Real Estate Data for the Rest of Us

articles about “Mortgages

Why Conforming Loan Limits Should Be Part of Housing-Finance Reform — But Aren’t

The current system of conforming loan limits falls far short of reflecting the actual differences in local home prices and ends up favoring borrowers in lower-cost markets. More than 30% of homes for sale in many California and Northeastern markets are above the conforming loan limit, versus less than 10% in much of the rest of the country. The leading housing-reform bill – the Johnson-Crapo bill – keeps loan limits intact, while making them matter less.

The U.S. housing finance system is national, but housing markets are local. Local markets often face different housing challenges: today, California’s coastal communities face an affordability crisis, while Florida is dealing with foreclosures and Detroit and other Midwestern cities are wrestling with neighborhoods of vacant homes. The housing finance system – as well as other national housing policies – needs to serve a country where local home prices in some markets are 10 times as high as in others, and where local and state laws affect how much new construction is allowed, how long foreclosures take, and more.

One way the national housing finance system explicitly accounts for local market differences is through the conforming loan limit – the maximum dollar amount of a home loan that Fannie Mae and Freddie Mac can guarantee (they purchase mortgages from originators, package them into securities that investors buy, and guarantee the payments in the event that a borrower defaults). Loans above this limit are “jumbos.” Until 2008, the conforming loan limit was the same throughout the country (Alaska and Hawaii are treated separately), but the 2008 Housing and Economic Recovery Act (HERA) granted “high-cost areas” higher conforming loan limits to reflect local price differences. Now, in 2014, the limit in high-cost areas is up to $625,500, which is 50% above the $417,000 limit that applies in most of the country.

For a borrower, one advantage of conforming mortgages is that they typically (though not currently) have a lower mortgage rate than jumbo loans; also, jumbo loans often require higher down payments, a higher credit score, or a lower debt-to-income ratio. Part of the rationale for having conforming loan limits – rather than allowing loans of any amount to be guaranteed – has been to target the benefits of conforming loans to borrowers buying more modestly priced homes and not to borrowers buying luxury homes.

This week, the Senate will markup the leading housing-reform proposal – the Johnson-Crapo bill – which would overhaul the housing finance system by replacing Fannie Mae and Freddie Mac, introducing an explicit but limited government backstop, and creating a new fund for affordable housing. Despite these major changes, the bill keeps current conforming loan limits intact. Is this implied vote of confidence in the current system of conforming loan limits warranted? Let’s assess how well these limits reflect local price differences.

Conforming Loan Limits Don’t Reflect Differences in Local Housing Prices
To see whether the conforming loan limits bind equally tightly across local housing markets, we calculated the share of for-sale homes on Trulia in each of the 100 largest U.S. metro areas that is above the local conforming loan limit, assuming an 80% loan-to-value (LTV) mortgage (that is, a 20% down payment — see endnote for details). If loan limits fully reflected local housing market differences, then a similar share of homes in every metro would be above the local loan limit. But the results show that a much higher share of homes is above the local loan limit in some metros than in others.

In the San Francisco metro area, 61% of homes for sale are priced above the conforming loan limit (the local limit, $625,500, equal the loan amount for an 80% LTV loan on a $781,875 home); the typical San Francisco home priced near the loan limit is a modest 1500 square feet. In several other California metros, as well as in New York and Boston and their respective neighbors of Fairfield County, CT, and Middlesex County, MA, 30% or more of the homes for sale are above the local conforming loan limits. These 10 metros with the highest share of for-sale homes above their respective local loan limit all are “high-cost” areas with limits above the national baseline of $417,000, but even with their higher loan limits, they have the highest share of homes for sale that would require jumbo loans.

