hey guys, what do you think about the future of finance in general and if junior bankers really face a real threat from technology, coding, automation etc… Not only speaking for IB but for most front/middle/back office roles in finance.
| +48 | Interviews Are So Fake | 29 | 21m | |
| +33 | 2024 UK Election - Tories finished? | 21 | 1d | |
| +28 | ADHD ! | 11 | 1h | |
| +28 | Being Christian in investment banking | 14 | 2d | |
| +26 | Non-Competes Banned | 28 | 1h | |
| How do I become Sigma | 15 | 16h | ||
| +19 | Moelis has the cutest Analysts? | 4 | 1d | |
| +18 | Best NYC neighborhood for single 30M | 12 | 3d | |
| +16 | Underage intern, drinking? | 7 | 2d | |
| +13 | Secretive vs Universal Prestige? | 7 | 1d |
Career Resources
Get instant access to lessons taught by experienced private equity pros and bulge bracket investment bankers including financial statement modeling, DCF, M&A, LBO, Comps and Excel Modeling.
They'll always be a job for investment bankers.
There’ll
Most things that can be automated away are not things you want to do anyways. They tend to be repetitive, basic tasks that aren’t enjoyable for human beings. If you want to be the guy who sends out the same spreadsheet report everyday that a machine can do, you aren’t aiming very high. Aim for something that humans add value to, and those tend to have the benefit of being more enjoyable as well.
Currently living in Shanghai and right now the CCP is making a lot of investments in AI and encouraging more start-ups in that industry. I was watching Msnbc (I think) and one of the CEOs from a software company in China essentially said that if your job has "analyst" at the end its going to be obsolete in the future 10 years or something long those lines...and this was like 2 years ago when he said it.
You really have to be able to combine analytical thinking, innovation, and real creativity in the future to get any kind of job in any industry in the future.
How did you end up working in Shanghai? Are you a local?
No. I attended a university here studied Economics graduated a few months ago. I'm looking to move to the UK soon actually.
True, but on the other hand startup CEO's are a little gung-ho about the future impact of their products
Short answer is yes.
Longer answer is no, kind of, the key job of an investment banker is relationship management. While this provides job security from automation, at least in the middle term, the problem it creates is that like most jobs you have to train the next generation to actually do the grunt work so they understand how to relay their value to clients. What worries me is that companies in their rush to add next quarter value to boost their bonuses they will shoot themselves in the gut by litterally gutting their junior workforce.
But don't think this only impacts jobs like IB, there are software engineers out there who are trying to put their own kind out of work. Any idiot can create a website these days with drag and drop templates. 20 years ago you needed highly skilled front end engineers to build basic websites.
I am saying this as someone who owns and founded an AI finance company.
Wow, I didn't know you did this. What's your AI company do? Trading, or something more unique? Are you a technical guy or on the business side of things?
We work on valuations for highly illiquid assets. I am on the business side of the company but I understand enough about the technical side to be far more informed than business people.
I didn't realize this till I moved into analytics and started teaching myself some basic programming, but automation is absolutely going to have an impact on a lot of professions in the not too distant future. Even with my very low-level knowledge of programming, I have been able to automate a lot of my workflow that I used to do in corp fin/accounting, so I'm sure people who do this full-time can do even more than I can, and much more efficiently.
As such, here's my prediction w/r/t automation:
1. There will be a material amount of entry-level work that get replaced by machines/coders going forward. A lot of entry-level work in Corporate America is basically just data aggregation, cleaning it (often the most time consuming part and why we need so much manpower), and then cutting it in pretty standard ways to send to people for interpretation. A python script can take data from multiple sources, aggregate it, and then slice/dice it however someone wants it infinitely more quickly than a human can. This is also a very scalable solution in that as datasets grow exponentially larger, the time to process only slightly rises, assuming it was scripted efficiently. I don't really see why companies will hire 3 people as "financial analysts", for example, when they can just hire 1 person, pay that person more than they would as one of those 3 financial analysts, and have that person play with Python/PowerBI/Alteryx, etc. to ETL data all day.
2. The decrease in entry-level workers will be at least partially offset by a need for people to either maintain these systems/automated solutions and to interpret data in an impactful way. Even ML, from what I understand, is glorified stats/probability and needs to be trained by an expert, so there will still be a need for humans to interact with machines to generate positive results. It's not a complete Skynet type of scenario where humans will be extinct. I see it as more so machine-enhanced humanity (whether that's good or has ethical issues is beyond the scope of this thread, so I won't opine on that) than absolute displacement.
