Happy US National Financial Literacy Awareness Month!
As AI and Large Language Models (LLMs) are top of mind, my focus this week is on algorithmic trading because it is where the largest impact will be realised sooner rather than later.
I set the scene with facts and figures with the aim to raise awareness around algorithmic trading.
Algorithmic trading is already a big sector and is growing in certain regions. In the US, according to the Wall Street Journal, algorithmic trading accounts for 60–73% of the overall US equity trading. As US equity markets are the most liquid globally, it makes a lot of sense that algo trading is so dominant. However, high growth (in terms of USD trading volume) is expected mostly in the East.
Types of algorithmic trading strategies
- Trend-following strategies: using technical indicators such as moving averages, relative strength index (RSI), and bollinger bands to identify trend changes.
- Mean-reversion strategies: using algorithms to identify assets that are trading above or below their historical averages and sell or buy with the expectation that the asset’s price will revert to the mean.
- Arbitrage strategies: using algorithms that aim to identify and exploit pricing discrepancies between different markets or instruments.
- High-frequency trading (HFT) strategies: using powerful computers and high-speed data networks to execute trades at extremely high speeds, often in milliseconds or microseconds. HFT strategies can take advantage of small price movements and market inefficiencies that may only exist for a fraction of a second. HFTs are frequently used to take advantage of arbitrage opportunities and to reduce the market impact of trades.
- News-based and or sentiment-based strategies: using algorithms to analyse news and social media feeds to identify trading opportunities. For example, a trader may use sentiment analysis to gauge the market’s reaction to a particular news event and take a position accordingly.
- Machine learning-based strategies: using advanced machine learning techniques to analyse large amounts of data and identify patterns that may not be visible to human traders. For example, a trader may use a neural network to identify correlations between different financial instruments and use this information to make trading decisions.
News-based and sentiment-based strategies and machine learning-based strategies are not new. They will be the top beneficiaries of the current advancements in LLMs.
Reasons for using algorithmic trading
The top reasons for using algorithms to trade have not changed much over the past decade. Whether it is institutional investors (e.g. hedge funds), retail investors, Day traders, or long-term traders, the main reasons are:
- Reduce the market impact of placing orders
- Consistency of execution performance
- Increase Trader productivity
- Ease of use
The reasons that have become more important recently (given the increased market volatility and uncertainties) are:
- Using Data on venue/order routing Logic
- Leveraging flexibility and sophistication of smart order routing (SOR) systems
Whether the algorithm is trading in highly liquid markets or not, algorithms will continue to dominate for the above main reasons. Traders will design their strategies which in several cases will be a mix of the strategies mentioned above, with rules or conditions that adapt to market conditions.
Recent reporting from Morgan Stanley shows that algorithmic trading is catching on even in the muni market which is only 20% trading electronically (Source).
Major players in algorithmic trading
The players in the spaces are either platforms or provide software tools. The algorithms may be deployed on-premises, on the Cloud, or hybrid.
Algorithmic traders must choose their strategies. Starting with basic choices around execution (Time-weighted average price (TWAP) or volume-weighted average price (VWAP) algorithms), all the way to the scope of the strategy.
According to Mordor Intelligence, the top five providers of algorithmic trading services are:
Nowadays, brokers like Interactive Brokers offer algorithmic trading capabilities. IB in collaboration with QuantInsti (algo trading executive education), is already suggesting how to experiment with ChatGPT in algorithmic trading.
FinTech players have been innovating in various ways in algo trading. Here are a few examples to get a sense:
- Quantconnect is one of the players that crowdsources algorithms to democratise finance. QuantConnect provides market data and a cluster computer so that engineers can quickly design an algorithmic trading system. They run monthly trading competitions to recruit the best engineers and give them access to capital. The Quantconnect open platform will soon become a testing ground for the unleashed advanced LLMs and ChatGPT.
- Numerai has its own cryptocurrency NMR which is used to incentivise tens of thousands of anonymous data scientists around the world to collaborate to create predictive models. The best machine learning models are used to trade the Numerai Global Equity (not crypto) hedge fund.
- Trality is for crypto retail investors. The Trality marketplace lists bots created by Python programmers using advancements in machine learning. Investors can rent these algorithms (a la copy trading), and programmers can earn income. The programmers on the Trality marketplace will surely be testing more bots with the programming help of ChatGPT.
The deployment of ChatGpt and all other advanced LLMs in Algorithmic Trading will explode. Whether in designing more strategies, programming, testing, or backtesting, the explosion will happen faster than any other financial services subsector.
Resources: ALGORITHMIC TRADING MARKET — GROWTH, TRENDS, COVID-19 IMPACT, AND FORECASTS (2023–2028)
Read the original article here.
