Algorithmic Trading: A Smarter Way to Make Money

For those who have watched The Wolf of Wall Street, the thought of stock markets, Wall Street, and finance, in general, might be dangerously warped. Most of us imagine a bunch of (pretend) alpha males going at it with crazy intensity. This was true a few years ago, but the rise of algorithmic trading has been seeing trying to put an end to that toxic culture. The “finance guys” have been replaced by the “science nerds” who keep staring at the numbers all day till those numbers start making some alpha. 



Since the coronavirus started wreaking havoc, there has been a significant market sell-off among hedge funds. According to data from Hedge Fund Research, average quantitative funds lost 5.25% in the first four months of 2020, whereas average hedge funds lost 6.56% during the same time. These numbers explain the rage over quantitative funds on Wall Street in a straightforward way. Even a marginal leg-up in trading can lead to substantial profits. 

Algorithmic trading strategies have been around for a while. But the amount of data and computing power available to hedge funds right now is nothing short of monstrous, which has led to most of the trades becoming automated. A study in 2016 showed that trading algorithms, rather than humans, performed over 80% of trading in the FOREX market. NYSE has managed to operate without a single trader on the floor during the lockdown period. In high-frequency trading, the precision and timing achieved by algorithmic trading are the only way profits can be made. There are various arguments in favor of and against algorithmic trading. Many people blame algorithmic trades for volatility in stock prices. For instance, some people link the 2010 flash crash($860 billion wiped out within 30 minutes) to algorithmic sell orders. One faction claims that algorithmic trades have made stock markets more liquid and efficient. So which argument is correct?


Some of the most basic algorithmic trading strategies include trends and arbitrage. A trend-based approach uses technical indicators like moving averages to make trades with the market momentum. Whereas, arbitrage provides a sort of risk-free income to traders. It takes advantage of a price differential in the stock price from two different exchanges. The price differences are usually not much, but due to high volumes, profits earned can be substantial, especially by algorithmic trading. Investment banks and hedge funds have developed several other algorithms. Few of them are Sniper, Guerrilla, Stealth and Dagger( hard to ignore the Wall Street “alpha-male culture” even in their naming patterns). For example, a strategy called “Iceberg” helps firms to make their trades anonymous and hide their original volume and plan by chopping up their orders into blocks.


In terms of AI and machine learning, hedge funds have started sensing the opportunity that data presents. This hype has led to a hiring spree among investment banks and hedge funds. WorldQuant, a hedge fund with $7 billion Assets Under Management(AUM), recently hired Yoram Singer as its chief AI scientist. Prof. Yoram Singer is part of the computer science department at Princeton and was also the principal chief scientist at Google. This recruitment follows another hiring of the former head of Goldman Sachs’ quantitative investment strategies. 

We live in a world where the President of the United States declares policy decisions on Twitter. For example, threatening China on twitter puts in motion a chain of events that can set off volatility in stock prices, thereby providing an opportunity for hedge funds to make money. Therefore, most hedge funds employ machine learning models and neural networks to predict stock prices from data. Apart from the financials, these models also perform sentiment analysis for a more coherent prediction. 


The market is not perfect, which implies that certain events lag behind each other in reflecting on the prices. This is where robust algorithms hit payday. The term payday is not entirely accurate as the profit margins on these trades are quite low. For example, statistical arbitrage might offer a few cents; however, due to the trades’ volume, hedge funds can pull huge profits. A lot of funds also consist of teams who only focus on event-based trading(aka investor irrationality). For example, these teams work on strategies that are applied to an earnings call or a day before options expiry. However, funds also have to take care that their algorithms do not make the market crash due to the reactivity of their algorithms to volatility.


Due to the availability of data, various hedge funds hire quantitative analysts to create innovative pricing strategies. Much of this work involves statistics, computer science, discrete mathematics(e.g., stochastic calculus), and financial modeling. It is not as “attractive” as what most people perceive of a career in finance. However, this groundwork is essential in building the best trading strategies. The movie,’ The Big Short’ fairly represents the kind of work that has to be done, especially the grueling numbers-work done by Dr. Michael Burry in the movie. It is essential to understand that, even after the hard work, some strategies are never deployed into the market. One of the most important reasons is that you were too late, and the market was already littered with such algorithms, thereby decreasing the profitability of the strategy.


A prime example would be quantitative researchers looking at satellite imagery to determine oil levels or weather data that affect the pricing of certain commodities. However, once Wall Street got wind of this strategy, it merely became a routine part of the groundwork, instead of the mind-blowing idea it once was. This makes it imperative for firms to hire people with strong statistical backgrounds[After all, ML is glorified statistics] to come up with these strategies regularly. (For the amount of money that firms pay, they should expect miracles)


As machine learning algorithms get smarter, the role of front end employees is diminishing. We should soon reach a stage where the financial system is dominated by Computer Science and Statistics majors. A large chunk of the tasks can be automated away to smart bots, ensuring that the stockbrokers need to spend less time on each client, which can help cut the workforce further. Firms like Worldquant and Tower Research tend to hire Computer Science majors for their front-end roles. Teaching finance to Computer Science/Statistics grads is much easier than teaching Computer Science to business majors. 


But none of the scattered few difficulties should be a deterrent towards algorithmic trading. The most logical reason is that algorithms are bereft of the emotional volatility of humans. These emotions are the most significant sources of bias. It still takes an immense amount of pure logic and discipline to make it big in algorithmic trading—those who do certainly deserve the ‘alpha’.



  2. 80% stat-




Leave a Reply