Housing Markets With Highest Share of
For-Sale Homes Above Local Loan Limit

# U.S. Metro Conforming loan limit % of for-sale homes above local loan limit Median size of for-sale homes near local loan limit, square feet
1 San Francisco, CA




2 San Jose, CA




3 Fairfield County, CT




4 Orange County, CA




5 Ventura County, CA




6 San Diego, CA




7 Middlesex County, MA




8 New York, NY-NJ




9 Oakland, CA




10 Boston, MA




Note: Among 100 largest U.S. metros

The metros with the smallest shares of for-sale homes above the local limit include El Paso, Little Rock, and Memphis in the South and Southwest; Dayton, Toledo, Akron, and Detroit in the Midwest; and the upstate New York metros of Rochester, Syracuse, and Buffalo. Relative to the low home prices in these metros, the $417,000 loan limit is very generous. In some of these markets, 4000 square-foot homes typically fall within the conforming loan limit. More broadly, in 48 of the 100 largest metros, fewer than 10% of homes for sale would be above the local conforming limit.

Housing Markets With Lowest Share of
For-Sale Homes Above Local Loan Limit

# U.S. Metro Conforming loan limit % of for-sale homes above local loan limit Median size of for-sale homes near local loan limit, square feet
1 El Paso, TX  $        417,000



2 Dayton, OH  $        417,000



3 Toledo, OH  $        417,000



4 Rochester, NY  $        417,000



5 Syracuse, NY  $        417,000



6 Little Rock, AR  $        417,000



7 Akron, OH  $        417,000



8 Memphis, TN-MS-AR  $        417,000



9 Detroit, MI  $        417,000



10 Buffalo, NY  $        417,000



Note: Among 100 largest U.S. metros

In short: despite higher loan limits in high-cost areas, conforming loan limits do not reflect the huge cost differences in housing markets across America. For loan limits to reflect the differences across local markets, the spread between loan limits in inexpensive markets and loan limits in expensive markets would have to be much larger than it is today. The gap between the national baseline of $417,000 and the maximum high-cost-area limit of $625,500 is much smaller than the difference in home prices; furthermore, markets with the national baseline loan limit of $417,000 include a wide range of local markets, from mid-to-high priced metros like Denver, Portland, and Miami to much cheaper metros like El Paso and Dayton.

But – you might ask — aren’t the people buying homes in expensive markets rich? Not necessarily: in expensive markets, households spend more of their income on housing. In fact, Census data show that median household incomes for homeowners whose homes are near the conforming loan limit were significantly lower in metros where more of the market is priced above the local loan limit. For instance, among households that own homes near the local loan limit, the median income of those households in San Francisco (where 61% of homes are above the local loan limit) is less than $120,000, while the median income of those households in Houston (where just 13% of homes are above the local loan limit) is $172,000. In other words, a San Francisco household is less likely to own a home that falls within the local conforming loan limit than a Houston household with the same income. More broadly, the correlation across metros between (1) the share of homes over the local loan limit and (2) median household incomes among homeowners whose homes are near the limit is negative (-0.37) and statistically significant. Therefore, the current system of conforming loan limits isn’t sufficiently aligned with local home price differences to give households at the same income level similar access to conforming loans.

The Complicated Politics of Loan Limits and Housing Policy
If conforming loan limits don’t reflect actual differences in local home prices, why not change the loan limits as part of overhauling the entire national housing finance system? Because changing loan limits would mean hurting or helping some local areas more than others, and that’s tricky politics. Lowering the limit in low-cost areas or raising it further in high-cost areas are both politically unrealistic:

  • Lowering loan limits in specific local markets would meet resistance from those areas’ elected officials. A lower loan limit would push more borrowers into the jumbo category, which, as explained above, means mortgages that are (at least historically) more expensive and more difficult to get.
  • Raising the loan limit further in high-cost areas would benefit borrowers in those areas but would increase the share of mortgages guaranteed or insured by the government – which runs counter to a central goal of national housing-finance reform, which is to reduce the mortgage market’s dependence on government.

