Dean Chisholm is regional head of operations for Asia Pacific at Invesco, an asset manager with over $900 billion in assets under management worldwide. He has been an outspoken proponent for automation in Asia among asset managers, distributors of funds and custodians for many years. Chisholm sat down with DigFin at the firm’s regional HQ in Hong Kong.
Data, building data lakes…priorities for your industry, but what’s really important here? What’s at stake?
Dean Chisholm: We have to provide data downstream to clients. We need a lot more information per transaction than we did before. Our analytics team needs a complete set of data, and it has to be cleaner. Our traditional systems won’t have a full set. So we as an industry are working to vet our data before it gets inputted into our systems.
How much data are we talking about?
Analytics draws on more data than a trading system. It goes into reports and understanding things like performance attribution. There are some vendors helping with this, but the entire industry is struggling.
Why, what’s so difficult?
It’s no longer just about accounting data. We’re adding unstructured data, such as company announcements that get reported in the press. We are using artificial intelligence to determine whether a report is positive in tone or negative. A.I. tools can analyze documents and reports down to the sentence level to understand that.
How confident are you that this information is true?
We try to reconcile it against other sources.
But can it be a tautology – is it from the same source, just in different words?
That’s what we’re trying to uncover. We have to check prices back to make sure they’re not all coming from the same feed. You have to rely on common sense. It’s easy to check the prices and yields of fixed-income instruments for a distortion.
Then you need to store this stuff, put it all somewhere.
You need a new generation of databases. If you’re working from a five-year old database, you’ve failed. There’s been a massive jump in technology over the past five years.
This reconciliation or checking of data – can you buy vendor tools to do it?
There are some tools that you can get off the shelf, but you need to add a lot of proprietary information. Now we are entering the realm of data architects.
Are there a lot of these data architects around?
Skills are in short supply. It’s a new area. This isn’t the I.T. of 20 years ago.
Has anyone cracked it – you guys, BlackRock, Fidelity?
If anyone’s solved it, it’s a miracle.
Where do you find the talent?
We need to acquire it. Some of it is at the banks, but of course they’re ahead of asset managers: Citi employs something like 200,000 people, while Invesco has 7,000.
That said, India seems like a good place to recruit those skills; there’s half a dozen cities full of technologists. And our own culture is attractive, in that it’s less structured than what you’d find at a bank. We recently added a data-quality group in Hyderabad, and they’re already making a difference. I see it in the quality of our fact sheets, for example.
Fact sheets? That’s a priority?
Take a large fund house with a range of, say, 70 funds, each with seven or eight fund classes, being sold in 15 countries across different distributors. Each of those needs a monthly fact sheet. And your neck is on the line with the regulators if you have errors due to bad data.
Oh. Okay. Let’s switch gears. What’s the vision for tech at Invesco? Because while you’ve convinced me that fact sheets are very important, it still sounds like firefighting rather than business transformation.
I think it starts with firefighting, but evolves to delivering solutions across different mediums. Some information can’t be released in real time because of regulation. But that’s not a problem with discretionary accounts, where we can give clients a straight look-through to the portfolio, rather than waiting for daily pricing.
So you create new possibilities with product and ideas.
Trades didn’t use to settle until T+5 in some cases. But if a client can have a real-time look at their portfolio – and this is still a couple of years away, depending on the asset class – you can configure downloads according to what the customer wants.
That allows the client to aggregate their information, whether they are an individual on their smartphone, or a sovereign wealth fund. And aggregation for big institutional clients is hard, because they have so many managers and products in their portfolio.
You gain a service edge.
It would let people see which funds actually add value. It would help financial advisors analyze performance.
Last year, Invesco acquired Jemstep, a Silicon Valley-based digital financial advisor. How does that fit into this picture?
Jemstep helps us sell product to mid-tier distributors in America. Their robo-advisory model offers a mix of Invesco and third-party products. It’s not a threat but an aide to our existing business.
And are you bringing it to Asia?
How and when we bring it here is an open discussion.
Based on…?
It differs by market. You have to look at your core proposition and what you want to be in a given market. In some places, our distributors already have a robo advisor. In others, the industry hasn’t become so closed, and a fund house could go in with a new proposition. And in some countries that are adopting electronic payments, this becomes easier. But there is no one formula to fit all markets.
We are evaluating some opportunities. Once we reach agreement, I think it’ll happen fast.
What’s the hardest part about introducing digital advice?
The challenge is the cost of customer acquisition. This is why direct business is so difficult in a place like Hong Kong and other developed markets. But in emerging markets, wealth management can go down the income scale, go down the asset scale – provided there are a lot of customers you can acquire.
Thank you, Dean.
