The invisible structures behind financial markets: complexities, failures, and solutions

🇬🇧 For the English version of this blog, read below. 🇳🇱  For the Dutch version, click here.

🌐 For other languages of this blog, click your flag below.

---

In recent years, I’ve immersed myself in the world of finance, macroeconomics, and the hidden financial flows that shape our society and business landscape. What started as a curiosity about basic economic principles grew into a deep dive into the complexity of economic systems and the ways they interact. Along the way, I noticed that many decision-makers, from regulators like the SEC to international brokers and policymakers, often have only a partial grasp of how these systems actually work and of the role technology platforms and conflicts of interest play within them.

This isn’t a criticism but a recognition of just how complex the subject truly is. Even with my background in technology, governance, economics, and complex problem-solving, plus more than eight years in the financial markets including hands-on experience in equity trading, I see how layered and intricate this field remains. Over the years, I’ve studied more than 10,000 pages of literature, reports, policy papers, and market analyses, and spoken with hundreds of full-time traders. Yet the system still holds many opaque corners. It’s no wonder that mistakes happen in the regulation and oversight of financial systems because the issues are too tangled for quick fixes.

Still, I believe it’s essential to share what I’ve learned. That’s why I’m writing this blog: to give more people a clearer view of this complex and often obscure system and to explore possible solutions.

Here in this blog, I’ll focus on the deeper layers of financial markets, where structural problems are emerging, and just as importantly, what solutions might work. My approach is always solution-driven because critical thinking should go hand in hand with constructive ideas. While my examples and analyses aren’t meant to apply to every country, as regulations and contexts differ, they highlight broader systemic challenges and opportunities for change.

If you have questions or would like to explore further, feel free to contact me.

---

From gold bars to digital currency: how our monetary system changed over five centuries

Before we dive into the structure of today’s financial markets, it helps to go back to the origin of money. Money did not appear out of thin air. It began as a practical medium of exchange, a stable measure of value that simplified trade and made it possible for strangers to trust one another. For centuries, gold and silver served as a neutral foundation because they were scarce and could not be manipulated by a single ruler.

In 1694, England founded the Bank of England. That bank was allowed to lend more money than it actually held in its vaults, a practice we now call fractional banking.

In 1792, the United States made the dollar its official currency, still linked to gold and silver. In 1933, the U.S. government issued Executive Order 6102, forcing citizens to surrender their gold to fight the Great Depression. Those who refused risked imprisonment.

During World War II, in 1944, the dollar became the world’s reserve currency under the Bretton Woods agreement, still backed by gold. In 1971, however, the link to gold was abandoned, and the dollar became a currency without physical backing. Since then, central banks create money through debt. Every unit of currency, such as the dollar or euro, in circulation is essentially a loan with interest, which requires the money supply to expand continuously.

The path to the euro began in 1970 with the Werner Report and continued in 1972 with the “snake in the tunnel” exchange-rate system, both efforts to bring European economies closer together. In 1991, the Maastricht Treaty was signed, committing member states to adopt a single currency. The euro went into digital use in 1999 and entered circulation as cash in 2002.

Today, the global financial system consists of layers of digital money, complex financial products, and ultra-fast trading computers that move trillions every day. Its connection to real economic value often weakens, risks are bundled and sold on, and debts grow ever larger.

---

A new era for money: the rise of digital currencies and the future of finance

A new phase of money creation is already taking shape. Central banks are developing their own digital currencies. In Europe, the digital euro is under study. Research began in 2021, and pilot tests will run until 2025 or 2026. In the United States, the Federal Reserve published a discussion paper on a digital dollar in 2022 and will start pilot projects with commercial banks in 2024, aiming for a launch thereafter. China introduced its digital yuan in 2020. Sweden is working on an e-krona, and Japan is exploring a digital yen.

Meanwhile, the idea of a World Currency Unit, or WCU, is gaining ground. This global digital reserve currency would be based on a basket of currencies and commodities. International institutions such as the Bank for International Settlements and the International Monetary Fund support this concept as a response to geopolitical fragmentation and currency instability.

