Week 2: Memos - Models of Innovation and Growth #6
Comments
|
Chapter 4 of The Middle-Income Trap sparked my interest in exploring the relationship between the prevalence of small firms in middle-income countries and size-dependent policies. In middle-income countries, most small firms do not grow, innovate, or exit the market, leading to stagnant productivity. For example, over 80% of firms in India, Mexico, and Peru employ fewer than five workers. size-dependent polices favor smaller firms (e.g., tax exemptions, subsidies, or reduced regulations) over successful firms, creating disincentives for firms to grow and become more productive. Net Benefit = P(S) - C(S) + PO(S) where PO(S) = { Benefit Amount if S ≤ T, 0 if S > T } This is a clear way of mathematically modeling why firms are incentivized to stay below the threshold T. These benefits incentivize firms to stay small, as the marginal advantages of remaining under the threshold often outweigh the profits of growing larger. Once a firm exceeds the threshold it loses these benefits, increasing effective costs and reducing the net benefit. I graphed the equation in Jupyter Notebook. The linear profit function reflects how revenue increases proportionally with firm size, while the quadratic cost function accounts for rising operational complexity and regulatory costs as firms grow larger. The conditional policy benefit provides a fixed amount for firms at or below a defined threshold, but this benefit disappears once the firm exceeds the threshold size. For the parameters, I set the firm size to range from 0 to 20 employees, with the threshold at 10 employees and a policy benefit amount of 50. The resulting graph demonstrates how these components interact and highlights how firms are incentivized to stay below the threshold to maximize their net benefit, thus explaining why there may be a relationship between size-dependent policies and smaller average firm size. |
The Impact of Political Connections on Innovation and Market Entry: An Empirical Analysis of the Chinese Technology SectorPolitical connections have significantly influenced the growth trajectories of firms in China's technology sector. A regression analysis is employed to empirically assess the impact of political ties on market dynamics and innovation from 2000 to 2020. Political connections can be a double-edged sword for firms. While they facilitate access to resources, financing, and favorable regulations, they may also diminish the incentive to innovate and hinder competition. Firms with political ties might rely on these connections instead of technological advancements or market competition. 
Where:
A significant proportion of large Chinese tech firms have executives with strong political ties, benefiting from preferential treatment in government policies, such as subsidies and favorable regulations. Political connections are more common in larger firms, with 40% of firms in the top 10% of market size having political connections.
Politically connected firms exhibit higher levels of patenting activity in the short run due to government-backed R&D funding. However, as these firms grow and establish their dominance, they tend to rely less on innovation and more on maintaining their market position through political ties. For example, Huawei initially focused on innovation but later shifted towards securing government contracts and regulatory advantages as it grew, leading to a slowdown in its R&D efforts relative to its competitors.
Politically connected firms benefit from subsidies and favorable tax policies, which provide them with an advantage over non-connected firms. However, this can lead to a focus on rent-seeking behavior rather than productive innovation. For instance, many firms that receive government subsidies invest in expanding their market share and reducing costs rather than improving their products or services. In the Chinese technology sector, political connections provide firms with significant advantages, including access to resources, financing, and regulatory leniency. The results from the regression analysis show that industries with more politically connected firms see lower entry rates, suggesting that political ties act as a barrier for new firms trying to innovate and compete. These findings underscore the trade-offs between short-term growth facilitated by political connections and long-term innovation that is hindered by the same connections. 
|
The Role of Mergers and Acquisitions (M&A) in Amplifying Non-Productive Strategies and Stifling InnovationMergers and acquisitions (M&A), particularly those of considerable size (e.g., $500M+), are often framed by C-suites and bankers as strategies to improve efficiency and consolidate resources. However, empirical evidence suggests that such large-scale M&A activities frequently signal a shift in focus from innovation-driven strategies to non-productive tactics aimed at preserving or enhancing market dominance. This behavior is consistent with the dynamics highlighted in Barriers to Creative Destruction, wherein dominant firms prioritize maintaining their position over fostering competition or innovation. A good example is the pharmaceutical industry is dominated by a few major players, including Pfizer, Johnson & Johnson, Merck, Novartis, and Roche. Over the past few decades, these firms have engaged in frequent large-scale M&A, often exceeding the $500M threshold. For instance:
Mechanism of Non-Productive Strategy Amplification
Proposed Formula: M&A and Non-Productive Strategy ImplementationTo measure the impact of M&A on a company's likelihood to implement non-productive strategies, a mathematical representation of this dynamic can be expressed as:
Where:
Hypotheses
Policy Implications
|
Patent and Innovation in the Era of AIThe Schumpeterian Growth Theory suggests that the value of innovation is equal to the profit flow divided by the risk-adjusted interest rate, where the risk lies in being displaced by a new innovator. According to this framework, patents can promote innovation by ensuring a relatively high profit flow for a specified period (by granting monopoly power over the commercialization of patented technologies) and reducing the risk of displacement by new entrants. 
However, alternative perspectives argue that patents may have minimal impact on innovation and could sometimes even hinder it. Historical evidence suggests that patents did not play a significant role in the invention boom during the Industrial Revolution in the United Kingdom. Many innovations occurred outside the purview of patent laws, relying instead on secrecy for protection. Furthermore, patent-induced litigation often protects incumbents, disproportionately affects late entrants, and may discourage innovation. In the era of AI, the role of patents in fostering innovation is uncertain. On one hand, the widespread availability of information and the open-source nature of many AI products leave less room for secrecy as a protection mechanism for emerging technologies. Research indicates that in sectors and countries where secrecy is less prevalent, innovation tends to rely more heavily on patents. This suggests that patents could play a more critical role in promoting innovation in this era. On the other hand, AI has dramatically increased research productivity and the speed of technological discovery. As a result, companies—both large and small—may arrive at similar or identical innovations within a short timeframe, complicating the fair distribution of patent rights. Moreover, controversies surrounding AI-related patent rules add to this complexity. For instance, the Biden administration’s Inventorship Guidance for AI-Assisted Inventions states that AI-associated technologies can be patented, but it must be proven that at least a human significantly contributed to the development process. This creates uncertainty about the scope of patent protection for certain technologies, potentially impacting how patents influence innovation. According to research article, two remedies have historically addressed issues within the patent system. The first is patent pools, which allow competing firms to combine their patents. The second is compulsory licensing, which enables competitors to produce patented inventions without the consent of the patent owners. Interestingly, data shows that patent pools slowed innovation during the textile era, while compulsory licensing has been shown to encourage innovation, particularly among new entrants, possibly by fostering greater market competition (see below for a graph). 
The evolution of patent laws globally remains uncertain, particularly in light of the challenges posed by AI. Policymakers should carefully weigh the positive impacts of patents on innovation against the potential challenges, especially those arising from the unique dynamics of AI. |
|
One of the Schumpeterian Growth Theory paper's topics is the effect of long-term tech waves on market and firm dynamics. It identifies three key results: 1) technological waves spur an increase in firm entry and exit as new firms innovate and obsolete firms exit the market, 2) they initially generate a productivity slowdown as resources are reallocated, and 3) they cause an initial increase in wage inequality, as workers who adapt quickly to new technologies receive higher wage premiums. These dynamics highlight the disruptive yet transformative nature of technological progress on the economy. I am particularly interested in the effect of new technology on productivity. At its inception, new technology often leads to a productivity slowdown due to resources and labor being diverted from normal productive activities into the learning and development required to integrate the technology. However, over the long term, this investment is expected to yield significant productivity gains. These gains arise from three primary sources: i) an increase in the quality of labor as workers acquire better education and skills tailored to the new technology, ii) an increase in capital stock driven by firms' investments in advanced machinery, computers, and infrastructure, and iii) an improvement in total factor productivity (TFP), which reflects enhanced efficiencies, innovation, and potential reductions in regulatory burdens. Traditionally, Productivity (𝐴) is measured using the Cobb-Douglas Production Function, which relates output (𝑌) to capital (𝐾) and labor (𝐿) inputs:
While this model captures the relationship between inputs and output effectively, it does not explicitly account for the nuanced effects of technological change. I propose an alternative measurement of productivity that incorporates the positive and negative shocks from new technology adoption:
Where: A = Productivity Hypothesis: delta_1 < 0: Re-education of workers diverts from normal activity This model captures both the short-term disruptions and the long-term benefits of technological waves. It emphasizes that while the initial stages of technology adoption can lead to inefficiencies and economic friction, the resulting improvements in labor quality, capital stock, and total factor productivity drive sustainable growth. |
In this week’s course material, I was piqued by the above graphs from Chapter 4 of the “World Development Report: the Middle-Income Trap.” In the first graph, we notice that there is a downward trend in percentages of firm entry rates in the United States while the second graph informs us that growth is driven by entrants in the United States. Even in the United States’ startup economy, the decline in firm entry rates over time demonstrates that the country is suffering from a decline in dynamism. In my memo I want to ask the question: why? What are the drivers of the United States’ decline in dynamism and how can the trend be reversed? What is the role of AI and industrial policies in reversing this trend? What data sets should we be looking at in order to better understand the nuances of this problem? One idea is that the decline in firm entry rate is related to a massive shift of innovative activity from large corporations, (as demonstrated in the figure B4.1.1 below), which in turn threatens small entrants. Another issue is the political connectivity of large firms. “Connecting to power” outlines that in industries with politically connected incumbents, there are fewer entrants by new firms. This is because small entrants face both productivity and bureaucratic challenges in the competition against established incumbents. In regards to AI, we know that its rise will cause massive displacements across the United States labor force. But AI also has the potential to reverse the US’ decline in dynamism by creating more space for entrants in the economy, fostering greater innovation between entrants and incumbents through creative destruction, and maximizing efficiency in productivity. However, some initial research has shown that if not harnessed with the right policies, there will be risks. For example, major tech giants already have control over the AI market in the United States: such as IBM and Google. Furthermore, the CEOs of OpenAI, Meta, Apple, and more have donated to the Trump inauguration, giving these incumbents more political power and increasing the relative regulatory barriers of entry for smaller firms. The United States Census Bureau demonstrates that AI usage is also more concentrated in larger firms, enhancing their productivity and efficiency and innovation of new products and services. These examples demonstrate how AI can be harmful towards the relationship between entrants and incumbents and stifle innovation. However, although the 2017 and 2018 censuses demonstrated that AI use in producing goods and services increased with firm size, this pattern might be changing with recent advances in generative AI. Researchers argue that generative AI might actually be able to help close the productivity gap between small and large firms, as it has a much greater marginal product of labor for smaller firms than large firms. 
