Outsourcing Currency Management with Lumint

When foreign exchange transactions executed for mutual funds, pension plans, and endowments don’t occur at the interbank rate, investors and beneficiaries lose out. A lack of oversight and a lack of transparency mean the costs of surprisingly wide FX spreads go unnoticed and undisclosed.

To illustrate the point that seemingly sophisticated institutional investors are receiving remarkably poor executions on their currency trades, Lumint Corporation and New Change FX recently teamed to study the performance of hedged and unhedged share classes of UK funds. They concluded that owners of the hedged share classes are indeed losing millions of pounds per year due to higher-than-necessary foreign exchange execution costs.

Lumint was founded in 2013 by a group of State Street FX veterans to drive innovation in currency hedging, increase transparency, and provide the analytics needed to meet the evolving needs of asset owners and asset managers. Lumint offers institutional investors both currency hedging sub-advisory services and outsourced agency FX execution services. Lumint currently has more than $22 billion in currency assets under administration. It also offers its cloud-based software to others wishing to manage their own currency hedging activities. 

Lumint has raised a total of $4.1M in funding over three rounds, most recently a $575,000 seed round in March of 2020, from undisclosed investors. The firm has six employees and is based in Boston. Alex Dunegan is cofounder, CEO, and CTO.

Alex Dunegan

Q.  Alex, is Lumint a currency hedging manager with a technology company at its core, or a technology company that also offers currency management?

A.  When we were fundraising, it was whatever the investors were most interested in (laughter).

We’ve become a technology company at our core. We started out wanting to be an outsourced overlay manager. We wanted to do fully outsourced hedging for clients but knew we couldn’t do it properly, or successfully, or with the right margins if we didn’t develop the technology. 

But if you’d asked me in 2015 if we’d ever license our technology to anybody, I’d have said no. I’d have asked why we would want to give our technology to competitors. But I’m glad that we’ve morphed. 

While we started as a currency management firm that relied on technology, the way things are going, and with what’s in store for us in the next year or two, we’ll become more a technology company that has currency management experience and has that business. But technology will be the primary business line for us. 

Q.  How is Lumint regulated?

A.  Because we’re an outsourced overlay manager, and because FX forwards fall under swap rules under Dodd-Frank, we’re registered by the CFTC. We’re a registered CTA — Commodity Trading Advisor. We don’t have any of our own accounts. When we’re trading for clients, we trade purely as agents. We trade on their behalf as an extension of their trading desks.

The software business and the CTA are the same entity today. There may be some point at which we split those for regulatory reasons, just to make things cleaner. 

Q.  Are clients turning to you to mitigate risk, to lower execution costs, or to generate alpha from currency trading? Or all three?

A.  We always hope it’s all three. Something that we talk about a lot, and that I believe in from an institutional investor perspective, is what’s called operational alpha. We don’t run alpha generating currency programs (although we have in the past) but we think there is a lot of operational alpha to generate in terms of process. It’s not just cost savings, but actually running a better currency management program. Whether in hedging or just transaction execution, there’s improved performance to be gained. 

Most of the time it comes down to costs, though. Often, institutional investors will think or know about extra costs they may have in the currency space. But what’s changed for us in our market over the last five to ten years is margin compression across the fund management world. Individuals at those firms are being asked to do more with less. Because often times currency is a byproduct of their core competency, they usually don’t have experts in-house focused on currency management or FX risk. As an equity shop or a fixed-income shop, or whatever they are, they look at currency as ripe to outsource. Or to automate by buying our software. 

If it’s a passive currency management mandate or program within a firm, you start at a score of one hundred percent and can only go down. You’re not going to make any more money. Your clients expect it to work great and it really becomes a source of operational risk in an overall cost picture. In a normal build vs. buy type of evaluation it makes sense to outsource. 

