House prices are rising at their fastest rate on record—and the pace of increases is accelerating. Monthly data from Zillow and from the S&P CoreLogic Case-Shiller Index both indicate that the current surge in U.S. home prices is unlike anything seen before, including the bubble of the 2000s.

In Phoenix and Austin, Texas, house prices have surged by a quarter since the pandemic began. Even in the metro areas that have had the biggest job losses and outflows of people, such as New York and San Francisco, house prices are nevertheless up about 10%.

So far, this hasn’t been showing up in the inflation data. In fact, the “shelter” component of the consumer price index has been growing at its slowest rate since the housing bust. That’s led some people to argue that inflation is worse than it appears.

The surge in house prices could be a harbinger of accelerating inflation if it’s being driven by savvy investors betting that the cost of renting is poised to rise. Higher house prices could also cause inflation to speed up by putting more spending power in the hands of consumers who could then monetize their extra home equity to pay for goods and services.

But house prices aren’t the same as the cost of housing—just as dividends aren’t the same as stock prices. For one thing, many Americans rent their home. And even home buyers are affected by factors beyond simply the headline price, such as the minimum down payment, credit requirements, taxes, interest rates, homeowners association fees, and mortgage insurance. Most people care about their upfront cost and their monthly payment. Those are the outputs of complex equations involving many different inputs.

Since the middle of 2019, and especially since the start of the pandemic, mortgage interest rates have plunged, inflating home values relative to rents. Home values have appreciated in parallel with the decline in yields even though rents have been flat or falling in most of the country.

The easiest way to see this is to look at the change in the gross rental yield of all U.S. housing as estimated by Zillow and compare it to the change in the 30-year mortgage rate calculated Freddie Mac. While the relationship isn’t exact, it’s clear that the decline in financing costs is linked to the drop in rental yields—which in turn explains the rise in house prices.

The impact hasn’t been uniform across the country, which partly reflects differences in whether housing is purchased as a speculative investment or is simply used as a place to live. In the greater New York City metro area, rental yields have dropped from 7% before the pandemic to 5.7% in April. But in the Miami and Atlanta metro areas, rental yields have declined only slightly from 7.4% to 7.0%. In the San Francisco Bay Area, the gross rental yield fell from just 3.3% before the pandemic to 2.8% as house prices rose by 8% even as rents fell by 10%.

Zillow also has price and rent data at the ZIP Code level, which makes it possible to compare changes in rents, house prices, and rental yields at a granular level. Some of the biggest differences have emerged between ZIP Codes in the same city. In general, rents have fallen the most in the priciest ZIP Codes within cities and fallen the least in the cheapest ZIP Codes.

New York City runs the gamut. In Manhattan’s 10002 ZIP Code, for example, which is on the Lower East Side, house prices have dropped 20%, but so have rents, with the result that it’s one of the only major ZIP Codes where rental yields have increased since the pandemic began, from 3.4% to 3.5%.

At the other end of the spectrum is 10458 in the Bronx, next to the Botanical Garden and home to Fordham University. There, rents have been flat even as house prices have jumped 17%, with the result that rental yields fell by almost 1.8 percentage points, from a little over 11.2% to just under 9.5%. That’s the single-biggest decline among America’s most populous ZIP Codes.

Paying more for housing today in the form of lower rental yields would make sense if rents are poised to rise sharply. In that scenario, rental yields would normalize without a decline in house prices, and the associated rise in housing costs would be a major inflationary impulse. That would presumably push up mortgage rates, dampening the returns for homeowners, but not by enough to hurt most people.

If mortgage rates stay low, however, rents wouldn’t have to rise sharply at all from their current levels. In that case, the national surge in house prices would have turned out to be a one-off re-pricing with no lasting macroeconomic impact.

Ultimately, the answer will depend on what happens in the rest of the economy—which is why housing costs, rather than house prices, are tracked in the inflation data.