The Impact of Automated or Algorithmic Trading on Financial Market Stability

Listen to the audio version of this article here.

If you had asked financial market traders just a few decades ago what the most important factor for success was, you would probably have heard answers like experience, speed of decision-making or knowledge of the market. But today a significant share of trades on stock exchanges and financial markets around the world is carried out not by humans but by computers; computers that analyze thousands of data points in a fraction of a second, identify trading opportunities and place buy or sell orders without the slightest emotion such as fear or greed.

This shift, known as algorithmic trading or automated trading, is considered one of the biggest changes in the history of financial markets, and its role in market stability or instability has drawn more attention than ever from economists, regulators and investors.

But do algorithms really make markets calmer? Or, on the contrary, can they become a factor that intensifies price crashes in times of crisis?

Contrary to what many people think, the answer to this question is not simple. Algorithmic trading has brought significant benefits to markets, but in some conditions it has also created new risks. Below, we look at this topic in more detail.

What is algorithmic trading and how does it work?

Algorithmic trading refers to a process in which assets are bought and sold based on a set of predefined rules, without direct human intervention. These rules can be very simple; for example:

• If the stock price rises above the 20-day moving average, buy.
• If trading volume suddenly doubles, close the sell position.

But at a professional level, algorithms use hundreds of different variables to make decisions, such as:

• Price
• Trading volume
• Interest rates
• Economic news
• Social media data
• Technical indicators
• Even satellite imagery or weather data

Put simply, an algorithm is a set of instructions that tells a computer what to do under any market condition.

Types of algorithmic trading

Trading algorithms come in different types depending on their purpose. In the table below you can see and compare the most important ones.

Among these methods, high-frequency trading (HFT) has drawn more attention than the other models, because its execution speed is so high that humans practically cannot compete with it.

The positive effects of algorithmic trading on the market

The most important question analysts raise about the effectiveness of automated trading is how, and to what extent, these trades may alter market stability.

 

From the perspective of financial economics, market stability means that prices change in an orderly, transparent way that reflects real information; not that they experience sharp jumps merely due to excitement or a lack of liquidity.

Algorithmic trading can affect this stability through several different channels:

1. Increasing market liquidity

One of the most important benefits of algorithms is increased liquidity. Liquidity means an investor can trade their asset whenever they want with the smallest possible difference between the buy and sell price.

Market-making algorithms (Market Makers) continuously place buy and sell orders, ensuring there are always buyers and sellers in the market. The result of this process is:

• Narrower gap between buy and sell prices (Bid-Ask Spread)
• Faster order execution
• Lower trading costs
• Greater investor confidence

For example, on the New York Stock Exchange and Nasdaq, a large share of daily liquidity is provided by these market-making algorithms.

2. Faster discovery of the true price

Another important benefit of algorithmic trading is an improved price discovery process. In financial markets, asset prices should reflect the available information. The faster information is reflected in prices, the more efficient the market will be.

By simultaneously analyzing news, economic data, financial statements and even market sentiment, algorithms react much faster than human traders. As a result:

• Prices reach equilibrium faster.
• opportunities Arbitrage opportunities decrease.
• Price differences between markets shrink.
• The likelihood of information abuse decreases.

Of course, this same high speed can also have negative effects in some conditions, which we will cover later in the article.

How can algorithmic trading increase volatility?

Financial markets are always influenced by news, economic data and investor behavior. In the past, an important piece of news usually took several minutes or even hours to affect prices, because human traders had to read the news, analyze it and then decide whether to buy or sell. But today this process has become almost instantaneous.

Many advanced algorithms process news published by the media, central bank reports and even social media content in a fraction of a second. If thousands of algorithms reach the same conclusion, they all send buy or sell orders almost simultaneously.

The result of such behavior can include the following:

• A sudden surge in trading volume
• Sharp price spikes or crashes within seconds
• A temporary drop in liquidity
• A wider gap between buy and sell prices
• Abnormal volatility

In such conditions, the main problem is not the algorithms themselves; rather, the coordinated and simultaneous reaction of a large number of them can throw the market out of balance.