Sources in addition to class readings and lectures: https://www.fastcompany.com/91257772/trump-inauguration-big-tech-donations-list-google-microsoft-meta-apple https://www.census.gov/newsroom/blogs/research-matters/2024/12/ai-use-small-businesses.html |
|
Hidden costs of SOEs A common argument used in favour of state owned enterprises is that it keeps prices low for everyone involved. This may certainly be the case in the short run where state owned enterprises may offer lower prices for their lack of need to generate a profit. The US Postal Service comes to mind as an example of a monopoly that offers relatively low rates while operating at a loss for extended periods of time. We can likely posit that in the long run it is likely that the lack of competition as well as the lack of incentives to improve efficiency can lead to stagnation. Furthermore, any benefits they bring in low prices can be said to be paid for by taxpayers through a variety of hidden costs. If a competitive, profit-seeking firm cannot match those prices, it is likely because it is unsustainable without some sort of support from the government which has its roots in tax dollars. I want to create a simple model that illustrates the effect of these hidden costs and how they lead to the persistence of state owned enterprises. Really, this argument can be generalized to any special interest group that by virtue of legislation/grants are able to keep competitors at bay be it airlines, mail, trucking, medical care, or subways. Let’s first consider the set of consumers: The firm imposes a cost For the sake of simplicity I incorporate the notion of “hidden costs” by applying Summing over the entire community we get: Which is equal to: Assuming that both parties are indifferent with regards to losses and gains, this shows that the obscurity of costs can have an impact on the perceived value of an issue. While exactly how this might be reflected in the political activities of the two parties is unclear, what is clear is that the upper bound imposed on the resources these two parties will deploy to either secure or prevent some sort of government sponsored advantage is influenced by the cost of getting informed. The model likely needs considerable refinement and additions to incorporate some of the more nuanced aspects of this procedure including the effect of lobbying on policy decisions, the interplay between consumers (such as incentives to free-ride), etc. It also makes some pretty critical assumptions, namely that the firm captures the full benefits of the costs imposed on the community which likely is not the case. Provided the deadweight losses arising from this issue are high enough, we can expect firms to gain significantly less than the burden they impose on consumers. The simple notion of the “enlightenment constant” nevertheless demonstrates how special interests have an advantage over dispersed interests and are able to leverage their informational advantage to rob the populace of its resources. |
A Regression Discontinuity Approach to the Effect of Immigration on Innovation in the USIn Chapter 4 of The Middle Income Trap, we learn that the US's status as an innovation powerhouse is at least partially a result of immigration. Sectors with a higher concentration of immigrants have experienced some of the largest growth since immigrant inventors generally have higher productivity than native-born inventors. I wanted to empirically test this theory, and I thought a good way to do this would be a regression discontinuity approach with the cutoff being the passage of the Immigration and Nationality Act of 1965, which abolished many US immigration restrictions and resulted in a wave of immigration to the US in the following years. If Chapter 4's argument is correct, we would expect a spike in patents (which act as a measure of innovation) following the passage of this new law. Using publicly available data from the US Patent Office on the total number of patents granted by year, I was able to run two regressions on the number of patents granted with respect to the years before 1965 and after 1965 with a lookback window of 30 years. This gave me a regression with 30 years of patents from the old policy and 30 years of patents with the new policy. Because the year the policy was passed is arbitrary, any difference just before 1965 and after 1965 gives us the treatment effect of increased immigrants on innovation. Doing so yields the following results: There is a clear discontinuity between the estimates before 1965 and after 1965, and looking at the counterfactual line, we can see that there is a clear spike in the number of patents granted. If we look at the size of the discontinuity (i.e., the treatment effect) based on the pre-treatment and post-treatment regression estimates for 1965, we get an increase of 9,070 patents thanks to the new immigration policy, which empirically supports Chapter 4's theory that immigration positively influenced innovation in the US. Of course, this assumes there were no covariates around 1965 that could have also led to the spike in patents. All code used can be found here. |
|
Political Lobbying and Harms to Creative Destruction in the Context of PG&E and California Wildfires Last week, I discussed how deep learning and AI could be utilized to detect and prevent future wildfires. With the wildfire devastations in Los Angeles, I am again thinking about how non-productive strategies, like political lobbying and connections, could have contributed to Californian wildfires in this week’s memo. The Schumpeterian Theory of creative destruction, as we learned about in the Aghion et al. text, emphasizes innovation’s role in driving economic and productivity growth by “destructing” outdated technologies. As seen in texts like the “Barriers to Creative Destruction” and “Connecting to Power” texts, many large firms rely on these non-productive strategies like garnering political connections, to maintain market dominance. PG&E, one of the largest utility companies in the U.S. that provides natural gas and electricity to millions of customers in California, has historically been at the center of California’s wildfire crisis. Experiencing lagging innovation and substantial improvements to their technologies, PG&E’s aging infrastructure has sparked catastrophic fires like the 2018 Camp Fire – which left 85 dead in California’s deadliest wildfire. Instead of investing in and proactively working on building innovative technologies, like underground power lines, AI-driven grid monitoring, or advanced weather forecasting technologies (which smaller firms and startups like Pano AI have been doing), PG&E has historically prioritized lobbying politicians for regulatory protections and liability limits. Having faced numerous lawsuits, like a $125 million settlement for the 2019 Kincade Fire, it makes sense that they need the support of politicians to sustain their company’s operations. 
(“Barrier to Creative Destruction” text, pg. 7) While lobbying provides immediate benefits (static benefits) like short-term cost savings and regulatory relief over many lawsuits, the long-term costs (dynamic losses) include higher wildfire risks, infrastructure failures, reduced public trust, and a multibillion-dollar bankruptcy filing in 2019. The “killer acquisition” analogy applies to PG&E here because their political lobbying efforts are actions to hinder competition from entering the market. This has led to slower market entry of smaller, innovative energy firms – a direct consequence of their non-productive strategies like we saw in Italy from the “Connecting to Power” text, where politically connected incumbents deterred market entrants and suppressed industry-wide innovation. Their focus on expanding market dominance reflects the idea about how politically connected firms display growth in revenue and size, but experience a stagnation in productivity. For the analytical element, I will interpret an equation from the Aghion et al. text (pg. 20) to illustrate PG&E’s resource allocation effects: 
Here, let Ze represent innovation contributed by new entrants (like Pano AI), Zi represent innovation contributed by incumbents, L represent the total resources available, and the rest of the variables representing parameters capturing efficiency, elasticity, and resource allocation dynamics. A higher Ze/(Ze + Zi) indicates that a greater share of productivity growth comes from new entrants. In PG&E’s case, smaller innovative firms like Pano AI represent Ze that focus on building wildfire prevention technologies. If PG&E allocates a large share of their total resources available (L) towards political lobbying, this would reduce Ze and suppress the potential for overall productivity growth. In other words, this could lead to higher wildfire threats and infrastructure failures in the future due to PG&E’s actions. |
|
Large foreign direct investment (FDI) projects, especially those valued at $500 million or more, are often praised for driving economic growth and bringing in cutting-edge technologies. By introducing both capital and managerial expertise, as well as connecting local businesses to global networks, FDI can help raise productivity and prepare domestic firms for international competition. However, recent research shows that when foreign investors gain majority ownership or exclusive partnerships, the initial benefits of FDI can shift from encouraging innovation to reducing competition. In such cases, foreign companies may rely on strategies that protect or expand their own market share at the expense of domestic firms, which can weaken the motivation for local innovation. This pattern reflects the issues highlighted in Barriers to Creative Destruction, where dominant market players—whether local or foreign—prioritize preserving their power instead of supporting vibrant, competitive markets. In some emerging economies, multinational companies leverage large patent portfolios, engage in lobbying for favorable regulations, or take advantage of easier access to infrastructure and financing to push aside local competitors. While foreign-owned firms can indeed boost initial economic indicators, they may also limit the opportunities for domestic businesses to develop new technologies and climb the “innovation ladder.” My research explores how foreign ownership influences innovation in local firms. I focus on whether a higher share of foreign equity leads to stronger research and development (R&D) efforts, or whether it simply encourages market consolidation. To do this, I will collect data on the percentage of foreign ownership in each firm, the level of market concentration across industries (using measures like the Herfindahl-Hirschman Index), and the availability of public support for R&D, such as subsidies or tax credits. The key question is whether stronger foreign participation enhances or undermines the domestic innovation process, and whether policy interventions—like targeted R&D incentives—can counter any negative effects of high foreign ownership. Ultimately, my goal is to provide insights for policymakers who want to ensure that large-scale FDI truly stimulates innovation. Such policies might include careful screening of major FDI deals so they do not create near-monopolies, along with training programs that strengthen the ability of domestic firms to learn from, and compete with, foreign investors. In addition, performance-based R&D incentives could help local firms stay committed to innovation, especially in industries characterized by a strong foreign presence. By balancing these interests, governments can make sure that FDI promotes sustainable economic growth and genuine technological progress. Simple Formula: Innova Intensity = R&D as a percentage of revenue |
|
How can AI help drive creative destruction and discourage non-productive strategy In the article “Barriers to Creative destruction: Large Firms and Non-Productive Strategies,” Baslandze analyses how incumbent firms uses non-productive strategies: Political connection, non-productive patenting and anti-competitive acquisition as barriers to creative destructions to maintain market power. With the increasing wide spread of AI in every industry, using Baslandze’s findings on why incumbent firms uses non-productive strategies and what they are, I want to highlight a few ways AI can help drive creative innovation and overcome non-productive strategies led by incumbent firms. The author highlights that one key factor driving incumbent firms toward non-productive strategies rather than innovation is that they are faced with greater potential losses from innovating compared to smaller firms. “The benefits from lower creative destruction obtained by means of patent protection are larger for market leaders simply because they have more to lose (14).” Any innovation that disrupts the market could reduce their profits significantly. As we discussed in class, the integration and widespread of AI can help help democratizes access to advanced tools for R&D, automate repetitive tasks, accelerating data analysis, and improving design processes. The cut in cost for innovating can incentivize incumbent firms to choose innovation over non-productive strategies. Simultaneously, it enables smaller firms to rethink the scale of their operations, increasing their likelihood of entering the market and competing with incumbents and lead to greater creative destruction. One real world example is in the health industry, we can already see clear evidence of how AI's contribution on R&D is reshaping the efficiency and cost to innovate. Another way AI might help with discouraging non-productive strategy can be through enabling quicker mass data analyses, and help policy maker and patent offices to pinpoint and regulate these non-productive strategies. For example, with Political connection, Baslandze highlighted that it is hard to observe connections between firms and politicians, and the strategies that have been taken is by matching employer-employee dataset to registered politicians. With AI automation and the ability to connect and read large amount of data quickly, more efficient and systematic analyses can be done through the datasets. This will be helpful for policy makers to detect patterns of favoritism, corruption, market trends and design effective policies accordingly. Similarly, as big firms uses patenting strategy to maintain their market power by filling numbers of patents, AI and machine learning can help quickly identify the non-productive patents to discourage mass licensing of patents, highlighting areas of excessive overlap and potential anti-competitive behavior. |
The Formula Industry is Lobbying to Prevent Creative DestructionThe theory of Creative Destruction emphasizes the need for innovation to drive productivity (and ultimately economic) growth. Unfortunately, market leaders are incentivized to protect their high market shares by resisting creative destruction. One avenue for doing so involves utilizing political power. Rather than continuously innovating, they use politics to prevent the rise of competitors. The infant formula industry is an example of this. The natural competitor to infant formula is breastfeeding. The formula industry is not innovating nor exiting the market if it cannot compete against breastfeeding. To protect themselves against breastfeeding, the formula industry is lobbying against policies that would benefit breastfeeding – namely against paid maternal leave. To understand the incentive of the formula industry to lobby against paid maternity leave, I calculated the loss in market share it would experience if a 12-week paid maternity leave were implemented in the United States. I assume the following: According to Allied Market Research, the U.S. baby infant formula market size was valued at $3,962.7 million in 2022. Table 1: Reductions in Breastfeeding if Paid Leave Were Implemented I calculated Reduced Formula Use = (Formula Use with 0 Weeks – Formula Use with 12 Weeks) / Formula Use with 0 Weeks. I calculated the market size of Abbot, Nestle, and Gerber using their market share. Then the market size of each subgroup using a 2:1 ratio between <6-month-old babies and >6-month-old-babies. Using the reduced formula used for each subgroup, I calculated the new market sizes. By adding the sub-market sizes, I found the new market sizes. Table 2: Calculation of Reduced Market Size The implementation of a 12-week paid leave would have resulted in a loss of $509 million for Abbot, $735 million for Mead Johnson, and $339 million for Nestle in 2022. In addition, the formula industry is projected to reach about $7,000 million by 2032, which leads to even greater losses for these companies. |
