The number in question detailed the percentage of mortgages that had been made in the United States that year that were subprime in classification. The number was 20 percent. “We knew that it was rising. But the whole [mortgage] market was essentially rising,” Greenspan told me when we met face-to-face in September 2008. He still sounded incredulous. “That it would be 20 percent of originations? It maybe got up to 11? Maybe. But 20?”

The numbers Greenspan had been shown that day were true. One in every five mortgages made in the United States was a subprime mortgage. For Greenspan, it was a huge revelation. “Everybody in retrospect now knows that the boom was developing under the markets for quite a period of time, but nobody knew it then,” contends the former Fed chairman. “In 2004, there was just no credible information on that. It wasn’t until we got well into 2005 that the first inkling that [it] was developing was emerging.”

It seems hard to believe. It’s not as if there weren’t countless anecdotes about home flippers in California and Florida and people who refinanced their mortgages once a year like clockwork, or that news of a housing boom wasn’t featured on the covers of countless magazines, in newspapers, and on CNBC. Greenspan dismisses those stories, asserting that they are not indicative of the true state of affairs. “One of the things that you become very careful about if you’re a supervisor or a regulator is to be careful of anecdotal evidence as a generic type of indication,” counsels Greenspan. “We have a vast country. The amount of economic activity that goes on is truly awesome. We’re dealing with hundreds of billions and trillions of dollars. And it is remarkable how big certain small things look.”

The data, says Greenspan, is what matters. And while it may be easy for people to look back now and say it was all so obvious, the man who ran the Federal Reserve during the greatest decline in lending standards our nation has ever seen is not willing to admit he missed it:

Clutching that report in late 2005, Greenspan, perhaps for the first time, was coming to grips with the fact that the problem of subprime mortgages was not going to disappear on its own. And as he dug in more deeply, Greenspan realized the reason it wasn’t going to go away. “The subprime market went from seven percent of total originations to twenty percent in three years . . . because it became securitized in a vast way and internationally,” explains Greenspan. “Certainly in retrospect that had a wholly unforeseen effect.”

Dallas had made it tougher for people to get a mortgage, but it didn’t stop the lesser-quality mortgages Ownit had already doled out from coming back to haunt it. Delinquencies were increasing by the day. “When we started to peel back the onion, you could see that all the things that we worried about were coming to bear,” recalls Dallas. Mortgages given to people with low credit scores whose incomes had not been verified were going bad fast. Dallas decided at the end of 2006 to shut Ownit down rather than meet a big margin call from Merrill Lynch. His was one of the first mortgage lenders in the country to go out of business. It would be far from the last.

The Wall Street firms that were buying mortgages from firms like Ownit also started to see signs of impending doom as 2006 came to a close. Banker Michael Francis explains:

Firms like Francis’s that assembled the mortgages and securitized them were at the end of the chain in the origination process. By the time Francis and his team were sending mortgage-backed securities out to the world, the mortgages they were created from had been around for four or five months. In more typical times, that was not a problem. But when you’re at the end of the biggest downturn in lending standards the housing market has ever seen and housing prices have stopped going up, things go bad pretty quickly. “We would put those loans into our deal and within sixty days of the security being sold, the delinquencies would rapidly be approaching the six, seven, eight, or nine percent range very, very quickly,” recalls Francis.

When 9 percent of the mortgages from a securitization are being paid late or not at all, it’s a very bad sign. “The wheels were starting to come off,” says Francis. Wall Street’s mortgage machine was sputtering as the links in the mortgage chain began to be pulled apart.

The forces behind that pull were apparent to people like Bill Dallas and Michael Francis. Mortgage brokers had long ago run out of people with good credit to sell mortgages to and now they were running out of people with bad credit to sell mortgages to. Some Wall Street firms began to pull back from the mortgage business, which meant it was that much tougher for mortgage originators to find the cash they needed to extend new home loans. Such was the case for Ownit, which, after trying to navigate a difficult market by tightening its lending standards, chose to give up and close down. The end result: There was less credit available for would-be homeowners.

The great ocean of credit that had fueled the housing boom started to dry up. Credit, it turned out, was not ubiquitous. And when it started going away, getting a home loan for those who couldn’t qualify for a traditional fixed-rate, prime mortgage became more expensive. That kept new buyers from being able to afford a home.

