2018. One should however note that this numerical assumption is purely based on the development of spreads in that period. The underlying decline in 10y and 3m yields in this case is not a reflection of investors losing investment appetite in China’s bond market, it rather relates to liquidity injections by the central bank.

Monetary policy

In both scenarios, we do not include any possible monetary policy rate hikes as a result of so-called cost-push inflation. The tariffs imposed do lead to upward pressure on prices, therefore NiGEM would endogenously relate this to an increase in interest rates, resulting in higher borrowing costs and thus lowering incentives to invest. This results in lower domestic demand and a negative impact on world trade. As such, this would lead to an additional negative knock-on effect aside from our assumptions regarding tariff and non-tariff add-ons. But such policymaking of central banks would be irrational, given the fact that these inflation effects due to higher tariffs are a one-off. For the forecast period after 2025Q3, we switch to endogenous monetary policy simulation.

Total factor productivity

Last but certainly not least we take stock of potential dynamic productivity effects in our study. This is a unique feature, as others so far did not incorporate any of these effects. It is based on the assumption that the trade tensions will have a permanent impact on the supply side of the economies involved, i.e. the US and China. For a detailed selection of variables incorporated in this analysis and the extrapolation of these variables, see Tables A.4 and A.5.

In this appendix we conduct two important robustness checks for the China TFP model. First, the model might be subject to serial correlation, as indicated by the relatively high Durbin-Watson statistic, which ideally should have a value near 2.0. To examine whether serial correlation leads to biased results, in our first robustness test we add a lagged dependent variable to our preferred Chinese TFP model. A second robustness check is to see whether our model remains stable when we substitute our TFP growth variable by the TFP alternative measure from the Conference Board. This is important, as TFP measures deviate to a relatively large degree between different data sources.

Serial correlation and omitted variable bias

In column (1) of Table A.6, we have included our preferred Chinese TFP model depicted in column (7) of Table 5. In column (2) we add a lagged dependent variable to test whether our results are biased due to serial correlation or omitted variable bias. Including a lagged dependent variable can reduce the occurrence of autocorrelation arising from model misspecification. Moreover, if omitted variable bias is a problem, a lagged dependent will pick up many variance and is likely to make other variables statistically less significant. Column (2) shows the results and show that the estimation results remain virtually unchanged. Moreover, the lagged dependent variable is insignificant. We also conducted equal test with our US TFP model, which leads to the same conclusions: the lagged dependent is insignificant and the other estimated effects are unchanged.

Alternative TFP measure

As discussed, there is much more debate about the accuracy of total factor productivity measurement for China. As a second robustness test, we have taken the alternative measures from the Conference Board based on the comprehensive work of Wu (2014) and ran two regressions. The results are depicted in columns (3) and (4) in Table A.6 and are remarkably close to the results based on PWT data. One important difference is that the magnitude of the elasticity for domestic private R&D and human capital is somewhat lower, whereas the coefficient for foreign private R&D capital is slightly higher.