This brief deals with the extent to which a more competent public bureaucracy can contribute to better economic outcomes. It addresses this question in the context of public procurement, governments’ purchase of goods and services from private contractors, which accounts for about 15% of GDP in most economies and is on the rise. The efficiency of the procurement process directly influences the prices and quality of many government-provided goods and services that are crucial to social welfare objectives and sustained economic growth. Several issues challenge this efficiency. Media attention is typically on episodes of corruption, which can of course be a major source of waste. Here, we focus on a less glamorous, often overlooked, but potentially even more important source of waste, the lack of procurement competence.
Public procurement is a complex task. Contracting authorities must know market characteristics, design and implement efficient award mechanisms, balance risks and incentives in drafting contracts, effectively manage the contracts in the execution phase, etc. Effective procurement, in particular for complex services or works, requires teams endowed with legal, marketing, engineering, and economic/strategic expertise. The World Bank‘s Benchmarking Public Procurement 2017 compares the quality of the legal and regulatory environments of 180 countries and reveals the existence of great heterogeneity in the quality of the procurement processes across countries. Saussier and Tirole (2015) focus on the case of France, documenting that 63% of the staff of French contracting authorities do not have a legal profile, and only 39% have qualifications specific for managing public purchases.
Recent research focusing on prices of standardized goods showed that (lack of) buyer competence among public buyers could make an even bigger impact on the waste of public funds than corruption. For example, Bandiera, Prat and Valletti (2009) estimate that Italian public buyers would save 21 percent of their expenditures if they all paid the same as the buyers at the 10th percentile of the estimated procurement price distribution. Savings could reach 1.6-2.1 percent of the Italian GDP per year. They then estimate that bureaucratic inefficiency also linked to incompetence is the main cause of waste, accounting for 83 percent of total estimated waste, compared to only 17 percent due to corruption. In a similar vein, Best, Hjort and Szakonyi (2017) report that over 40 percent of within-product price variation on standardized goods in Russia in 2011-2015 can be ascribed to the bureaucrats and organizations in charge of procurement. They estimate that if the least effective quartile of bureaucrats and organizations had the effectiveness of the 75th percentile, the Russian government would save around $13 billion per year – roughly one fifth of the total amount spent on health care by the Russian government at federal, regional, and municipal level combined.
The role of competence in complex procurement
This problem is becoming even more serious now that, being under fiscal pressure after the crises, many governments are promoting the use of public procurement not only as a tool to save budgets – sometimes at the expense of quality – but also to achieve more complex objectives like fostering innovation, protecting the environment, and promoting social objectives, a multiplicity of goals that per se makes the procurement mission even more complex.
Little is known about the importance of procurement competence in more complex procurements, not least because it is very difficult to measure performance in these environments. In our paper (Decarolis et al. (2019)) we try to make a step in this direction by focusing on works and services, typically more complex than goods. We use data from the US, probably the country with the most well-developed system of production and certification of procurement competences. Thus, our estimates of the effect of lack of competences should provide a lower bound of most other countries.
We combine, for the first time, three large databases: contract-level data on procurement performance in the Federal Procurement Data System (FPDS); bureau-level data from a survey conducted by the Office of Personnel Management since 2002 on federal employees, the Federal Employee Viewpoint Survey (FEVS); and Federal Workforce Data (FedScope) containing information on characteristics of the public workforce at the employee level.
To quantify the extent to which the government-bureau-level competencies determine procurement outcomes, we use the first database to construct procurement performance measures and the second dataset to build measures of procurement offices’ competence. We then use the third database to construct instruments that help us addressing important endogeneity issues. Our identification strategy exploits the exogeneity of death events involving public officials to allow for a causal interpretation of bureau competence on procurement performance.
Indeed, there are three main challenges that our analysis needs to overcome. The first is how to measure procurement performance. Unit price comparisons have been used for standardized goods, but they are not suitable for the more complex procurements we focus on as they are heterogeneous in many non-recorded dimensions and their contracts are often incomplete. We use FPDS instead to construct three proxies of performance based on time delays, cost overruns, and the number of renegotiations. Although the first two measures are widely used in the literature, we are careful to take into account that cost overruns and delays may be due to new or additional work requested by the public buyer, in which case they should not be viewed as indicative of a poor outcome. We, therefore, consider only those which have occurred to deliver the work or service that was originally tendered. The third performance measure, the overall number of renegotiation episodes, is new and aims at capturing Williamson’s “haggling costs,” which are a pure deadweight loss present whatever the reason behind the renegotiation and have been shown to be economically sizeable for complex contracts. Our data reveals a surprising and persistent heterogeneity along these three dimensions across US federal bureaus.
The second challenge is the measurement of bureaucratic competence. Other papers in the field have measured it using buyer fixed effects. We use a novel approach based on the mentioned survey of employees’ subjective evaluations (FEVS). The survey is extremely rich, and we chose the most general question as an overall measure for competence (How would you rate the overall quality of work done by your work unit?). Responses to this question should be seen as measures of the overall efficacy of the workflow and processes within the bureau, hence proxying for the ideal measure of competence on the many different aspects relevant to procurement. An extensive set of robustness checks support our idea of measuring competence through the FEVS data.
The third measurement problem is the association between more complex contracts and more competent buyers: the most competent buyers may consistently produce poor performance because they are allocated the most complicated procurements. This point is well illustrated in a case study showing that the performance of the agencies that are worst in terms of competence (the Department of Veterans Affairs and the Department of Justice) is superior to that of the two most competent agencies (the NASA and the Nuclear Regulatory Commission) in terms of both delays and cost overruns. This striking inversion indicates that any straightforward regression of performance on competence would grossly underestimate the impact of competence.