Every so often we do see very acute risk management stuff as the key factor. I had a client when I worked at State Street and we pitched them every six months for four years, gave them research, and they had a big book of business but never moved on it until one day we got a call out of the blue and they were ready to hire us. We later learned that they’d discovered a seven-digit operational error related to FX and that was the catalyst.

So sometimes it’s cost, sometimes it’s risk, and — probably less often than I want it to be — it’s operational alpha.

Q.  Why should institutional investors spend client money on managing and mitigating the impact of currency volatility in global investments? For long-term investors, don’t currency fluctuations come out in the wash over time? 

A.  I learned most of what I know about currency management from a portfolio allocation or risk allocation perspective from Mark Kritzman at Windham Capital. This was something he would talk about. While maybe that’s actually true, most investors — particularly fund managers — don’t have a long enough time horizon. You probably have three to five years. If your track record stinks, you’re not going to make it, so why suffer what could be very significant drawdowns in currency and career risk in the meantime?

The area we focus on mostly is passive currency mandates for asset managers, but on the investor side of things. It’s a niche part of the financial world and differs from your question where you’re talking about the international investments that are generating an additional currency return stream that adds risk or diversifies, depending on what’s going on.

But the area we’ve spent a lot of time on because it’s so operational is the investor side. Fund manager A is running their international equity fund. They think about FX the way they think about it, or don’t, but that’s part of their portfolio management process. And maybe they offer that fund to U.S. investors. But then along comes a European investor who says, “I like this track record and I want to invest in it, but I want the same returns as that U.S. investor, so hedge my currency risk between dollars and euros.” It doesn’t actually have anything to do with the underlying investment, they just want to replicate that return. It’s called share class hedging most of the time, but it could be feeder fund hedging in the hedge fund sense. But it’s a purely passive scenario, there’s one currency risk that needs to be hedged and the math isn’t hard — it’s 100% hedged. Where we focus is on making sure the operations are smooth. What’s hard there is the replication, making sure you’re dealing with fund accountants and transfer agents and that it runs on a tight work-flow schedule day-in and day-out without issue. On that side of things, the real goal for the asset manager is to offer that replicated return to that European investor. Because of that, they need to hedge that very, very strictly. Here, it’s not a question of if the returns wash out over the long-term; it’s just part of the mandate they have from that client.

That kind of thing is definitely ripe for outsourcing because it’s not part of the portfolio management process at all. It needs to be done usually on a daily basis (depending on data), or a couple of times a day, and it can only hurt you from an operational risk perspective. It needs to be done all the time regardless of macro-economic changes that impact currency rates. That’s really where we focus, because it’s a steady business but also because it has a very specific rules-based approach that suits software well.

Q.  How is it possible that sophisticated institutional investors aren’t receiving transparent pricing and best execution on FX trades?

A.  This has been THE hot topic in the small world of foreign exchange for 20 years now.

For a while I thought this was really irresponsible of institutional investors. But what I’ve come to realize — and this is a strong argument for outsourcing too — is that often times it’s just at the bottom of the list. For example, as a fund manager, I have a fiduciary duty to my clients. But at the top of the list is making sure I’m not overpaying for my short interest rates or making sure I’m improving my program from a diversification or alpha generating perspective. Currency, from a magnitude perspective, is just lower down on that list. When there isn’t time or resources at a firm, it just falls down the priority list and never gets addressed. And that’s a problem, but it’s also a strong reason to give it to a specialist because it’s usually not worth it financially for a firm to have their own specialist in-house. At the small to medium end of the market, where they’re going to have the worst pricing, they may have an agreed pricing model, but it would still be much higher than what we would think is a fair price.

Q.  But isn’t it the responsibility of a plan trustee or fund trustee to ensure they are getting the best price? Where are these folks?