The phenomenon of «algorithmic herding»

One concept that has drawn the attention of financial economics researchers in recent years is algorithmic herding. In traditional markets, herding occurs when investors imitate others without independent analysis; for example, they buy simply because everyone else is buying. Something similar can happen among algorithms as well.

Suppose most investment firms use similar models to analyze the market. If a particular indicator falls below a certain level, thousands of algorithms may conclude almost at the same moment that they should sell their assets.

As a result, selling pressure builds and prices fall faster than economic fundamentals would justify.

One of the best-known examples of the impact of algorithmic trading on market instability is the event known as the Flash Crash. On 6 May 2010, the US stock market experienced one of the strangest crashes in its history within just a few minutes.

The Dow Jones index plunged nearly 1000 points in a short time, equivalent at the time to about 9 percent of the index's value. Interestingly, much of this drop was recovered just a few minutes later. The event was not caused by an economic crisis but by a combination of several factors:

• A very large sell order placed by an investment fund
• A chain reaction among trading algorithms
• The temporary withdrawal of market makers from the market
• A sharp drop in liquidity

As a result, many stocks and investment funds traded for a few moments at prices far below their true value.

After this incident, US regulators developed new rules such as the «Circuit Breaker», or «emergency trading halt mechanism», so that in the event of abnormal volatility, trading is halted for a few minutes and the market has a chance to regain its balance.

Artificial intelligence: a new chapter in algorithmic trading

While the first generation of algorithmic trading operated on a set of predefined rules, the new generation of this technology has moved toward artificial intelligence (AI) and Machine Learning.

The main difference between the two approaches is that traditional algorithms execute exactly the rules the programmer defined for them, whereas AI-based algorithms can learn from historical data, discover new patterns and even adapt their behavior as market conditions change.

For example, a classic algorithm might issue a buy order only when the price crosses the 50-day moving average. By contrast, a machine learning model analyzes, alongside this indicator, hundreds of other variables at the same time, such as trading volume, interest rates, economic news, user sentiment on social media and even the correlations between different markets.

How do regulators control algorithmic trading?

As automated trading has spread, almost all of the world's major exchanges have drawn up special rules for risk management of this technology. The aim of these rules is to prevent disruption to market functioning and to maintain investor confidence.

Among the most important measures are the following:

1. Automatic trading halts (Circuit Breaker)

2. Testing algorithms before deployment

3. Real-time monitoring of orders

4. Limits on order submission speed

These measures show that rather than fighting algorithmic trading, financial markets are trying to build a framework that preserves the benefits of this technology while keeping its risks under control.

The state of algorithmic trading in Iran

Iran's capital market has also taken steps toward algorithmic trading in recent years. Although the technology is not yet used as widely as on US or European exchanges, many financial technology (fintech) companies, brokerages and professional traders have become interested in developing automated tools.

In past years, the Securities and Exchange Organization, aiming to prevent disruption to trading, has imposed restrictions on the use of certain algorithms and made their use conditional on obtaining a license and complying with specific rules. This approach shows that developing algorithmic trading in Iran, as in other countries, requires striking a balance between innovation and maintaining market stability.

The future of financial markets: humans and machines working together

It is sometimes assumed that as artificial intelligence advances, the role of human traders will disappear entirely. But the available evidence paints a different picture.

Algorithms perform far better than humans at analyzing huge volumes of data, executing orders quickly and managing thousands of trades simultaneously; but they still have limitations in understanding qualitative factors, analyzing political events, making decisions in entirely new situations and assessing long-term consequences.

For this reason, many experts believe the future of financial markets belongs neither to humans nor to machines, but will depend on smart collaboration between the two. In this model, algorithms take on repetitive, high-speed and data-driven tasks, while humans focus on strategy design, risk management, oversight of system performance and high-level decision-making.