We, therefore, develop an instrumental variable strategy exploiting exogenous changes in competence. We use FedScope to build instruments for bureaus’ competence based on the deaths of specific types of employees: bureau managers and white-collar employees who are relatively young and earn a relatively high wage. The idea is that more competent offices adopt better managerial practices, routines and processes that are more resilient to risks, such that of an unexpected loss of a key employee, and less dependent on specific individuals. This is precisely what the first stage of our IV strategy documents. Our instruments perform well in terms of their statistical properties and they allow us to estimate a causal effect of bureau competence on procurement outcomes that is an order of magnitude larger than the corresponding OLS estimate.
We find that one standard deviation increase in competence reduces the number of days of delay by 23 percent, cost overruns by 29 percent and the number of renegotiations by half. This implies that if all federal bureaus were to obtain NASA’s high level of competence – corresponding to the top 10th percentile of the competence distribution – delays in contract execution would decline by 4.8 million days, and cost overrun would drop by $6.7 billions over the entire sample analyzed. We also find a consistently negative effect of greater competence on the number of renegotiations: one standard deviation increase in competence causes 0.5 (39%) and 0.8 (71%) fewer cost renegotiations and time renegotiations, respectively.
Finally, we try to understand what exactly makes a bureau ‘competent’ using the FEVS data to identify three different components: cooperation among employees, incentives and skills. Separately estimating their causal effects is unfeasible with instruments like the two described above as the validity of the exclusion restriction, which can be argued to be satisfied when measuring a broadly defined notion of bureau competence, is unlikely to hold for more specific components. However, we provide multiple pieces of evidence suggestive that cooperation is the key driver behind the positive effects of bureau competence. This finding conforms with the view that successful procurement requires appropriate coordination of a multiplicity of tasks involving different individuals. We also consider the extent to which the role of cooperation is due to the presence of capable managers, able to lead a group to effective cooperation, exploiting the heterogenous effects obtained through instruments considering the deaths of different subgroups of employees. We find that the deaths that matter the most are those of relatively young and best paid white-collar employees.
These results point at the large potential improvement in the performance of public contracts that could be achieved by investing more resources in increasing the competence of contracting authorities, even in a country with long-established procurement training and certification institutions such as the US. In Europe, recent policy initiatives see the introduction of qualification systems for public procurers as a necessary response to the generally lower procurement competence coupled with the greater discretion granted by the 2014 Procurement Directives. Our results on the role of cooperation suggest that certification programs would be also useful at the level of the procuring office, and should include features such as the organisation of the acquisition process and the prevailing management practices, as is often done for private firms.
- Bandiera, Oriana, Andrea Prat, and Tommaso Valletti. (2009). “Active and passive waste in government spending: Evidence from a policy experiment.” American Economic Review, 99(4): 1278-1308.
- Best, Michael Carlos, Jonas Hjort, and David Szakonyi. (2017). “Individuals and Organizations as Sources of State Effectiveness, and Consequences for Policy.” NBER Working Paper 23350.
- Bucciol, A., Camboni R., and P. Valbonesi. (2017). ”Buyers’ Ability in Public Procurement: A Structural Analysis of Italian Medical Devices.” Working Paper n.17, Department of Economics, University of Verona.
- Decarolis, F, L Giuffrida, E Iossa, V Mollisi, and G Spagnolo. (2019 revision), “Bureaucratic Competence and Procurement Outcomes”, NBER Working Paper 24201.
- Saussier, S, and J Tirole. (2015). “Strengthening the Efficiency of Public Procurement”, French Council of Economic Analysis, April.
- World Bank (2017), Benchmarking Public Procurement 2017, The World Bank.
 See also, Bucciol et al (2017) who study procurement of standardized medical devices purchased by local Italian purchasing bodies, finding that the price for the same medical devices paid by Italian public buyers differ substantially, and that the differences are explained by ‘buyers fixed effects’ capturing all specific buyers characteristics, including their competence levels.
Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.
The last 30 years has witnessed a shift of the world’s manufacturing core from Europe and North America to China. A key question is what impact this has had on manufacturing workers in other developed economies, and also on innovation, patenting, IT adoption, and productivity growth. While a rigorous data analysis on these variables for developing economies, particularly in Eastern Europe, is not yet available, this brief examines the impact of the rise of China on innovation in Western Europe, and also reviews the evidence on the impact of the rise of China generally. Recent research by Bloom, Draca, and Van Reenen (2016) found that Chinese competition induced a rise in patenting, IT adoption, and TFP by 30% of the total increase in Europe in the early 2000s. Yet, we find numerous problems with the Bloom et al. analysis, and, overall, we do not find convincing evidence that Chinese competition increased innovation in Europe.
Few events have inspired the ire of economists as much as Brexit and the rise of Donald Trump, two events seen as related as both were a seeming reaction to both globalization and slowing economic growth, particularly as some (such as Trump himself) saw the former as a key cause of the latter. Both Brexit and the trade war spawned by Trump do seem to have had negative economic effects – US equities have suffered every time the trade war has escalated, while anecdotal reports and more sophisticated economic analyses seem to suggest that Brexit has cost the UK jobs.
And yet, there is a need for policy makers and economists to hold two ideas in our heads simultaneously: Trump’s trade war and Brexit may be policy disasters, and yet globalization can create both winners and losers, even if it is clear that, generally speaking, the overall gains are likely positive and large. This is likely also true of the rise of China – one of the most dramatic events in international economics in the past 50 years. Figure 1 shows the increase in trade with China from the early 1980s to 2017, a period in which US imports from China grew from 7 to 476 billion dollars.