Trustbusters: An Antitrust Policy Case Study to Protect Future Innovation and GrowthWhat we knowThe focus of this set of readings revolves around this implementation difficulty of innovation related policy. While innovation is incentivized by the prize of monopoly rent, it is not the only way to achieve it. Plenty of anti-competitive practices accomplish the same, whether it be mergers and acquisitions, stifling newcomers, or political connections. Worse yet, without intervention, previously innovative firms may converge to these behaviors once they have received the rewards for their innovation. One policy is suggested in Chapter 4 of the World Bank report, and that is antitrust regulations. In order to better inform policy on clamping down on these anti-competitive practices, we might examine historical examples of antitrust policy and enforcement. Two notable examples of trusts are Standard Oil and AT&T, and the Sherman Antitrust Act that was used to bring them down. Standard Oil
AT&T
How do we proceed?The lesson we can learn from existing antitrust history in the US is that it is to somewhat effective at breaking up monopolies, and could definitely be improved. Policy makers should be aware of the “re-merging” effect and keep tabs on the company and the industry as a whole, even decades after the passage of the policy. If we wanted to take a closer look, we could try to use patent/research output from these companies, and compare that to figures before antitrust enforcement. This way we could in some way estimate how much antitrust policy is able to encourage competition and innovation. This would better inform future antitrust policy, and our effectively designed policy could strengthen innovative forces and increase economic growth greatly. Even today, the same Sherman Antitrust Act is being used in United States v. Google LLC (2020), (2023). While it has had an important role in ensuring competition, we would also like to adapt this policy to the modern day as our understanding of trusts and economic growth increases. |
Modeling the impact of foreign direct investment (FDI) on innovation in ChinaHow does foreign direct investment impact innovation in China? Previous literature suggests mixed results. FDI-driven competition can have contrasting impacts on innovation (Aghion et al., 2005). On one side, increased competition from foreign firms may push domestic companies to innovate more aggressively to maintain their competitive edge—an effect known as the escape-competition effect. On the other side, the presence of foreign competitors can shrink the market share of domestic firms, lowering their profitability and, consequently, their incentive to innovate. This is referred to as the business-stealing effect. A key milestone in China’s foreign investment policy framework is the passage of “Foreign Investment Law” (FIL) in March 2019. Since its accession to WTO in 2002, China has become the world’s top recipient of FDI, with over 2.1 trillion US dollars of accumulated FDI by the end of 2018. The FIL calls for a gradual elimination of caps on foreign ownership in key Chinese industries, marking an important shift toward a more unionized and liberalized policy stance on FDI injection. I propose that the enactment of FIL can be treated as an exogenous variable, given that it was not influenced by the behavior of domestic firms and that the changes were externally imposed. Hence, we can employ a Difference-in-Difference (DiD) approach to isolate the causal relationship between FDI and innovation in China by leveraging the 2019 FIL legislation as an instrumental variable. Two-stage regression model:The first stage regression estimates the relationship between the FIE legislation changes and FDI intensity
where:
Using the predicted FDI_Industry from the first stage, the second stage examines the impact of FDI on firm innovation:
Where:
Key Assumptions
|
Capital Misallocation by Venture Capital FundsIn Creation, the World Bank proposes that American start-ups face pressure from a “get in, get out” culture, and that value-add firms in the US attract investment to grow without government intervention that can lead to capital misallocation. This reasoning is also used to justify the high flow rate of small firms, as they either develop into middle-sized firms or exit the market; and ultimately, the authors recognize this trend as evidence of the selectivity of private markets in the US. Implicit in this rationale is that firm lifetime in the US is an indicator of value, as it must have attracted sufficient investment to survive. I believe that this implication is overly optimistic of early-stage investments in US private markets. Venture capital (VC) funding is instrumental to the development of start-ups to middle-sized firms, and yet VC is riddled with its own pressures that can contribute to investment into––and growth of––non-value-add companies, like the role of governments in middle-income countries. VC firms need to carefully identify investments within bubbles and face the competitive politics of leading series rounds. A misallocation of capital through VC in a non-value-add firm may only be realized in a later stage of a company––when its deficiencies grow larger than its persuasive power toward VC partners in winning investment. Suppose we define a “wasted deal” as a venture capital funding round for a company that eventually goes out of business by 2025––either from bankruptcy, defaulting, or liquidation. A majority of start-ups fail in the seed round, which is a healthy indicator of VC selectivity and inevitable capital misallocation to discover value-add firms. Instead, suppose we only consider wasted deal sizes of above $1M, when firms have escaped the earliest vetting round by VC funds. For each year and size of funding round, we can define the wasted deal ratio as the number of wasted deals divided by the number of total deals. I gathered data from Pitchbook on ~120,000 VC deals and funding sizes since 2009 and plotted the data:
One caveat is to understand the lag time for a wasted deal to be realized; for example, if a start-up raises in a series B round in 2014 but liquidates in 2018, it is still considered a wasted deal in 2014. This is why the downward sloping lines for each funding size may not represent improved vetting by VC funds or less misallocated capital, but instead that some subpar start-ups have yet to go out of business by 2025. A better indicator of improved capital allocation by VC funds is YoY slope; the steeper the line, the less the rate at which wasted deals are made, which can mean either i) start-ups are supporting and investing in fewer subpar companies, or ii) the lag time for a group of start-ups in that year is longer than before. But even in the case of ii), this can represent a marginally improved vetting process by VC funds, as companies taking longer to fail may have offered value-add at some point. This is yet another caveat, since classifying a failed business as a “wasted” deal ignores any short-term value they added during their lifetime. The magnitude of waste ratios are larger for smaller funding rounds, which is expected––the 1M - 4.9M phase is still early for a start-up to encounter market difficulties. What’s more interesting is the closeness of waste ratios for deal sizes of 5M - 9.9M and 10M - 24.9M, and 10M - 24.9M and 25M+. These represent mature start-ups that likely take longer to go out of business by token of funding received from the deal. However, apart from 2012 and 2015, each pair of consecutive funding sizes have a maximum absolute distance of ~0.025, which suggests that VC funds faced difficulties in identifying value-add start-ups once they attracted enough attention to begin large fundraising rounds. These failed firms also represent the most significant amount of wasted capital. This trend hints at a larger tendency in private markets for investors to follow the crowd, which will eventually amplify inefficiencies––and wasted capital––in start-ups that originally went unnoticed. |
|
Political connections play a significant role in shaping firm dynamics, particularly in industries where regulatory burdens create barriers to competition. While firms may leverage political ties to bypass these constraints, the long-term consequences on innovation and market efficiency are detrimental. This relationship can be expressed through the following equation: This equation illustrates how political connections negatively impact innovation , particularly for larger firms, older firms, and industries with high regulatory burdens. Larger firms often seek political connections to maintain market dominance rather than investing in new technologies, thereby reducing overall economic dynamism. 
The left graph supports this by showing the share of high-rank connected large firms in Italy from 1985 to 2014. A sharp decline in the early 1990s coincides with the Mani Pulite (Clean Hands) anti-corruption crackdown, suggesting temporary disruption in political-business ties. However, post-1995, the share of politically connected large firms surged, showing that firms rapidly rebuilt these connections to retain competitive advantages. Despite some fluctuations, political ties among large firms remained significantly higher than pre-1990 levels, reinforcing the persistence of rent-seeking behavior. The right graph compares the trend of political connections among firms in Germany, France, and Italy from 1990 to 2015. While Germany and France show a steady rise in politically connected firms, Italy follows a different trajectory—peaking in the early 2000s before experiencing a decline. This suggests that Italian firms relied heavily on political influence initially but later shifted strategies, possibly due to regulatory changes or diminishing returns from political ties. However, this decline does not necessarily indicate increased innovation; rather, it may reflect economic stagnation or firms struggling to adapt. Ultimately, political connections create an unequal playing field, where connected firms grow in size but reduce their innovation efforts. This weakens creative destruction, discourages new entrants, and results in lower productivity growth. While political ties may provide short-term advantages, they distort market competition and hinder long-term economic progress. Reducing regulatory complexity and limiting rent-seeking behavior are crucial steps toward fostering a more competitive and innovation-driven economy. |
Competition and Pharmaceutical InnovationOur readings this week highlighted the debate over whether competition fosters or stifles innovation, and which barriers might shape that outcome. Searching for a specific industry with unique characteristics—and thus potentially more nuanced reactions to competition and innovation—I found that the pharmaceutical industry stands out due to its high R&D costs, stricter regulations, and longer development cycles, driven by the critical need to ensure drugs are thoroughly proven safe and effective. In exploring the industry, I came across a particularly interesting article by Jung and Yoo. Rather than supporting the inverted-U relationship described in “What Do We Learn from Schumpeterian Growth Theory”—which states that “competition and productivity growth display an inverted-U relationship”—Jung and Yoo’s work suggests a positive competition–innovation connection, although it still underscores implications for inequality (see Aghion et al., 2013). Their study indicates that more intense competition correlates with more active R&D pipelines, effectively eliminating the notion that excessive competition necessarily undermines innovation. This view differs from the U-shaped theory because there is no “turning point” at which competition shifts from being productive to harmful. Jung and Yoo also examine moderating effects related to firm size and age, which supports the idea we have discussed at length: incumbents—especially larger, more established ones—tend to act as barriers to entry for smaller firms, thus limiting competition and, by extension, innovation. Their findings specifically highlight that overly dominant leading firms (particularly older, bigger ones) reduce the pro-innovation influence of competition, while smaller, younger firms respond more strongly to competitive pressures and hence drive innovation more effectively. These claims are reflected in the figures below. 
The moderating effect of the leading companies’ size on the relationship between competition measured by 1-HHI or the number of competitors and innovation activities measured by the number of drugs under development. 
The moderating effect of the age of leading companies on the relationship between competition and innovation activities. (a) 1-HHI is used as an indicator of competitive intensity, (b) the number of competitors is used as an indicator of competitive intensity. The attributes of the pharmaceutical industry and its market dynamics seem to favor an “escape competition” effect—where firms ramp up investment in innovation to gain an edge—over Schumpeterian theory, driving more potential for innovation. As a result, Jung & Yoo show that smaller players are more sensitive to competitive pressures, suggesting that targeted R&D incentives or anti-monopoly measures could amplify breakthroughs and spread those benefits more broadly. I would be curious to learn more about this divergence from Schumpeterian theory, particularly which industry characteristics (such as high R&D requirements and strict regulatory frameworks in pharmaceuticals) make it more resistant to the U-shaped relationship posited by many economists. It raises questions about how we might approach these industries—and their unique market dynamics—in a more nuanced and carefully tailored way, potentially leading to more effective strategies for innovation, competition and development overall. |
Malaysia, Once Again: Politics and the Middle Income TrapIn my previous memo, I talked about how Malaysia has made remarkable progress toward escaping the middle-income trap, but it still faces significant challenges in transitioning to a high-income economy. A major barrier is the role of political connections, which have led to inefficiencies and hinder innovation in businesses. Despite these challenges, Malaysia’s consistent efforts through targeted policies and investments show that progress is possible. Political Connections and Resource MisallocationPolitical patronage has long influenced Malaysia’s economy. Firms with strong political ties often benefit from easier access to financing, government contracts, and regulatory leniency. While these advantages can yield short-term gains, they frequently result in resource misallocation. Companies spend more time and resources maintaining political relationships than improving productivity or pursuing innovation. The New Economic Policy (NEP), implemented in 1971, is a clear example. It aimed to increase Bumiputera participation in the economy, but it unintentionally created “Ali Baba” arrangements. In these arrangements, Bumiputera individuals (Ali) acted as the official business owners, while non-Bumiputera partners (Baba) handled operations. This led to resources being allocated based on political or ethnic considerations rather than merit. A study of 834 Malaysian companies between 2000 and 2022 found that politically connected firms overinvest in less productive projects because of their access to preferential resources. This trend has reduced the efficiency of Malaysia’s resource allocation system. link Here, Malaysia’s performance in patents granted (Figure 9) remains far behind Taiwan and South Korea, emphasizing the inefficiencies from politically connected businesses. 
