Thirty years ago, a group of maths and engineering graduates from pariss elite grandes coles changed the direction of one of frances oldest and most important banks.

Under antoine paille, a 31-year-old software engineer, the small team was given a basement office a few streets from the palais garnier opera house in paris with instructions to build a new business for socit gnrale, the lender founded in the 19th century.

Mr paille believed that socgens dive into options and equity derivatives, which would eventually be packaged up for professional and retail investors into so-called structured products, could provide the bank with an advantage over its larger global competitors. he proved to be right.

Now, however, after many of these products unravelled spectacularly throughout the coronavirus-induced market ructions, the paris lender has had to take stock of its financial engineering.

Socgen recorded a 326m loss in the first quarter of this year after revenues in its equities trading division collapsed almost 99 per cent to just 9m. another shock loss landed in august after structured products were hit for a second quarter in a row.

In response, frdric ouda, socgens chief executive, said the bank would slash the risk being taken by the structured products teams, resulting in up to 250m of lost revenue, to be offset with a 450m fall in net costs by 2023.

Column chart of revenues (m) showing equity trading revenues collapse at socgen

There was no training in universities in financial products. we had to learn on the job, says an early member of the group.

In the 1990s, that job for a team that included a fresh-faced jean pierre mustier, now head of italys unicredit involved constructing products which limited losses while offering the chance to make money. from a $100 investment, $70 might be put into a zero-coupon bond, while the rest, after fees, would go into bets on stocks, such as a call on the cac 40.

A steady stream of talented maths graduates churned out by french universities, along with a plan to build centralised teams and sell the products at scale, allowed socgen to steal a march on rivals. insiders say a start-up culture allowed them to beat us banks which were selling bespoke products to institutional clients. one ex-socgen banker said the us banks were not as interested in structured products because they were making enough money elsewhere.

By the early 2000s, socgens equity derivatives unit employed close to 2,000 people and had grown into the banks profits engine, accounting for 95 per cent of its investment bank earnings.

Since then, french banks including socgen, bnp paribas and natixis have become the global leaders in structured products, promising investors steady returns by using derivatives to smooth out the highs and lows of market moves.

But the original products have changed, becoming increasingly complicated structures that tried to maintain returns in a world of low interest rates and copycat products.

Normally when we created a product, wed have it to ourselves for about a year and a half, then others would start to copy it, said one former member of the structured products team.

Simple plain vanilla guaranteed return products morphed into correlation offerings based on baskets of stocks, and on to autocalls, which pay out a coupon similar to a bond as long as losses or gains are within a certain threshold.

Everest was beautiful, another banker reminisced, remembering a correlation product that was sold to retail investors and named by christophe miann, one of the leading lights in equity derivatives in the 1980s, after he saw a film about the mountain projected near socgens paris headquarters in la dfense.

Everest had a duration of 10 years and guaranteed that clients capital would be returned. additional returns were based on the performance of 10 or more stocks, with redemption based on the performance of the weakest member of that chosen basket.

We told investors that if you invest 100 then after ten years we would give you 200, minus the performance of the worst stock you chose, said someone who used to sell the products.

If the worst performing stock went to zero, then clients received only their capital back and if all the stocks performed well they got the upside. the risk that all the stocks would move in the same direction at the same time was the banks to wear.

The crash of 2008, when correlation products were hammered as markets did move violently in one direction, led to regulators stepping in and telling banks to simplify the opaque financial instruments which had become too complicated for retail investors to understand cutting the number of factors the products tracked from 10 to three.

Meanwhile, the jrme kerviel rogue trading scandal of 2008 also forced socgen to pull back from riskier corners of the market and forced mr mustier out as head of the investment bank, while mr ouda took the chief executive job, changing the banks path.

The departure of mr mustier is seen by many inside and outside of the bank as a watershed moment, with a generation of bankers fundamental to socgens reputation for financial engineering following him out the door.

Some critics and rivals believe the bank now lacks the expertise it was once famous for.

Ouda was never comfortable with market activities. he didnt understand them, he put the brakes on...there was also a lot of pressure from regulators to reduce the risk in the investment bank, said one senior ex-socgen executive.

But socgen did not give up on structured products, moving into autocalls, which one banker called a start product every bank had to have on its shelf and which became very successful after 2008 as it allowed retail investors to bet on volatility.

Structured products were hit hard again at the end of 2018 when volatility across a range of asset classes caused losses for socgen, bnp and natixis on derivatives trades. natixis was the worst affected, losing 260m mainly because of its range of autocalls sold to retail investors in south korea, and promising to stop offering some products.

Structured products have undoubtedly been the jewel in the crown for the likes of socgen and natixis so the market is wondering whether these cuts are an overreaction to short-term dislocation, said omar fall, an analyst at barclays. the cuts to dividends this year really feel like a once in a lifetime event.

Socgen is not giving up on structured products, despite the hit from covid-19 forcing it to pull back. instead it is trying to walk a tightrope between reducing risk and remaining true to its core identity.

It will do that, said jean-franois grgoire, head of global markets at the bank, by trying to push new ranges of products that reduce the potentially large losses the bank faces in market turmoil, leaning once again on its ability to engineer.

The innovation that weve been pushing [this time] is completely different. its innovation that goes towards creating products that are easier to manage for us and that are still very profitable for the customers, mr grgoire said.

Line chart of share prices rebased showing socit gnrale has fallen behind peers this year

The bank sees it as an acceleration of an existing strategy one product called evolution was launched before the pandemic struck and can, said mr grgoire, reduce volatility risk in any kind of situation.

The broader challenge for socgen and mr ouda is how to rebuild an investment case for the bank which is priced at just 18 per cent of its book value after its shares shed close to 60 per cent this year as it pulls back from the products that formed the bedrock of its reputation.

Retreat too far and he will be accused of squandering a traditional strength. but ramp up too much and socgen will be seen as courting danger.

As one equity derivatives expert at a european bank put it: when you make too much money out of products like these, its not because you are more clever than others its because you dont see the risk.