Figure 1. Chinese Imports (in logs, deflated)
Source: World Bank WITS
The academic literature tends to show that this impact, the rise of China, may have cost the US as much as 2.2 million jobs directly (Autor et al.), and as much as 3 million jobs once all input-output and local labor market effects are included. While approximate, these numbers are large enough for the China shock to have played a role in the initial onset of “secular stagnation” – the growth slowdown which began around 2000 for many advanced nations, including the US and Europe. In addition, Autor et al. (forthcoming) found that Chinese competition also resulted in a decline in patent growth. In the European context, however, other authors have found that although China did do some damage to certain sectors, overall, it does not appear to have been quite as damaging, particularly in Germany, which also benefitted from exporting increased machine tools to the Chinese manufacturing sector. And, in a seminal paper, Bloom, Draca, and Van Reenen (2016) find that Chinese competition actually led to an increase in patents, IT adoption, and productivity in Europe from 1996 to 2005, along accounting for nearly 30% of the increase. This is important, as it implies that without the rise of competition with China, the slowdown in European growth would have been even more pronounced than it was. It also implies that, far from being a source of stagnation, Chinese competition has been a source of strength. It also makes it more likely that the slowdown in growth since 2000 was caused by supply-side factors, such as new inventions becoming more difficult over time, as is perhaps the leading explanation among economists, notably Northwestern University business professor, Robert Gordon (2017), and also supported by others (see this VoxEU Ebook featuring a “who’s who?” among economists). It would also be evidence that contradicts the “Bernanke Hypothesis” that the former US Fed Chair first laid out in a 2005 speech at Jackson Hole, in which he suggested that international factors – particularly the savings glut and US trade deficit – were behind falling interest rates in the US. Since then, Ben Bernanke has followed up with a series of blog posts suggesting that these international factors were the cause of the initial onset of secular stagnation.
Figure 2. European Growth Relative to Trend
Source: World Bank WDI
In this brief, I present new research in which my coauthor and I test the robustness of the research finding that China had a positive impact on innovation in Europe (Campbell and Mau, 2019). We find that these findings are very sensitive to controls for time trends and other slight changes in specification. We also find that the number of patents matched to firms in the sample shrinks over the sample period (from 1996 to 2005). Overall, we conclude that, unfortunately, it is unlikely that the rise led to a significant increase in innovation in Europe, although more research is needed. Our research also sheds light on the so-called “replication crisis” currently gripping the social sciences, as researchers begin to realize that many published findings are not robust.
Trade-Induced Technical Change?
Bloom, Draca, and Van Reenen (2016) – hereafter BDV – tried to isolate the impact of the rise of China on Europe using several methods, using firm-level data for Europe. They placed each firm in a 4-digit sector, where they measured imports from China over time. First, they just looked at changes in patents, IT, and total factor productivity (TFP) at the firm level for sectors in which Chinese imports increased a lot vs. other sectors. But, because economists are always weary of the difficulty of isolating a causal relationship from non-experimental data, the authors, worrying that the sectors which saw increases in Chinese imports might differ systematically from the others, the authors also used what is called an instrumental variable. That is, they used the fact that when China joined the WTO in 2001, they also negotiated a reduction in textile quotas. Thus, BDV reason that textile sectors which had tightly binding quotas prior to removal were likely to have had fast growth in Chinese imports after China’s accession to the WTO. Thus, they end up comparing textile sectors in which the quotas were binding to sectors in which they were not binding. We went back and compared the evolution of patents in these same groups (sectors with binding textile quotas vs. not binding) below in Figure 3.
Figure 3. Patent Growth in China-Competing Sectors (Quota Group) vs. Other Sectors
Notes: The vertical red lines are dates when textile quotas were removed. The blue line shows the evolution of patents in the sectors without binding quotas (non-competing sectors), and the red line is the evolution of patents in the China-competing sectors. The dotted lines are 2 standard deviation error bounds.
What is immediately obvious in Figure 3 is that patents are declining rapidly over the whole period in both groups. The overall level of patents was falling in both groups for the full period. There is a 95.8% decline in patenting for the China-competing group, vs. a 96.2% decline for firms in the non-competing (“No quota”) group. By 2005, average patents per firm are close to zero in both groups (.04 in the China-competing sectors vs. .11 in the others). However, in the “No quota” group, the initial level of patents – close to three per firm per year – was much larger than in the quota group. Since patents are falling rapidly in both groups but bounded by zero, the level of the fall in patents in the non-quota group is larger, but one can easily see that much of this decline happens before quotas are removed. If we control for simple time trends, the effect goes away. Also, given the tendency of patents to decline, we can also remove the correlation between Chinese competition and patent growth in some specifications by simply controlling for the lagged level of patents. The overall declining share of patents in the BDV data also raises questions about data selection issues, as patents granted in the BDV data in the later years were a smaller share of the total patents actually granted in reality.
BDV also look at the impact of the rise of China on IT adoption. However, here they proxied IT adoption by computers per worker, but they did not collect enough data to control for pre-trends properly in the data, so we cannot be sure whether this correlation is causal or not. (For what it is worth, on the data we do have, from 2000 to 2007, including trends in the data renders the apparent correlation between Chinese import growth and computers-per-worker insignificant.)