Impact on the Middle-Income TrapMalaysia’s reliance on political connections has contributed to this issue by limiting competition and discouraging innovation. When firms focus on political affiliations instead of market-driven strategies, they lose technological advancement and diversification, which are essential for economic growth. However, Malaysia has not been completely stymied by these challenges. As noted in last week’s memo, the World Bank recognizes Malaysia’s transformation into a key manufacturing exporter. Initiatives like the Economic Transformation Programme (ETP) have addressed some inefficiencies, emphasizing public transport improvements, high-value manufacturing, and innovation hubs. For example, industrial parks like Batu Kawan and infrastructure projects like the MRT system have enhanced Malaysia’s economic efficiency. link Export Sophistication and GDP (1980 vs. 2003): 
The broader implication is that government mismanagement has hindered Malaysia's economy, but other growth factors and strategic investments have gradually driven the country's progress. To overcome the middle-income trap, Malaysia must address the structural inefficiencies caused by political connections. Resources should be allocated based on merit rather than political ties. Recent high-growth sectors like semiconductors and AI require must prioritize innovation over short-term political gains. Malaysia’s globalist strategy under the 13th Malaysia Plan (13MP) shows its commitment to reform. Economy Minister Rafizi Ramli emphasized the importance of international partnerships and innovation in breaking free from the middle-income trap. Initiatives like Technology Park Malaysia (TPM) show how Malaysia can foster domestic innovation while attracting foreign investment. link Political connections have definitely hindered Malaysia’s economic growth by misallocating resources and slowing innovation. However, Malaysia has been adapting through strategic investments and policy reforms. I believe that by gradually shifting away from government ties, Malaysia can realistically achieve high-income status and (fully) escape the middle-income trap. *This is in response to Professor Akcigit's article "Connecting to Power: Political Connections, Innovations, and Firm Dynamics", where political ties is said to hinder growth and resources allocation. |
|
The process of creative destruction is fundamental to economic growth and innovation. By allowing outdated firms and practices to be replaced by new and more efficient competitors, creative destruction drives technological progress and market dynamism. However, this process can be influenced by various factors, including venture capital (VC) investment and government support. In class discussions and readings, we explored the role of political connections in hindering creative destruction. The question I posted for the readings was on whether VC firms are a help or hindrance to creative destruction. Thus, for my memo, I thought it would be interesting to dive deeper into this topic. |
|
Hypothesizing the Role of Market Dynamics and R&D in Patenting Strategies In today’s economy, the way companies approach patenting has a significant impact on competition and innovation. However, not all patenting is about driving progress. Some firms engage in non-productive patenting—using patents strategically to block competitors rather than protect real innovations—which can stifle industry growth and creativity. To explore what shapes these patenting strategies, I developed a regression model that examines how much emphasis a firm places on innovation-driven patenting (represented by α) in its approach. The proposed model is as follows:
Where:
For each of the regression coefficients, here are the hypotheses:
This model hypothesizes that firms with lower market share and higher R&D intensity are more likely to adopt patenting strategies driven by genuine innovation, whereas firms operating in rapidly growing industries may gravitate toward defensive practices. In order to test this model, you could use data from firms' patenting activity, innovation output, market share, industry characteristics, and R&D investment. It could provide empirical evidence for the relationship between firm characteristics, industry dynamics, and the extent of non-productive patenting, which might be valuable in shaping policies aimed at addressing the potential negative consequences of non-productive patenting on competition and innovation. |
|
Argentina provides an interesting case to examine the interplay between innovation, political connections, and economic growth, paralleling studies like "Connecting to Power: Political Connections, Innovation, and Firm Dynamics." Using Argentina-specific data on patents granted, economic freedom, and corruption perception, I analyze how these variables correlate and draw insights from Schumpeterian growth theory and the concept of creative destruction. From 2008 to 2011, patent approvals in Argentina remained constant at around 1,300 a year. At the same time, Argentina declined a bit in terms of economic freedom. Then, in 2012, patent approvals sank to 932, coinciding with a big drop in economic freedom that caused Argentina to be classified as a “repressed” nation for the first time. The correlation between these two variables weakened a bit over the next 4 years, during which patent approvals increased gradually while economic freedom decreased gradually. Corruption levels also worsened during this period, with several high-profile scandals involving Argentina’s president and state-sponsored businessmen. However, in 2017, when a new, more market-friendly administration took office, some regulations, such as foreign exchange market controls, were relaxed, and Argentina jumped back up to pre-2012 levels in the economic freedom index. This year also saw corruption decrease, and patent approvals surged to 2,301. Finally, from 2017 until the present, Argentina’s economic freedom and corruption rating have dipped slightly while patent approvals have fallen dramatically to a level almost as low as in the 2008-2011 period. The interventionist Peronists returned to power in 2019, exacerbating adverse economic conditions. All in all, throughout all these years of data, patent approvals, economic freedom, and corruption perception showed a notable relationship. The fluctuation in innovation activity reflected the broader economic volatility and regulatory landscape faced by Argentine firms, as well as the growth of politically connected companies that received billions in government contracts. High levels of regulation and weak property rights contributed to slowdowns in innovation by stifling competitive markets with policies that favored preserving government-protected industries instead of creative destruction. The barriers to creative destruction in Argentina bear similarities to those observed in Italy. High corruption levels and regulatory capture create a challenging environment for new entrants, discouraging competition and limiting resource reallocation to more innovative firms. This dynamic aligns with Schumpeterian growth theory, which emphasizes the necessity of competition for fostering creative destruction and driving technological advancement. In Argentina, however, entrenched incumbents often benefit from their connections, further stifling opportunities for innovation. The decline in patent grants after 2020 suggests a reduction in investment in innovation, potentially due to economic instability and the prioritization of rent-seeking behaviors by politically connected firms. This pattern mirrors findings in Italy between 1993 and 2014, where political connections offered firms short-term survival benefits but discouraged long-term innovation. The correlation between declining economic freedom and stagnation in innovation highlights the detrimental impact of excessive regulation on entrepreneurial activity and technological diffusion. Argentina’s experience greatly resembles Italy’s and reinforces the importance of limiting political interference and fostering market competition to achieve sustainable growth with creative destruction.
|
Cruise and Waymo -- application of the political economy of creative destructionIn October of 2023, a woman was hit and dragged by a Cruise AV in San Francisco, hospitalized with serious injuries. This event eventually led to the California DMV suspending Cruise's license to operate AVs, forcing Cruise to recall its entire fleet from San Francisco, and GM to stop funding the Cruise unit (Cruise was acquired by GM pre-crash for $1b). With Cruise and GM seeming to pull out entirely from the AV industry, Waymo, Google's AV unit, was essentially left as the last (developed, deployed, and fully autonomous) man standing. What's interesting in these details was that the CDMV's decision to revoke Cruise's operating license comes only in part from deeming the technology as unsafe -- the other part was believing "the company misrepresented the incident." (source here) While going into a data analysis of Cruise versus Waymo crashes might be helpful and lend a more nuanced view to the CDMV's decision, it might be summarized as such: Waymo has a higher crash rate than Cruise, but Waymo claims that higher rate is due to others hitting their stationary vehicles (~55% of the time), and from having more cars driving more miles than Cruise. Cruise has a higher injury rate from their crashes, at ~20%. In this sense, Cruise might have a more dangerous technology compared to Waymo, but still outperforms humans in comparable driving environments, here. I don't wish to claim that Waymo and Cruise were equals in technology or safety because I'm not certain that they were (and it seems that Waymo is generally considered better). Instead, I'm trying to look at the CDMV's decision from a political connections lens -- that Cruise and Waymo data seems somewhat comparable, and neither's technology was necessarily bad. At least one piece of evidence that might undermine this claim that both were in good standing (particularly because of the timing) is this incident that occurred in August 2023 where the DMV required Cruise to cut half of its operating fleet (roughly two months before this second decision to revoke the license entirely). In a report published by a law firm Cruise and GM had hired, it was "found that while the [Cruise / GM] executives had not intentionally misled state officials, they had failed to explain key details about the incident...The summary of the report was a long list of reasons to explain why regulators accused Cruise of misleading them." source here) One reason includes state officials claiming that Cruise did not share the entire video of the crash; another relates to the Cruise team not explaining the technology / decision-making of the car as it went through the incident. Therefore, it seems that the decision to revoke the operating license was somewhat personal -- one made off of principles of honesty instead of safety. A consequence that could have been potentially avoided had Cruise constructed better political relationships. Local officials (those in San Francisco) have been the most critical of AVs (especially Fire Departments) but they also have no control of AV activity -- that power lies completely in the California DMV and Public Utilities Commission. Since Jan 2021 to June 2023, Cruise spent ~$800k on lobbying in Sacramento (California capital). In that same period, Waymo spent ~$1.2m, with half of that coming in the months before an August 10th vote that gave the two companies 24/7 operating licenses. In 2022, Cruise gave $50k to support Gavin Newsom's reelection as governor, and it is believed that Cruise's state-level strategy was organized by one of the governor's closest associates. Cruise also has ties to the CPUC, as one of their former managing counsels now serves on the state utilities commission, having been appointed by Newsom. However, Waymo still outspent Cruise in total lobbying, and more recently in 2024 got Newsom to veto a bill they were against. It seems intellectually dishonest to make the claim that Cruise was shut down due solely to a lack of quality political relationships. But it doesn't seem incredibly intellectually dishonest to claim that the CDMV's decision to revoke Cruise's operating license had some human relationship / trust element to it. Either way, Waymo is now the only one standing (though there are new startups, notably Zoox and even Tesla), and it seems that some competitive environment that kept some innovating flame in the two companies has been extinguished. |
Incumbents and Cannibalization: Investigating when Incumbents InnovateIn "Barriers to Creative Destruction", Baslandze notes that incumbents often benefit less from innovation due to the risk of "cannibalizing the firms' own rents". Thus, they have an incentive to block creative destruction through the use of strategic patenting. Building upon the innovation technology model discussed in class, this phenomenon can be represented simply as follows for a given incumbent originally making monopoly profits: Given Standard Case (No action): If creative destruction occurs, the firm is pushed out of the market by competitors. Otherwise the firm makes standard profits with old technology Innovation Case (Firm chooses to innovate): Since new products are direct substitutes to existing ones, successful innovation with probability
Strategic Patenting Case: By obtaining patents with cost denoted by
Hence, opting for strategic patenting can be viewed as a form of insurance for the incumbent. Although innovating can potentially protect incumbents from externally occuring creative destruction, cannibalization acts as a ceiling against the profits the incumbent can earn. Thus, risk averse incumbents and incumbents operating in markets with high costs of research and development may see strategic patenting as an attractive option. Since cost of research is a function of Special Case: SoftwareSoftware is an area that I would like to examine more closely. Since software IP is traditionally governed by copyright law rather than patents (due to difficulty of meeting patent criteria), an industry specific model should either remove the option to strategically patent or significantly increase |
|
Corporate Venture Capital: Creation or Preservation? Chapter 4 of the World Development Report and many of the other readings this week highlight the significant tensions that incumbents face between maintaining the status quo and investing in innovation. At face value, corporate venture capital (CVC) arms seem to reconcile these tensions. CVC arms can theoretically produce win-win effects for both the incumbent and the entrant via matching of resources. Incumbents get exposure to new ideas, early access to future strategic options, and financial upside. Entrants get financing, professional guidance, and access to business networks. Depending on the degree of control embedded in the CVC investment, however, there can be significant harms to innovation such as the loss of intellectual property and reduced entrepreneurial freedom. To investigate the recent trends in the space, I collected and graphed Pitchbook data on CVC investments (see figure). Note that the “capital invested” metric aggregates the check sizes of each syndicated round, with CVC funds contributing only a portion of that amount. Clearly, the 2010s spurred a tremendous expansion in CVC capital deployment, followed by the industry-wide contraction in 2022. Given the degree of new deals funded by CVC arms it is clear that incumbents have recognized some form of strategic or financial returns from deploying capital to entrants. This capital allocation strategy is probably strictly preferred to directing funds to more explicit “preservation” strategies such as political connections, non-productive patenting, and anti-competitive acquisitions. Nevertheless, incumbents could easily direct their CVC investments to their internal R&D efforts, which may produce more fruitful returns. CVC investments could also stymie the disruptive nature of new entrants as they are forced into complementing their corporate investor or are eventually subsumed into the incumbent ecosystem. One particular dynamic that I find compelling is that the knowledge transfer from entrant to incumbent may actually cap the potential incentives for the entrant to innovate because the incumbent can maneuver their R&D organization to imitate and capture the entrant’s new ideas. CVC deals could thus be interpreted as hidden "killer acquisitions". These second-level impacts on innovation merit further research given the significant uplift in corporate venture capital dealmaking as a creation (or preservation?) strategy by incumbents. 