If this had been any other period in the history of the U.S. housing market, the disappearance of these previously “marginal” buyers would not have had a great impact on the vast market for homes in the United States. But this was the housing market of 2006, in which over 40 percent of all new mortgages (subprime and Alt-A) were made to non-prime borrowers. The marginal buyer had become a major player, and when these buyers could no longer get a home loan, the impact was sizable. Home prices stopped rising.

Rising prices had been the elixir that everyone connected with the business of residential mortgages relied on to keep the machine running. When home prices were rising, the freedom to refinance was manna from heaven. You could refinance to pay off your credit card bills or put in a new pool.You could refinance in hopes of using your newfound equity to start a business, as Arturo Trevilla hoped to do. But most important, the freedom to refinance allowed people to escape exotic mortgage products before they became toxic. There was no concern about the higher rate that was coming if the buyer knew she could simply step into yet another mortgage that would delay those higher payments for two or three more years.

When home prices stopped rising, the freedom to refinance got taken away. People with mortgages they could not afford suddenly found themselves without an exit strategy. And so they began to stop paying their mortgages back.

The call was like a lightning bolt. “David,” said the Texas twang on the other end of the phone. “Do you realize what is happening?”

“Kyle? Is that you?”

“David, do—you—realize—what—is—happening?”

Obviously, I was not at all sure of what was happening, but it seemed pretty certain that I was about to find out.

I still have my notes from the call. The page is filled with words and phrases that I was hearing for the first time: CDOs; $100 billion in subprime mezzanine tranches; ABX Index; 1100 is mid-06-BBB on the index; ’06 CDO issuance—$200 billion. And there were some things that I understood pretty well: Mortgage market imploding; Someone is losing a lot of $; Merrill has the most exposure to CDOs; Bear is in trouble; It’s very bad; Can’t ramp the warehousing line.

In the months to come, I would slowly begin to make sense of the avalanche of information Bass unleashed on that February afternoon. The reason for his call was due to the release of so-called remittance data earlier that day. If you own a piece of a mortgage securitization, you receive data on the 25th of each month telling you how many of the mortgages in the securitization pool are 30 days delinquent, 60 days delinquent, 90 days delinquent, or in default.

On February 25, 2007, the remittance data was bad. It showed a sudden spike in delinquencies. It was a spike that Kyle had been expecting for months, but it caught most investors off guard. “All of Wall Street said, ‘Wait a minute, these loan delinquencies are not supposed to spike like this,’ ” Bass said. An index that had been created to track the value of the BBB and BBB- pieces of mortgage securitizations fell sharply in value. Kyle’s credit protection on those same securities rose sharply in value. He was about to become a very wealthy man.

Delinquencies did not trend much higher in March 2007 and the uneventful data from the spring months, when people were likely getting tax refunds that helped them pay their mortgages, convinced many that February had been an aberration. The index that tracked the various slices of mortgage securitizations rebounded sharply from February. People were breathing a sigh of relief.

But not the people who ran the largest subprime lenders in the nation. They were done. In March 2007, in rapid succession, Fremont General, New Century, People’s Choice, and Ameriquest announced they were closing their lending operations.The funding these firms had relied on from Wall Street was drying up.

In 2006, New Century had made roughly $60 billion worth of subprime loans. Now, facing mounting defaults on those same loans, the banks that gave it funding pulled away. In the good times, New Century sold its mortgages to Morgan Stanley, Goldman Sachs, Barclays, Citigroup, Credit Suisse, and other Wall Street firms, and they in turn provided the credit lines that kept its origination machine humming. But that February’s remittance data was scaring away the investors who used to buy the mortgage-backed securities these firms sold, and so they were now pulling back as well. Michael Francis would soon lose his job on Wall Street. He now runs ExMor Capital, a firm that counsels mortgage lenders.

New Century, Fremont, and Ameriquest were among the top-10 subprime lenders in the nation in 2006. By March 2007, they were out of the business. Soon they would all collapse. Over the ensuing months, more than 40,000 people who worked in the subprime industry would lose their jobs, including Lou Pacific from Quick Loan Funding.