Lastly, BDV look at the impact of the rise of China on TFP growth. Here, unlike before, we find that their measure is robust across various estimation methodologies. However, when we look at changes in a commonly used alternative measure of productivity, value-added per worker, instead of TFP (as TFP needs to be calculated using strong assumptions about the functional form of technology), we find no impact (see Figure 4 below).
Figure 4. Value-Added per worker Growth: China-competing sectors vs. others
Figure 4 above compares the evolution of value-added per worker in the most China-competing sectors vs. the others. Trends look similar for firms in either group of sectors (China-competing or otherwise), and we do not find a correlation. We also do not find that Chinese competition led to an increase in profits, nor an increase in sales per worker (in fact, we found a significant decrease in most specifications).
All in all, we find that the BDV findings suggesting that the rise of China had a large impact on innovation in Europe is not robust. However, in most specifications, we also don’t find a negative impact as did Autor et al. (forthcoming) for the US. This might have to do with data quality, although it does seem to be closer to other work, such as Dauth et al. (2014), which suggests that the rise of China had a smaller impact in Germany than in the US.
We also felt it was a bit alarming that a simple plot of the trends in patents for China-competing and not-competing sectors was enough to seriously question the conclusions of BDV, as their paper was published in the Review of Economic Studies, a top 5 journal in academic economics. If influential articles published in the most fancy journals can exhibit such mistakes, this underscores the extent which the profession of economics may suffer from many published “false-positive” results. The reasons why this could be the case are obvious: researchers are under pressure to find significant results, as top journals don’t often publish null results, and replication is exceedingly rare in a field in which one needs to make friends to publish. However, there are signs that replication is becoming more mainstream, and as it does, we can certainly hope that voters around the world will turn back to science.
- Autor, D., D. Dorn, G. H. Hanson, G. Pisano, and P. Shu. Forthcoming. Foreign Competition and Domestic Innovation: Evidence from US Patents. Forthcoming: AEJ:Insights.
- Bloom, N., M. Draca, 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.
- Campbell, Douglas and Mau, Karsten. 2019.. Trade Induced Technological Change: Did Chinese Competition Increase Innovation in Europe?”, mimeo
- Dauth, W., S. Findeisen, and J. Suedekum. 2014. “The Rise of the East and the Far East: German Labor Markets and Trade Integration.” Journal of the European Economic Association 12 (6): 1643–1675.
- Gordon, R.J., 2017. The rise and fall of American growth: The US standard of living since the civil war (Vol. 70). Princeton University Press.
Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.
The trade war between the United States and China has become one of the main events in the global economy this year. What could be its consequences for the US and China, and how might it affect other countries – for example, Russia?
Chronology of the trade war
Donald Trump started the war, raising import tariffs on solar panels in January 2018, of which the main supplier is China. In response, on April 2nd, China raised import duties on 128 commodities originating from the United States. On July 6th, the US increased tariffs on Chinese goods by 25 pp., imports worth $34 billion. China responded symmetrically. In August, the United States increased the tariffs on another $16 billion of imported goods from China, to which a symmetrical response again followed. In September, the United States again applied higher tariffs for $200 billion of Chinese exports, and China for $60 billion of US exports. At each stage of the conflict escalation, China appealed to the WTO with complaints about the actions of the United States, pointing to the inconsistency of their actions with the obligations and principles of the WTO. There were several meetings of official representatives from the United States and China – without any significant results.
What are the main reasons for this unprecedented escalation?
Imbalance and intellectual property
The economies of the US and China today are by far the largest in the world, and the trade turnover between the two countries is one of the most important. A remarkable feature of these trade flows over last decades is their imbalance. In 2017, the United States imported $526 billion worth of goods from China, while China’s imports from the United States amounted to $154 billion. Part of this imbalance is offset by trade in services, but it is not enough to even it out: in the same the year the United States delivered $57 billion worth of services to China while importing services of $17 billion from China.
Experts have different views on this imbalance. On the one hand, there is a perception that it is a source of world economy vulnerability, a source of potential crisis. Therefore, it is necessary to reduce the trade deficit. Another point of view is that this imbalance merely reflects the fact that the US economy and its assets are very attractive to investors from all over the world, including Chinese – and that, in turn, requires that the surplus of capital flows biased to US side, was compensated by the corresponding deficit of trade in goods and services. One such investor is the Chinese state itself, which for many years has been pursuing a policy of exchange rate undervaluation in order to promote foreign trade. It led to an enormous accumulation of foreign exchange reserves and as of January 2018, China held $1.17 trillion of US bonds and was the largest creditor of US government.
US President Donald Trump referred to this trade imbalance as one of the reasons for the outbreak of this trade war against China. Trump aims at reducing the deficit by $100 billion from the current $375 billion. The unilateral increase in import tariffs applied to Chinese goods was the first action of the US administration in this direction.
The second, no less important, formal reason for the trade war is the inadequate protection of intellectual property rights in China. China’s production of counterfeit products, the lack of adequate practices and laws to protect foreign technologies from illegal dissemination in the country, is not news to anyone. And although the almost two decades since China’s WTO accession have meant a largely modernized legal framework in this regard, a number of important provisions are still inconsistent with international practices, and the implementation of existing intellectual property rights leaves much to be desired. Established in 2012, The Commission on the Theft of American Intellectual Property identifies China as the most malicious violator of US rights. The exact damage is not known, but the commission assessment of the losses to the American economy due to the forced transfer of technology to Chinese partners – which is an unspoken condition of foreign manufacturers access to the Chinese market – industrial espionage, contradictions in legislation, requirements for the storage of sensitive data in China are in the range from $225 to $600 billion per year (Office of US Trade Representative, 2018).