A.  It’s been an uphill battle, but one of our core goals is providing visibility into that. One of the things you hear talked about — I’ve heard it talked about at every FX conference I’ve been to in the last five or six years — is FX transaction costs analysis. Plan sponsors that are used to seeing equity TCA reports and evaluating them across their managers can receive FX TCA reports. But they’re all a little different, there’s no central exchange, there are a lot of other little considerations that come into play in terms of where you can actually execute — do I have enough credit lines to trade at multiple banks in competition or do I have to trade with a single bank, maybe the margin rules are different. One of the things we’ve really tried to do with our analytics is push those considerations and the impacts of those decisions on pricing to the fore so that clients can make those decisions.

Our challenge as FX market participants is to do a lot of educating. Part of what we were doing with the New Change study is to say that we can draw some broad conclusions regarding what you are paying for FX and highlight what you need to be working on.  I think that’s where you’ll see the plan sponsors say, “Now I have something. Maybe I wasn’t certain before, but now I know.”

Q.  By relying on exchange rates provided by custody banks, are investors subsidizing other clients of those banks? Paying indirectly for other services? Fattening bank margins? Where is the money going?

A.  It’s a good question. You would have often heard that argument during the pricing scandals before 2010. “Big Pension Plan X is suing us but really we have this handshake agreement that they were going to give us FX flow and we were going to give them low custody bank fees.” There’s probably still some of that, but most of the time now it’s just a lack of awareness of what the costs are, or how those costs accumulate.  And in many cases, for most custody banks, the jumbo spread costs uncovered by the standing instruction pricing scandals are a thing of the past.

Oftentimes, I think, people look at it as a cost of convenience. They go to their custody bank and their fund admin who might be providing nine or ten services for them, and just check the box for a one-size-fits-all currency management service and say give me a pricing arrangement. Honestly, for some firms that may be fine. They may pay higher than average, but then they don’t have to have an outsourcing agreement with someone like us. They also may not care about tailoring their approach for an optimal outcome.  They can look at it as a fair price for convenience in some cases. I don’t look at it exactly that way all the time, but typically when you have a small firm that doesn’t have the resources to do anything else, that may be a premium they pay when they’re small and they hope to move onto something better when they grow. 

On the other hand, hiring a custody bank like Northern Trust, which uses our software, or any custody bank, really, that provides an automated and tailored service at a fair price, opens the markets for institutional clients to get the convenience of working with their custodian without sacrificing on customization or price. This is something I think about a lot, in terms of corporate strategy. Namely, our advantage of scalable tech and efficient process won’t last forever. The custody banks still using obsolete software and processes today will not sit idly by and we need to keep innovating to stay relevant.  

Coming back to your original question, I still think the root of it, when a client is being overcharged, is just a lack of awareness and visibility. In some cases, they may actually know, in fact, but they can’t get the data and don’t have a dedicated TCA provider to evaluate it for them.

Q.  You’ve made a strong case for why firms should outsource this. Why don’t more firms delegate FX management? What holds asset managers back? Is it resources?

A.  It runs the gamut. At some firms it’s a personnel thing — there’s someone whose role it is to do this, and that person makes the decisions around it. I can’t go to that person and say outsource your job to me.

 At others, it’s a lack of awareness because they can just check the box with their principal dealer to handle it for them. They don’t know they can get an agency-based solution, or a solution tailored to them by their custody bank or other specialist provider.

Sometimes, too, it’s part of the firm’s operational synergy. Funds of a certain size may have a lot of FX activity from transactions that they manage in-house, and so they want to lump all their FX transactions in together. If they have an internal trading team it can help them to do it all in aggregate from a pricing and execution standpoint. We’ve seen other firms retain it to reduce their margin requirements.

Small firms or startups may find someone like us or maybe just go with their prime broker or custody bank. The vast middle that doesn’t have the resources and wants to outsource because they know that it’s impactful. Then there’s the top, where the very largest firms have the resources, and it can make sense to insource for cost savings. An example would be a superannuation fund or big pension fund in Canada or Australia that has the resources, the expertise, and the scale to do these things in-house and do them well.