With this background in mind, we can now examine how financial markets are built and what challenges they face.

---

Architecture of financial markets: why complexity leads to systemic failure

When we inspect financial markets, we see a system that has drifted far from its original purpose: allocating capital efficiently to productive investments. Instead, there is a tangled network of financial products, continuously operating trading algorithms, and banks working out of public view. This obscures the underlying real economy.

The main problem is what experts call institutional information asymmetry. Large financial firms have access to better data, faster algorithms, and deeper insights into market structures than the regulators charged with overseeing them. This gap widens when staff move between regulators and financial institutions, sometimes taking lucrative jobs at the very firms they once supervised.

How do these algorithms work? An algorithm is simply a set of instructions that tells a system how to decide. In finance, an algorithm might automatically buy a stock when its price falls and sell it again when the price rises. Simple algorithms follow fixed rules, while more advanced versions use machine learning to spot patterns and make predictions. As a result, algorithms play an ever-larger role in analyzing data and making split-second trading decisions.

Here is one common scenario. A senior regulator spends five to seven years building expertise in a specific market segment. Then a major bank or hedge fund offers that person a salary three to five times higher than their government pay, plus generous bonuses. Faced with limited career advancement or heavy workloads in the public sector, the regulator may accept. They bring with them detailed knowledge of oversight procedures, system vulnerabilities, and how rules are applied in practice. Within two years, they may be advising financial firms on how to operate right up to the legal boundary.

Another critical issue is high-frequency trading. These traders use ultramodern computer systems and algorithms to execute massive numbers of trades in fractions of a second. Their strategy exploits tiny price differences between exchanges or between bid and ask prices, opening and closing positions in extremely short time windows.

A microsecond is one millionth of a second, or 0.000001 seconds. It is hard to imagine, but in electronic trading, this interval is crucial. In that brief moment, high-frequency traders can react to price movements before other market participants even see them.

Firms gain this edge by laying thousands of kilometers of fiber-optic cable as straight as possible between trading venues, for example between New York’s exchange and a data center in Chicago. Shorter, straighter cables mean lower latency. This allows firms to receive price information milliseconds before their rivals and send orders faster. In practice, they can place or cancel buy and sell orders almost instantly, before others have the chance to respond. This hidden advantage creates an uneven playing field, systematically disadvantaging ordinary investors and long-term savers.

About 80 percent of daily price movements on the U.S. stock market are driven by automated trading systems, according to one fund manager who spoke with CNBC. Technology and automation now play a dominant role in financial markets.

---

Complexity of algorithms in financial markets

It is important to highlight that the complexity of algorithms in financial markets is not only a technical issue but also a societal and regulatory challenge. Today’s algorithms are so advanced that even highly educated professionals such as quants with a PhD in financial markets struggle to fully understand them. It takes years of study and in-depth research to grasp their underlying logic and structure. Only after a significant investment in knowledge and experience can these experts intervene at the core of such systems.

What makes this even more complicated is that these algorithms do not operate in isolation. They continuously integrate data from diverse financial markets and economic sectors. The connections between stock markets, bond markets, currency markets, and commodity markets influence one another in a constantly shifting web of interactions. This creates complex dynamics at the macroeconomic level, where not only the functioning of individual algorithms matters but also how they interact and respond to global economic developments, geopolitical events, and policy decisions. As a result, it is extremely difficult for both specialists and policymakers to appreciate the full impact and interdependence of these systems.

Given this complexity, it is understandable that policymakers and regulators, who generally lack the same technical background, find it hard to grasp the operation and implications of these algorithms and their interactions. The rapid pace of financial technology development only makes it harder to keep up with market realities. This gap increases the risk that decision-making and regulation will lag behind technological and economic change.

To address this challenge, policymakers and regulators must actively seek ways to bridge the complexity. Collaboration with industry experts is crucial, as is sustained investment in research and the development of new oversight methods. Only then can we ensure that algorithms are deployed in a transparent, fair, and secure manner, supporting a stable and just financial sector and safeguarding broader macroeconomic stability.