The implications if my prediction is correct:
1. People need to learn how to get more tech-savvy. I see a lot of people in finance/accounting that basically think it's not their job to fiddle with data, but unless you're good at managing people or are one of the most knowledgeable people in your domain where a company can't afford to lose you, you might find yourself out of a job soon (within 10 years IMO).
2. There's going to be even further inequality from automation and we need to either equip people to be the automators rather than the automated or figure out how to help people that are going to be let behind by this shift - probably both. Think about the scenario I mentioned earlier where 3 financial analysts were previously hired to do the work of 1 in the future. The winners in this scenario are the company (more efficient and more profitable from shedding payroll) and the employee who was hired (gets paid above what the previous norm was). However, the 2 people who were automated away are pretty SOL because they invested in a career they thought would give them a middle class life, but now either have to re-tool (potentially expensive nowadays) or are going to be stuck doing menial work that pays a lot less.
Curious what "data cleaning" means to someone that's not in tech. To me it's just running a script to get rid of nulls, outliers, etc.
It's basically taking messy data and putting it into a usable form for analysis, which is the most time consuming part of pretty much any data job. Messy data doesn't just mean nulls, outliers, etc., although those are included in the process (I'd argue outliers are more "data science", but I digress).
An example that someone in finance can relate to is if you're trying to recreate financial statements, you'll often get transaction data from a client/target in multiple excel files (i.e. one for each month). Within those files, the columns could change but mean the same thing (i.e. maybe in 2019 they had year as "Year" but in 2020 they changed it to "Fiscal Year") and you have to account for things like this when aggregating. On top of that, you may have dates listed as 08/01/2021 for certain files/periods, but have them broken out by month, year, date in separate columns in other files. To do your analysis, you'd obviously want to have one or both types of dates for whatever reason, so you have to transform this data to fit your requirements.
That's a very rudimentary example, as it can get way more complicated than that, but the overall point is when you get data from multiple sources, systems, etc., things will be formatted/created in various forms. Data Engineering is such a big field because data is going to become more and more prominent going forward and they need people who can design systems/processes to extract all this data from various sources, transform it so that it's usable, and then "load" it into reports for end-users (hence the acronym ETL - extract, transform, load)
I agree with a lot of this. Most of the entry-level reporting jobs will eventually be replaced by tech, but there will still be a need for finance people who (1) can interpret the financial results and make decisions (ie more senior levels) and, (2) junior employees who can create ad-hoc analyses that are difficult to standardize & automate (ie strategic thinkers). Net, the amount of entry-level corporate finance roles will be reduced in the future and will be replaced by people that are IT-like.
This is the key. I've unfortunately been a part of a 'strategic direction' group at my former role and it encompassed outsourcing/automating reporting/streamlining the aggregation of data for strategy-driven thinkers (analysts to c-suite) to interpret and make decisions off of. Forward thinking/offensively focused individuals won't be impacted by this, but definitely back-office/reporting roles will
Bureaucracy is like a gas that expands to fill the size of its container. The container is the practical constraints imposed by technological limitations. If new technology makes it easier to produce paperwork, more paperwork will be required so that the overal amount of human work is maintained.
I've had to endure years of the Kool-aiders here telling me "IB will NEVER be automated, Its a RelAtIOnshiP dRivEn BusINeSS". Most of it clearly can. And more obviously it can be offshored. That said, I work in one of the most inefficient markets out there, Ive partnered with fintech companies (one which has been gradually bought over by more reputable fintech names over the years) to try to merely aggregate data (market data, order data, cash data, whatever), and its just too difficult to be perfected. and also too many changes to keep up with. oh new investor in our fund that has x, y and z requirements? how was that going to be coded in to the investment strategy? oh new investors wants a and b ESG methodology? better get damn automating. it gets to the point where you spend the same amount of time processing the work manually as you would hiring a nerd, I mean a coder, to 'automate it'. I brought this company i'm talking about on to this new place I work at, and its taken 4 months (still ongoing) to get the basic requirements built in to the software solution they offer, I'm still having daily half hour calls with every stakeholders that wants/needs to see certain data in this system - for example compliance would argue over segregation of duties, which disrupts the workflow of an investment etc etc . and the financial markets are not just a case of equity, fixed income and advisory - there is an infinite amount of markets out there, some very complex, which will never be fully automated - thats why I need to roll my eyes at people that start replies with "AkshuAlly ItS vErY sIMpLE".
