Key Takeaway: Eight layers of double marginalization costing homeowners $90 billion a year in preventable losses.
Friction is the product. The mortgage industry is a cost-plus system -- every participant adds their margin on top of the margin before them. Economists call this double marginalization: when sequential monopolists each mark up independently, the final price is higher than any single integrated firm would charge. Except in mortgages, it's not double. It's eight layers deep. And the beautiful perversity of the whole thing is that every participant knows this. I've never met a mortgage executive who couldn't explain the inefficiency. They just can't fix it, because fixing it means eliminating their own role in the chain.
Origination, servicing, delinquency, default, loss mitigation, foreclosure, REO, recovery. The industry treats each as a separate business with separate technology, separate regulation, separate economics. It's one pipeline. Everyone knows this. Nobody says it out loud because two million people's livelihoods depend on the fiction that these are distinct industries.
Follow one mortgage from birth to death
A lender underwrites a loan, capturing extensive data. The loan gets pooled and sold as a mortgage-backed security. Origination data compresses into a 50-100 field data tape. Richness gone. A servicer — often a different company — collects payments with the data tape but not the origination file. Borrower misses a payment. Collections follows a scripted process with no origination context. Loss mitigation re-underwrites the same loan from scratch. If mitigation fails, foreclosure — different department, different system. Every transition is a handoff where data is re-entered, re-keyed, or lost. Eight companies doing the same underwriting on the same borrower, each pretending they're the first.
Why it stays fragmented
Regulatory architecture. CFPB oversees origination and servicing. SEC oversees securitization. State regulators oversee servicer licensing. Each framework was built independently in response to different crises. Nobody owns the whole pipeline because no regulator does.
Economic incentives. Origination fees are transaction-based — more loans, more revenue. Servicing fees minimize delinquency cost. Foreclosure attorneys get paid by the file. No single participant is aligned with the system-optimal outcome. Everyone optimizes their piece. That's the point.
What connecting the data would do
Early intervention. Origination data combined with servicing behavior can predict default 12-18 months before a missed payment. Pre-delinquency intervention reduces default rates by 30-50%. Nobody does it systematically because the data sits in systems servicing can't access. Found money, sitting in a database nobody queries.
Better pricing. Default and recovery data should inform origination pricing. If loans with specific characteristics default at 3x the average, price that risk into the rate. This feedback loop barely exists.
Lower total losses. The cost of a mortgage default from first missed payment through REO sale averages $75K-$100K. Modifications keeping borrowers in homes reduce total losses 40-60% versus foreclosure. Decline the foreclosure. Keep the borrower in the home. The math is obvious.
A unified pipeline could reduce annual mortgage losses by $30-40 billion. Add savings from better pricing and reduced operational redundancy, and total structural savings approaches $90 billion annually.
Conway's Law in action
The organizational boundaries mirror exactly the system fragmentation. Nobody has the scope, incentive, or data access to build the unified pipeline. The solution won't come from within the existing structure. The $90 billion in annual waste is the measure of the opportunity. One pipeline. That's what it is.