There is a selling frenzy on Wall Street. Not of stocks but of businesses-and in the dramatic cases of Bear Stearns and last month Merrill Lynch, entire firms.

Mergers and acquisitions, of course, are a mainstay of life on Wall Street. But in this period of economic turmoil, as firms desperately try to raise capital to offset losses and write-downs brought on by the ongoing financial crisis, or to jettison those businesses that may in fact be harmful to them, they have been shopping portions of their enterprises to the highest bidder.

Before it lost 94 percent of its value on the stock market and was forced into Chapter 11 bankruptcy when potential buyers Bank of America and Barclays backed away, Lehman Brothers was looking to sell off pieces of its organization in an attempt to keep the core safe. Its prized asset management arm, Neuberger Berman Management, was a strong candidate to be spun off and sold separately in a bid to raise capital. At press time, Barclays had agreed to purchase Lehman's US investment banking and capital markets divisions for $1.75 billion, a deal which includes a significant amount of Lehman's technology infrastructure. Japanese firm Nomura is buying Lehman's European and Middle Eastern equities and investment-banking operations as well as its entire Asian franchise.

Meanwhile, UBS announced plans in August to separate its relatively healthy wealth management arm from its investment bank. Banks and broker-dealers are susceptible to the ups and downs of the markets whereas wealth management and asset management arms have a steadier rate of income thanks to the fees they charge their clients. Over the years, many broker-dealers have acquired asset management and wealth management firms in an effort to offset the volatility they are exposed to on the trading side, says Maureen Callahan, CEO of consultancy Callahan Co.

While UBS is not ready to speak about the technology implications of the split publicly-they and other major banks declined to speak with Waters-industry observers speculate that the move is an effort to reassure wealth management customers that if the investment bank goes down the proverbial toilet it will not drag their money down with it. There has been no talk so far of either business being sold, but the firm has deemed it strategically beneficial to hold them at arm's length from one another.

The trend for divesting-but hopefully not wholesale liquidation-looks set to continue as the economic downturn progresses.

Banks that are thinking about selling various groups want to get the appropriate value for them, and they don't want to make a hasty sale if they can help it-although of course in some cases it is unavoidable, says Peter Bond, partner at boutique consulting firm Milestone. At the same time, sovereign wealth funds, foreign banks and private equity shops sense that they can gain some significant value, he says. "Most of the time the two sides are fairly far apart so it's slow moving, but I think for various reasons-either to shore up their capital base, or to spin off businesses that aren't really part of the bank's core service or product offering-we will see an increase in that activity," Bond says.

These dramatic company divisions happen out of business necessity, but the technology infrastructure and operations implications are not trivial.

Just how easy it is to split up or sell off businesses depends on how tightly integrated they were to begin with. For years the talk at financial technology conferences has been about breaking down silos, integration, virtualization, sharing and reusing internal resources and reducing redundancies. And these things make sense-they lead to considerably reduced costs. For example, in an interview in the June issue of Waters, Richard Greenbaum, Barclays Capital's CIO for the Americas, said that streamlining efforts-including virtualization, and efforts to share resources across silos-had created efficiencies of $60 million in 2007.

It is ironic, however, that if a firm has been successful in breaking down silos it could actually make it harder to cleave off a particular unit when difficult times demand it. "If a firm has really gone down the path of virtualization then it potentially becomes more difficult to extract that business, depending on how they have implemented it," says Octavio Marenzi, CEO of analyst firm Celent. "You have to figure out where the interfaces are to the existing systems, and whether this unit can really even run as a stand-alone if they have a technology dependence on some other part of the business," he says. Unless these carve-outs are identifiable stand-alone units, it can be unclear what the purchaser is actually buying, he adds.

Defining and carving out the business unit or the piece of the business it is looking to sell is one of the biggest challenges for the separating entity. "Depending on how integrated or stand-alone that business is you run into issues around which clients, staff, systems, infrastructure and buildings are going to go," says Roger White, managing director at buy-side consulting firm Citisoft. If it is very tightly integrated, that can create some big challenges. "And once they have been able to define what it is they want to sell or spin off, coming up with the value for that organization presents another challenge," White adds.

If the divested arm has large technology and operational dependencies-including clearing and settlement-on the separating parent, there are a number of arrangements that may take place. "In some cases, the separating firm might create some sort of service level agreement (SLA) to support the business from an operations or technology standpoint until the receiving entity can make their transition either onto their own platform or service provider arrangement. In others, they go straight away to their own environment," says White.

If a firm has moved to having single instances of infrastructure, data, logic and applications supporting multiple business lines-an oft-stated goal-this could make separating two businesses far more complex, says Mary Knox, research director, banking and investment services at Gartner.

And a service-oriented architecture (SOA), designed to aid integration, if not governed properly can add to the complexity, Knox says. If SOA were taken to its full potential, a firm might have a universal fee or interest calculation engine, or a shared account opening mechanism. "If the logic for opening an account has been moved to a shared layer so that the same component is being used for account opening for checking accounts as for wealth management accounts and so on-firms will need to think about what happens to that logic when they split the firms out," says Knox.

With SOA, good governance and a clean registry of application components is key. Provided this is present, some argue that having a well-defined SOA will actually make it easier to separate a business. "In my opinion, having a SOA is the best scenario as far as divestiture is concerned," says Kiran Garimella, vice president at SOA consultancy Software AG. "When you have a SOA, the business and technology services are encapsulated not only from a technical perspective but also from the perspective of making them intelligible to all the parties concerned. Each SOA component is well understood, we know what it's doing, and we know exactly how to break it up and carve it out," he says.