While both the trade deficit and the intellectual property rights issue were recognized for many years, it was in 2018 that Trump started acting on them. Therefore, in order to discuss the potential impact of the conflict between the world’s largest economies on themselves and other economies, such as Russia, it is important to understand what drives the actions undertaken by Trump’s administration.
Trump won the elections in 2016 with a minimum margin against the Democratic rival. To provide support for his decisions and to increase the chances of being reelected for the next term in 2020, it is crucial to maximize the pool of his supporters. Trade policy measures aimed at import substitution are very effective populist policies in any country. One of the first steps made by the US toward trade war was the increase in import tariffs on steel and aluminum – for all countries. Metallurgy and coal industries are among the most organized and strong lobbyists in any country. The European Union as an economic organization started with the European Coal and Steel Association. By aligning interests with these sectors much can be achieved in relation to trade liberalization, and vice versa – by increasing the level of protectionism, a significant popularity increase can be among voters whose incomes depend on the success of companies in these industries.
China works hard raising the technological level of its economy. In recent years the Chinese government and Communist party launched a number of ambitious programs aimed at achieving a technological breakthrough, lessening the dependence on imported technologies by substituting them with ones produced by domestic innovation centers. These programs specify the priority sectors, in which state subsidies are provided for the acquisition of foreign technologies by Chinese companies and their adaptation. One of the common arguments was that the United States believes that powerful state support for technology sectors in China, along with the existing problems in protecting intellectual property rights, increases the risks and potential losses of American companies.
However, while these concerns seem reasonable at first, they should not be taken at the face value.
China’s ability to push out American companies in the high-tech sector on the world market seems rather limited. So far, China has only succeeded in increasing its share in the middle and low technology segments. Instead, in recent years, China is rapidly increasing its defense spending, which in 2017, for the first time, reached a level of 1 trillion yuan (about $150 billion). China’s defense spending is the second highest in the world after the United States. Moreover, it’s growing very fast. While in 2005 the Chinese nominal defense expenses were only 10% of American expenses, in 2018 they are already around 40%. The dominance of state enterprises in the defense industry in China implies that the real purchasing value of these expenditures is quite comparable. New and existing Chinese industrial policy programs target military and dual-use industries among others. Therefore whilst addressing the intellectual property rights problem in China now, Trump’s administration also aims at preserving US leadership position in the military sector, which finds widespread support in Trump’s main voter groups among Republicans.
Historically, trade wars implied tariff escalations to protect domestic industries from foreign competition. Today, the Trump administration behaves in a similar manner. However, the circumstances now are fundamentally different from those in the first half of 20th century and earlier. Firms not only trade in final goods, but more and more they trade in intermediate products and within firms themselves (Baldwin, 2012). The distribution of the production process to many companies across different countries of the world leads to two important effects, which were not observed in previous trade wars.
First, it is the effect of the escalation of tariff protection in the framework of the value chains. The import tariff is applied to the gross value of the product crossing the customs border. However, the exporting firm’s contribution to the gross value might be quite small. So the effective level of the tariff will be higher than the nominal level of the tariff, known as a so called amplification effect (World Bank, 2017, page 98). It means that the effective growth of the tariff by 25 percentage points in relation to Chinese imports will significantly exceed 25 % and in some cases can even become prohibitive. So, the tariff warfare will result in significantly greater losses for the sectors involved in the value chains, compared to the sectors less exposed to them. It means that foreign investors and multinational companies in China will suffer bigger losses compared to purely domestic Chinese companies. The Peterson Institute for International Economics made an assessment and confirmed these observations (Lovely and Yang, 2018).
Second, China’s participation in international multinational companies most often occurs in the assembly segments, while developed countries’ companies contribute at other stages, such as with innovation, design, financial and consulting services, marketing, and after-sales services. Then, the protectionist measures against goods produced in China by multinational companies will hit an American economy, generating losses in the service segments. A similar episode happened, for example, in 2006, when the European Union introduced anti-dumping duties on imported footwear from China and Vietnam, which in turn lead to a decline in the services sector in Europe – imported footwear contained a significant share of the value added created by European designers and distributors (World Bank, 2017). Obviously, we will observe the same consequences in the United States now, since the role of the American services sector in creating and promoting Chinese goods on the American market is significant and according to World Bank estimates in 2011, the contribution of value added generated by foreign services in China’s gross exports amounted to about 15% (World Bank, 2017).
Thus, not only the economy of China, but also the US economy itself will suffer from the growth of import tariffs in the USA. The USA is not an exception here – the governments of most countries continue to live in the paradigm of trade policy, which suits the structure of the world trade as at the beginning of the 20th century, while trade has gone far ahead and requires much more elaborate effective regulatory tools than tariffs on imported goods.
Consequences for Russia
The consequences of the US trade war with China for the Russian economy depend on what the main goals of the war are. If the motive is primarily electoral – to secure enough support in 2020, one can expect that the protective measures will be short-lived, and the geographical distribution of investment flows will remain almost intact and that China will remain an important location for global value chains transactions. The trade war will in this case lead to some economic slowdown in the short term. The main effects will be related to the redistribution of income within economies, where protected sectors will benefit on the expense of all other sectors. In these circumstances, Russia would suffer direct losses from the growth of tariffs on their exports to US (now it is predominantly steel and aluminum), but for the economy as a whole, the losses will not be significant, especially relative to the losses Russia bears because of sanctions.