Practical example: May 6, 2010 — the flash crash

On May 6, 2010, the famous “Flash Crash” took place. Within minutes, the Dow Jones Industrial Average plunged nearly 1,000 points, about 9 percent, and then recovered almost as quickly. In just twenty minutes, billions of dollars’ worth of shares changed hands, and some companies briefly lost almost 99 percent of their market value.

Investigators found that a single large sell order from an institutional investor triggered an uncontrolled reaction by high-frequency trading algorithms, which reinforced each other in a downward spiral. There was no significant economic news behind the drop. It was purely the result of algorithmic trading that ignored broader economic consequences. This incident painfully illustrated how vulnerable and disconnected markets had become from the real economy and how major systemic risks can arise from uncoordinated algorithmic reactions.

---

Dark pools

Another major factor is the use of “dark pools,” private trading venues where large orders are executed anonymously without immediate visibility in public markets. In an era where data is treated as “the new gold,” the information gathered in these venues is crucial. Because orders in dark pools are not publicly reported, valuable trading information remains hidden from most investors and market participants. High-frequency traders can exploit this concealment to their advantage. By combining extremely fast access to trade data with these secret venues, they can effectively front-run large institutional orders.

In practice, we see that ordinary traders are willing to invest millions to gain these small but strategically crucial advantages, seriously undermining market fairness.

To illustrate, imagine you place an order through your bank to buy 100 shares of Apple at €150 each. You assume your trade executes immediately, but first a high-frequency trader buys those shares at €150 and then sells them to you at €150.05. You end up paying 5 cents more per share without realizing it. If you repeat this every month with €1,000, you incur €60 in hidden costs per year. It gets worse. When you place an order with an online broker, it often routes first to a market maker that aggregates many orders. This market maker has access to dark pools, where large institutions such as pension funds quietly sell shares. The market maker buys those shares cheaply and then sells them to you at a profit, adding a few cents per share that can amount to hundreds of euros over a year.

Technological platforms thus play a dual role. On one hand, they increase transparency and efficiency. On the other, they introduce new risks that current rules and oversight struggle to manage. A clear example is the rise of decentralized finance. While traditional banks grapple with outdated systems, new players build entirely new financial networks outside conventional supervision. This creates a complex landscape of both opportunities and challenges.

---

The myth of the rational market and the reality of algorithmic chain reactions

One of the greatest misconceptions in economics is that financial markets are always rational and efficient. In reality, modern markets are often dominated by algorithms that respond to each other in microseconds. This triggers chain reactions that are disconnected from the real economic value of assets.

Behind the scenes, a large quantitative hedge fund might use machine learning to detect a small price anomaly, say in the options market. Within microseconds, it places thousands of orders to profit from it. Other algorithms, trained to spot the same patterns, automatically adopt similar positions. This snowball effect amplifies the initial price movement without any fundamental analysis.

What goes further wrong is that the market now operates at three distinct speeds: high-frequency traders act within microseconds; traditional investors operate in seconds or minutes; and the real economy moves over days, weeks, or months. This separation creates a parallel universe in which the majority of trading activity has little to do with underlying economic value. Some high-frequency trading firms even use spoofing, placing large orders and then canceling them to mislead the market about supply and demand.

“The market is a rigged game where the house rules are secretly changed. Only radical transparency can solve this.”
— Haim Bodek

The concentration of influence makes the problem worse. Approximately seven major quantitative hedge funds and ten large banks account for more than half of all trading in U.S. equities. These firms use co-location: their servers sit literally beside exchange computers. As a result, they can send orders microseconds faster than others. Although this time difference seems tiny, in high-frequency trading it is more than enough to yield consistent profits, not because they make better investments but simply because they are faster.

To illustrate this further, suppose you invest €10,000 in an index fund. On a normal trading day, the shares in your fund might suddenly drop by 2 to 3 percent even though there is no news about the underlying companies. This happens because algorithms trigger each other into a downward spiral. You lose €200 to €300 in a single day purely due to technical trading activity. The next day, the price may recover, but by then you have already taken a loss.

