When AXA Investment Managers (AXA IM), the investment management arm of the French insurance giant AXA, bought a 75% stake in the US based Rosenberg Group in 1998 it appeared an unusual partnership. Rosenberg, based in Orinda, California, is the quantitative investing business built by Barr Rosenberg, the US academic who gave his name to Barra risk management. AXA IM, based in Paris, is the asset management arm of AXA, the global insurance and financial services giant.

The intention was that Rosenberg would become the exclusive quantitative equity manager in AXA IM, and that AXA would provide a global distribution network for Rosenberg's expertise in modelling and implementing structured equity strategies.

The deal has paid off handsomely. Last year, AXA Rosenberg added €15bn of net new money and now manages over €72bn of AXA IM's €432bn assets under management.

Stéphane Prunet, the global chief executive officer of AXA Rosenberg, says the partnership has succeeded because AXA IM has always allowed AXA Rosenberg to do what it does best.

"It's something which has worked very well. First, because the dealmakers of AXA IM, Nicolas Moreau and Donald Brydon, understood very well that what they were buying was a specific investment process and a specific culture that underlies it.

"They have been clever because they have never tried to tamper with it. AXA Rosenberg has been completely left alone as regards investment process or culture, because they understood that the incremental benefit of trying to tamper with this would have been very limited compared with the downside."

The approach has been to provide AXA Rosenberg with a broader market for its ideas without compromising these ideas, Prunet says. "What we are trying to do at AXA Rosenberg is to develop the firm, leverage the intellectual capital, but within the core competencies, belief and positioning, and not try to do everything.

"My strong belief is that investment teams are usually good at doing the one thing that matches their belief and their culture, and the minute you begin to tell them to do something different, it goes wrong.

"It's like a chef who offers you Asian and French and American cuisine. Unless the chef is exceptionally gifted, the chances are you are not going to have the meal of a lifetime."

Prunet says that AXA Rosenberg offers a cuisine in which it knows it can excel. It models and predicts company fair value, future earnings, and risk in portfolios that aim to produce higher future earnings than the markets. AXA Rosenberg embeds this analysis is in an ‘expert system', a proprietary software that has been created to embody the firm's collective knowledge. This enables it to apply its analysis in a consistent and repeatable fashion anywhere in the world.

"The idea was and still is very innovative," says Prunet. "Our belief is that companies that deliver superior earnings will get rewarded over time. So our job is to build portfolios that create more earnings than the market.

"There are two ways we can do that. The first is to acquire the same amount of earnings in the portfolio than the average of the market, but for a cheaper price. That is the value side of equity investment.

"The second way is to try to get some insight in which companies will actually deliver superior earnings compared with the rest of the market. And the way we have built the model is so it serves both sides.

"So we have a valuation team, which is looking at the different business line of each company, looking at what the market has paid over a long period of time, and comparing it with other companies. Then we use our analysts' reports to forecast a one year outlook of the company's earnings."

These two approaches are combined to create a single investment style. AXA Rosenberg also concentrates on a single asset class - global equities: "We've got a very peculiar way of managing global equities. We don't do any asset allocation. We just stay true to the company weighting in the benchmark," Prunet says.

"If you do asset allocation by countries you have maybe 20 important opportunities. So you need extremely precise and reasonably reliable proprietary information in any of the those opportunities to generate consistent alpha.

"We prefer to operate at the stock picking level because the opportunities are just more diverse. We've got 18,000 reasons to be right. Our belief is that with our investment style, you don't need lots of extra information to generate alpha.

This means that AXA Rosenberg can be wrong about some stocks, he concedes. "We are consistently wrong, and that's fine. But it's like a batting average. If we are close to 60% right we will generate a lot of alpha.

"This is why we build portfolios which are highly diversified. We typically have been 300 and 500 names in the portfolio. We don't take a huge amount of bets on any particular stocks. So when an Enron happens, it's not pleasant but it's not a huge deal for us.

"We are just trying to stick to where we think we have a proprietary insight into valuations.

"The model for these insights is run from a resource centre in Orinda, California in that collects large quantities of data from principle suppliers. "Here we crunch the data, enrich it and then model it. This allows us to cover 18,000 stocks and rank them individually every three minutes," says Prunet.

The model's recommendations are then passed on to the investment management companies that operate in the US, Japan, Asia, Europe and the Middle East regions, based in Orinda, Tokyo, Singapore and London respectively. These companies turn the recommendations into portfolios.

A consistent application of the recommendations is important, Prunet emphasises. "Being present in a number of locations could lead to fragmentation, doing something different in Singapore than we are doing in the US. We cannot afford to have any differences in investment style or in investment implementation. First because we want to remain very focused, and second because clients, such as global consultants, are often the same clients in Europe as in Asia."

Yet AXA Rosenberg has had to approach each market differently in terms of business development, says Prunet. Europe has required a steady, long term commitment. "We opened in London more than 10 years ago, so I think we have paid our dues. It was a big commitment at the time, when the firm was not as large as it is today, and initially it was a loss making operation. But it paid off and Europe has been big success story for us."

AXA Rosenberg has prospered principally in consultant-driven markets like The Netherlands, Scandinavia and the UK, which have more appetite for the specialist approach, says Prunet. Elsewhere, it has relied on the muscle of its mighty parent, AXA. "I don't think we, as US money managers, would have been as successful in France, or even in Germany, if it was not for the backing of AXA," he says . "We have now gained a large market share, both institutional and retail, in Germany and France, mostly with European and global equity mandates.

"For us, Europe has been the perfect success story and I think it was really the combination of Rosenberg and AXA that made a complete proposition for European markets."

Prunet's other hat, besides global CEO of AXA Rosenberg, is AXA IM's head of Japan and Asia Pacific. "We have actually merged our offices in Japan and in Asia Pacific, because these markets, despite all the talk, are not yet huge for pure institutional products. They are growing fast but they are not very large - apart from Japan, of course," he says.

AXA Rosenberg's direct business development efforts have always been focused on institutional rather than retail business, says Prunet.

This has meant a change in focus, away from a direct involvement in retail business in the US. In 2003, Charles Schwab adopted AXA Rosenberg's entire family of US mutual funds in one of the largest fund adoptions in mutual fund history.

Under the deal, Charles Schwab Investment Management has become the investment adviser to the existing AXA Rosenberg Funds with $1.4bn (€1.1bn) in assets under management. AXA Rosenberg continued to be responsible for the investment management of the funds through a sub-advisory relationship.

"This mirrors what we do with AXA IM in Europe because AXA IM handles the wholesaling and marketing our UCITS range," Prunet says. "So the only difference is that we are still running the products in Europe which we don't do any more in the US. We only sub advise them"

AXA Rosenberg has not been afraid to change. It has a well-established investment process, but that does not prevent it from rolling out new products on the back of it, says Prunet.

In August, it launched a global long/short and a 130/30 limited shorting strategy for institutional clients. These types of portfolios allow managers to extend a traditional long-only portfolio to include a limited number of short positions (130% long and 30% short).

AXA Rosenberg has over 17 years of experience shorting stocks, both in the US and internationally, Prunet points out. "A 130/30 strategy is the perfect bridge between the alternative space and a traditional long-only strategy. We feel this is a very important product, as the opportunity for enhanced alpha and information ratio is dramatic."

For Prunet, this is a good example of how AXA Rosenberg can stay ahead of the game without having to change the rules.