Quick Loan, which at its height employed 700 people and had handed out about $4 billion worth of subprime loans, somehow managed to hang on until the summer of 2007, when it, too, was forced to close its doors. Today, its founder, Daniel Sadek, is broke. He told me so himself. After refusing numerous requests for an interview for more than a year, Sadek called me at my CNBC office in February 2009 to tell me that Quick Loan was far from the worst offender when it came to subprime mortgages. “We didn’t do loans where you weren’t allowed to make a pre-payment and we would agree to reduce people’s interest rates,” said Sadek.

Sadek knows about that kind of forbearance. After he fell two months behind in his mortgage payments to Citigroup, the bank’s Residential Lending unit offered to modify its loan terms so he could meet his mortgage payments. Sadek, according to the reporting of John Gittelsohn of the Orange County Register, did not make a single payment after the loan modification. Citigroup subsequently issued him a notice of default and is in the process of foreclosing on his home.

One might argue that the credit crisis began on June 20, 2007, when two highly leveraged hedge funds run by Bear Stearns found themselves near collapse. The funds were heavily invested in CDOs, and, because of the recent drop in value of those securities, they were in danger of getting those CDOs sold out from under them by the investment banks that lent the hedge funds money.

It was one of those occasions that warranted a call from Kyle Bass. I had been a poor student. Despite his erratic efforts to educate me on the structure of CDOs and the subprime securitization market, I had not advanced far from our first conversation on this subject in February. On this day, I was to learn of still-more-esoteric CDO-related products, such as CDO-squareds (A CDO made up of other CDOs), and was quickly informed that the prices of all CDOs based on subprime mortgages were not worth anywhere near what their owners were saying they were worth. Bass’s hedge fund and the fund he raised to bet exclusively against the mortgage market were soaring in value. His investments would ultimately be up 600 percent over the 18 months he held them. The profits for his investors would total over $2 billion. It was a once-in-a-lifetime trade.

In late June 2007, Bear Stearns initially managed to stave off the collapse of its two hedge funds by infusing capital into the funds and convincing many of the banks that had lent the funds money to allow them to keep operating. A little more than a month later, with the market for all subprime-related securities falling still further, the hedge funds collapsed anyway. The Bear Stearns High-Grade Structured Credit Strategies Master Fund and the High-Grade Structured Credit Strategies Enhanced Leverage Master Fund filed for bankruptcy in the Cayman Islands and sought an orderly liquidation of assets. The bankruptcy did limited damage to Bear Stearns’s stock price at the time, which hovered around $120 a share.

Alan Greenspan is unequivocal. He says the credit crisis began on August 9, 2007, when the French bank BNP Paribas suspended trading in three of its mutual funds.The funds had been large buyers of securities backed by U.S. subprime mortgages; unable to find willing buyers for those securities and therefore establish a value for them, BNP Paribas halted withdrawals from the funds. That event set off a panic in the short-term credit markets.

I prefer to date the start of the credit crisis to July 10, 2007. On that day, the credit rating agency Standard & Poor’s announced that it would shortly downgrade its credit ratings on billions of dollars’ worth of mortgage-backed securities comprised of subprime mortgages. Also likely, said S&P, was a wave of downgrades of CDOs made up of any of those mortgage-backed securities. And in a monumental admission that it had misjudged the creditworthiness of such mortgages (my words, not theirs), S&P said it was revising the way it went about judging just how likely subprime borrowers were to pay their mortgages back. And why did it do so?

Here are S&P’s rationale and my translation of what the firm was actually saying:

The ignorant foreign buyers who had been voraciously feeding on anything with a triple-A rating suddenly realized they had been had.They had never read the fine print in a CDO or RMBS prospectus.They had looked only at the triple-A rating.They had relied on a corrupt system, little of which they understood. Now they understood one key truth. The reason triple-A credits never default isn’t because the initial credit rating is accurate. It’s because the rating agencies will always downgrade them before they default. Triple-A doesn’t mean triple-A for life—not even close.

When that lesson was learned, the buyers Wall Street had counted on to keep devouring all those securities it concocted out of mortgages left the scene. No one wanted to buy this stuff anymore—not the banks in mainland China or the Chinese government, not the banks or insurance companies in Taiwan or Korea or Germany or France or the United Kingdom. They were done, and not just with products structured from mortgages. All those buyers with their “excess pools of liquidity” started to have a well-deserved crisis of confidence in any security that had previously been given a strong credit rating. After all, if the rating agencies did such a poor job of ascertaining the risk of subprime mortgage securities, why would anyone believe their failure was an isolated event?

And that was the start of the credit crisis.

Did Merrill’s business really change that much in a year? Of course not. It was the same company doing the same things. But in 2007, Merrill had to start telling its investors about CDOs because it was losing so much money from them. In the second half of 2007, as the credit markets seized up and any security with a mortgage in it could no longer be sold, Merrill found itself in a tough position.

It had begun the year with all guns blazing as it sought to dominate the market for securitized mortgage products. It was not prepared in any way to deal with the devastating impact from the credit crisis. It takes months to put together a CDO, given its components. A firm needs to keep thousands of mortgages in-house to create mortgage-backed securities and then needs even more time to rejigger them to create CDOs. As well, many firms kept certain parts of the CDOs they could not sell in the belief they were money-good. Merrill had played a terrible game of musical chairs. It wasn’t left standing when the music stopped. It was left in the fetal position.

The writedowns came swiftly and kept getting worse. By the end of 2007, Merrill Lynch had absorbed losses of $23.2 billion from its business of buying, packaging, and selling investment products made from subprime mortgages. The vast majority of those losses came specifically from CDOs made up of asset-backed securities. But the damage wasn’t confined to those products. The absence of buyers for any fixed-income security led to a decline in the value of all fixed-income securities and there were plenty of those on the balance sheet of every financial company in the world.

Citigroup, Wachovia, Washington Mutual, UBS, Bank of America, Morgan Stanley, Lehman Brothers, Bear Stearns, Fannie Mae, and Freddie Mac were among the most notable financial companies that would end 2007 in a seriously weakened state thanks to their participation in the U.S. mortgage market. It wasn’t just that those firms owned mortgages and mortgage-related securities. It was also what they had chosen to do with their balance sheets.

These public companies, in their race to enhance returns (the compensation of their senior executives was often linked to those returns), had mightily increased their use of leverage. Many of them had equity (the capital controlled by the institution) that was as little as 2 or 3 percent of the total amount of assets on their balance sheets. The way they financed those assets was by borrowing money from other financial institutions. That borrowed money is what is known as leverage. Leverage is like alcohol. It makes the good times better and the bad times worse. And times were about to get a whole lot worse.

That’s exactly what began to happen to Merrill Lynch. It ended 2007 with a pre-tax loss for the year of $12.8 billion (while it lost far more than that in the second half of the year, it had earned money during the first half ) and would need to quickly replenish its equity capital. The firm, under the leadership of new CEO John Thain, raised roughly $13 billion during the last quarter of 2007 and the first quarter of 2008, through the issuance of common and preferred stock. It would prove to be far too little.

In 2008, Merrill Lynch would lose $41.19 billion before taxes. In 15 months it had managed to lose $54 billion, the vast majority of it coming as a result of Stan O’Neal’s fateful decision to take more and more risk and plunge Merrill Lynch into the deep end of the mortgage securitization market. O’Neal is a very smart man. Anyone who knows him will tell you that. And one of his strongest suits is math. He understands numbers in the way few people can. Yet the risk management model at Merrill Lynch completely failed, as it did at financial giants such as Citigroup, Lehman Brothers, UBS, AIG, Bear Stearns, Fannie Mae, and so many others.

Former Federal Reserve Chairman Alan Greenspan says there is a reason those models failed so miserably. “The mathematics of it and the economics of it are very sophisticated.” The trouble, says Greenspan, is that the formulas the banks used to extrapolate risk were based on data from a period of economic expansion.The banks used periods of euphoria rather than periods of fear on which to base their risk management decisions. “The coefficients that you will get in a period of fear would have told them that the degree of leverage they were taking was very risky,” says Greenspan. “The problem was in the data, not in the mathematics.”

Wherever the problem may have been, Greenspan believes the executives who ran our nation’s large financial institutions ultimately did understand the risk they were taking. “I spoke to them,” says Greenspan. In those conversations with the leaders of the financial system, Greenspan says he would remind them of how easy it had become to borrow money for even the riskiest of credits. Risk, says Greenspan, had become severely underpriced. He cites as an example, and one he shared with many CEOs, the fact that companies with an atrocious credit rating of CCC could borrow at only 4 percent more than the U.S. Treasury. In far different times, triple-C-rated companies would have to pay 24 percent more than the U.S.Treasury for their money.

“It was extraordinary. Everyone was aware of it [the underpricing of risk]. But what they all recognized is if they didn’t participate, or as Chuck Prince [former CEO of Citigroup] said, ‘if they don’t participate and dance on the floor, they will irretrievably lose market share,’ ” argues Greenspan.

Executives such as Chuck Prince (CEO, Citigroup), Jimmy Cayne (CEO, Bear Stearns), Dick Fuld (CEO, Lehman brothers), and Stan O’Neal had all been around for a long time. They had lived through the short-term crisis in the debt markets brought on by the collapse of the hedge fund Long Term Capital in October 1998. They had seen plenty of history and Greenspan submits that they knew exactly what they were doing.The problem, says the former Fed chief, is that they all thought they knew when to get out.

Greenspan may have been warning the CEOs to be careful, but he did little more than that. As their chief regulator, he believed they would make the right decisions to safeguard their institutions. It was a significant error. And whereas Alan Greenspan may have believed the CEOs all understood the risk they were taking, after speaking with people close to Stan O’Neal, it is clear to me that while he may have understood his firm was taking too much risk, he didn’t fully appreciate the nature of that risk.

Up until the summer of 2007, insiders tell me O’Neal didn’t ask many detailed questions about Merrill Lynch’s mortgage business. He didn’t ask, and given the culture he had created at the firm, there wasn’t anyone who was going to volunteer even the idea that the business could run off the rails.

O’Neal had fired Jeff Kronthal, who oversaw much of Merrill Lynch’s fixed-income business, in the summer of 2006, after Kronthal resisted attempts to take more risk and had the temerity to tell O’Neal that to his face. The people who inherited Kronthal’s responsibilities weren’t about to tell Stan O’Neal something he didn’t want to hear, especially given how much money they were all making. It was not until late in the summer of 2007 that O’Neal started “digging in” to the mess he’d made of Merrill Lynch, according to people who worked closely with him.

Stan O’Neal ran a vast company with over 60,000 employees and tens of thousands of shareholders. I realize the hectic nature of our world and the myriad distractions that can keep an executive from focusing on one part of a large enterprise. But I still can’t imagine how it is possible that Stan O’Neal wasn’t “digging in” to fully understand the mortgage market of which his firm was such an important part, particularly because this was not a matter of experience or intelligence. Ask people who know Stan O’Neal to describe him and the two adjectives they come up with are smart and mean. Those two characteristics are usually a good combination for running a financial services company.

By the middle of August 2007, with the credit crisis underway, Stan O’Neal was dug in. Perhaps it came to him when he learned of the complexity of synthetic CDOs and started to appreciate the risk Merrill Lynch had taken on with that product. Or maybe he dug in on the credit side of the spectrum and came to grips with the fact that mortgages his firm had been buying were being given to people who couldn’t pay them back. Whatever the curriculum in his studies, he finally knew the truth: Merrill Lynch was in deep trouble. He would probably lose his job. And the fact of his knowing this probably meant it was too late to alter either outcome.

Those lonely rounds of golf in the late summer of 2007 may have offered O’Neal the opportunity to focus on what he could do to try to save Merrill. And to his credit, insiders tell me he returned after Labor Day determined to sell the company to Bank of America. A deal may have been in hand for as much as $80 a share, but Merrill’s board vetoed it.

O’Neal stepped down as Merrill’s CEO on October 30, 2007. In the waning days of that month, all but certain that he would soon be asked to leave, O’Neal retreated to his executive suite. A senior manager recalls walking toward O’Neal’s office to ask a question. He approached, but at seeing the room was dark, began to back away. Still, he gave the slightly ajar door a knock and was surprised to hear a reply from within. He entered. Stan O’Neal’s lights were off. He was sitting in the dark. And as the visitor came closer, he heard O’Neal repeating something over and over.

“I should have known better . . . I should have known better. . . .”