|
|
“Brazilian Miracle” - an exception to the Schumpeterian Growth Model? The "Brazilian Miracle" refers to the period of extraordinary economic growth experienced by Brazil between 1968 and 1974 under the authoritarian rule of the military regime that took power in 1964. During this period, annual GDP growth rates exceeded 10%, a remarkable achievement driven by pro-market policies designed by technocrats within the military government. These policies included trade liberalization, fiscal reforms, and measures to attract foreign direct investment, all of which were complemented by large-scale state-led infrastructure projects and industrialization efforts. Export diversification also played a crucial role, with favorable global conditions boosting manufacturing and agricultural exports. However, this rapid growth came at significant social costs. The regime’s centralized approach prioritized economic expansion over equity, leading to deepening income inequality, labor rights suppression, and a lack of attention to the broader welfare of the population. Furthermore, the growth model’s heavy reliance on foreign capital and loans made the economy vulnerable to external shocks, contributing to an eventual debt crisis in the following decades. The Schumpeterian Growth Theory argues that political involvement can foster economic growth to some extent but warns against excessive interference. Policies such as protectionism, subsidies, or authoritarian control, Schumpeter posits, can entrench incumbent firms and shield them from competition, stifling innovation and productivity. This interference disrupts the natural cycle of creative destruction, where outdated businesses are replaced by more efficient and innovative competitors, a process critical to sustained growth and technological progress. While state intervention can address market failures or support infrastructure development, Schumpeter emphasizes that sustained growth requires a careful balance to avoid the overreach of centralized power. Without this balance, the processes that fuels productivity gains and entrepreneurial innovation may be undermined, resulting in stagnation rather than dynamism. Using data from the Penn World Table, I analyzed the relationship between Total Factor Productivity (TFP) and Real GDP to examine Brazil's economic performance during the "Brazilian Miracle" period (1962–1980). The first graph compares TFP Level and Real GDP. TFP Level reflects the relative efficiency of Brazil's economy compared to the U.S. (used as a benchmark where the U.S. TFP level equals 1), while Real GDP measures the country’s total economic output in constant 2017 dollars. The second graph displays the TFP Growth Rate and Real GDP Growth Rate annually. TFP Growth Rate captures how efficiently inputs like labor and capital are utilized to produce output, while Real GDP Growth Rate reflects the year-over-year percentage change in total economic output, adjusted for inflation. The two graphs reveal that during the "Brazilian Miracle," both TFP and Real GDP experienced significant growth, though GDP growth was more consistent. The presence of TFP growth might initially appear to challenge Schumpeter's theory, which predicts that authoritarian regimes stifle innovation through excessive control. However, Brazil's case underscores a unique blend of factors that complicates this view. The military regime implemented a centralized, technocratic economic policy that combined state-led initiatives with market-oriented reforms. Infrastructure investments, export incentives, and policies to attract foreign investment drove rapid industrialization and productivity improvements, suggesting that, under specific conditions, state-driven strategies can promote productivity growth in structurally inefficient economies. The coexistence of state control and market mechanisms allowed for limited innovation and capacity-building, even within an authoritarian framework, positioning Brazil's experience as a partial exception to Schumpeter's emphasis on free-market competition and entrepreneurial dynamism. Nonetheless, Brazil's broader economic trajectory during and after the "miracle" aligns with Schumpeterian theory, reinforcing its long-term relevance. While the military regime achieved short-term productivity gains, the volatility of TFP and its decline in the late 1970s exposed the limitations of state-driven growth. The regime’s suppression of labor rights, concentration of benefits among elites, and reliance on capital accumulation rather than systemic innovation stifled the creative destruction necessary for sustained growth. Schumpeter’s warnings about the risks of protecting entrenched incumbents and the lack of entrepreneurial dynamism became evident as Brazil faced stagnation and a debt crisis. This outcome demonstrates that while directed industrial policies and central planning can yield temporary success, they do not foster the systemic innovation required for long-term productivity. Thus, Brazil's "economic miracle" serves as a partial exception to Schumpeterian theory, showcasing the potential of state-led growth in specific contexts but ultimately reinforcing the importance of competition and innovation for sustainable economic performance. 
|
|
Political connections influence the market by giving the incumbent a cost advantage: lower marginal cost, by removing regulatory taxes. Due to technological improvements, new entrants must create a good with a significantly higher cost advantage to replace the incumbent’s good. When the improvement is smaller than the value of the regulatory tax, the incumbent’s good is not replaced. Perhaps we live in a world where there will be an immediate public outcry for tax evasion due to political connections. Then we can still approach this problem from another angle and reach a similar conclusion. Instead of having the value of the tax as the reduction in marginal cost for an incumbent, we could argue that there are frictions in supplying goods that are overcome by bribing distributors. For example, say a distributor must decide between supplying Firm A or Firm B’s goods, and A is politically connected, but B is not. Firm A uses its political connections to bribe the supermarket with a one-time lump sum transfer in order to supply good A instead of good B. Further, bribes can be cheaper with greater political connections. The model is different because now, both face the same marginal costs, but whether a firm can get profits is determined by which good the supplier decides to supply. In this world, both firms are supplying the exact same good at time 0, with the same marginal revenue and marginal cost, and monopoly profits for whichever firm gets its goods distributed by the supplier (reflecting the idea of a dominating incumbent). Both firms have political connections, but one has stronger political connections than the other. Stronger political connections allow a firm to have their goods distributed at a lower bribery cost, because the politicians influence the decision of the supplier through some benefits/other measure (could be a range of things). The decision that the supplier makes in choosing between either firm can be modelled by an equation: Benefit from supply = bribe * political connection Thus, for firms that differ only by political connection, the firm with greater political connections can profit. In each time period, the incumbent firm risks being overtaken with probability p. Therefore, the cumulative profit that a firm gets from supplying its goods can be graphed as below:
Diminishing expected profits in each period reflects constant profits discounted by a higher probability of another firm being able to offer a higher bribe due to a technological improvement that allows them to offer a higher bribe, which eventually convinces the supplier to choose to distribute that other good instead. The maximum bribe a firm is willing to pay at time t is the area under the curve. This value is higher with greater profit margins created by technological increases. At time 0, when goods are identical, the firm with stronger political connections benefits by being able to offer a lower bribe than the firm with weaker connections and still having its goods distributed. Hence, they profit. The incentive to politically connect is no longer related to the ongoing tax but is a result of a desire for future profits decided at discrete time periods right before the distributor chooses whose goods to supply. |
|
Higher Education and Innovation: The Formal Qualifications of Global R&D Personnel Since 2000 Both “Connecting to Power” and “Barriers to Creative Destruction” explore the role of large firms in creative destruction. While some incumbents might adopt a pro-competitive model in support of creative destruction and innovation, other incumbents choose a more cost-efficient, anti-competitive model that drives entrants out of the economy. What is the role of higher education in innovation? On one hand, we think of higher education as the hub of innovation. Students who have been through academic training are equipped with various skills necessary for R&D research. On the other hand, aren’t universities also large, old organizations that establish themselves primarily through history, prestige, and even political connections? They seem to be potential candidates for the preservation force or dominant structure in innovation. This week, I began exploring datasets on the relationship between innovation and higher education. As my first step into the topic, I conducted basic data analyses, which yielded some initial graphs related to this topic. To be specific, I gathered data about R&D personnel by sector of performance and formal qualification from The OECD (Organisation for Economic Co-operation and Development). This dataset provides various statistics about around 60,000 individuals employed directly in the field of R&D from 2000 to 2022. I summarized the distribution of R&D Personnel by country in the graph below. I also provided a graph summarizing the distribution of R&D Personnel by Education level (excluding the education level denoting “Total,” which can mean any level of education and not informative). It is clear from the data that among all educational backgrounds, R&D personnel who obtained a doctoral degree or its equivalent form the largest group. To investigate this subgroup more closely, I plotted the percentage of doctoral-level personnel against time. I also created a regression line visualizing the trend. Here, we see a rising proportion of R&D personnel who obtain doctoral-level over time. The outliers in 2021 and 2022 are curious but might be explained by issues during data collection. In “Tapping into Talent: Coupling Education and Innovation Policies for Economic Growth,” Ufuk Akcigit et al. highlight the development of scarce talent and career choice as a crucial thread in understanding and possibly intervening in innovation and growth. Although this paper focuses directly on inventors instead of R&D personnel, my result resonates with one of the facts discovered in the sense that PhDs are more likely to engage in innovation compared to the average person in society. Based on its discoveries of the development of scarce talent, the paper proceeds with the discussion of the impact of R&D and Education policies on the development of innovators. I wonder if I could also start with an exploration of innovative talents (R&D personnel) and continue with more policy discussion. 
|
Scoring Entry Feasibility in Indonesia’s Biotechnology MarketI develop a potential model that assigns a creative destruction “score,” which can be used to determine whether the market conditions for new entrants are favorable. Here, new entrants are used as a proxy for creative destruction, as successful disruptors are a key part of that process. I apply information from The World Bank Group’s “Chapter 4: Creation” and Akcigit, Baslandze, and Lotti’s “Connecting to Power” to define model parameters. The The World Bank group claims that middle income countries have higher barriers to entry and are less likely to have effective creative destruction (hence the middle income trap). In these countries, incumbents are more likely to block entry and prevent innovation. So, in this case, we would expect something like InnovInt is the innovation intensity of an industry, which contributes positively to our score as more innovation should make it easier for a new entrant to gain market power. Based on data I could find, I set For competition, I will use Regulatory burden and political connections in an industry generally would have a negative effect on new entrants, hence the negative signs. Regulation may be quantified by the number of laws/policies for new entrants. For the PolConn score, we could find the ratio of total incumbents in an industry with political connections. In the future, I’d include a term that accounts for the interaction between RegBurden and PolCon. Akcigit's model shows that a firm with political connections allows it to mitigate regulatory red tape, and by combining these two terms, we can account for the combined effects. I was not able to find good data for these values, so I’ll leave these as placeholders. To expand on this model, I would first need to develop methods for RegBurden and PolConn values. Also, the score itself must be compared against other industries or countries to make sense of the scale. Still, this serves as one way to view the connections between innovation, growth, and competition with creation. |
|
After reviewing "Barriers to Creative Destruction: Large Firms and Non-Productive Strategies" and "Connecting to Power: Political Connections, Innovation, and Firm Dynamics," it's clear that both papers critique non-productive strategies like political alliance employed by large incumbent firms to protect themselves from creative destruction and that the process of creative destruction is hampered by the consolidation of capital and political power at the higher end of the firm size distribution. This lagging innovation as a result of political connectedness with large firms is demonstrated by the data collected in Italy pertaining to the political influence of large incumbent firms and correlated election effects on expected firm size. I wanted to explore this concept of political connectedness negatively affecting innovation across the global economy to see if this effect is generalizable internationally, beyond the data we've seen in Italy. I found data from the American Economics Association detailing the "Country Distribution of Firms with Political Connections" in a paper that similarly explores political connectedness of firms titled, "Politically Connected Firms." The data shows different countries' connected firms as percentage of market capitalization, which I combined with the most recent Real GDP Growth per Capita data to unveil any trends between productivity and political influence as a proportion of market capitalization across several countries. The trend line in the following graph suggests that countries with smaller capitalized representations of politically allied firms usually had higher Real GDP growth per capita based on GDP data ranging from 2022-2023. Therefore, the graph below suggests that the connection between political connectedness and lower innovation can roughly be generalized to global economies. This extension confirms that large incumbent firms representing significant shares of a country's market cap and exercising non-productive strategies like political alliance as opposed to R&D spend, decreases recent real GDP growth per capita, a proxy for recent advances in labor productivity. 
Sources: American Economic Review https://www.aeaweb.org/articles?id=10.1257/000282806776157704 |
Cultural Influences on InnovationFrom this week’s readings, two findings stood out to me. In “Barriers to Creative Destruction:Large Firms and Non-Productive Strategies,” Baslandze discusses the difficulty in measuring innovation using just patents due to the significance of unproductive patents as a barrier to creative destruction. I was critical of last week’s readings use of patents as a measure of innovation since there are are nuances when it comes to their effectiveness. I found it interesting to read more about novelty indices and patent-to-product ratio per product category, and I wonder what results would look like for conducting similar analyses in industries other than CPG, such as in Tech or Finance. Secondly, I was skeptical of the notion that “ideas are harder to find” (Jones, 2009; Gordon, 2016; Bloom et al., 2020) since “low-hanging fruits” have all been grabbed.” This struck me as a very presentist perspective that is biased that our current time period is special in terms of how likely innovations are to arise. Rather than interpreting declines in patent productivity as an indicator of less productive societies, I agree with Baslandze’s point that it implies other motivations coming into play, such as large firms seeking to bar new entrants. In class today, we discussed the sociology of innovation such as the impact of individualism on willingness to embrace technological change. Cultural influence is a component of lambda, the synergy between science and practical applications. I am curious if there is a way to quantify cultural influence, or study it more comprehensively than focusing on specific dimensions such as individualism. I’m interesting in exploring the relationship between intangible cultural influence and tangible innovations. One approach to doing so could be identifying more potential factors (other than individualism) that might impact willingness to embrace innovation, using data to uncover potential correlations, then extracting themes across multiple factors. I propose a potential multivariate regression to measure the relationship between cultural factors and openness to innovation: 
|
|
In this week’s lessons on Schumpeterian growth, we learned of a formal model that endogenizes technological progress into a system where economies grow through the creative destruction of incumbents and new entrants. Within the paper What Do We Learn From Schumpeterian Growth Theory?, we learn of the concept of Kondratieff cycles of long-term accelerations and slowdowns in economic growth that occur at times in an economy. This explanation provided for these waves is the general-purpose technology which offer tremendous benefits to the output and productivity of an economy in the long run. As a short-run drawback, the GPT also causes cyclical fluctuations since an economy must adjust to use these technologies which will cause restructuring and adjustments. This cycle can be represented in this figure. 
