As congress prepares for an intense round of questioning of Enron directors and officials, there is a growing suspicion that at the heart of the once-mighty energy trader was a financial hole.

Evidence is accumulating that the Houston-based group, which boasted of being asset-light, may also have been light on profitability at core operations.

A combination of aggressive accounting, off-balance-sheet deals and brow-beating of employees and advisers, allowed Enron management to create a virtual company with virtual profits.

A Financial Times examination of company documents, interviews with former employees and analysis by accounting and legal experts show that Enron inflated revenues, for example by treating the turnover of trading through its online subsidiary as revenue.

Enron bolstered profits by booking income immediately on contracts that would take up to 10 years to complete. It shifted debts into partnerships it created and in effect controlled, even though defined by auditors as off balance sheet. It used such entities to manipulate its accounts at the end of each quarter and employed financial derivatives and other complex transactions aggressively to the same end. It masked poorly performing assets with rapid deal-making.

The group also employed an aggressive tax avoidance strategy. According to Citizens for Tax Justice, a Washington advocacy group, Enron's pre-tax profits between 1996 and 2000 totalled $1.79bn and it received net US federal tax rebates of $381m. In only one year did Enron pay federal tax at all - $17m in 1997.

The report issued by the special investigation committee appointed by the Enron board highlights some of these strategies, many of which are used by other large companies.

Enron, however, used them more aggressively and comprehensively than most to create the image of a dynamic corporation with lower debt, higher revenues and bigger profits than justified by its businesses.

To do this, Enron's senior management and sales team placed enormous pressure on middle managers and on the outside auditor, Andersen.

When Enron's risk managers questioned generous profit forecasts, senior managers insisted they would be borne out over time.

Enron's off-balance-sheet vehicles played a central role in efforts to flatter results. Those who have examined Enron's accounts and partnership documents say it struck deals with the partnerships in the final weeks of each quarter, reversing them early the next quarter after Enron had filed accounts.

The company's presentation on October 26, 2000, to partners in LJM2 CoInvestment, one of the central partnerships, listed 24 investments since formation in December 1999. Half came in the final month of a quarter and nine in the last week of the accounting period.

The investigation said the rapid reversal of many of these deals and the fact that LJM partnerships "made a profit on every transaction . . . call into question the legitimacy of the sales".

Such transactions, many with other Enron-related entities, may have helped the group in two ways, according to accountancy experts. They helped to reduce debt temporarily, but also created price benchmarks - the so-called "estimated fair value" used in its accounts - so the company could assign generous values to assets ranging from power plants to derivatives.

Frank Partnoy, a professor at the University of San Diego School of Law who has studied Enron's derivatives transactions, told a committee last month: "Enron entered into derivatives transactions with these entities to shield volatile assets from quarterly financial reporting and to inflate artificially the value of certain Enron assets."

Derivatives, he added, were also used to "hide speculator losses it suffered on technology stocks, hide huge debts incurred to finance unprofitable new businesses, including retail energy services for new customers, [and] inflate the value of other troubled businesses, including its new ventures in fibre-optic bandwidth".

For some assets, such as shares trading in transparent markets, there is little or no leeway in defining the "fair value" of the holding. The asset is "marked to market" at the quoted price and a loss or gain is included in net income.

"Mark-to-market items are usually things you can determine the value of by picking up today's FT," says Robert Verrecchia, accounting professor at the Wharton School at the University of Pennsylvania. "The problem with all the mark-to-marketing is that at the end of the day you have to have some valuation in place."

In the case of assets, such as long-term energy contracts, in which there was no transparent trading, Enron had to estimate fair value.

"There was no independent market check [on values of long-term energy contracts], beyond about 24 months," says Jim Chanos, president of Kynikos Associates, a short-selling hedge fund that profited from the decline in Enron's shares. "There was no publicly traded market."

As Enron pointed out in the footnotes to its annual report for 2000: "Judgment is necessarily required in interpreting market data and the use of different market assumptions or estimation methodologies may affect the estimated fair value amounts."

Enron boasted about its role in creating some of these markets and even claimed to have pioneered the accounting treatment of such assets.

Bold use of mark-to-market accounting by Enron Energy Services (EES), a unit that signed long-term energy supply deals with outside companies, was one of the worrying signals identified by Sherron Watkins, the senior Enron official who warned Kenneth Lay, then chairman and chief executive, about apparent irregularities in an August 2001 letter.

A contract between EES and Quaker Oats, signed last February, reveals how a few aggressive - but not necessarily illegal - accounting sleights of hand allowed the company to book tens of millions of dollars in up-front profits on a deal that might otherwise have yielded none.

Under one of its so-called "bundled contracts", EES agreed to supply 15 Quaker plants with their energy needs, from natural gas and electricity to workers who would maintain boilers and pipes and procure spare parts. Enron guaranteed Quaker it could save $4.4m from its 1999 energy bill.