Furthermore, aligning loan limits with home prices probably wouldn’t get bipartisan support. The current system of conforming loan limits favors “red” America, where home prices tend to be lower than in “blue” America. In the reddest housing markets – the 32 of the 100 largest metros where presidential candidate Mitt Romney got more votes than President Barack Obama in the 2012 Presidential election — only 10% of homes for sale are above the local conforming loan limit. But in the bluest housing markets– the 28 of the 100 largest metros where Obama beat Romney by at least 20 points – nearly twice as many homes for sale (18%) are above the conforming loan limit. Six of the 10 metros with the highest share of homes over the local limit (from the first table, above) are deep blue, including San Francisco, San Jose, Middlesex County, MA, New York, Oakland, and Boston.

Looking across the 100 largest metros, the correlation between (1) Obama’s margin over Romney in 2012 and (2) the share of homes for sale above the local loan limit is 0.39 (and statistically significant). Because the current system favors red housing markets, a change that would get loan limits more in line with home price differences would disproportionately help blue areas – and might not get much Republican support. (Side note: not all national housing policies favor red America. For instance, the average household in an Obama-voting state claims 66% more for the mortgage interest deduction than the average household in a Romney-voting state. The mortgage interest deduction benefits people where home prices are higher and where incomes are higher, which tend to be blue states.)

What will happen if the conforming loan limits don’t change, even as the housing finance system gets overhauled? Ironically, by removing the implicit subsidy and therefore reducing the cost advantage to borrowers of conforming loans, the Johnson-Crapo bill and other housing reform proposals might end up making conforming loan limits largely a moot point.  But the passage of these reforms is far from a done deal, and their implementation would be years into the future.

Until then, the failure of conforming loan limits to reflect local housing differences is likely to get worse, not better. In the housing recovery, home prices have been rising faster in the less affordable markets, so the gap in home prices between lower-priced markets and higher-priced markets has been widening. Our current system of conforming loan limits will fail to account for the large and growing differences in local housing markets.

Note: The conforming loan limits for 2014 are available here. Loan limits are determined separately for each county, but counties within a metropolitan area are all assigned a uniform loan limit. We calculated the share of for-sale homes, excluding foreclosures, for which an 80% LTV loan would exceed the local conforming loan limit. We also calculated the median square footage among homes for which an 80% loan would be within 10% of the local conforming loan limit. Both of those calculations were based on for-sale listings on Trulia as of March 22, 2014. We used the 2012 American Community Survey Public Use Microdata Sample to calculate median household income among owners of homes for which an 80% loan would be within 10% of the local conforming loan limit. Sample sizes limited the analysis of income to a few dozen metros. “Statistically significant” means at the 5% level.


Shutdown Hasn’t Hurt October Asking Home Prices So Far

Asking home prices in the first half of October are up 1% versus September. Prices changes are no different in metros more dependent on the federal government, like Washington DC, compared with metros less directly affected by the shutdown.

Jed Kolko, Chief Economist
October 16, 2013

How has the two-week shutdown of the federal government affected home prices? The main sales-price indexes won’t tell us until 2014: homes going under contract in October will close in November (or later), and November sales prices will get reported starting in January. But the Trulia Price Monitor shows how asking prices – a leading indicator of sales prices – are trending almost in real time, adjusting for both the mix of listed homes and for seasonality. This morning we analyzed asking prices between October 1 and October 15.

Finding the Effect of the Shutdown on Asking Home Prices
Nationally, asking home prices are up 1.0% between September and the first half of October, seasonally adjusted. This partial month-over-month increase is roughly in line with the month-over-month increases over the past few months. Before the shutdown started, several factors were already cooling down price gains, including expanding inventory, higher mortgage rates, and declining investor activity. Therefore, comparing how much prices have risen in October to date with previous months can’t, by itself, show whether the shutdown has affected asking prices.

Instead, to tease out the effect of the shutdown on asking home prices, we looked at price trends across individual metros. We compared price changes in metros where the local economy is more dependent on the federal government – like Washington D.C., of course, but other metros around the country as well – with prices changes in metros where the local economy is less dependent on the federal government. (Our measure of dependence on the federal government – and therefore likely impact from the shutdown – is the share of local wages coming from the feds.)