However, if the main reason for the trade war has a long-term perspective, the investors will be forced to adjust the geography of their investment plans and China will face a significant outflow of foreign investments, which will significantly affect Chinese – and global – economic growth. In this case, both for Russia and for the whole world, the indirect effect of the US-Chinese trade conflict will be quite noticeable and it will take years to create new trade links and restore world trade and global value chains.
- Baldwin, Richard, 2012. “Global supply chains: why they emerged, why they matter, and where they are going”, CTEI Working papers 2012-13, The Graduate Institute, Geneve
- Lovely, Mary E., and Liang Yang, 2018. “Revised Tariffs Against China Hit Chinese Non-Supply Chains Even Harder.” PIIE Policy brief, Peterson Institute
- Office of the US Trade Representative. March 22, 2018. “Executive office of the President findings of the investigation into China’s acts, policies, and practices related to technology transfer, intellectual property, and innovation under section 301 of the trade act of 1974.” https://ustr.gov/sites/default/files/Section%20301%20FINAL.PDF
- World Bank, 2017. “Measuring and analyzing the impact of GVCs on economic development”. World Bank, Washington DC.
A longer version of this brief has been published in Russian by Republic: https://republic.ru/posts/92217
Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.
This brief finds that whereas in the 1980s richer women had fewer children than women near the middle of income distribution in the US, it is no longer true today. It argues that the rise in inequality is the main driver for this change. Greater income inequality enables high-income families to outsource household production to lower-income people. Changes to minimum wage laws are thus likely to affect the fertility and career decisions of the rich.
“I have frequently been questioned, especially by women, of how I could reconcile family life with a scientific career. Well, it has not been easy.”
– Marie Curie, 1867-1934
Much has been made of women “leaning in” at work at a cost to their families. Indeed, this discussion has become more prevalent as women have surpassed men in higher education in most developed countries, and have entered prestigious careers en masse, a fact reinforced by public policy. For example, in 2012 the European Commission published a special report on women in decision-making positions, suggesting legislation to achieve balanced representation of women and men on company boards. One natural question to ask is, how high is the cost of a woman’s career to her family? This is a difficult, multifaceted, and even sexist question to ask.
High-income women have historically had fewer kids (Figure 1 for the year 1980). Social scientists’ leading explanations rely on the difficulty of combining children and a career. Under this view of the world, as more women focus on their careers, they have fewer children. On the other hand, the evidence shows that more educated (or wealthier) women produce more educated children. Given these two regularities, the majority of children are born to poorer mothers, and thus receive an inferior education. Moreover, this creates a feedback loop that depresses the average education through time making us question our ability to sustain a satisfactory average level of education.
Figure 1. Fertility rates by income deciles, 1980 and 2010
Notes: Calculated using Census and American Community Survey Data. The sample is restricted to white, non-Hispanic married women. Fertility rates are hybrid fertility rates, constructed by age-specific deciles. Deciles are constructed using total household income.
However, the negative relationship between family income and fertility ceases to hold after the 2000s. Figure 1 shows that for the year 2010, the cross-sectional relationship between income and fertility has flattened or even become a U-shape. Today, high-income women have higher fertility rates than those of women near the middle of income distribution. This is a result of a substantial increase in fertility among women in the 9th and 10th decile of family income: they increased their fertility by 0.66 & 0.84 children, respectively. The rise in fertility of high-skilled females was first documented in Hazan and Zoabi (2015), discussed in a previous FREE Policy Brief. The implications are profound; children are more likely to be born to wealthier or more educated mothers than in the past. This has a far-reaching impact on the future composition of the population.
How can we understand the change in fertility patterns over time? We argue that rising wage inequality played an important role. Data for the years 1980 and 2010 show that average real hourly wages, quoted in 2010 $ grew from $28 ($51) to $50 ($64) for women (men) in the 10th decile of the income distribution. This increase was accompanied by stagnant wages for women (men) in the 1st decile, precisely the people who are most likely to provide services that substitute for household chores (Figure 2). Thus, growing wage inequality over the past three decades created both a group of women who can afford to buy services that help them raise their children, and a group who is willing to supply these services cheaply. In a recent paper, we found that the increase in wage inequality from 1980 and 2010 can actually explain the rise in high income fertility (Bar et al. 2017). Moreover, this rise in inequality has resulted in a large increase in college attendance through the changing patterns of fertility. This is because more children are now born to highly educated mothers.
Figure 2. Wives’ Wage by Income Decile 1980 & 2010
Notes: Calculated using Census and American Community Survey Data. The sample is restricted to white, non-Hispanic married men. Deciles are constructed age-by-age, using total household income. Representative wages for each decile is the average of these decile-specific wages from ages 25 to 50.
Our new understanding of the interrelation between income inequality, the relative cost of home production substitutes, fertility pattern and educational choice induces us to rethink some typical economic debates. For instance, consider the minimum wage. The typical debate about the minimum wage is focused on how it affects lower wage individuals in terms of income and their ability to find work. However, if people who earn the minimum wage are disproportionately also those who help raise wealthier families’ children, or simply make running a household easier, then a higher minimum wage can make home production substitutes more expensive for high wage women, making it harder for them to afford both a family and a career. While indirect, this effect can be significant. Figure 3 shows the distribution of the real wage, relative to the minimum wage, both for the industries of the economy associated with home production substitutes and other sectors of the economy. The figure clearly shows that workers in industries associated with home production substitutes are concentrated around the minimum wage and thus are much more likely to earn wages that are close to the minimum wage.
Figure 3. The distribution of real wages, relative to the effective real minimum wage in each state and year, by sector of the economy
Notes: Data from Current Population Survey, 1980–2010, using all workers.