As shown, after the arrival of a GPT, the economy will begin an economic slump, as firms shift labor towards R&D and necessarily away from output in order to capitalize off of the discovery of a component that can utilize the GPT. Afterward, the economy will reach a boom phase where the GPT-enabled productivity increase drives greater output. In this memo, I would like to analyze inter-firm dynamics within each of these phases to understand how each slump and boom is decomposed in terms of actions from firms on the frontier and firms behind the curve. More specifically, I want to raise the idea that firms below the frontier of innovation may choose to shift greater labor resources into output and away from R&D in response to the action of frontier firms in diverting resources away from production. In this way, although the overall effect of the economy may be a slump in output, smaller, and less frontier firms, will increase their market share, while larger, frontier firms will decline in their market share. Firstly, I would like to describe the non-frontier decision-making scheme to research or to increase output. The value of research would be the gains from the discovery of component After a GPT technology is discovered and frontier firms shift their labor away from production and into R&D, the supply curve of the good will shift leftwards, causing price to increase. Breaking down the effect on the maximization problem, non-frontier firms will need to determine whether or not to change their labor decisions. From this point, the positive effect on labor output would be the increased From this analysis, we can claim that in the slump phase of Schumpeterian growth, some firm's may be better suited to dedicate more labor to increasing output rather than research and development. This will allow firms to capture a larger share of the market in the short-run which may be more beneficial compared to the cost of lower output from R&D due to lagging behind research-driven firms that can be first-movers in using GPT-compatible components. |
|
Memo: Barriers to Creative Destruction and Their Implications for Innovation To summarize the key themes of our lectures this week, the Schumpeterian concept of creative destruction emphasizes how innovation by new entrants disrupts existing markets, fostering economic growth. However, as outlined in Baslandze’s review, large firms often employ three common non-productive strategies, such as political connections, anti-competitive acquisitions, and non-productive patenting, to stifle this process (p. 2). These behaviors are particularly prevalent in heavily regulated environments like Italy (p. 4). By suppressing the reallocation of resources toward innovative entrants, such strategies hinder economic growth, contributing to declining business dynamism and stagnant productivity in advanced economies like the U.S. (p. 5). This reflects a troubling shift in incentives, where incumbents focus on protecting existing rents instead of fostering innovation, resulting in long-term economic inefficiencies (p. 7). However, as we transition to the digital era, the landscape of creative destruction has evolved significantly, with technology giants like Meta and Google playing a dominant role. These companies embody a duality: on one hand, they drive innovation through groundbreaking advancements in AI and digital platforms; on the other, they engage in practices such as frequent acquisitions of small, highly innovative SaaS companies. While these acquisitions can integrate new ideas into dominant platforms, boosting short-term innovation, they also contribute to market consolidation, reducing opportunities for smaller competitors and potentially stifling broader market dynamism. This duality makes it more difficult for policymakers to navigate the tradeoff between monopoly power and innovation - can a corporation balance both of these dynamics? These acquisitions can simultaneously exhibit elements of creative destruction, integrating new ideas into dominant platforms, and market consolidation that limits broader competitive dynamics. Economic Model of Innovation and Market Concentration: I(A,C)=αA−βC Where: Market Concentration Impact: Market concentration (MC) is positively related to acquisitions (A): MC=γA Where: Net Innovation Effect: The net effect on innovation (NI) accounts for both integration and competition: Where: Hypotheses for Value Interactions The relationship between acquisitions, market concentration, competition, and innovation highlights the delicate balance needed to maintain a dynamic and competitive market. For companies like Google, whose innovation drives progress but also consolidates power, it is clear that thoughtful policy decisions are essential to encourage creativity while preserving fair competition. |
|
Political Influence in Sustaining Unproductive Structures The Schumpeterian framework emphasizes growth and creative methods; thus, relying on political connections to grow is not conducive to development, as outlined in “Connecting to Power: Political Connections, Innovation” by Ufuk Akcigit, Salome Baslandze, and Francesca Lotti. An example of such a scenario not in the paper is the ZFM (Zona Franca de Manaus), a special area in the Brazilian Amazon, where extensive fiscal incentives are in place. The ZFM’s incentives were established to develop the region; however, they often maintain firms that lack competitiveness in national and global markets. They survive only because of the subsidies and the protective measures put in place to prevent foreign competition, thus the companies make sure through political pressure that they remain in place. Large firms, particularly multinationals, benefit disproportionately due to their legal teams’ ability to comply with the complex incentive regulations. Companies that wish to receive the benefit must comply with the Processo Produtivo Básico (PPB), a set of requirements that must be met and verified every time production processes change. Thus, due to the high costs and bureaucratic problems associated with renewing subsidy eligibility, most companies do not change their outdated production strategies. This discourages the entry of smaller, potentially more innovative firms, slowing the Schumpeterian process of creative destruction. One piece of evidence of such a lack of innovation is the productivity drop in the state where the zone is located relative to other Brazilian states. The following graph I produced shows how other states are improving their productivity, while firms in the ZFM are preserved through subsidies. I obtained data from the World Bank from 2003 to 2018 and focused on the manufacturing sector, as it is the one influenced the most by the subsidies: 
This other graph I got from another world bank report highlights the diminishing of the productivity of the state compared to the others in absolute terms: 
Also, the region is detached from the main consumer markets in Brazil and is surrounded by the Amazon forest. Firms spend money that could have been used for R&D on increased logistics and legal fees to obtain the subsidies. The profit of these companies is directly linked to good relations with the local government and the ability to incur extra costs; thus, productivity stagnates. A study performed by the World Bank, titled “Urban Competitiveness in Brazil’s State of Amazonas: A Green Growth Agenda,” shows that if we could hypothetically move the state of Amazonas (where the zone is located) to a region closer to the population centers, its GDP would increase by 27.2%. The problem is not only the existence of these benefits but also the political connections that ensure they are enshrined. These benefits are too expensive for local politicians to revoke, as the companies do not want to lose their advantages, and their employees want to maintain their jobs. These companies are also notorious for exerting heavy influence on the House and Senate, ensuring that import taxes remain high for everyone except them. |
|
Political Connections and the Innovation Trade-off Political connections allow firms to bypass regulatory hurdles, often giving them an edge in survival and growth. However, evidence suggests that these advantages come at the cost of innovation. In South Asia (developing world), for instance, data shows a high prevalence of expected "gifts" or bribes to secure government contracts. This raises a critical question: do political connections serve as a catalyst for efficiency, or do they entrench rent-seeking behavior, hindering broader economic progress? Firms in highly regulated industries are particularly inclined to seek connections, as these reduce the regulatory wedge (tau) and provide access to lucrative contracts. Yet, this strategy redirects resources away from innovation (R&D i.e. new products development) toward rent-seeking activities (e.g., bribes). Empirical evidence from South Asia highlights that firms in this region spend a disproportionately high share of resources on gifts or payments to navigate government contracts, exacerbating the trade-off between short-term efficiency and long-term dynamism. 
Model: Extended Profit Function for Politically Connected Firms To analyze this trade-off, I extend the baseline profit function of politically connected firms to explicitly incorporate both innovation investments and rent-seeking costs: 
This equation captures the trade-offs faced by connected firms. While reducing regulatory burdens (interaction of gamma and tau) enhances short-term profits, high rent-seeking costs (beta . bribery) and foregone innovation returns (alpha . R&D) limit long-term growth. The high level of bribery observed in South Asia (beta) suggests that political connections in this region may disproportionately favor rent-seeking over innovation. For example, data from the World Bank Enterprise Surveys indicate that firms in South Asia report a significantly higher likelihood of paying gifts to secure government contracts compared to firms in developed regions like Europe. This aligns with the model's prediction: in environments with high regulatory burdens (tau) and weak institutional frameworks, political connections may prioritize immediate gains, sidelining innovation. The returns to R&D denoted by alpha also vary across regions. In developed economies, firms with political connections often balance rent-seeking with substantial investments in innovation, resulting in higher productivity. In contrast, in developing economies, limited R&D intensity amplifies the negative impact of rent-seeking. Policy interventions should aim to recalibrate these trade-offs. Reducing the regulatory wedge in heavily burdened sectors could diminish the attractiveness of political connections. Simultaneously, incentivizing R& D through tax credits and penalizing excessive bribery could shift firms' focus toward innovation-driven growth. |
|
In “Connecting to Power”, the link between political connections, firm survival and innovation rate was established. A leadership paradox became clear – market leaders tended to be politically connected, leading to higher employment and revenue, but coming at the expense of innovation. One of these mechanisms that I (very reductively) understood was that they could use these political connections to maintain their market dominance through uncompetitive measures like raising the barriers to entry by restricting the ease of obtaining a license to operate etc. Without political connetions, the main way through which these incumbents would maintain their dominance would be through out-innovating their competition. Instead, they spend money that could have been used in R&D to instead lobby politicians to help them use government regulation to uphold their market dominance. This got me thinking, above and beyond having political connections per se, would it not also be imperative to look at the political situation of the country? The assumption underlying the analysis is that politicians who join the companies would necessarily be willing and able to help the companies enact anti competitive measures at the risk of being seen as corrupt. Hence, I thought it might be interesting to also look at other variables that might impact this mechanism linking political connections to the enactment of anti competitive measures. In essence, is there also a correlation between the political climate and innovation? Namely, I looked at the extent of political turmoil and the extent of corruption. For the former, my intuition was that the more political turmoil, the less the extent of correlation between political connections and innovation as even with political connections. If these connections can’t stay in power long enough, they might not be able to enact long-standing, impactful policies that effectively deter entrants. For the latter, I would think that the more corrupt a country is, the more likely it is for these political connections to enact anti competitive measures and hence dampen innovation. I chose to compare Singapore and Venezuela, countries that have notably different political situations. Here are the graphs I came up with using the Fragile State index and Perceived Corruption Index.
And here’s what I have for Corruption Perceptions Index.
