Investment bankers are fond of saying that the task of pricing and allocating an initial public offering “is more art than science”. It's also often surmised that artificial intelligence won't be able to find success in art because it will never be able to replicate human creativity.
That might be wishful thinking, but assuming it's true, does that mean that AI can never replace the role of bankers in IPO pricing and allocation?
The answer is probably not.
Capital markets and investment banking are driven by data and data is growing at a massive pace. This suggests that the industry is poised to benefit directly by applying AI, which can filter, process and analyse data at a pace human beings can never reach.
And yet, investment banking has been slow to embrace AI. The business remains heavily labour-intensive and decisions are made mostly by a small group of senior bankers and advisers.
Th e lucrative salaries they earn perhaps helps to explain their reluctance to accept AI. After all, AI may cost only a fraction of their packages yet be able to deliver more efficient results in certain investment banking functions.
To be sure, there is a long way to go before AI can fully replace human investment bankers because bankers are currently irreplaceable in some functions like on-site due diligence, client presentations and dealing with stock exchange enquiries.
But there are certain functions that AI could gradually replace them in. IPO pricing and allocation is one of them.
THE TASK
Investment bankers say pricing an IPO is an art because underwriters have to strike a balance between achieving the maximum possible proceeds and ensuring the share price doesn't plunge immediately after trading begins.
These two elements are, for the most part, contradictory because the higher an IPO price, the more likely it will plunge. While the company typically wants to find the highest price at which investors will snap them up, they would also want to avoid the embarrassment and potential lawsuits if the price drops precipitously.
IPO underwriters, to some extent, minimise such a possibility by allocating the bulk of the shares to institutional investors that they believe will hold the shares for the long term.
But in practice there is no guarantee that these “long-term investors” won’t sell the shares immediately once trading starts.
Bankers would argue that they use their collective experience, as well as study the track record of investment funds, to determine the likelihood of them holding shares for longer, before deciding how to allocate shares.
APPLYING AI
Intriguingly, data analysis of such a kind is exactly what AI is good at. Machine-learning algorithms help with this task by analysing historical trading data, in bulk, and coming up with data sets that reveal the trading patterns of any particular mutual fund or hedge fund.
These reports can even show, in detail, customer classification by product characteristic, product interest, or any other kind of input.
Needless to say, these data sets are deemed more accurate than the judgement of bankers.
One of the reasons why AI is most suited to assist in IPO pricing and allocation is that there is currently no regulation governing how IPO shares are allocated. After all, the decision is made at the full discretion of the company owners and therefore the application of AI is easier than in other regulated fields like corporate due diligence and client data management.
Last year, British auction house Christie’s sold the world’s first piece of AI-generated artwork for $432,500 – almost 45 times higher than its anticipated price – in a clear sign that artificial intelligence is developing into new areas much quicker than ever.
As such, the day of applying AI in the art of IPO pricing and allocation may also come sooner than expected.