There are some big firms that don’t want to outsource but may not have the specialized technology to help them do it. With our software offering, whether you want to outsource or insource, we have something that hopefully can help you. 

Q.  What are the risks involved in outsourcing?

A.  Some people think that when they outsource FX management, they no longer have oversight responsibility. But in fact, particularly in Europe and in the UK, that oversight responsibility goes up in terms of monitoring compliance with the hedging mandate and the daily responsibility of making sure transactions are done. So there’s sometimes additional but at the very least the same responsibility for managing the oversight program.

There’s operational risk in outsourcing in that you carry the operational and reputational risk of that firm you’ve outsourced to; there has to be a due diligence effort that is non-trivial in today’s environment. Cybersecurity is a significant part of the due diligence process we do. We’ve created a questionnaire with almost 1,000 questions on cybersecurity and operational security. The onus is on the hiring firm here.

There’s also a bit of a risk in the handoff. Whenever we talk to a firm that does it themselves, they always say their process is one, two, three. But when you peel the layers back, it’s not really one, two, three, it’s one through fifteen. We spend a lot of time onboarding in parallel to be sure we understand what all those steps are.

Q.  How do you suggest clients evaluate the effectiveness of their outsourced currency management program?

A.  We like to provide a full parallel run at the start. We know there are competitors that will show a demo that looks like a fully automated, scalable system but it’s just a façade, and on the back end there’s a spreadsheet.

By doing a live parallel, we accomplish several things. It’s transparent and demonstrates the process and the workflow to the client. They know that they’re seeing an apples-to-apples comparison. It also gives them comfort that their mandate is being captured and run properly.

Most currency managers will say they’re running the best program for you but don’t have the analytics to back that up. But that’s been core to our process from the beginning. We show clients exactly how the hedging program has performed, exactly how risk was managed and mitigated in their portfolio. All these things are important to demonstrate to clients so they can fully understand and evaluate us, but it’s also part of a feedback loop that allows us to continuously improve the program.

Q.  Is the decision to delegate foreign exchange trading different for active versus passive managers?

A.  Yes, passive is the leader in terms of delegation. If you look at ETF managers that are trying to track MSCI World Hedged or FTSE Hedged or Barclays Global Aggregate Hedged, those currency management programs are very rigid in terms of structure so that they can track their index as best they can. Those managers live and die by low tracking error so they are very apt to outsource to specialists who can structure a very specialized hedging program for them.

Q.  Who is more likely to outsource currency management, an equity or a fixed-income manager? Are managers based overseas more likely to outsource than managers based in the U.S.?

A.  Fixed income managers are generally more focused on currency management because, historically speaking, from a proportion of return perspective currency has a bigger impact on fixed income returns than on equity returns.

That’s changed a little bit as awareness has grown. The financial crisis resulted in a generally greater focus on transparency. Whether an error resulted in a gain of 100 basis points or a loss, institutional investors now look at it the same way. Equity and fixed income managers alike are today more aware of currency risk and currency impact. The result has been more outsourcing by both.

For a small to medium sized firm that’s based in Europe, because there’s so much more cross-border investment activity there, and because they don’t have nearly as much of a “home” investor bias as U.S. investors do (true of Australia, as well), there’s a greater reliance on international assets in general as part of their overall allocation, making it a riper market for us. 

Q.  All in — meaning technology and systems costs, headcount, and including better execution prices — what can Lumint save for its hedging clients? How do you measure it?

A.  It always depends on what the assets are. It could be that we’ll save you, at the very least, the price differential versus what you’re paying by doing it in-house or using a custodian with an obsolete process — three or five or whatever basis points off your transaction costs. 

But on the other end, we’ve seen where running a program we’ve designed or run through our platform for improved speed to execution can mean hundreds of basis points saved. That could be over time, because there’s accumulated slippage from executing slowly, or sometimes it’s because of acute issues where a firm that had a systematic process in place would make a mistake and give up a couple hundred basis points in a single day.