Doing a really simple summary statistics, we can see that for the FSI, there seems to be an inverse relationship between the FSI and innovation. The more fragile a state becomes, the less innovation comes out of it. This makes intuitive sense, even if there are many politicians who are connected to the company, if the political situation is fragile, they might not be able to effectively enact the anticompetitive policies. The trend is less obvious for Singapore, perhaps because our political situation has not experienced much change in a while. For the corruption index, due to a lack of data, we’ll focus instead on comparing the general levels of innovation. Obviously, Singapore with it’s relatively much less corrupt government, enjoys a much higher rate of innovation than Venezuela, which consistently ranks in the bottom quartile of the corruption index. Of course, Sum Stats are often subject to noise and can be but spurious correlations, and so we must be cognizant not to conclude any causal relationship just yet. That said, there is definitely room to use these as variables in our regressions to confirm the robustness of our conclusions from the “Connecting to Power” paper. Apart from simply having politicians on board, what are the other factors (political instability, degree to which the politicians are willing to enact anti competitive measures) that we should factor into our regressions, to correctly isolate the causal relationship between political connection and innovation outcomes. Sources: |
|
Through Barriers to Creative Destruction: Large Firms and Non-Productive Strategies, we are presented with the lack of progress that is driven by incumbent’s fear of being replaced. There are various strategies put into place from using patents as an overbearing defense mechanism, using political connections to abuse power, and even acquiring potential rival firms: nipping any competition in the bud. |
Amakudari: The Employment of Elite Bureaucrats in JapanIn “Political Connections, Innovation, and Firm Dynamics,” the authors argue that politically connected firms often grow larger and generate higher revenue but invest less in innovation and productivity. These firms use their connections to maintain a quasi-monopolistic position, which reduces competition and stifles creative destruction at the aggregate level. Political connections provide advantages such as easier access to permits, reduced regulatory oversight, and preferential contracting. In Japan, the practice of amakudari (directly translated as “descent from heaven”) exemplifies these dynamics. This term refers to retired high-ranking bureaucrats leveraging their government connections to secure positions in private firms, often in industries like public works and construction. While it can be argued that these bureaucrats bring valuable expertise from years in ministries like the Ministry of Finance or the Bank of Japan , their primary role is often to act as a communication channel between firms and the government. This makes amakudari executives an effective but controversial means of bypassing bureaucratic obstacles. Japan’s bureaucratic structure fosters this practice. Promotions within the system follow a rigid, tournament-like progression. Bureaucrats typically face mandatory retirement in their early 50s as opportunities for advancement diminish. Ministries often arrange post-retirement jobs for officials as a form of reward, reinforcing government–industry ties and consolidating influence. These bureaucrats are frequently hired by companies with prior connections to their ministries. For example, data from the Internal Affairs and Communications Ministry shows that nearly 70% of the 1,968 bureaucrats who secured post-retirement jobs from 2004 to 2006 were employed by firms linked to their former ministries. A notable case involves the Japan Green Resources Agency, where two retired bureaucrats were implicated in bid-rigging for public works projects. The agency also employed 256 former officials, most of whom came from the Forestry Agency. While such connections enable firms to secure lucrative public projects, they distort market competition. Companies may offer excessive compensation packages to attract bureaucrats, effectively functioning as a form of bribery. This undermines market efficiency, misallocates resources, and discourages innovation. Moreover, these practices weaken the competitive landscape, allowing less efficient firms to dominate. To better understand the decision-making dynamics behind these practices, I propose a simple equation that reflects how ministries evaluate project allocation. The decision involves comparing the value each firm provides and awarding the contract to the firm with the highest net value: 
• x represents the societal gain if the project is successfully completed. The government attempts to balance receiving a high bribe (the compensation for the ex-official) against minimizing the loss of societal value by avoiding overly inefficient firms. This model illustrates that even a highly inefficient firm can win contracts if their bribe is large enough. |
Memo: Contract Institutions, Creative Destruction, and Robot Adoption in ChinaThis memo integrates Schumpeterian growth theory with contract institutions to propose a framework for understanding how contract incompleteness affects robot adoption and creative destruction in China's manufacturing sector. By linking institutional quality, technological innovation, and productivity growth, the framework provides insights into the mechanisms underlying technological upgrading. Building on the Schumpeterian framework of creative destruction and incorporating contract institutions, I propose: Where:
This equation combines: The classic Schumpeterian growth equation ( This model helps explain key empirical patterns observed in China's manufacturing sector. First, it accounts for the inverted-U relationship between competition and innovation, as industries with moderate competition foster optimal conditions for creative destruction. Second, it explains why robot adoption is more prevalent in industries with complex production processes and weak contract enforcement; in such settings, firms turn to automation as a substitute for institutional reliability. Finally, the framework emphasizes the critical role of creative destruction in driving technological progress, illustrating how firms with robust institutional support more effectively implement disruptive technologies. Chinese manufacturing firms have been at the forefront of global robot adoption, accounting for 52.48% of installations in 2022 according to the IFR. However, this trend reveals significant regional disparities driven by differences in contract enforcement quality. In coastal provinces such as Jiangsu and Zhejiang, stronger institutional frameworks enable firms to rely on external contractors, reducing the necessity for automation. In contrast, interior provinces like Sichuan and Gansu, where contract enforcement is weaker, experience higher rates of robot adoption. This variation reflects a strategic response by firms in weaker institutional environments, where automation serves as a substitute for unreliable contractual arrangements, particularly in industries with high production complexity. The impact of contract enforcement on robot adoption is further evident in sectoral differences. In high-complexity industries, such as electronics and automotive manufacturing, weak institutions impose higher transaction costs, resulting in a stronger incentive to adopt robots. Empirical data shows that weak enforcement quality significantly reduces productivity growth in these sectors, as these industries rely less on external contracting and are less sensitive to institutional quality. Over time, Chinese firms in regions with weaker contract enforcement have consistently adopted robots at higher rates, particularly in industries with greater production complexity. This trend underscores how institutional weaknesses shape firm behavior, driving technological adoption as a way to internalize production processes and mitigate risks associated with external contracting. For example, during the early 2000s, industries such as automotive manufacturing saw a sharp rise in robot adoption in regions with weak contract enforcement, where firms faced mounting challenges from unreliable external partners. The case of Chinese manufacturing underscores the broader implications of institutional quality on technological adoption. Weak contract enforcement drives firms to adopt robots as both a tool for improving efficiency and a strategic response to institutional shortcomings. Strengthening contract enforcement could reduce firms’ reliance on automation as a substitute for external contracting, fostering more balanced growth across regions. |
|
Salome Baslandze’s paper draws specific attention to how large firms leverage their use of non-productive strategies to increase their market share, increase their profitability, and maintain their ranks as the unstoppable giants within their particular industry. For my memo, I want to take a deeper dive into one of the large firms belonging to one of the most prevalent figures in the United States, and the world, today, and that is none other than Elon Musk’s Tesla. Particularly, I will take a look at recent trends in patenting activity within Tesla, and determine whether this aligns with Baslandze’s arguments suggesting decreasing innovation stemming from anti-competitive decisions. In June of 2014, Tesla announced that it would open its patents to the public. This move was initially highly criticized by Tesla’s own employees, as their mentalities aligned with that of the ruthless giant seeking to protect its own technologies and innovations against competitors. This move seems to contradict the intentions of large firms illustrated in Baslandze’s paper, however, upon further research, it was determined that Tesla in fact did have some alternative anti-competitive strategies up their sleeve after all. Tesla announced that they would not initiate a lawsuit for those who wanted to use their technology, however this only applied to about 14% of Tesla’s total patent portfolio, and furthermore, Tesla did not continue to update the public’s access to new patents created after opening patents to the public, as seen in the graphs below highlighting the differences between patents accessible to the public, and total patents in Tesla’s portfolio. Of particular importance to note is the increase in patents (and, one could argue, innovation) seen in the 4 years immediately after 2014. Thus, what initially appeared to be a move to propel industry-wide innovation, and thus promote creative destruction, actually has resulted in Tesla’s own profitability as they have leveraged their position as the “good guys” in the industry in order to advance their own innovations behind the scenes without making these newer innovations accessible to the public. In essence, Tesla said, “Here are our patents, have fun with them, but we will keep advancing technologies as well, and most likely, to a better degree than you.” Not surprisingly, Tesla has continued to grow to be the world’s leading producer of electric vehicles. To put Tesla’s patent status and innovative capabilities in perspective, the last graph below illustrates Tesla’s patent portfolio valuations, which amounts to roughly 4 times as much compared to the industry’s average. Now, Tesla stands in prime position to improve their market dominance given Musk’s continued allegiance to president-elect Donald Trump. While Baslandze’s paper highlights Ufuk’s analysis on regional political connections with large firms leading to decreased innovation, research on this front can now extend to the next four years when the CEO of Tesla now utilizes his political connection at the national level to his advantage. Will the trend of increasing patents and innovation continue, or will we finally see stagnation thanks to his relationship with Trump? 
|
|
Innovation as the Key to Escaping the Middle-Income Trap The Innovation Challenge Case in Point: The Republic of Korea The Innovation-Growth Link Y(t)=A(t) * K(t)^α * L(t)^(1−α) Here: We assume productivity grows exponentially with innovation: A(t)=A_0 * e^(γt) Where: This equation tells us that for a country to keep growing, it needs to continuously improve productivity through innovation (γ). For middle-income countries, boosting innovation through better policies, more R&D, and stronger global connections is key to reaching high-income status. The chapter makes it clear that countries aiming to escape the middle-income trap need to prioritize innovation. It’s not just about making small improvements but about creating an environment where big, transformative ideas can thrive. Korea’s experience shows how a deliberate focus on innovation can change a country’s future, offering valuable lessons for others stuck in the same spot. |
|
Market structure and firm dynamics play an integral role in innovation. In this memo, I review the relevant literature and offer an argument based on available data. Aghion et al (2014) describe a Schumpeterian growth model in which, unlike the quality ladder model, challengers proceed to the frontiers occupied by incumbents in a non-instantaneous manner—step by step instead of “leapfrogging.” As a result, R&D investments made by these firms essentially pose competition to incumbents. This challenges the simple view of incumbents as necessarily obstacles to innovation that must be overthrown via creative innovation to bring about growth. Of course, incumbents can also profit through rent-seeking behaviors like acquiring potential competitors to avoid changes to the market or fostering political connections that allow them to avoid regulations that are imposed on their peers, see Akcigit, Baslandze, and Lotti (2023) for further discussion and empirical evidence of these latter behaviors. A key policy question is the optimal antitrust enforcement for growth and innovation. Using dynamic modeling, Cavenaile et al (2021) and Mermenstein et al (2021) find support for stricter enforcement in the aggregate. Mermenstein argues that the expectation of stricter enforcement may lead to less anticompetitive acquisition practices or stalled R&D investment, while Cavenaile calls for different antitrust guidelines that would capture more cases in which innovation is stalled. However, as Baslandze (2023) notes, this is an empirical question where courts and policymakers must consider the aggregate implications of both the synergistic effect of mergers and the potential of anti-competitive and anti-innovation acquisitions. The empirical evidence is mixed. Bena and Li (2014) study exogenous variation in the success of merger bids to infer that firms with significant technological overlaps are more likely to be in an M&A, and when they do, they increase output. On the other hand, Cunningham et al (2021) find an opposite effect in the pharmaceutical industry, where larger incumbent firms acquire smaller firms on the verge of new innovations. Outside of the research Baslandze compiles, it is worth noting other studies find that antitrust enforcement causally results in greater economic output, as per Babina et al (2023). See below Figure 2 of that paper, which shows the treatment effect estimated via two-stage DID before and after enforcement on business formation. Their overall result is that hiring and expansion is improved by antitrust enforcement, which provides plausible causal evidence for stricter enforcement. 
A possible continuation in this direction of research would be to decompose their analysis into the types of antitrust actions taken by courts to determine what the sources of innovation and anticompetitiveness are in terms of types of antitrust complaints. This could in turn help shape the details of future merger guidelines or the economic arguments made in antitrust cases. |
|
In “Connecting to Power: Political Connections, Innovation, and Firm Dynamics” Ufuk Akcigit, Salome Baslandze, and Francesca Lotti, we learn that firm level political connections are widespread, especially among large firms, and that industries with large share of politically connected firms are likely to innovate. I propose that the innovation outcome of a firm i in country c at time t could be modeled as: Where:
Hypotheses: Prevalence of family ties and a social network with politicians, and a healthy corruption score are positively connected with growth in size, and associated negatively with future growth in productivity measures. I am also curious if other variables could be added to analyze firm growth:
And if we can analyze the effect of political connectivity on:
|
|
In Barriers to Creative Destruction: Large Firms and Non-Productive Strategies, the argument is made that overtime, the patents firms apply for become increasingly similar as their speed of innovation slows down: “with the help of textual analysis and machine learning tools, patent applications that are too similar to their predecessors could be singled out and their necessity and applicability could be scrutinized” (15). It would be interesting to use NLP to analyze the similarity between different patents over time. This is empirically very feasible because all patent applications are public. Below is an illustration of the basic process that would occur in order to analyze similarity between different texts. First, the texts need to be tokenized and vectorized. Then, generate world and sentence embeddings in order to capture semantic meaning. Lastly, run a similarity model on the texts using Jaccard, Levenshtein, or cosine distance. I do not have sufficient understanding of NLP to know which type of distance is most suitable in this instance, but all of them should give a score between 0 and 1, loosely understood as percent similarity.
The graph is meant to illustrate that the firm’s age and similarity scores (of the last n patents, however many one should choose to analyze) should be positively correlated. The relationship could be exponential, linear, or quadratic, which we can interpret as the rate at which innovation slows down. In theory, the older firms will produce more similar patents than younger firms. In discussing this with a friend who wants to work in patent law, she informed me that oftentimes a company’s new patent will refer to previous patents rather than restating what was previously written. To what extent this would affect the similarity score is unclear. Along these lines, another way of testing for originality would be counting the number of times the new patent refers to previous patents, with less references being more originality. The other thing to consider is how to select which patent to compare. Is it the last two patents? Or could we compare patents from the same company overtime? For example, we could take four patents from Meta—2010 vs. 2015 and 2019 vs. 2024. The gap of five years is arbitrary—I am not sure what is the optimal number of years to compare the difference. The idea is that the first pair should have a lower similarity score than the second pair, or the second pair references more previous patents than the first two pairs. Doing this across a set of 1000 firms over 40+ years old should generate a pretty solid picture of how innovation in these companies are slowing across time. |
|
Political Connections, Non-Productive Strategies, and Creative Destruction: Implications for Growth and Policy Role of Political Connections: Politically connected firms grow in size and revenue but exhibit lower labor productivity and innovation intensity (Akcigit et al.). Beyond political connections, large firms utilize non-productive patenting and acquisitions to deter competitors. Both papers emphasize the dynamic costs of these strategies, showing how they reduce the pace of creative destruction and aggregate productivity growth. Key Observation: Firms with higher market ranks (e.g., leaders) exhibit lower innovation intensity but greater reliance on political connections. This inverse relationship underscores the shift from productive to non-productive strategies as firms grow dominant. Policy Implications Transparency in Political Connections: |
The Implications of Innovative Efficiency on Patent Reform and Market StructureIn "Barriers to Creative Destruction: Large Firms and Non-Productive Strategies", Baslandze finds that large incumbent firms engage extensively in non-productive patenting to raise entry costs for competitors, thus hindering creative destruction. This phenomenon offers an alternative explanation for why TFP growth has stalled since the 1980s despite a massive surge in patenting, contrasting with Bloom et al's thesis that "low-hanging fruit" have been picked and ideas are becoming harder to find. While Baslandze's analysis of non-productive patenting practices is compelling, it doesn't necessarily preclude decreasing research productivity. Even if Bloom's study didn't fully account for non-productive strategies, Matt Clancy's preliminary analysis suggests that science itself is getting harder, using metrics relatively insulated from corporate activity like Nobel prizes and top-cited papers. Given the relationship between scientific advancement and innovation, this supports the notion that at least some productivity stagnation stems from declining research efficiency. During class, Professor Akcigit suggested considering patent reform in light of AI tools potentially increasing innovation efficiency, which could reduce justification for prolonged monopoly rents. This presents an interesting counterpoint to the declining innovation efficiency documented in the above research. To explore this, I used the Schumpeterian growth model we learned in class to investigate how revenues and costs change given either increasing or decreasing innovation efficiency parameters (lambda). Here is my simulation graph:
The results are interesting—changes in innovation efficiency lead to markedly different profit trajectories, supporting Professor Akcigit's argument that extended monopoly profits become harder to justify as innovation becomes more efficient. However, under decreasing efficiency, profits remain relatively stagnant. Under this world, the arguable merits of the current patent system might be maintained. For instance, even with pharmaceuticals, where patenting practices face legitimate criticism, Ji and Rogers found that Medicare's regulated price cuts led to a 29% decline in new product introductions, suggesting monopoly rents play a role in enabling innovation. The second graph is particularly interesting—costs rise substantially even as revenues increase. This is particularly noteworthy because some economic models suggest that higher entry costs reduce the free-entry equilibrium number of firms, as each firm must capture greater market share to recover costs.