Any separation requires a greater or lesser degree of replication of functions. At the very least there will be shared real estate-for both offices and data centers-telecommunications contracts and procurement teams. "That's a stable piece to centralize-it's a no-brainer to save on the procurement side and reduce headcounts," says Alois Pirker, senior analyst at Aite Group. "But if you want to spin off an investment bank or asset management arm, you'll have to get a new telecoms provider, a new procurement department and it becomes a bigger transition project than otherwise," he says.

Firms may have unified human resources, payroll and accounting systems across their business units. A divested or spun-off unit may have to create new agreements with these providers and ensure the appropriate data is transferred without any loss of data integrity. Perhaps their entire operations department was outsourced, or perhaps IT development and testing. Firms find they have to start replicating functions where they had put very efficient processes in place, Pirker says. "Automatically, when they start separating units they start adding costs-they need to hire more people to do the same tasks that one person did," he says.

Many firms aim to have one firm-wide agreement with data providers-although this is not always the reality. "It's not uncommon to find organizations that have grown rapidly through acquisition and have multiple data sourcing contracts," says White. These organizations realize they are buying the same data from the same provider in three different parts of the organization and there is a significant opportunity to reduce cost and manage data in a more efficient manner.

It is usually especially beneficial for firms to bring their data together to be able to support functions that require consolidated information, such as risk management, compliance, and financial reporting needs, Pirker says. If they had consolidated, they would likely need to replicate these arrangements for the piece of the business being sliced off to make it an appealing package for the acquirer, he adds. If too much work needs to be done before a business can be purchased, some buyers might shy away as it becomes a riskier undertaking, Pirker says.

All in all, the more stand-alone a business is, the easier it will be to separate it should the markets require it.

Fortunately for most financial services firms, the majority had not got as far along with their plans to centralize IT functions as they would have liked. "Many firms have actually been quite bad at integrating these things-either by design or because they never got around to it," says Marenzi.

Many of the synergies and economies of scale that one would expect these large universal banking operations to achieve are simply not there, because they have in essence kept everything separate, Marenzi says.

While firms are loathe to admit it, analysts and consultants agree that redundancies abound with financial services firms on both sides of the Street. "Where organizations have grown rapidly through acquisition we see environments where there are redundant systems, operations, and units. They might have seven trade order management systems or 10 accounting systems globally," says White. Many firms are taking a step back and trying to reduce the number of platforms they are maintaining, but there is a way to go.

How integrated an acquired firm is with the parent may depend on the reason for the original acquisition. If a broker-dealer acquired an asset management firm to offset the impact of market downturns, the likelihood is that it has chosen to keep the two entities fairly separate. "The clients of the asset management firm would want to be sure that the asset manager is not overpaying in terms of trading fees, so it's going to insist that the asset manager shops around in terms of brokerage services. That means there is a clear separation that takes place," Marenzi says.

"The way the integration takes place is in the end of the quarter financials being aggregated together, and you feed data into the common accounting systems. But that's relatively easy to break out," Marenzi says.

In some cases, bolt-on acquisitions are never properly integrated because the cost associated with full integration outweighs the benefits provided by the synergies that would be gained.

If an asset management arm operates in a completely standalone fashion then it should be relatively easy to sell on. The less it shares from a back-office perspective with other business units, the easier the separation will be.

However, if a firm was originally acquired to expand the parent organization's reach into a new asset class, for example, there would be more opportunities for synergies. Units might have joint outsourcing agreements, or be served by captive offshore centers of excellence in Bangalore, Singapore or Canada. It will be a painful process combing through agreements and working out how to make the amputated arm a stand-alone entity. But the more complete the package, the easier it will be to attract buyers, says Pirker of the Aite Group.

So does all this mean that firms should avoid integrating businesses in the first place? Should they stick to their comfy business and technology silos, hug their servers, and cordon off their own developers?

Tom Redman, president of Navesink Consulting Group, puts it this way: Merging two companies is like a marriage. Both parties come with their own belongings and furniture, maybe even their own real estate. Once they are married, the couple has to decide what to keep and what to throw or give away. "After a while they throw out all the old furniture and get new stuff and then kids come along," he says.

But what happens if it doesn't work out and leads to a divorce? The couple is faced with a struggle to divide up their assets. It will be difficult-in fact it will seem near impossible, says Redman. But that doesn't mean they shouldn't have thrown out their own belongings and given the marriage a chance in the first place. "None of us is going to argue that when you move in together that you ought to live separate lives. The whole idea of putting things together in our personal lives as in business is to make the whole more than the sum of the parts," he says.

It is unlikely that the potential challenges of separation will dissuade firms from trying to achieve some form of integration and centralization in the future. You can't plan for failure, says Bond. "It's an interesting point that if you ended up not going down that path that was recommended by so many for so long and ended up siloed it's a lot easier to do a lift-out," says Bond. But he believes firms will-and should-still look for efficiencies and synergies where they can.

There is a lot of work that can be done within business units and even within silos in terms of SOA, virtualization and so on. Usually, SOA is not really enterprise-wide, but is used for things like deploying an Internet front end on an asset management solution that pulls in data from the back office systems, Marenzi says. "Frequently, interfaces are defined at that level within the lines of business as opposed to cutting across the lines of business," he says.

Most most agree that silos are here to stay. Silos came about because of political and technology fiefdoms and because the needs of fixed income are different from the needs of equities and derivatives and the cash markets. "These areas tended to get built up by business groups and the ones making the most money spent the most money and then didn't want anyone in their silo," says Callahan.

This is unlikely to change.