---

Conflicts of interest in self-regulation

Conflicts of interest are not always immediately visible, but they can be structurally embedded in the way financial oversight is organized. It still often occurs that regulators later take jobs in the very sector they once had to supervise. This can affect their independence and the way decisions are made in the public interest.

One example of this can be seen with credit rating agencies. These institutions normally assign scores to financial products to indicate their risk. However, they may be paid by the very parties they rate. In practice, this can lead to “ratings shopping,” where a company with an unfavorable score simply switches to another agency in the hope of receiving a higher rating. This practice can undermine trust in the reliability of these assessments.

Within the market itself, there can also be shifts from genuine value creation to forms of trading that focus primarily on profit through speed rather than substantive analysis. Some firms build extremely costly networks so they can execute trades a fraction of a second before others. These practices may add little to the real economy, while their costs are indirectly borne by the broader market, especially retail investors.

Large financial institutions can also influence legislation and policy. Every year, vast sums are spent worldwide on lobbying, not only to advocate interests but also to help draft policy documents or proposed laws. In some cases, parts of official regulations originate from texts written by industry representatives.

For further illustration, an investor might choose a bond fund because it carries a so-called “safe” AAA rating. At first glance, this seems sensible. Yet that rating may have been influenced by a financial relationship between the fund provider and the rating agency. During the 2008 financial crisis, it became clear that many highly rated products still carried significant risk. Even today, complex products with top ratings may conceal true risks that remain opaque to the average investor.

---

AI in financial markets

Recent studies show that artificial intelligence (AI) outperforms human traders in making trading decisions. AI algorithms can analyze vast amounts of data in real time, recognize patterns, and execute trades far faster than any person. For example, an IMF analysis found that AI can generate a trading signal within seconds based on the Federal Reserve’s detailed minutes while human traders need much more time. This shift signals that a financial system built largely on human intuition and traditional methods is becoming increasingly unsustainable.

The exponential growth of AI in finance is driven further by current economic conditions, rising inflation and the continual expansion of central-bank balance sheets, especially in U.S. dollars. These factors make it all the more important to manage markets more efficiently and transparently, with AI playing a central role.

In practice, AI is already widely deployed. XTX Markets, one of the world’s largest algorithmic trading firms, uses advanced AI models to process over $250 billion in transactions daily. The firm runs a supercomputer with 25,000 GPUs and 650 petabytes of storage. Brokers, too, integrate AI to provide better client advice and to analyze market sentiment via social media and news feeds.

AI’s key advantage is its ability to reduce human bias and emotional decision-making through objective data analysis. It can perform sentiment analysis to gauge market reactions to news and automatically adjust trading strategies. AI systems can also deliver personalized investment advice tailored to each investor’s risk profile. A trader survey even found that algorithmic systems boost productivity by 10 percent.

Yet AI has limitations. It may struggle to predict unexpected events such as natural disasters or sudden leadership changes in companies. It also lacks the intuitive judgment that human traders sometimes rely on in uncertain markets. Data errors or algorithmic bias can lead to flawed predictions.

The Bank of England has warned of AI’s potential risks to financial stability. Widespread use of identical AI models can lead to highly correlated trading positions, which may destabilize markets during stress. AI systems are also vulnerable to cyberattacks or operational failures at service providers, events that could have severe consequences.

In summary, AI offers enormous opportunities to enhance the efficiency, speed, and accuracy of financial markets. At the same time, it is crucial to manage its risks and maintain human oversight alongside AI-driven systems. Only by doing so can AI contribute to a more stable and fair financial system in a fast changing world.

---

Focus on the short term and lack of broader economic insight in the markets

What further strikes me is that many regular day traders (who take positions daily), high-frequency trading (HFT) firms, and investors seem barely interested in what is actually happening in the economy. Their focus is mainly on short-term profit: an edge of microseconds, a fraction of a cent per share, or a quick swing trade. The broader economic consequences such as employment, inflation, or the stability of the financial system are often overlooked.

