Costs rose consistently, values only half as often. That, in a nutshell, is why the latest Cost vs. Value Report’s overall cost-value ratio ended two years of gains by slipping to 62.2% for 2015. That’s a 3.9-point drop from last year but still the second-best number in the past five.

This 2015 edition of the Cost vs. Value Report, which compares changes in job costs with Realtors’ perceptions of what those jobs bring to a home’s price at resale, reflects the mixed messages that were delivered by the housing market in 2014. In housing, it took until October before monthly sales of existing homes topped their year-earlier numbers, and the increase in median price paid for a home dropped by roughly half between last winter and last summer, according to the National Association of Realtors. New-home sales also were lackluster.

That ho-hum performance might have influenced Realtors nationwide when they sat down last fall to predict the boost a remodeling or replacement project would bring to a home’s resale value. Compared with the previous year’s survey, the Realtors increased by up to 11.6% the value of 17 projects, but they reduced 18 projects’ predicted payback by as much as 8.32%. The average change was a gain of just 0.29%.

Meanwhile, the estimated cost of a pro doing those projects rose 4.22% from what it was in the 2014 Cost vs. Value report. As a result, only five projects saw their cost-value ratios rise in the 2015 report. The midrange roofing replacement led the way, rising 5.9%. The other four gainers were the midrange garage door replacement, up 5.6%; the 20-gauge steel replacement entry door, up 5.4%; vinyl siding replacement, up 3.2%; and the sole upscale project: a fiberglass replacement entry door, up 1.7%.

The other 30 projects declined. Ironically, this year’s biggest drop was the same project that had the biggest gain in last year’s report: installing a backup power generator. Its cost-value ratio sank 11.3% after having jumped 28% in the 2014 report. That project’s cost-value ratio now stands at 59.9%

In both cases, weather may have played a role: Events such as 2013’s Superstorm Sandy on the East Coast were likely fresh in Realtors’ minds when they assessed how installing a backup generator to supply electricity during a utility outage would affect a home’s value at resale. In the absence of a weather event like Sandy in 2014, Realtors had perhaps begun to see the generator’s value in a new light.

Two-story additions saw the second-biggest decline in this year’s report, falling back 10.7%. Then came composite deck additions, an upscale project that lost 9.5% of its expected return, and two big projects that both fell 8.6%: master suites and major kitchen remodels.

Our Cost vs. Value Reports in recent years have made note of how big-ticket projects were enjoying ever-bigger payoffs. For instance, in our 2014 summary we noted that the cost-value ratio for attic bedrooms had risen 15.6% and basement remodels went up 10.4% from the previous year. “The fact that their comparatively high initial cost is balanced by a higher value at resale than at any time since the peak year of 2005 signals a return of confidence in the value of remodeling,” we wrote last January.

Not so fast, it now appears. The cost-value ratio for attic bedrooms sank 8.4% this past year, while basement remodeling projects lost 6.2%. At 77.2% and 72.8% respectively, both are still above where they were two years ago, but the drop does suggest that the bloom is off the rose.

One project was added this year: manufactured stone veneer. It joined Cost vs. Value with a splash, ranking second among all projects with a cost-value return of 92.2%. The only project that beat it was for a replacement steel entry door, which delivered a cost recouped of 101.8%.

The cost-value ratio expresses resale value as a percentage of construction cost. When cost and value are equal, the ratio is 100%; when cost is higher than value, the ratio is less than 100%; when value is higher than cost, the ratio exceeds 100%. That means the replacement steel entry door is the only project that, on a national basis, more than pays back its investment in the form of a better home resale price.

  Within the 102 markets surveyed in the 2015 report, 269 of the 3,672 total projects—that’s 7.3%—had payoffs topping 100%. The steel entry door job was most common, recouping all its costs and more in 43 markets. Following that were midrange garage door replacements, which topped 100% in 28 markets, manufactured stone veneer jobs (27), upscale garage-door replacements (17), midrange wood window replacements (16), and minor kitchen remodels (15).

All told, 26 of the 36 projects could claim a cost-value ratio topping 100% in at least one market. More often than not, that market was San Francisco, where the average payback for all projects was 100.3%. Honolulu was the second-best bang-for-the-buck market at 94.6%. Then came San Jose, Calif., at 87.1%; Riverside, Calif., 84.1%; Sarasota, Fla., 80.7%; Austin, Texas, 78.6%; and New Orleans 77.6%.

