The Swedish model: Hans Lindh

The Swedish model offers the prospect of complete unbundling between commission payments for research and execution. Champions of the new model believe that it offers the only means to offer true best execution. Hans Lindh, head of trading at Swedbank, played a key role in the development of the new model.

“When I began work on this area back in 2007, we quickly realised that the old bundled model wasn’t benefitting the end consumer,” Lindh told the Buy-side Perspectives. “We realised we had to do something to reach best execution. I wasn’t happy having the payment for research linked to the execution of the trade.”

Lindh has been in the industry since 1985, when he was part of the equities team. In the late 2000s, Lindh began to lobby within the firm for the adoption of CSAs. However, at the time it was something the firm did not want to do. The breakthrough came in 2014 when Swedish regulator FI released a letter to asset managers that stated that research and execution payments must be separated. Lindh and a few others in the industry started to question the CSA model. It seemed to have limitations – why create a limited number of accounts and then grow them? Might this not undermine best execution?

Instead, Lindh began work on an alternative that would avoid linking execution with research payments completely. The idea was that whether a buy-side firm traded with one or a thousand counterparties, it would not affect the research costs in any way. “It was a huge relief”, he said. Under the new model, Swedbank sets aside a fraction of a basis point every day and transfers the total every quarter to an RPA, which pays the firms the portfolio manager finds useful.

“We didn’t have to hire any people,” added Lindh. “Taking the CSA route would have been more of a burden due to reconciliation and things like currency trades and run offs, which we don’t have to worry about. CSAs would require surplus money on the account, which is less efficient.”

In terms of best execution Lindh is directly involved in reviewing reports and signing them, checking outlier trades and picking up on any unusual performance. His role covers FX and fixed income as well as equities – but it is fixed income that is the most difficult asset class in which to demonstrate TCA due to the huge range of products, the lower liquidity per instrument and the complexity of the market.

In line with this focus on efficiency, Lindh reports that one of the impacts of this new model has been to shift more flow to electronic channels, and to increase the use of TCA. The flip side is that the firm has made less use of research-only firms than expected, and particularly less maintenance research – a factor that Lindh believes does improve the quality of the research being consumed compared to some of the material produced in the past.

One of the challenges to the Swedish model is that if a fund turns over nothing at all, the trader might ask why they are paying for research. Under the old model, they would not have paid anything in that situation. But under the new model, they will still pay for research if they consume it. The benefit compared to the old system however is that if there was high turnover, the trader would pay far more for research, whereas under the new model this is not the case.

“The research might tell you not to trade,” said Lindh. “Staying put might be the best choice. So the new system is more fair, because every end client pays exactly the same.”

There are areas however where the model can still be improved- notably, the Swedish model does not track each and every fund. Swedbank currently argues that it is not necessary to go into each individual fund because of the linked nature of the process across the whole asset management team. In future, it is possible that RPAs could be created for every fund.  

In addition, while large firms have the resources to make the model work, Lindh acknowledges that smaller ones might not. Fund owners may find it easier to adopt the model, since they don’t have to negotiate with clients and can act independently. However, Lindh is upbeat about the prospects for overcoming these obstacles in the future.

“The weaknesses can be solved,” he said. “There are issues, but there’s always room for improvement – it is not a problem.”

Payment for research does have complexities – Lindh explains that the value of a piece of research might be $10,000 one firm, but ten times as much to another. Then there are other contextual factors like the time of day – a piece of research might be more valuable at 08.00 in the morning versus 16.00 in the afternoon. Demand also varies by subject from day to day – a steel analyst for example may be in demand one day, but overshadowed by the healthcare analyst the next. At Swedbank, providers are asked to name their price, and then a decision is made.

“The end customer has definitely benefitted,” said Lindh. “If you look what we paid bundled and what we pay today, we are paying less than we did before. But I still think people doing a good job on the research side are receiving more or at least the same as before, because we are concentrating our research payments in the best way possible.”

In the Swedish RPA model, the entire research budget gets used – 100% is paid out each quarter and there are no surplus funds. If the amount is too high, the budget can be changed, but in general Swedbank aims to keep to the budget decided at the beginning of the year.

From a regulatory standpoint, MiFID II is likely to provide the final say on the role of CSAs. Lindh expects that Swedish regulator FI will accept whatever the conclusion of the MiFID II process is – a scenario which most likely means that in addition to the RPA model, there will also probably be CSAs. But Lindh is sure that the RPA is the better model.

“I hope the Swedish RPA model is here to stay,” he said. “I wouldn’t like to take the CSA route, because RPA is much better for us. It might be difficult for some firms, for example UK asset managers with diverse client bases, but the other option is to pay for research from P&L, and I think that’s bad for the smaller asset managers – and it doesn’t change the cost for the client anyway. Our estimated market impact looks good, we feel we are doing the right thing. Not having to focus on anything but best execution helps.”