Yet their trading is directly connected to the real economy. When algorithms sell en masse on the basis of negative news, they can trigger a chain reaction: stock prices plummet, wealth evaporates, companies struggle to raise capital, and consumers lose confidence. An algorithmic “flash crash” can destroy billions in value within minutes, even when there is no underlying economic crisis.

If you would like to read more, see my previous blog post (click here), in which I explain how stock-market prices can affect our daily lives.

The disinterest in how the economy actually works partly stems from the fact that some traders operate in silos. They optimize for risk and return within their own portfolios, without regard for broader system-wide risk. HFT firms treat markets as a speed game, not as a reflection of economic health. Investors chase sentiment and momentum, not the true value of companies or the social impact of their actions.

To raise awareness, the government should at least ensure the following measures are adopted:

1. Introduce mandatory systemic-risk education for all professional traders and investors similar to the existing financial-certification exams, but focused on broader economic impact.
2. Require real-time exposure to systemic risk: traders should see live dashboards on trading platforms that show how their actions contribute to market volatility.
3. Impose a tax on ultra-short transactions (such as a Tobin tax) to discourage extreme short-term speculation and promote long-term thinking.

These steps can encourage the sector to consider not only profit, but also the health of the economy as a whole.

---

Technological disruption as an opportunity for system reform

Despite the sometimes bleak analysis, positive developments are also emerging in financial markets. In particular, technological progress offers promising paths toward greater transparency, efficiency, and fairer access to financial systems. Blockchain technology and decentralized platforms, for example, can help break existing power concentrations. These technologies enable greater transparency, lower transaction costs, and new forms of democratic oversight.

One solution is to create a “fair access” infrastructure, meaning that all market participants receive the same data and speed, regardless of size or budget. Technically, this could be achieved by imposing a uniform “speed bump” an artificial delay of, say, 350 microseconds on all orders. This would remove the edge of high-frequency traders without harming market liquidity.

Another concrete proposal is to mandate “lit” trading venues where all orders are visible to everyone, unlike today’s dark pools. This could be paired with a “trade-at” rule, requiring orders first to be posted on public exchanges before they can move to private platforms.

---

Increased transparency and fairness through peer-to-peer technology

Peer-to-peer (P2P) technology in the world of cryptocurrency such as Bitcoin can also boost transparency and fairness. P2P means people can trade directly with one another, without an intermediary like a bank or an exchange. This allows for fast, low-cost transactions. In Bitcoin, for example, all transactions are recorded on a public blockchain that anyone can view but no one can alter. This ensures transparency and prevents rule-manipulation.

Decentralized Autonomous Organizations (DAOs) further this vision. DAOs run entirely on blockchain technology without a central leadership. Decisions are made by members typically through voting so no single person or group can unilaterally change the rules. This collective decision-making makes it harder for algorithms to skew outcomes.

Tokenization is another key element: converting physical assets such as real estate or art into digital tokens on a blockchain. These tokens can then be traded P2P, enabling wider access to investments even for those with limited funds. Tokenization broadens investment opportunities and allows more people to benefit from financial markets.

An additional advantage of P2P technology is its potential to strengthen local economies by facilitating direct connections between consumers and producers. This reduces reliance on large intermediaries and stimulates local trade. When people buy and sell locally, money stays within the community, supporting jobs and growth. P2P platforms can also help small businesses and entrepreneurs reach customers more easily, fostering innovation and diversity in products and services.

Moreover, P2P technology can help mitigate the impact of international shocks and so-called “black swan” events unpredictable crises like financial meltdowns or natural disasters. By reinforcing local networks, communities become more resilient. In times of crisis, local P2P platforms can ease the distribution of goods and services, reducing dependence on global supply chains. This preparedness enhances local stability and speeds recovery after unexpected disruptions.

Together, P2P technology, DAOs, and tokenization create a financial ecosystem that is more transparent and equitable, making it much harder for algorithms to manipulate outcomes. This gives many more people the chance to participate in the economy.