This year’s report includes 10 new cities that were brought in to help ensure that the overall Cost vs. Value Report includes the top markets nationwide. From east to west, those 10 new markets are: Augusta, Ga.; Chapel Hill, N.C.; Daytona Beach, Fla.; Winter Haven, Fla.; McAllen, Texas; Ogden, Utah; Spokane, Wash.; Santa Rosa, Calif.; Bakersfield, Calif.; and Stockton, Calif. Their collective Cost vs. Value ratio for all projects was 61.3%.

(Important note: Cost vs. Value uses cities as shorthand for that metropolitan area’s market. Some are modestly sized, such as Chapel Hill, N.C. Others are enormous: Los Angeles encompasses all of L.A. and Orange counties, while the New York metro area takes in all of Long Island and half of New Jersey plus parts of Pennsylvania.)

As in past years, the West Coast stands out when one breaks down the Cost vs. Value numbers by census division. The Pacific division (encompassing California, Oregon, and Washington plus Alaska and Hawaii) posted an overall cost-value ratio of 74.0%. That’s nearly 10 points higher than the next division and 21 points better than the payoffs available in the Middle Atlantic states of New York, New Jersey, and Pennsylvania.

It’s no surprise that replacement jobs—such as door, window, and siding projects—generated a higher return than remodeling projects. That’s been the case since at least 2003. (See our historical charts for various project categories.) But the gap between the two categories widened by 3.8 percentage points this year even as both declined in value: Replacement projects showed an average return of 73.2% in this year’s report, just a smidgen below its 73.7% last year, while the cost-value ratio of remodeling projects sank to 60.8% in this year’s report from 65.1% last year.

When grouped by job type, siding jobs fared better than most, perhaps because of a rising perception nationwide of the value of curb appeal. Midrange vinyl siding replacement jobs were one of only five projects to rise in value, to 80.7% from 78.2%. A replacement job involving foam-backed siding slipped just half a point in value, to 77.6%, while the cost-value ratio for a fiber-cement replacement job dipped to 84.3% from 87.0%. Similarly, window jobs were no more than 2.1 points lower this year than in the 2014 report, and they ranked between ninth and 16th in overall payback.

In contrast, kitchen remodels declined as much as 6.6 percentage points, while the drop for bathroom additions and remodels was more modest, slipping 3.8 points or less.

As a general rule, the simpler and lower-cost the project, the bigger its cost-value ratio. Three of the four projects that cost less than $5,000 for a pro to do were ranked in the top five for cost recouped, and the other two were in the $5,000-to-$25,000 price range. No project costing more than $25,000 ranked better than 14th.

One of the Report’s great values is that it shows how the new-home market and competition for jobs influence remodeling projects. For instance, when Realtors took the Cost vs. Value survey in the pre-housing-boom year of 2003, the average project cost $38,286. The Realtors said that those jobs delivered $31,591 in resale value, producing a cost-value ratio of 82.5%.

Two years later, as the housing boom began, the cost-value ratio climbed to 86.7%. During the next two years, in 2006 and 2007, the rush to build new homes meant that remodelers and materials both were in short supply. Project costs rose 17.3% in 2006 and another 10.5% the next year, while the value of that work—possibly because of the glut of new homes entering the market—rose 2.9% and 1.8%, respectively. That pushed the overall cost-value ratio down to 70.1%.

Then came the crash. Costs sank 10.4% in 2010, another 1.9% in 2011, and 6.0% in 2012, in part because remodelers were scrambling for work. But Realtors’ valuation of the payback of those projects fell even more, dropping 22% between the 2010 and 2013 reports. Since then, a rebound in housing prices pushed the ratio in last year’s report to 66.1%. And now, the so-so market brings us back to 62.2% in this year’s report.

One could argue that this general decline might have been influenced by the fact that, over the years, more than 20 projects have been added to a Cost vs. Value Report that in 2003 had just 15 jobs. As it turns out, the rise and fall of those original 15 projects is virtually the same as for the Cost vs. Value as a whole, as is its end point: 62.0% for the 15 projects in this year’s report.