Assuming that this model is applicable, this suggests that in a world of increasing innovative efficiency (or even just constant innovative efficiency), patent reform alone might not advance creative destruction, as market structures may naturally evolve toward oligopolies or monopolies. This implies that we may need more innovative policies to enable new entrants to continue creative destruction cycle (while being careful to avoid growth-inhibiting preferential policies). |
|
The pharmaceutical industry is placed in the crossroad between innovation and regulation. Heavily investing in R&D can yield groundbreaking therapies that improve public health outcomes, yet companies often also choose to invest heavily in lobbying to shape regulatory processes and gain market advantages. The work by Baslandze (2023) and Akcigit (2023) highlights this dilemma of “political-tiedness,” where firms shift resources from genuine R&D toward lobbying to secure favorable policy decisions. The scatterplot below is based on data from pharmaceutical firms that have spent more than $1,000,000 lobbying3 in the U.S. last year, and it illustrates a moderate correlation (R² = 0.47) between total lobbying expenditures in 2024 and FDA drug approvals between 2021 and 2024. While higher lobbying spending appears linked to more approvals, this pattern raises questions about whether such approvals represent transformative innovations or primarily reflect political leverage. Creative destruction suggests that entrenched players might use political connections to maintain dominance, thereby discouraging smaller firms from pursuing new treatments. Indeed, these findings echo Akcigit’s “leadership paradox,” where established incumbents can outmaneuver genuinely innovative challengers. A related concern is regulatory capture wherein regulatory agencies serve the interests of powerful industry actors. If lobbying funds simply accelerate approval for “me-too” drugs rather than groundbreaking therapies, then the public risks missing out on real medical advances. A more detailed examination of each approved drug’s therapeutic impact—perhaps by comparing clinical trial outcomes or novelty in patent filings—could help clarify whether lobbying expenditures drive genuine breakthroughs. These results pose important questions about how to balance lobbying activities with the pursuit of valuable innovations. Policymakers might consider reforms that ensure the regulatory process is tied more closely to verified clinical gains or introduce stricter transparency requirements around lobbying efforts. This could help curb any chance of political connections overshadowing the pharmaceutical industry’s central purpose: delivering truly novel and effective treatments to the patients who need them most.
|
|
Regulatory Burdens and Anti-Competitive Strategies: Impact on Innovation and Market Dynamics The regulation of firm sizes has long been a double-edged sword: while intended to promote fair practices and oversight, these policies often result in a number of unintended consequences that distort competition and discourage innovation. For example, France's 50-employee regulatory threshold imposes significant costs on firms that exceed it, leading many firms to stagnate just below the threshold to avoid additional regulatory burdens. For this week's memo, I extracted insights from an American Economic Association paper titled "The Impact of Regulation on Innovation", which highlights these dynamics and provides a critical lens through which to evaluate size-dependent policies.
Impact on Innovation The figure above ("Total Innovation by Firms of Different Employment Sizes") displays the sharp decline in innovation activity just before the regulatory threshold-what one might call an "innovation valley." This sharp decline occurs because firms below the threshold seek to avoid crossing the threshold beyond which costs will skyrocket-for example, where additional labor regulations must be adhered to.
Similarly, Panel A (Innovation) makes a more general comparison between the regulated and the unregulated environments: whereas in the unregulated contexts, innovation rises smoothly with firm size; in regulated ones, the threshold disrupts this natural process. The message is blunt: regulations with the purpose of 'fairness' may have an unplanned consequence for discouraging innovation, especially of smaller firms around critical size thresholds. Distorted Firm Size Distribution
Another spillover of such regulations is how they affect the firm size distribution, which Panel B (Size Distribution) shows rather clearly. The chart dramatically illustrates how regulatory thresholds distort the natural size distribution of firms. Many firms deliberately remain below the threshold of 50 employees, which creates a visible "bulge" in the data, while the share of larger firms declines. But this practice undermines competition because large firms are protected from the pressures of smaller ones that might otherwise grow and challenge them. The result is a less dynamic economy, with less creative destruction- a key driver of growth. Policy Implications These findings have crucial implications. Rigid, size-dependent policies create sharp disincentives for growth and innovation in their efforts to address very legitimate concerns. Policymakers need to consider more graduated approaches, such as phasing regulatory requirements in as firms grow, rather than simply placing binary thresholds. A flexible framework could allow firms to grow and innovate without the fear of a sudden jump in costs. This approach would preserve the original goals of regulation while removing barriers that discourage growth. Conclusion The visual evidence here speaks volumes about how well-intended regulations can backfire, suppress innovation, and alter the competitive landscape. Revisiting such policies with a focus on flexibility and scalability is essential in fostering a vibrant economy. Without reform, size-dependent regulations risk holding back the very dynamism they aim to protect, leaving firms stuck in a pattern of stagnation. Meeting these challenges is critical for sustaining long-term growth and productivity. |
|
ARE RUSSIAN MICRO FIRMS STRUGGLING TO SCALE OR FINDING NEW PATHS TO GROWTH? In the readings and during class, Professor Akcigit highlighted that productivity growth, often proxied by the number of patents per capita, can be driven by both new entrants (e.g., the U.S.) and established incumbents (e.g., Germany). The World Development Report 2024 further showed that younger firms, which are typically small, account for the majority of net job creation. Building on this framework, I wanted to conduct a simple analysis in the Russian market, focusing specifically on micro and small firms to see if there are similar trends.
Unfortunately, there is no publicly available information on the number of the employees and the age of the firms, so I decided to look at the general trend of how the number of micro and small firms has been changing over years. I was able to find data on the number of “newly born” organizations, which are very vaguely defined in the literature and in the datasets themselves but assumed to refer to the number of newly registered firms in a given year. This term typically applies to micro firms, as it is challenging for a micro firm to grow into a small firm within a year. Interesting, however, 4 newly born small firms were recorded in 2023. In Figure 1, the percentage of newly born micro firms relative to the total number of micro firms has remained stable over the past eight years, fluctuating between 18% and 20% (excluding COVID years). This suggests that approximately 20% of micro firms are entering the market each year. However, Figure 2 reveals that the growth rate of micro firms is highly volatile, remaining below 6%, even turning negative in 2019-2020. Two possible scenarios could explain the trend before 2022: 1) a significant percentage of micro firms fail to survive or 2) some micro firms grow and transition into small firms. However, the decline in the number of micro firms correlates with a simultaneous decline in the number of small firms, with the percentage of small firms decreasing even more sharply. This indicates that the likelihood of micro firms transitioning into small firms is low. It means that first, it is difficult for micro firms to grow into larger firms with more than 15 employees. Second, the high attrition rate of micro firms suggests that most are probably exiting the market rather than scaling up. Unfortunately, due to the lack of data on the age of the firms, it’s challenging to determine whether these exits are due to inability of young, innovative startups tp survive or the long-anticipated decline of older micro firms. After 2022, the number of small firms has gradually increased. This growth over the past three years could be attributed to one of two factors. The first possibility is that medium-sized firms (those with fewer than 250 employees and profit of less than 2 billion rubles, ~ 20 million USD) may have downsized into the small firms. However, this is unlikely, as Figure 2 indicates that the number of medium-sized firms has been growing during this period. The second and more plausible explanation is that some micro firms actually managed to grow and transition into small firms, which could signal the potential for small, innovative firms to compete with well-established ones. However, these are just potential explanations, and further analysis is needed to draw more definitive conclusions. 
Figure 1 
Figure 2 Data Source: Federal Tax Service of Russia. (2025). Statistics on small and medium-sized businesses. Retrieved January 10, 2025, from https://ofd.nalog.ru/statistics.html?statDate=10.01.2025&level=2&fo=&ssrf=&t=1737096973584&t=1737096973585 |
|
In Chapter 4 of the World Development Report 2024: The Middle-Income Trap, the authors emphasize how middle-income countries often fail to transition to high-income status when their economic growth model becomes trapped in sectors with limited scope for innovation. Simultaneously, Schumpeterian growth theory (Aghion, Akcigit, & Howitt, 2014) underscores the essential role of creative destruction—the entry of innovative firms displacing unproductive incumbents—in sustaining long-run growth. Integrating these perspectives, I argue that artificial intelligence can serve as a potent lever to overcome MIC stagnation, provided institutions avert policies that dampen the dynamism of new entrants. Despite AI’s promise, incumbent firms often use non-productive strategies (Barriers to Creative Destruction, Baslandze 2023) to defend existing market power—especially in MICs where policies may favor large, politically connected actors. For example, firms might lobby for size-dependent subsidies or preferential treatment, which can distort incentives for widespread adoption of AI-based innovation. This inertia stifles creative destruction, preventing smaller AI-driven startups from scaling up and challenging incumbents. However, if AI is democratized—through accessible machine-learning tools, supportive regulatory frameworks, and open innovation platforms—smaller firms in MICs can more effectively harness technology to boost productivity and disrupt entrenched incumbents. By lowering the barriers to data processing and R&D, AI can raise the likelihood of leapfrogging in sectors from agriculture (e.g., automated crop monitoring) to finance (e.g., digital lending services). These transformations become particularly powerful when accompanied by pro-competition reforms that reduce bureaucratic red tape, increase transparency in procurement, and limit preferential policies for incumbent firms. Analytical Element: Modeling AI-Enhanced Creative Destruction
As Ultimately, aligning AI investments with competitive, pro-innovation policies can help middle-income countries break free from the trap and cultivate sustained Schumpeterian growth. |
|
The Role of Market Concentration in Stifling Innovation: Evidence from the U.S. Tech Sector Market Concentration and Innovation in the U.S. Tech Sector Using Herfindahl-Hirschman Index (HHI) data from the U.S. Federal Trade Commission (FTC), we observe that the tech sector's concentration levels have been rising steadily. An HHI above 2,500 indicates high concentration, and recent estimates suggest that the software and digital services sectors have reached levels exceeding this threshold, signaling reduced competitive intensity. The Effects of Market Concentration on Innovation 
Firm Entry Rate captures new business creation as a share of total firms in the sector. Policy Implications |
and others
Post your memo in response any (or all) of the week's readings and an empirical case regarding artificial intelligence, innovation, and/or growth:
Post by Thursday @ midnight. By 1pm Friday, each student will up-vote (“thumbs up”) what they think are the five most interesting memos for that session. The memo should be 300–500 words (text) + 1 custom analytical element (e.g., equation, graphical figure, image, etc.) that supports or complements your argument. These memos should: 1) test out ideas and analyses you expect to become part of your final projects; and 2) involve a custom (non-hallucinated) theoretical and/or empirical demonstration that will result in the relevant analytical element. Because these memos relate to an empirical case students hope to further develop into a substantial final project and because they involve original analytical work, they will be very difficult to produce with generative AI and we strongly discourage you from attempting it. Some of the top-voted memos will form the backbone of discussion in our full class discussion and break-out room sessions.
The text was updated successfully, but these errors were encountered: