I Thought It Was a Safe Port. It Turned Out to Be the Eye of the Storm.

After I left the mortgage business in 2005, I didn't take a vacation. I took on a turnaround.

The Latin Business Association

A friend I'd met on a Pepperdine international trip, Ricardo "Rick" Sarmiento, was chairman of the Latin Business Association in Los Angeles, then about 30 years old. Rick was about to become Governor Arnold Schwarzenegger's Secretary of Appointments, the youngest person ever to hold that cabinet post in California, while still serving as chairman. Not surprisingly, the new job was eating up his time. He asked me to come in as executive director and turn the place around. Rick's charisma was profound. People were naturally, magnetically drawn to him.

I walked into turmoil. Several of the big national banks that sponsored us had decided not to renew. For a nonprofit, losing sponsors is like a restaurant losing its gas line. A handful of dedicated board members, our staff, Rick, his brother Antonio and I got to work.

We brought Pepperdine's entrepreneurship classes, taught by Dr. Demos Vardiabasis, to LBA members. We rebuilt the sponsor relationships one meeting at a time. Then came the gala. That year Mel Gibson came to our annual event at the Beverly Hilton to talk about his film Apocalypto, and it got us national exposure, all the way to Jimmy Kimmel's show. When the dust settled, we had restored financial commitments from Bank of America, FedEx, Office Depot, Union Bank, Wells Fargo and others. In about six months, the turnaround was done.

I'll always be grateful to Rick for bringing me into the LBA, and for the behind-the-scenes look at government he gave me. Rick passed away unexpectedly on January 18, 2026. I still think fondly of our long strategy sessions, working alongside one another in the trenches, and the many laughs we shared along the way. Rest in peace, Rick.

A safe port in a storm

Then I got a chance to go back into mortgages, at Countrywide, my biggest former competitor. I joined in 2007 to manage the Strategic Planning for the wholesale division, one of the company's three main lending channels. The division had a P&L of about $1.6 billion on roughly $100 billion in fundings, around 7,500 employees and about 36,000 broker relationships.

At its peak in the fall of 2006, Countrywide made about one of every six mortgages in the country, counting the loans it bought from other lenders. It was the biggest lender in America, and even in 2007, as the market cracked, it still originated more than $400 billion in loans. I thought it would be a safe port in a storm and a chance to see behind the curtain.

It turned out to be the eye of the storm. Some people would later call Countrywide the epicenter of the financial crisis. That's an overstatement, but I wasn't far from the shaking.

No second set of eyes

At my old company, we checked the quality of every single file we underwrote, so the loans could be pooled and sold to Wall Street with no fear of buybacks. We made sure at least three buyers would bid on every pool we put out.

At Countrywide, I learned that when things got backed up, sometimes no second person reviewed a file at all. While I was there we went through at least seven rounds of layoffs. Corporate drew on its credit lines to keep the company alive. In January 2008, Bank of America agreed to buy us for about $4 billion in stock, and the deal closed that July.

I want to be fair. The people I worked with were exceptional, hardworking professionals. They ran into the fire and made the best of an extremely difficult situation.

Were brokers the problem?

During the merger, the broker channel was a sore subject. Inside the industry and on the news, everyone "knew" brokers brought in the worst borrowers. Countrywide's own retail channel was considered quality, and the correspondent channel, which bought loans from banks, was considered top-notch.

I ran an analysis on one of the largest mortgage data sets in the country, and it told a more nuanced story. Brokers weren't bad. They had adverse selection. If you couldn't get a loan at a bank, or didn't feel comfortable walking into one, you went to your local broker, and those borrowers tended to have slightly lower FICO scores.

That kind of problem has a fix. We pulled a few levers, and the biggest by far was writing more loans to higher-FICO borrowers. Afterward, wholesale loans performed better than both retail and correspondent on measures like early payment defaults. That improvement in our broker channel became one of the key focal points that helped the Bank of America merger come together.

The fees hiding in plain sight

My old company published its rates, every pricing adjustment and every fee. Countrywide didn't publish its fees. They showed up on the good-faith estimate, and they were far below market. We could charge at least $1,200, while Countrywide charged $300 to $500. Because brokers were choosing a lender without seeing the fees, there was room to raise them.

The CFO worried about price elasticity, meaning we'd lose volume. Many of the managers had never worked the street. They'd never originated loans or sat with loan officers, realtors and everyone else actually doing the business. I argued the worry was unfounded, and even if it wasn't, we could test it on a small slice first.

We raised the fees, and volume didn't drop. Contribution margin to the parent company rose meaningfully, part of a roughly 10% bottom-line improvement I identified for the division.

The assumption that broke

Angelo Mozilo, Countrywide's co-founder, worked from an assumption that had held up for a long time: if borrowers got into trouble, they could refinance, and that ability would keep them out of foreclosure. The long-term data supported him, right up until it didn't. When home prices fell and credit froze, the refinance escape hatch slammed shut for millions of people at once.

The cautionary tale isn't that the data was wrong. It's that nobody synthesized it. Nobody steel-manned the other side, debated it hard, or asked what unintended consequences perverse incentives might be creating. Correlation in the data is not the same as understanding. You can't run a company on correlation without synthesis.

Inside the storm

It was a stressful stretch, and I don't want to dress it up. Round after round, I watched good people get let go. At the same time, my own net worth was shrinking as property values and stock prices fell. For a while the jumbo loan market froze completely. There were no jumbo loans and no jumbo refinancing, at any price, and that locked out anyone who needed one, including me.

After the storm

After the merger, the market had little use for mortgage experience. I took a job at a commercial bank that was soon hit by the crisis too. In 2008 my son was born by emergency C-section. Not long after, I was laid off.

It was a real valley, and I had to start over. That climb back eventually led to the insurance companies and to Oak Stream. But that's another log entry.

The lesson from the trenches

Inside a big company, conventional wisdom is often just untested assumptions. Look at the data, ask why, and when someone says "we can't," offer to test it small. Some of the most valuable fixes are hiding in plain sight, sometimes right on the good-faith estimate.