… continue reading


When Will the Jump in Mortgage Rates Bite?

Recent history shows that spiking mortgage rates take a big chomp out of refinancing immediately and smaller nibbles out of sales three months later. Longer term, the impact of rising rates is typically offset by stronger economic growth. The overall effect of rising rates may turn out to be more bark than bite.

Jed Kolko, Chief Economist
September 11, 2013

Ever since mortgage rates started their steep climb in early May, we’ve all been on high alert, watching how higher rates will affect the housing market. For a would-be buyer calculating the mortgage payment on their dream home, the effects are obvious: the increase in the 30-year fixed rate from 3.59% in early May to 4.73% at the end of August (according to the Mortgage Bankers’ Association, or MBA) means a 15% increase in the monthly payment on a $200,000 mortgage. That should deter homebuyers and reduce mortgage applications, sales, and prices, right? In theory, yes, but of course the real world is much more complicated. Mortgage rates aren’t rising all on their own: other housing and economic shifts are happening at the same time.

Fortunately, the recent past is a useful guide. The 30-year fixed rate jumped .47 points in May 2013 and .51 points in June 2013, comparing the levels at months’ end (MBA). (Side point: the 30-year fixed reached 4.80 this morning, September 11, .22 points higher than at the end of June, which means July, August, and early September have seen much milder increases compared with the May & June spike.) But this year isn’t the only time when mortgage rates have jumped up: they also climbed at least .4 points in seven other months since 1999. With some simple time-series regressions, we traced out the typical paths of mortgage applications, sales, and prices in the months immediately after a mortgage rate spike. 

The Month-by-Month Impact of a Rate Spike
Our analysis of mortgage rates and other housing data from January 1999 through April 2013 – just before the current spike – shows that mortgage rates hit refinancing applications (MBA) earlier and harder than any other measure of housing market activity. (Not all of the data series are available back to 1999.) Here’s the timeline of what typically happens when rates spike by half a point in a month:

  • The month when rates spike: Refinancing applications typically fall by 45% in the month of a spike, with further falls one and two months after mortgage rates jump, compounding the effect. The drop in refinancing applications this year was roughly 50% cumulatively over two months, which actually looks small compared with similar rate jumps in the recent past.
  • 1-2 months after the spike: Pending home sales and home-purchase mortgage applications typically decline slightly, though the effect isn’t statistically significant. New home sales also decline modestly.
  • 3 months after a spike: New home sales and existing home sales drop. That means that the May mortgage rate spike should show up most strongly in August new home sales and existing home sales, both of which will be reported later this month (on September 25 and September 19, respectively).

Compared with the impact on refinancing, the impact of a rate spike on home-purchase mortgage applications and sales volumes is very small and not always statistically significant.

Housing indicator

Month of biggest mortgage rate impact

Effect in month of biggest impact*

Statistically significant?

Which report will show biggest impact of May 2013 rate spike

Refinance mortgage applications (MBA) Same month as rate spike (plus additional impact 1-2 months after)


Yes May data (already reported)
Pending home sales (NAR) 1 month after


No June data (already reported)
Home-purchase mortgage applications (MBA) 2 months after


No July data (already reported)
New home sales (Census) 3 months after (plus modest impact 1-2 months after)


Yes August data, to be reported Sept 25
Existing home sales (NAR) 3 months after


Yes August data, to be reported Sept 19
Sales prices (Case-Shiller, FHFA) No short-term impact


Note: The “effect in month of biggest impact” equals the month-over-month change in the indicator for a 0.5 point rate spike, relative to when the mortgage rate doesn’t change, in percentage points. 