Interestingly, we calculate a change in the cost of home production substitutes following an increase of the Federal minimum wage from $7.25 to $15/hour, as suggested by Bernie Sanders during the 2016 presidential election. It turns out that this increase in the minimum wage would increase the cost of market services that substitute for household chores by about 21.1%. Indeed, the minimum wage has a strong impact on the average wages of workers producing home production substitutes. However, how does this increase affect the economy?
According to our theory, higher costs of home production substitutes would affect women’s choice of how to allocate their time between labor force participation and home production, including raising children. The higher cost of these substitutes induces women to buy less of them and spend more of their time producing home production goods. Indeed, we find that the increase in the minimum wage decreases fertility and increases mothers’ time at home, and more so for higher income households. The magnitudes are large. A 10th (5th) decile household decreases fertility by 12.8% (9.4%), while the mother spends 9.7% (2.5%) more time at home. Notice that these numbers are calculated under the assumption that women can adjust fertility. What about those who are “locked in” their fertility choice? We recalculate changes in mother’s time at home for these mothers using the model’s fertility in 2010 with the increased cost of market services that substitute for household chores. A 10th decile mother increases time at home by 25.9%, while a 5th decile mother increases it by 13.1%. These numbers are larger as the family has not had a chance to scale back fertility. The short run effect on labor supply is also very large. The average reduction in labor supply by women in the 9th and 10th deciles is 3.5%.
Whether an increase in the minimum wage is good or bad for the society is a big question. Not only does it lie beyond the scope of our theory, but also beyond the scope of social sciences. However, the one modest contribution we try to make is in observing that an increase in the minimum wage heightens the rivalry between a woman’s career and family. As such, it forces women to forgo one in order to opt for the other.
The sexist nature of our question lay in the implicit assumption that it is the mother’s responsibility to look after the children or home production in general, rather than the father’s. While once this was a nearly universal attitude, it is now increasingly common for fathers to take a more central role in childcare rather than leave everything to the mother. How does this change in gender roles affect our analysis? In modern times, both spouses’ careers are potentially affected by children, as both parents take a role in child care. Fathers are now facing the same tradeoffs as mothers did in the traditional gender role story: children vs. careers. As a result, marketization is more important than ever for career oriented parents.
Talk to a high wage family and no doubt that they’ll readily tell you how important their ability to purchase daycare, prepared food, or other help at home is to their success as parents. Perhaps parents don’t realize that the price of these goods are so intricately linked to inequality or the minimum wage, but the policy maker should bear in mind that these are key factors for career women and the family.
- Hazan and Zoabi (2015), “Do Highly Educated Women Have Smaller Families” The Economic Journal
- Bar, Hazan, Leukhina, Weiss, and Zoabi (In progress) “Is the Market Pronatalist? Inequality, Differential Fertility, and Growth Revisited”
Governments often take unpopular measures. To minimize the political cost of such measures policy makers may strategically time them to coincide with other newsworthy events, which distract the media and the public. We test this hypothesis using data on the recurrent Israeli-Palestinian conflict. We show that Israeli attacks are more likely to be carried out when the U.S. news are expected to be dominated by important (non-Israel-related) events on the following day. In contrast, we find no evidence of strategic timing for Palestinian attacks.
The role of media in today’s conflicts is enormous. Parties to conflicts use propaganda in state-sponsored media and enroll state-sponsored trolls in social media to gain domestic public support for their military campaigns and, more generally, to raise own popularity. Involvement of Russia in Syria and Eastern Ukraine and its coverage on Russia-sponsored TV is a forceful illustration of this. Some most devastating conflicts used state media to enroll paramilitary. For example, Yanagizawa-Drott (2014) estimated that 51,000 perpetrators in Rwandan genocide were persuaded to participate in mass killings by RTLM radio.
Not all the media are under control of parties involved in conflicts. What is the role of independent media during conflicts? It is one thing to use the dependent media to portray one’s participation in conflict in a slanted manner; it is another to change one’s military strategy in order to improve one’s image in the independent media. Do military choose the timing and the weapon for their offences depending on the expectation of how their actions will be portrayed by the independent media? A statement on June 4, 2002, by Major General Moshe Ya’alon, then the Israel Defense Forces (IDF) chief of staff designate and until recently the defense minister of Israel, strongly suggests this is the case for the Israeli-Palestinian conflict. Mr. Ya’alon said: “This is first and foremost a war of ideology, and as such the media factor, the psychological impact of our actions, is critical. If we understand that a photograph of a tank speaks against us on CNN, we can take this into account in our decision as to whether or not to send in the tank. We schedule helicopter operations for after dark so they cannot be photographed easily. … Such considerations are already second nature to us. Officers … must understand that there are strategic media considerations. The tension between the need to destroy a particular building or to use a tank or helicopter, and the manner, in which the world perceives these actions, can affect the ultimate success or failure of the campaign. Even if we triumph in battle, we can lose in the media and consequently on the ideological plane.”
Our recent paper “Attack When the World Is Not Watching? U.S. News and the Israeli-Palestinian Conflict” (Durante and Zhuravskaya, 2017) forthcoming in the Journal of Political Economy investigates how Israeli military changes the planning of its operations in Gaza and the West Bank in the face of coverage by US media. In particular, we test whether Israeli authorities choose the timing of their attacks strategically to coincide with other newsworthy events so as to minimize the negative impact of their actions on U.S. public opinion by avoiding U.S. media coverage of their military operations, especially when they might lead to civilian casualties.