---

Challenges with Bitcoin and Cryptocurrencies

However, a caveat is necessary: the biggest problem with Bitcoin and other cryptocurrencies today is that they are heavily influenced by the same financial algorithms used in traditional markets (as described above). These algorithms can react instantly to news and market movements, causing price volatility and manipulation. As a result, prices may not always reflect the true value of assets, but rather the short-term strategies of algorithmic traders.

P2P technology can help counter this by making transactions transparent and public. When everyone has access to the same information and trades, algorithms cannot exploit hidden advantages. Furthermore, if the majority of traders in a P2P network adhere to fair-trade practices, manipulative algorithms will find it much harder to exert influence.

To ensure this protection, it is important to establish clear guidelines and rules for P2P platforms. The government can play a crucial role by:

• Regulation and supervision: Enact clear rules for P2P platforms and ensure compliance. This may involve creating a dedicated oversight body to monitor platform activity and guarantee transparency and fairness.
• Education and awareness: Launch public-education programs to inform users about the risks of algorithmic trading and teach them safe practices on P2P platforms. This fosters a culture of honesty and openness.
• Reporting mechanisms: Set up systems that allow users to report suspicious activity. Quick identification and response to manipulation or unfair practices will follow.
• Technology support: Invest in technologies such as improved blockchain frameworks and smart contracts that enhance the security and transparency of P2P platforms. This preserves market integrity.
• Promote open, competitive markets: Ensure many participants can trade fairly so no single actor gains excessive power. Transparency and oversight help prices form fairly and risks distribute evenly. Stable, reliable markets then support sustainable economic growth.

On the other hand, users can also play an active role by:

• Ethical conduct: Commit to fair-trade practices and remain mindful of how individual actions affect the market. Avoid manipulative strategies and champion transparency.
• Self-education: Learn how P2P platforms work and understand the risks of algorithmic trading. Better knowledge leads to wiser decisions and greater protection against manipulation.
• Community engagement: Participate in P2P forums and discussions to share experiences and support one another in promoting fair-trade standards.

By having governments and users together implement these measures, we can safeguard the integrity of cryptocurrency markets and minimize the influence of algorithms and AI on prices.

---

Towards a new economic paradigm: proposals for systemic reform

Based on in-depth analysis, several additional concrete directions for systemic reform can be envisaged.

First, there is an urgent need to redesign oversight structures with a focus on independence and expertise. This involves not only adjusting rotation limits but also investing in the technological competence of regulators.

Concretely, this could mean that regulators receive a “technology bonus” empowering them to work with the same tools as the institutions they supervise. For example, the U.S. Securities and Exchange Commission (SEC) should have access to real-time data feeds and artificial intelligence analysis tools comparable to those used by the largest hedge funds. In addition, there should be a prohibition on direct movement into the financial industry for at least five years after leaving a regulatory position.

A specific solution to the problem of high-frequency trading is the introduction of a “Tobin tax” on extremely short-term transactions. A levy of 0.1% on positions held for less than one second would render most high-frequency trading strategies unprofitable without harming market liquidity for ordinary investors. This mechanism can be implemented by applying precise timestamps to all transactions.

Second, it is essential to break up excessive concentrations of market power. This can be achieved by promoting open-source alternatives for critical financial infrastructure. The European Union has already taken a first step with the Revised Payment Services Directive (PSD2), which requires banks to grant third parties access via application programming interfaces (APIs). This requirement could be extended to cover all core functions of the financial system.

Another concrete measure is to mandate “speed bumps” on all trading venues. These are artificial delays typically around 350 microseconds that eliminate speed as a competitive advantage. IEX, a relatively new stock exchange, has successfully implemented this approach, demonstrating that it does not harm the market but instead makes trading fairer.

Third, we must develop new indicators of economic health. Current metrics fail to provide an adequate picture of real economic wellbeing. One promising alternative is to create a “Broad Wellbeing Index” that incorporates income inequality, environmental impact, and systemic risks.

---

Overview of key challenges and additional aspects

Below is an overview of the challenges outlined above, supplemented with possible solutions and further considerations.