The Longer-Term Impact of Sustained Rate Increases
Even if the immediate impact of mortgage rate spikes is small – aside from the huge effect on refinancing – shouldn’t sustained rate increases should depress housing activity? Again, recent history tells a more complicated story. Since 1999, mortgage purchase applications and all measures of sales activity – NAR pending home sales, NAR existing home sales, and Census new home sales – have actually been higher when mortgage rates were higher. Sales prices were also the same level or higher (depending on the sales price index) when mortgage rates were higher compared to periods of lower rates. Of all the measures of housing activity, only refinancing applications were lower during periods of higher mortgage rates.

Here’s the missing piece of the puzzle: over the past decade and a half, mortgage rates have been higher when the economy was doing better. Since 1999, the correlation between the monthly unemployment rate – a good, if imperfect, measure of how the economy is doing overall – and the 30-year fixed rate was -0.8, making it a very strong relationship.

Furthermore, every measure of housing activity (except refinancing activity) improved when the overall economy did better. That means that a stronger economy is associated with BOTH higher mortgage rates AND more sales, higher home prices, and more home-purchase mortgage applications. That’s why these measures of housing activity go up when mortgage rates are higher.

If we statistically remove the effect of changes in the overall economy (by including the unemployment rate as a control in a simple statistical regression), then we see exactly what we’d expect: mortgage applications, sales, and home prices are all lower when mortgage rates are higher. In other words: all else equal, higher mortgage rates do depress housing demand.

As Rates Rise, All Else Won’t Be Equal
When it comes to mortgage rates, all else is never equal. Three other factors will complicate or even offset the impact of the recent rise in mortgage rates, even if rates continue to climb: the strengthening economy, expanding inventory, and looser mortgage credit:

  1. A post-recession economic recovery tends to push interest rates higher as demand for credit increases and if investors start to worry more about inflation. Furthermore, the Fed has said it will taper its bond-buying only if the economy seems strong enough to weather it. Both through market forces and the actions of the Fed, rising rates should be accompanied by a strengthening economy.
  2. Inventory has been expanding for the past six months on a seasonally adjusted basis. More for-sale inventory on the market slows price gains: in fact, the Trulia Price Monitor and other price indexes have been slowing down before the May rate spike could have affected prices, pointing to expanding inventory as a likelier explanation for the price slowdown. While rising rates and expanding inventory should both slow down prices, these same two factors should pull sales in opposite directions. All else equal, rising rates should slow sales, but expanding inventory should boost sales – since more homes can be sold if there are more homes for sale. Therefore, even though this month’s sales data should be slowed by sales, it could be lifted by rising inventory.
  3. Mortgage credit, though still tight, shows signs of loosening for two reasons. First, as they face diminishing demand for refinancing, banks might look to expand their home-purchase lending instead. Furthermore, new mortgage rules coming into effect next year will give banks more clarity about which loans are considered risky, hopefully making banks more willing to write mortgages deemed to be safer. The negative impact of rising rates, therefore, could be partially offset by looser mortgage credit.

All told, the housing market and the economy have a lot of moving parts. Aside from the sharp and immediate effect that rising mortgage rates have on refinancing, the impact of rising rates on the housing recovery is hard to pinpoint. This month’s sales reports, covering new and existing home sales from August, should show some decline from the May rate spike, but mortgage rates are just one of many factors affecting the housing recovery.


Rising Mortgage Rates Giving Would-Be Homebuyers Jitters

Rising mortgage rates are the top worry for people thinking of buying a home someday, and 56% of Americans say they would be discouraged from homeownership if rates reach 6%. But pay more attention to what consumers do than what they say.

After years of low-and-lower mortgage rates, the 30-year fixed rate shot up from a near-historic-low of 3.35% in early May to 4.46% in late June before settling back to 4.29% last week, according to Freddie Mac. The rate increase was sudden and steep, but not a surprise. Economists and forecasters have been waiting for rates to go up for two reasons: (1) the strengthening economy tends to push up rates, and (2) the Fed is expected to pull back on bond-buying and other measures that have kept rates low, which they reaffirmed in mid-June. By historical standards, rates are still low: remember that mortgage rates hovered around 6% for most of the 2000s, 7-9% in the 1990s, and above 10% in the 1980s. Nonetheless, the recent rate climb has been steep. 