So, the answer really is it depends. We’ve spent a lot of time building analytical tools to put that question in context for clients. It’s a different answer for a commodity fund versus a fixed income fund. It’s different if we’re talking about switching from a competing currency manager versus going from doing nothing to hedging. But I’m still looking for that one score that would make it a lot easier to compare (laughter). 

Q.  Unusually, you are both CEO and CTO. Any thoughts of splitting the positions?

A.  It can’t happen soon enough! We were hoping to move on a CTO hire in the midst of our latest fundraise in March/April, but with the pandemic — assets dropped in March and we get paid on assets under administration — we decided to be very conservative with our cash. The current plan is that this is a Q1 hire for us. We’ve got a couple candidates identified and I think we can move quickly now. It’s high time to split the roles.

Q.  What made you select AWS to host your platform rather than one of the other cloud providers?

A.  One of the things that led us to AWS was that our team was a bit more familiar with it, but there were other factors. One is we use Heroku for platform as a service capabilities and underneath they use Amazon Web Services. They’re a layer on top, so if you need to scale your servers 10X you hit one button as opposed to doing something more complicated. That was a great model for us, especially without a dedicated CTO or a development ops team in place. We paid a premium for that, but it made it much easier to scale up production cloud applications. We may move down to direct AWS at some point in the very near future and take Heroku out of it just for cost savings, when we hire a CTO.

In general, we like AWS because of the broader set of cloud capabilities they have, and it’s a bit more full-service, although we do go back and forth because we run a trading application that is very time sensitive for execution. AWS has had outages, as they all have. If that odd outage hits at the wrong time that’s going to be a significant problem for us even if it’s only out for 20 minutes. We have talked a good bit about whether we need to have a duplicate version with Azure, or whatever. It’s a big lift to do that, so we’re not sure about that, but, more and more, redundancy will be built into our process

Q.  What is the role of machine learning in your offerings?

A.  We introduced a machine learning component — the Robotic Oversight System (ROSY) — in 2019. The main goal is to further automation. The general workflow in passive currency management is you take in a bunch of data on that client’s account whether it’s custody data listing all the holdings in a portfolio and the currencies associated with those holdings, maybe it’s a share class net asset value that you know is euro and you hedge it back to dollars, or new subscription and redemption activity. All this data comes in, you apply some static rules to it, i.e., hedge 100% for one month forward. You rinse and repeat and you true-up the hedge as you go.

One of the challenges is that even if you manage to automate the data in-take piece, because there’s data that’s coming in from varied sources there may be issues. For instance, one thing that often happens for subscription-redemption activity into a new fund is that an investor may cancel the subscription at the last second, or maybe they slide a subscription in at the last second. Some transfer agents still use fax order forms, so you may get something with an extra zero added that didn’t get corrected. It’s very much going to be garbage in, garbage out, even if you can do some validation up front. 

The challenge we always encountered, and I’m sure our competitors encounter, is that you still want a human to come in and check before anything happens and say if things look right and look qualitatively okay. Yes, I’ll trade it or no, I need to talk to somebody about it. We spent a lot of time automating that intake piece and doing the validation — is there duplicate data, did we get a PDF when we thought we were going to get a CSV —standard things you’d expect. But we’d still been doing that human check at the final stage before we hit transmit for execution. We’ve started to use machine learning to help us make that determination, that it looks okay.

Instead of a rigid checklist, or a static threshold, now we use adaptive machine learning models to say given the account, the account’s history, the currencies involved, the assets classes, then we can make a determination in a very context-subjective way for that particular account or transaction, whether or not something looks unusual. We use machine learning to help identify that on our path to fully automating the hedging cycle. We are already automatically managing data. We want to be able to automatically do that human-esque qualitative review and then automatically trade. That’s where we’re going.