For its own part, Enron forecast a $36.8m profit over the 10-year deal and used mark-to-market accounting to book $23.4m of that before it had ever turned on Quaker's lights.

Under accounting rules, such treatment is permitted for commodities, such as natural gas and electricity. But the rules are more restrictive when it comes to services - such as boiler maintenance and parts procurement, for which no forward markets exist. Profits from these activities are supposed to be booked on a more conservative "accrual" basis, whereby a fraction of the profit is realised each year as it comes in.

Enron's problem was that almost all the profits projected for the Quaker deal were derived from services, not commodities. How did it manage to book them up-front?

The company used a questionable method called "revenue allocation". The net effect of this highly complex treatment was to redefine as commodities some of the money Quaker was paying for services and therefore create more profits that Enron could book up front.

Under the system, Enron's internal accountants created a new category called "allocated revenues". These were based not on what Quaker had historically paid for energy commodities and its service contracts, but on figures that Enron claimed reflected the open market value of the commodities and services.

This revaluation made a significant difference to the reported worth of the contract. Enron would have earned only a small margin supplying gas and power to Quaker based on the original revenue figures it used to calculate the deal. These activities could have actually been loss-making with Quaker's annual discount. Instead, revenue allocation allowed the company to claim an immediate hefty profit on the deal. Asked if such a move is illegal, the former Enron accountant says: "It's certainly skirting the edge. It's very, very aggressive."

Altogether, former Enron employees claim the company managed to mark-to- market $85m in services profits from a dozen deals, including Quaker, that should have been accrued. In some cases, those profits came from such services as changing light bulbs and air conditioning filters.

Former EES employees say that recognising these profits up-front was crucial because the assumptions that underlay them would often not have panned out over the life of the deal.

For example, they say EES would routinely under-estimate commodities prices in the latter years of a contract to lower Enron's cost. This was difficult for auditors to contest because no one could say with certainty what the price of natural gas would be years into the future.

Also, some of the prices Enron quoted, such as electricity distribution, were in highly-illiquid markets that were Enron-dominated. It could virtually name whatever price it pleased.

Much larger - and also dubious - projected savings may have come from the service estimates for the Quaker deal. Enron predicted it could operate and manage the 15 plants for less than half the cost Quaker had been paying. This estimate accounted for almost all the profit on the deal.

Former employees say it was easy for Enron to inflate services margins, because no one could accurately predict them. Perhaps Enron's boldest assumptions had to do with something called "efficiency projects".

As part of the Quaker deal, Enron was permitted to spend its own money to upgrade equipment such as boilers and lighting if it believed the investment would improve its margins over the life of the deal.

In total, EES identified efficiency projects at Quaker that it forecast would save $25.3m over the life of the deal at a cost of $14m.

These projections were highly speculative, former employees say. They took into account hard-to-predict variables, including future energy prices, construction costs and the client's future energy use. In spite of this uncertainty, before the projects had proved their savings Enron booked $11.3m in profits based on efficiency projects at Quaker.

Enron realised $250m in profits from these projects during the last three years, according to a former employee, even though many of the projects were never fully implemented.

How accurate were Enron Energy Service's projections at Quaker? We will never know - Enron collapsed just months into the deal. Quaker says it has since made "other arrangements".

According to Charles Mulford, accounting professor at the Georgia Institute of Technology and co-author of The Financial Numbers Game, due to be published this month, one indication that Enron's core business was suffering was the decline in the ratio of its operating cash flow to income.

Such a decline normally occurs when a company reports earnings that are not backed by cash flow. In Enron's case, according to Prof Mulford, it declined in 1997, 1998 and 1999. Although it improved in 2000, bolstered in part by a sharp increase in "accounts payable" by Enron, the ratio worsened markedly in the first two quarters of last year, the last filings before Enron restated its accounts.

One last piece in the puzzle was to complete flurries of deals, which helped blur the poor performance of existing assets.

One former executive in the London office of Enron's Azurix subsidiary says she sensed pressure to do deals to make the financial numbers look better. "The feeling [in London] was that the deals were being done just to make deals," says Clare Spottiswoode, a former senior vice-president of regulatory affairs in Europe.

Ms Spottiswoode says she had assumed Azurix, an ill-fated venture created to own water assets around the world, was an exception: a poorly run subsidiary, in an industry not understood in Houston, within a company that excelled. In retrospect, she feels it was typical of Enron's culture.

As the board-appointed committee concludes, the partnership structure was doomed by "a flawed idea, self-enrichment by employees, inadequately designed controls, poor implementation, inattentive oversight, simple (and not-so-simple) accounting mistakes, and overreaching in a culture that appears to have encouraged pushing the limits".

The FT's broader investigation suggests such flaws extended throughout the company.