3 3. Depositories: Two major depositories were made to keep a record of stock holdings and their owners and also to keep a track of buying and selling activities by these depository account holders. NSDL (National Securities Depositories Ltd.) and CDSL (Central Depositories Services Ltd.) were the two biggest depositories which hold the maximum accounts of the nation [7].
4 4. T + 2 Rolling Day Settlement: [8] Rolling settlement is a settlement process where a trade executed in the stock market is settled in trading day plus 2 more working days. In the ring trading system, this process used to take days at minimum 5–6 days, but in online trading with the help of [9] depositories and back accounts and their sync with the stock exchanges, the time to complete the process of transferring shares from the account of seller to buyer and also monetary transaction of money from seller to buy’s account took only 2 days after the day of trading. This increased liquidity to a large extent as the money exchanged hands quickly.
5 5. Investor Registration Norms: The two major parties in stock broking are the investors and the brokers. The registration norms for both parties were made stricter. Brokers could not be registered unless they passed exams from NCFM to show that they have required knowledge of stock broking and its process. The clients could not trade unless they presented documents of their address, identity, and financial standing. These strict norms were not there earlier. These helped in making the stock broking forgery free.
6 6. Order Verification by SEBI: In this stage, a very elaborative order verification system was put in place. The brokers were directed very strictly that they need to get the order slips signed by the clients on a daily basis, and also after the end of each day, they were required to make a call to the clients and tell them the exact order which was executed on behalf of him. Stock exchanges and DPs were also advised to keep sending the order details and statements on a regular basis. All this helped in making the system very transparent and full proof where investors do not feel cheated. This all led to more satisfaction in the clients/investors [10].
7 7. Strict Regulation of Brokers: Brokers were very strictly monitored by the regulatory authority SEBI [11]. Regular audits were conducted by SEBI to see that all norms were followed by the broking firms and also the companies. SEBI also issued regular guidelines for the brokers laying out very clearly the process to be followed by the brokers for client registration and executing their trades and also communication with them. These norms were also updated from time to time if technology or economy changed.
8 8. Investor Protection: [12] determined that norms for investor protection lead to investor confidence and thus helped in boosting investor morale and increase in their participation. Participation over a period of time happened in two ways: direct and indirect mode. Directly was when investors participated in buying and selling from your own account and indirectly was when they invested in mutual funds and other such instruments. Investor protection was taken very seriously by regulators and new act was passed with strict adherence norms to be followed by each and every broker and intermediary.
9 9. Operators/Punters: Though the online system of stock trading is totally nameless, but still there were ways in which stocks could be maneuvered to an extent with lot of money and connections. Operators driven scripts are the type of scripts which are controlled by the people who maneuver the stock price according to their will. These stocks could be found even in the index companies. This was one of the hind sides of this system. It was believed that when stocks do not follow the fundamental analysis, it might be moved by an operator.
2.5 Generation 3: Algo Trading
Today is the age of AI which has made its way in every aspect of our lives whether shopping or investing. So, stock markets globally are no different. AI has taken over all the aspects of trading in stock broking and investment including surveillance, monitoring, compliance, and controlling price and volume of stocks. AI has been a major tool to predict trends and also in managing investment portfolios and selecting good return generating stocks. While large companies have been making use of AI for years to mine huge amounts of data together with not only stock performance but also corporate commentary, social media trends, consumer behavior, credit card trends, etc., the advent and rise of AI-based technology have set international stock markets in a new age.
Until recent times, stock market data and price movements were analyzed via quantitative analysis was but it was time-consuming and only a few could do it and extract meaningful information, so it was used only by a few major players like Goldman Sachs and J.P Morgan, which managed nearly 20% of its portfolios with AI. Now that AI is nearly everywhere and the barriers to entry have decreased, small-time brokers and startups have started looking to leverage this tech into building a new model for investors to pick stocks.
Let us create an understanding as to what exactly is AI and how it is being used in trading and analyzing stocks, and some controversy surrounding AI’s mass acceptance.
2.6 Artificial Intelligence
The simplest definition of AI was given by famous professor McCarthy (1950), dating back to the 1950s by Dartmouth professor Joseph McCarthy, which is a process of using software to mimic aspects of learning and decision-making so that a machine can be made to simulate it. Since the starting of AI, its applications have modified and process has scaled to accommodate growing technology. At present, modern world has stuck to the concept, AI is being used nearly everywhere:
Google’s Map application uses AI to predict traffic patterns to offer the quickest route and also tells about the precise amount of traffic and congestion one would find after an hour or two on any required route.
Online shopping at retailers like Amazon use AI to make price changes and product recommendations to meet customer’s demands.
Uber and Lyft use AI to determine fair pricing based on peak usage.
Banks use AI as part of their fraud protection and prevent identity theft.
Credit card companies use AI to determine whether a customer is eligible for a credit increase.
Every flight in the world uses AI-powered autopilot to steer the vehicle (humans only account for ~7 minutes of control, reserved for take-offs and landings).
Spam filters on your email sort out behavior patterns of junk mail and scammer tactics.
Plagiarism checks in professional and academic settings can quickly analyze papers for stolen or redundant content.
The list goes on. So, if you think that AI is something new and has come up in recent years, then it is important that you know that it is an old concept which was there for many decades though not in the same shape as it has evolved over years. Aspects of AI have been refined in recent years which have made it smarter to the current status, with machine learning and deep learning being popular buzzwords.
Machine Learning means training a machine to perform a particular action by programming algorithms set by a programmer. This includes things like recognizing a particular any patter of expenditure made by a person by her credit card and then indentifying any inconsistency to trace fraud. As more algorithms are added and data accumulated, the AI becomes more precise and capable of processing data to make better-informed decisions.
Deep Learning, which is almost like machine learning, is a process of training a machine to perform actions and become more precise over time. However, deep learning goes further, involving an approach that involves artificial neural networks—similar to how human brains learn patterns of behavior (for example, someone falls and you automatically extend your hand to help without any conscious thought). With researchers making new breakthrough in this concept every year, deep learning becomes a type of responsive intelligence that learns as it goes. Usually, deep learning enhances machine learning by being able to adapt to new data