Q.  Are you selling your foreign exchange solutions to asset managers and asset owners directly, through Northern Trust, or both?

A.  A partner like Northern Trust, our keystone client in this space, has a global sales team, which I’m not going to have, maybe ever. They’re out talking to clients all the time and — even better — they’re selling to asset managers who have their own sales team out there raising assets so there’s a little bit of an exponential growth path there which suits the business really well. 

What’s surprised me as the company has developed is that the attraction to our software offering is far greater than I expected. That’s where we’ve really found our sweet spot. I can envision a time when we’re not actually selling currency management services to anybody because our partners would be the ones out raising the assets.

Today, though, we still do a fair amount of direct selling. We focus on smaller accounts that partners such as Northern Trust are probably not calling on.

Q.  What can you tell me about your partnership with New Change FX?

A.  As FX trade executors, we think it is definitely best practice to not have our own TCA. We can’t evaluate ourselves and have our clients assume that’s a valid benchmark. 

I’d known Andy Woolmer at New Change FX and had met him a couple of times. What Andy and I started talking about was a partnership with a couple layers. One was we could use their TCA to create a really great technical API for TCA. A lot of FX TCA stuff is you report your data out and it comes back tomorrow, or if you’re working with an incumbent TCA provider it might come back once a month and it might be a PDF. That might be fine for an asset manager or a pension plan that doesn’t need to look at these things any more frequently than on a monthly or quarterly basis. But we need that real time input to feed our analytics, to feed our trading process, and that’s a core part of how New Change does TCA, so working with them really made a lot of sense for us from a TCA perspective.

Then, from the New Change side, they’re out talking to clients all the time about FX TCA and execution. If there’s an issue, or someplace to improve, the client is asking New Change how to fix it. So New Change was looking for a solution to that. We can go to their clients as two distinct firms, but we can have a combined offering. They can say, look, this is the firm we can recommend that can help you improve the deficiencies you’ve seen in your program. And for us, their TCA platform has been tremendous.

Q.  How did WFH and COVID-19 travel restrictions impact Lumint in 2020? It appears as though you were pretty active on the product development front.

A.  It’s been very productive. I moved to Brooklyn several years ago and our team was spread out along the East Coast already, and we would all get together periodically in Boston. As a result, the remote working setup was seamless for us. 

 It did impact our prospecting. We weren’t flying around the world, of course, but there was a significant uptick in prospective clients looking at this. As the dust settled and people got used to working from home, people did a lot of institutional tidying up. Some of that has been addressing traditionally back-burner items like FX. As the year went on, prospecting really picked up much more than we thought it would and much faster than we thought it would.

But that middle time allowed us to get a lot of work done, not least of which is the New Change analysis which my colleague Ashish Agarwal worked on. And we continued to invest a lot of time and resources in growing out the platform. We’re really excited about the stuff we’ve been working on and our plan is to announce new partnerships in the coming months.

Q.  That leads us to my next question. What can we expect to see from Lumint in 2021?

A.  We’re finalizing the paperwork on some exciting announcements! Some of those will look like the partnerships we have now with Northern Trust and New Change, but some will be a little different. We’re planning to move into the corporate hedging space as well. It’s long been a goal of ours to work with corporate treasuries that may not have access to customized programs or that have a lot of data to wrangle to understand their exposures so they can make decisions about hedging them.

And then continuing down the machine learning path. From a business scale perspective but also from an efficacy perspective, end-to-end automation will help us stay in front of our competitors.

Something folks aren’t always aware of is that they might think they have a good hedging program, but they may get data today and not act on it till tomorrow or get data this morning but not execute until the end of the day. Those hours add significant risk to their process. At Lumint, we can shrink that lag down to ten or fifteen minutes, and I want to get that to be totally seamless, including for data that comes in overnight. With machine learning, I think we can get there and ultimately it will be a better outcome for clients and their hedging programs. 

# # #

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.