Tracking Informal Mortgages on Rental Properties

In past articles, we have discussed a number of informal homeownership issues, including land contracts or contracts for deed and heirs' property. In this piece, we look at informal multifamily lending and rental housing issues. While "informal" does not connote "undesirable," there are certain inherent risks when property is sold or financed outside of regulated, supervised markets such as the traditional mortgage market.

Hard money or informal mortgages are a type of private party lending used by rental property owners in a variety of situations. Owners of multifamily rental properties may use hard money loans for maintenance or upgrades on the properties. Hard money may also be used for property acquisition. Lenders may include family, friends, fellow investors, or private financiers. The loans may be relatively quick and easy to attain, but they are more likely to incur higher costs, including fees and interest rates. The loans may present risks for tenants if they are used to "flip" properties or overleverage property owners, placing the owners at risk of foreclosure and their tenants at risk of eviction.

In order to better understand trends in multifamily informal mortgages and the market conditions in which they are likely to occur, we analyzed real estate transaction data and compared lending activity with poverty data nationwide.

About the data
Our analysis relied on transaction data from CoreLogic to identify informal mortgages. The data set includes deeds and mortgage originations recorded at the county level by local registries or recorders of deeds. First, we identified multifamily properties using CoreLogic's property type field for apartments and other multifamily rental properties with more than four units. Duplex, triplex, and quadplexes were not included in these tabulations. To identify likely informal mortgages, we used CoreLogic's code for transactions involving a private party lender. Due to the nature of these transactions, it is quite possible that many are not reported. The following is our best approximation, given available data.

Private party lending
Overall, we found that among 3,962,984 deeds recorded, 148,318 (4 percent) involved a private party lender. Lenders were identified in the data by a first and last name (51 percent), as an unnamed "private individual" (5 percent), or as an "institutional investor" (43 percent); percents do not sum to 100 due to rounding. Of this latter group of corporations and agencies, 38,276 unique entities were reported as private party lenders with an average of 1.68 transactions per company.

A few relatively high-volume lenders existed, with eight lenders listed as the seller of more than 100 transactions. A subset of private party transactions (4 percent) were explicitly coded as "intrafamily sales." Family members are commonly the source of hard money loans. This figure likely underreports the phenomena of intrafamily lending, as coding is uneven across counties and often unreported.

The majority (94 percent) of 148,318 deeds involving a private party lender were recorded between 2000 and 2015. As shown in figure 1, the volume of private party lending varied over this time frame, with an average of 8,717 transactions per year. The annual volume ranged from a peak of 10,849 in 2005 to a low of 6,397 in 2009. In 2015, the volume was above average for this period, or 9,549 transactions.

Next, we analyzed the transaction amount. As shown in figure 2, we found that the largest share (23 percent) of these transactions were under $50,000 and the majority (50 percent) were under $150,000. However, a fairly sizable proportion (12 percent) were over $1 million. An institution or corporate entity was the lender in the majority of (73 percent) of these large-dollar transactions, a relatively high percentage compared to the percentage of corporate lenders among all private party transactions (43 percent).

Other aspects of the transactions also varied considerably. An interest rate was missing or zero for 86 percent of the records; however, the geographic distribution of available data was roughly similar to the full sample, with Florida somewhat overrepresented and California somewhat underrepresented. While the number of missing values requires cautious interpretation of the data, known interest rates varied between 0.5 and 70 percent. The majority (67 percent) were between 4.01 percent and 10 percent interest. The median interest rate was 7.5 percent. The loan term was also missing or zero for most records (63 percent). Terms that were recorded ranged from one to 100 years, with a median of five years.

The geography of private party transactions
Private party transaction records were geographically distributed fairly evenly, with larger volumes in high-population states, including California, Florida, and New York, and metropolitan areas, including Chicago, Los Angeles, and New York City. Within these geographies, we also sought to understand local patterns. Our underlying assumption was that private party lending would be more prevalent in submarkets where it is more difficult to secure traditional financing. Thus, we compared block group level lending with poverty data to better understand neighborhood-level trends. Overall, we found that high-poverty block groups had larger concentrations of private party lending.

The analysis was conducted using the same CoreLogic deeds data as well as CoreLogic tax records and 2015 Census American Community Survey (ACS) five-year estimates at the block group level. Since years of available deeds data vary by county, data from before 2012 were discarded. Based on CoreLogic metadata, all counties reported data back to at least 2012. Thus, we were able to analyze 36,240 private party transactions from 2012 to 2015 out of 148,318 total records (24 percent).

In order to quantify the universe of block groups with multifamily rental properties, CoreLogic tax records were used to identify the locations of 10,686,328 multifamily rental properties with more than four units (using the same land use categories as previous analysis of deeds records). We relied on CoreLogic's block group code to place each record, although the block group variable was somewhat noisy. For instance, 4 percent of all tax records had no reported block group value, 0.1 percent were invalid values, and an unknown number of values may be incorrect. In addition, a relatively small number of records (545 or 0.01 percent) were located in block groups that are unpopulated, according to the U.S. Census. The ACS table "Poverty Status in the Past 12 Months by Household Type by Age of Householder" was used to determine the percentage of households in poverty by block group.

To compare the levels of private party lending activity by poverty rate, CoreLogic data were aggregated to the block group level. For the analysis, we eliminated block groups that had no CoreLogic multifamily rental properties as well as any invalid or unpopulated block groups found in the CoreLogic data. Using the Census figures, three poverty thresholds were used: a 20 percent threshold denoting a high-poverty neighborhood, a 40 percent threshold denoting concentrated poverty, and a midrange threshold of 27 percent poverty, selected based on results of a previous analysis conducted by Desmond and Wilmers (2017).

At all three thresholds, the difference in private party transactions per 100 multifamily properties between high- and low-poverty block groups was statistically significant at a 95 percent confidence interval (p < 0.0001). As shown in figure 3, at least twice as many transactions occurred in high-poverty neighborhoods versus low-poverty neighborhoods. While the average number of transactions per block group is still quite low, the differential is troubling from both a social equity standpoint as well as a risk standpoint. In addition, as previously noted, the number of transactions is almost certainly underreported, particularly in lower-resourced counties where informal mortgages are more likely to occur.

Most rental properties are financed through formal lending institutions or are purchased outright. This suggests that researchers who wish to calculate landlord expenditures can be fairly confident that public mortgage and tax records cover the vast majority of property owners' mortgage burden. However, the volume of informal mortgages on rental properties is increasing and on par with levels witnessed before the foreclosure crisis. We also found that the rate of informal mortgages in poor neighborhoods is more than double that in non-poor neighborhoods. This suggests that property owners in low-income communities may face barriers to securing conventional loans for their buildings. These owners might be undercapitalized, which could cause them to cut costs through property disinvestment or move quickly to evict tenants who fall behind on rent (Desmond 2016). More research on the financing of rental properties, particularly in high-poverty areas, is needed.

By Ann Carpenter, senior CED adviser, and Matthew Desmond, professor of sociology at Princeton University


Desmond, Matthew. 2016. Evicted: Poverty and Profit in the American City. New York: Crown.

Desmond, Matthew, and Nathan Wilmers. 2017. "Is Housing the Poor Lucrative? The Profit Margins of Urban Landlords." Working Paper: Princeton University.