We compile a list of fully exogenous events from forward-looking political and sports calendars in the U.S. between 2001 and 2011 and verify which of these events actually dominate US TV news, leaving little or no time to coverage of other events. Then, we compare the timing of these events to the timing of Israeli attacks on a daily basis.
We also use another, more continuous measure of whether the U.S. media and the public are distracted by other important events, namely the length of top three non-conflict-related news stories during evening news on three U.S. TV networks, where the evening newscasts are limited to 30 minutes, namely ABC, CBS, and NBC. As Eisensee and Stromberg (2007) point out, due to the competition between networks for audience, we can measure the importance of newsworthy events featured on the evening broadcasts because more important stories appear before less important stories, and they are longer.
Timing of Israeli attacks and their coverage in US media
We find that both the incidence and the severity of Israeli attacks increased sharply when U.S. news were dominated by other events, such as US primaries and caucuses, general elections, and Presidential inaugurations. The probability that Israel carried out an attack against Palestinians rose to 53.2% one day before these important U.S. events from 38.7% on days that did not coincide with these events (over our observation period of 11 years, which includes heavy fighting during the Second Intifada). Figure 1 illustrates this finding. Attacks which coincide with the major political and sports events are also more deadly; as a consequence, the number of victims of Israeli attacks per day is 1.51 times higher during the days that coincide with major political and sports events compared to days that do not coincide with major events.
Figure 1. IDF attacks and exogenous predictable newsworthy events in the U.S.
Using another measure, the length of top three non-conflict-related news stories during evening news on three U.S. TV networks, we also find that Israeli attacks are significantly more likely to occur and are more deadly when top three non-conflict-related news are longer on the following day.
Does it matter which military operation?
As some military operations are more costly to postpone than others, one should expect that only attacks that are less costly to more be strategically timed to other important events. This is exactly what we find: the timing of special targeted-killing operations, which are considered as extremely urgent by IDF, is not related to U.S. news cycle. In addition, one should expect military operations to be timed to other newsworthy events only when they are likely to generate negative publicity. As negative publicity about the conflict is mainly associated with civilian casualties, and civilian casualties are more likely when the operations are executed with heavy weapons, we find that the relationship between occurrence and severity of Israeli attacks and U.S. newsworthy events on the following day holds only for operations that involve the use of heavy weapons. We also check that the attacks are only timed to predictable newsworthy events.
Why tomorrow’s coverage matters more?
Israeli attacks get news coverage in U.S. media both on the day of the attack and one day later. Why, then, Israel times its attacks to news pressure on the following day rather than on the same day? To answer this question, we analyzed the content of news broadcasts and found that the type of coverage of Israeli attacks differs substantially between same-day and next-day reports. While the same-day and next-day news stories are equally likely to report information on the number of victims, news stories that appear on the day after the attack are much more likely to present personal stories of civilian victims and include interviews with their relatives or friends. Furthermore, next-day coverage is significantly more likely to include emotionally charged visuals of burial processions and scenes of mourning. Anecdotal evidence suggests that it is both easier and safer for a foreign journalist to get details of the story on the next day; and that the next day affords an opportunity to produce emotionally charged videos of funerals. Figure 2 illustrates these findings.
Figure 2. Comparison of the content of news casts about attacks that aired on the same day as an attack and on the day following the attack.
Since people react more strongly to personal stories than to statistics and facts, and since information transmitted only through words is less likely to be retained than information accompanied by images, it is not surprising that Israel times its attacks to predictable international newsworthy events expected on the following day, as the next-day news stories are more damaging to Israel’s public image.
These results have broader implications. Policy makers in other policy domains and other countries may also strategically manipulate the timing of their unpopular actions to coincide with other important events that distract the mass media and the public. Examples of unpopular policies characterized by suspicious timing abound: Silvio Berlusconi’s government passed an emergency decree that freed hundreds of corrupt politicians on July 13, 1994, the day Italy qualified for the FIFA World Cup final. Russian troops stormed into Georgia on August 8, 2008, the opening day of the Beijing Summer Olympics. Political spin-doctors often release potentially harmful information in tandem with other important events. This is exemplified by a notorious statement from the former UK Labour Party’s spin doctor, Jo Moore, who, in a leaked memo sent to her superiors on the afternoon of 9/11, said that it was “a very good day to get out anything we want to bury” (see http://www.telegraph.co.uk/news/uknews/1358985/Sept-11-a-good-day-to-bury-bad-news.html (accessed on July 7, 2015) and http://www.theguardian.com/politics/2001/oct/10/uk.Whitehall (accessed on July 7, 2015)).
Overall, policy makers’ strategic behavior may undermine the effectiveness of mass media as a watchdog, thus reducing citizens’ ability to keep public officials accountable
- Durante, Ruben; and Ekaterina Zhuravskaya, 2017. “Attack When the World Is Not Watching? U.S. News and the Israeli-Palestinian Conflict”, Journal of Political Economy (forthcoming)
- Eisensee, Thomas; and David Stromberg, 2007. “News Droughts, News Floods, and U.S. Disaster Relief,” Quarterly Journal of Economics, 05, 122 (2), 693–728.
- Nevo, Baruch; and Shur Yael, 2003. The IDF and the press during hostilities, Jerusalem: The Jerusalem Democracy Institute, pp. 84-85, available at http://en.idi.org.il/media/1431355/IDFPress.pdf, accessed on May 18, 2016.
- Yanagizawa-Drott, David, 2014. “Propaganda and Conflict: Evidence from the Rwandan Genocide,” Quarterly Journal of Economics, 129(4), pp.1947-1994.