What Consumers Think of Rising Rates
Consumers are anxious about rising mortgage rates. Trulia surveyed more than 2,000 people during June 24-26, after rates rose sharply. We asked what their biggest worry would be if they were to buy a home this year. Among all consumers who plan to buy a home in the future, 41% said their top worry is that mortgage rates would rise before they actually bought. The next biggest worries were that prices would rise before they actually bought (37%) and that they wouldn’t find a home for sale that they like (36%).

How high do mortgage rates have to rise before consumers are discouraged from buying a home? Among consumers who plan to buy a home someday, 13% said that mortgage rates of 4% (which is what the rate had climbed to when the survey was conducted) were already too high for them to consider buying a home. Another 20% said they’d be discouraged from buying a home if rates reach 5%; yet another 22% said they’d be discouraged from buying a home if rates reach 6%. Combining these groups, 56% of consumers who plan to buy a home someday would be discouraged from doing so if rates reach 6%. Among renters who plan to buy a home someday, 62% would be discouraged from doing so if rates reach 6%.

… continue reading


Buying Cheaper Than Renting Til Mortgage Rates Hit 10.5%

Nationally, at today’s prices and rents, buying would be cheaper than renting until the 30-year fixed rate reaches 10.5%. San Jose has the lowest mortgage rate “tipping point” at 5.2%, followed by San Francisco and Honolulu.

The recent rise in mortgage rates has made buying a house a little more expensive: the increase in the 30-year fixed rate over the past month from 3.4% to 3.9% (Freddie Mac) raised the monthly payment on a $200,000 mortgage by $56, or 6%. However, because mortgage rates are still near long-term lows, and because prices fell so much after the housing bubble burst and remain low relative to rents even after recent price increases, buying is still much cheaper than renting. That means that the recent jump in rates doesn’t change the rent-versus-buy math much.

Rates are likely to keep rising, but how far must rates rise before buying a home starts to look expensive relative to renting? To answer this, we updated our Rent vs. Buy analysis with the latest asking prices and rents from March, April, and May 2013. Following our standard approach, we calculated the cost of buying and renting for identical sets of properties, including maintenance, insurance, taxes, closing costs, down payment, sales proceeds, and, of course, the monthly mortgage payment on a 30-year fixed-rate loan with 20% down and monthly rent. We assume people will stay in their homes for 7 years, deduct their mortgage interest and property tax payments at the 25% tax bracket, and get modest home price appreciation (see the detailed methodology and example here). Here’s what we found:

Buying remains cheaper than renting so long as mortgage rates are below 10.5%. At 3.9%, the current 30-year fixed rate according to Freddie Mac, buying is 41% cheaper than renting nationally. With a 5% mortgage rate, buying is still 34% cheaper than renting nationally. Mortgage rates would have to rise a huge amount – to 10.5% – to tip the math in favor of renting, which isn’t impossible. Rates were that high throughout the 1980s, but have been consistently below 10.5% since May 1990.

Each local market, of course, has its own mortgage rate “tipping point” when renting becomes cheaper than buying a home. At 3.9%, buying is cheaper than renting in all of the 100 largest metros, which means the tipping point is above 3.9% everywhere. The tipping point is lowest in San Jose, which would tip in favor of renting if rates reach 5.2%. It’s between 5% and 6% in San Francisco and Honolulu, and between 6% and 7% in New York and Orange County, CA.

10 Metros with the Lowest Mortgage-Rate Tipping Point

# U.S. Metro Mortgage rate below which buying is cheaper than renting
1 San Jose, CA


2 San Francisco, CA


3 Honolulu, HI


4 New York, NY-NJ


5 Orange County, CA


6 Los Angeles, CA


7 San Diego, CA


8 Ventura County, CA


9 Sacramento, CA


10 Oakland, CA


… continue reading