Looking to identify the nation’s most statistically undervalued multifamily properties, Enodo, an automated underwriting platform for multifamily real estate, recently applied its price and cap rate prediction algorithms to a database of 1.7 million assets. The properties identified by the research are statistically underpriced for their respective markets, and rents in these properties can be raised by as much as 20 percent with minimal loss of occupancy.
Using machine learning, Enodo developed a method to systematically identify undervalued assets based on their individual performance among competing properties. The algorithm flags properties that are either prime value-add candidates or properties that are charging rents below market rate. After identifying the undervalued assets, Enodo applies its machine-generated cap rate to calculate the market value for each asset as it is currently performing, and identifies the potential yield if acquired.
There were common features among many of the properties that this research uncovered as being underpriced for their market. Age, for example, was a determining factor since the highest numbers of undervalued properties were those built in the early ‘70s or mid-’80s.
Size, too, played a part as large, garden-style properties with approximately 500 units were the most likely to be undervalued. There are two reasons for this; first, these properties are often distanced from competing properties, thus there may be fewer comps to study when setting prices. Second, properties with high unit counts often optimize pricing based on other units in the same property, which may make them less likely to adjust prices upward for market trends.
To identify the most statistically undervalued assets nationwide, the researchers filtered out properties based on investor input, including anything that was built or renovated before 1960 or after 2006, as well as properties with fewer than 25 or more than 600 units. Affordable, public, senior and student housing properties were also struck from the list, as these behave differently than market rate multifamily properties.
“By focusing our efforts on identifying undervalued assets, we’ve created an entirely new line of business for Enodo,” said Enodo COO, Thomas Delaney. “Our clients are continually searching for the needle in the haystack, and Enodo just invented the metal detector. The implications are huge and we will be cautious and diligent as we explore this new opportunity.”
The methodology excluded other properties, including those with units that are erratically small or large for their markets, as well as assets whose amenities are not accurately represented in online listings. These properties usually represent special cases that are not necessarily good candidates for value-add renovations.
“Enodo currently predicts market rents nationally with under 5.5 percent median error, and cap rates within 0.35 percent of actuals. This gives us all the insight we need to determine when an asset is renting below market, and what the incremental income from rent increases will be worth to investors in each market,” said Enodo CEO, Marc Rutzen.
The algorithms were trained using data from more than 21 million apartments, 12,500 operating expense statements and over 3,000 closed multifamily transactions nationwide to develop its automated valuation model. By predicting rent, operating expenses and cap rate separately, Enodo is able to accurately replicate the income approach used by multifamily appraisers.
“This advancement by our data science team has created the ability for Enodo to monitor the entirety of the multifamily market and automatically identify opportunities to acquire properties at below market value,” Rutzen said.