How the American bonding machine works

Developers are probably the best known players for investors. They lead the process and in many cases deal directly with lenders. But for a mortgage with typical terms and size, it is usually not the player who ultimately owns the loan.

One important reason is the unique system of taxpayers in the US housing market through government-subsidized enterprises. Fannie Mae FNMA 1.27%

and Freddie Mac FMCC 0.87%

buy loans of origin, guarantee them and resell them to investors as agency bonds. In turn, the economy of many origins is ultimately driven by the amount of loans they produce and sell via Fannie or Freddie. This business model also avoids loan risk and requires less capital, which makes it attractive to investors.

But selling loans is quite complicated. To get someone else interested in buying or trading loans negotiated by third parties, many things need to happen to commodify a 30-year mortgage loan. Originators sell mainly in standardized pools of mortgage loans organized in half-point buckets interest rates, such as 2.5% or 3%. Investors buy slices of these pools in the form of a securitization.

The rate is not the same as what the borrower pays. A 3% bond can end up in a 2% pool. This is because parts of the interest pay other transformation services to further standardize the loan. One part is for Fannie or Freddie to cover their base costs to guarantee the mortgage, plus various adjustments based on the individual mortgage. Another part is for a server that handles the collection of the borrower and then pays out to investors, tax authorities, and so on.

In return for these lengthy stream of fees, conscripts carry certain risks. First, as interest rates fall, more mortgage lending is refinanced and repaid early, causing service providers to lose payment flows. Servers also cover some missed payments before a mortgage is actually defaulted. In an economy where many people miss payments, it can bite. The increase in deferral of payments during the pandemic, for example, has weighed heavily on service workers.

Developers may also need to use private mortgage insurance if the loan-to-value ratio for a guarantor is too low, perhaps because the borrower drops less than 20%. Borrowers can pay this fee directly or indirectly through a higher mortgage rate.

Even after paying for service and credit risk, a founder still cannot always count on a predictable selling price for every mortgage. Mortgage rates or the relative prices between buckets may fluctuate during the long closing period, but lenders like ‘offers’ on the offered rates. There is a large market for future delivery of mortgage loans, known as the TBA market, or “To Be Annonce”, which is used to effectively hedge the rate risk for borrowers. But it has a cost that can vary with how long the protection lasts.

An emerging technology component of the business is the use of data and analytics to reconcile the rate offered on a mortgage with how it can be hedged and sold, explains Vishal Garg, CEO of Better, a digital homeowner business. “You can be a much better market participant by meeting the demand of the end investor to the consumer,” he says. ” A traditional lending officer may not consider all the scenarios. ‘

Originators have some natural counterparties that accept interest rate risk. The demand from investors such as investment trusts in mortgage lending, informed by how cheaply they can fund themselves, helps raise prices.

A major way that rate risk manifests is the rate at which people pay upfront. This in turn can affect what investors are willing to pay, because bonds obtained from the mortgage loan are in fact shorter. Although the origin of the volume enjoys benefits when many people refinance, they can earn less if they sell mortgage loans. Of course, if the Federal Reserve buys mortgage bonds, and if the rates on other fixed income are so low, the profits that the mortgage lender sells can remain quite large.

Smart investors will understand how changes in the market would make their portfolios succeed.

Write to Telis Demos by [email protected]