Lol @ this shit about automation.
There's not enough people to fill 14.50/hour jobs.
Doubtful.
The analysis and data modelling in traditional finance is far too bespoke to be chucked away by any automation models / programs.
There are so many micro decisions about what assumption to tweak, which line items matter for certain industries, how to interpret a company’s story / strategy (an NLP model can’t just glean what is BS or not, you need human intuition for that), how to “massage the numbers” to fit a narrative, etc etc.
It’s the same in law. Even though contracts are / should be boiler plate. They almost never are. Over the course of any legal client work you will be going back and forward with in-house counsel / management to understand what terms matter and why. How to obfuscate language that favours those terms. The exact terms themselves.
Really, the only fields that have to be concerned about automation are routine white collar jobs like those in back office like settlements, reconciliations, onboarding, support / client services (at least low level tickets not major ones). Most other things are just too reliant on dynamic human judgement and interpersonal liaising to standardise into a model. Not even a sophisticated ML model trained on real conversations could get any of this stuff right - at the end of the day, analytics is just automated processing and pattern recognition.
Where I see this heading is there will be a BUNCH of workflow / operational task software targeted at fields like financial services or consulting or law for routine data collection and data processing. Also, likely a more central piece of software to iterate faster on doc reviews between higher up the food chain professionals and lower down the food chain professionals. Basically, enablement software. Like Salesforce for sales teams or Zendesk for support teams.
Automating high fidelity professional fields would literally take solving the artificial general intelligence problem to get right. People in general I think put too much faith into technology without really knowing what it does or what’s it’s useful for.
Depends on what role. I work in LO AM and only know one guy who knew how to code (R). He's quit now and there's no one on the I-team who who codes (at least for anything on the job). As growth investors, a company will look substantially diff in 5yr vs. today -- there are limitations to modeling / data read-thru and most of the important data is just not available easily. You can only get so much via the internet / filings, most of the important insights are found in expert calls / creative thinking and analysis that is just not modelable today as scenarios for each company are totally different (then you have geography / regulations / politics / etc which muddle the picture further). We are custom building Ferraris vs. mass producing Toyotas -- now both can make money, but the former has far less standardization and thus far less to automate
Investment Banking will be less affected as it is relationship driven. At the end of the day, CEOs want to make deals on the golf course not via some MIT nerds programming analysis that got sent to him on valuation. Trading will continue to be automated however more esoteric products such as CLOs, Mortgages, Leveraged Loans, exotic derivatives, interest rate products will continue to thrive. The boring repetitive tasks that most bankers complain about, rightly so as they should not have to wait hours just to receive comments where they added in very little and wasted their night, should be automated eventually.
Possibly, but from what I saw during the summer: many tasks are very difficult to automate. Not impossible but will likely take time or not be automated. There are a lot of things that are very customised, the way data is presented, the inputs etc. Lots of MDs, VPs want things their own way and there are a lot of nuances. Even when you are using CapIQ, Eikon or Bloomberg there are inevitably going to be pieces of missing data or data that needs scrubbing. The AI or technology would need to be super advanced to replace it and I think it is is possible in the longer-term but imo I dont see it happening in the next 5-7 years.
I feel like a lot of these pro-automation hot takes come from people who have never worked inside an IB or have any idea what it is we do all day. As if crunching models and PPT is our sole raison d'etre.
Et aut ducimus iusto maiores nam. Ea veritatis vitae laboriosam perspiciatis beatae.
Autem at eius doloremque quos sed fugit. Velit eaque consequatur rem delectus temporibus qui. Eum eaque qui eos tenetur incidunt. Illum omnis qui debitis ipsa quisquam amet debitis.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...
Rem culpa in et eos magnam dolorem ea esse. Qui aliquam officiis exercitationem nobis.
Voluptates laboriosam a vel id ut officia. Soluta sint rem ea libero quam. Rerum officiis et nihil. Facilis maxime quidem voluptas quia voluptate aliquid.
Omnis reiciendis eum quis ut et officia. Sed aspernatur eos laborum sunt voluptas. Voluptatem dicta facilis rem voluptates repellendus voluptatibus eum numquam. Vero laudantium quibusdam cumque vel beatae. Culpa et delectus distinctio aut ad dolorum.