Federal Housing Finance Agency Print

Research

​Research

 

Our economists conduct research on a range of topics in housing finance, including analyzing data and uncovering emerging trends.  In addition to presenting their research to policy makers, they share their research at academic conferences and publish in journals and other scholarly outlets.  Our work enables those interested in housing finance to make decisions based on the best information available.

In particular, our researchers focused on housing trends in house prices, housing market conditions, and mortgage lending activity.  In addition, we analyze the risk and capital adequacy of the housing government-sponsored enterprises and publish papers aimed at improving public understanding of the mortgage finance system.

See FHFA's Copyright and Permissions Notice.

 Papers

 

 

The Size of the Affordable Mortgage Market: 2022-2024 Enterprise Single-Family Housing Goals - December 202136648<p><span style="font-size&#58;13px;">In establishing benchmarks for the single-family home purchase and refinance goals for Fannie Mae and Freddie Mac (the Enterprises), the Federal Housing Finance Agency (FHFA) is required to measure the size of the affordable mortgage market.&#160; This FHFA technical report documents the statistical forecast models that the modeling team has developed as part of the process for establishing the affordable housing goal benchmark levels for the Enterprises for 2022 through 2024.​</span><br></p>12/22/2021 3:30:33 PMHome / Policy, Programs & Research / Research / The Size of the Affordable Mortgage Market: 2022-2024 Enterprise Single-Family Housing Goals - December 2021 1489https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
Working Paper 21-03: The Riskiness of Outstanding Mortgages in the United States, 1999 – 201936688<p>​Abstract&#58; This paper introduces summary measures of credit risk for the stock of all outstanding mortgages in the United States for each quarter between 1999 and 2019. Mortgage terminations play a fundamental role in offsetting risk introduced by the flow of new originations because of refinance activity and the often dual nature of home buyers as concurrent sellers. To illustrate these concepts in a policy setting, I show the Home Affordable Refinance Program increased origination risk metrics but reduced overall risk due to the associated terminations of even riskier loans. Generally, book-level risk tends to lag behind originations&#58; while origination risk peaked in 2006, the risk of outstanding mortgages peaked in 2007, and while origination risk bottomed out in 2011 and has been rising since, book-level risk continued its downward trend in 2019. Other results highlight previously rarely-examined market segments, including credit unions, the Federal Home Loan Bank system, and loans guaranteed by the Farm Service Agency/Rural Housing Service.<br></p>11/1/2021 3:00:17 PMHome / Policy, Programs & Research / Research / Working Paper 21-03: The Riskiness of Outstanding Mortgages in the United States, 1999 – 2019 To illustrate these concepts in a 2245https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
Working Paper 21-02: Borrower Expectations and Mortgage Performance: Evidence from the COVID-19 Pandemic35980<p>​We assess issues related to borrower beliefs and mortgage performance using new individual panel data that simultaneously cover borrower expectations, forbearance status during the COVID-19 pandemic, and a wide array of demographic characteristics. First, we establish the determinants of borrower expectations, with local experiences and those of social networks playing important roles. We then show that households who, at origination, were optimistic about future house price appreciation or pessimistic about the possibility of future unemployment were more likely to enter forbearance in 2020. However, by early&#160;2021, appreciation-optimistic borrowers who were in forbearance were likely to have cured or prepaid their loan, while those who expected unemployment were likely to still be in forbearance.&#160; We offer three channels by which expectations affect forbearance behavior&#58; choices of initial loan terms, associations with actual future events, and factors related to belief formation that are also plausibly associated with forbearance. Our findings highlight the crucial role borrower expectations play in both leverage choices and mortgage performance.<br></p>10/29/2021 6:00:12 PMHome / Policy, Programs & Research / Research / Working Paper 21-02: Borrower Expectations and Mortgage Performance: Evidence from the COVID-19 Pandemic 2096https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
The Size of the Affordable Mortgage Market: 2022-2024 Enterprise Single Family Housing Goals36164<p>In establishing benchmarks for the single-family home purchase housing goals for Fannie Mae and Freddie Mac (the Enterprises), the Federal Housing Finance Agency (FHFA) is required to measure the size of the mortgage market. This FHFA technical report documents the statistical forecast models that the modeling team has developed as part of the process for establishing the affordable housing goal benchmark levels for Fannie Mae and Freddie Mac for 2022 through 2024.<br><br></p>8/18/2021 3:08:43 PMHome / Policy, Programs & Research / Research / The Size of the Affordable Mortgage Market: 2022-2024 Enterprise Single Family Housing Goals Ken Lam, Omena Ubogu, Jay Schultz 1770https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
Working Paper 21-01: A Flexible Method of House Price Index Construction using Repeat-Sales Aggregates33182<h4 style="font-size&#58;13px;"><span class="ms-rteThemeForeColor-2-0" style="line-height&#58;22px;display&#58;none;"></span></h4><h4 style="font-size&#58;13px;font-style&#58;normal;"><span class="ms-rteThemeForeColor-2-0" style="font-style&#58;normal;line-height&#58;22px;"><span style="font-style&#58;normal;line-height&#58;22px;">*Revised June 2022</span></span></h4><span style="color&#58;#444444;font-style&#58;normal;">&#160;​</span><span style="color&#58;#444444;font-style&#58;normal;"></span><h2>​Abstract&#58;<br></h2><p>The major issue which we address in this paper is the one-size-fits-all nature of the typical city-level house price index. In this vein, we make two contributions. First, we develop a new algorithm to ensure feasible estimation of geographically granular repeat-sales house price indices in cases of low transactions counts. This facilitates the estimation of a balanced panel of 63,122 Census tract-level repeat-sales house price indices (2010 definitions) at an annual frequency between 1989 and 2021, which we release alongside this paper. Second, we use these indices to estimate city-level house price indices that are robust to heterogeneous submarket appreciation and non-random sampling, two issues that confound classic approaches. Numerical simulations show this algorithm uncovers population indices even when house prices, quantities, and transaction sampling vary across locations and over time. This approach can be used in a flexible manner to calculate canonical price indices such as Lowe and Laspeyres, and more tailored summary indices on a variety of topics, including collateral valuation, climate risk assessment, or tracking changes to minority housing wealth over time.​<br></p><p><span style="font-style&#58;normal;"><span style="font-style&#58;normal;">T</span><span style="font-style&#58;normal;">he supertract-based census tract HPI</span><span style="font-style&#58;normal;">s</span><span style="font-style&#58;normal;">&#160;</span><span style="font-style&#58;normal;">constructed in this staff workin</span><span style="font-style&#58;normal;">g paper are available&#160;below.&#160;</span></span><span style="font-style&#58;normal;">Our&#160;</span><a href="/PolicyProgramsResearch/Research/PaperDocuments/wp2101-FAQs.pdf" target="_blank" style="font-style&#58;normal;font-size&#58;14px;font-family&#58;&quot;source sans pro&quot;, sans-serif;">FA​​Qs</a><span style="font-style&#58;normal;">&#160;address common questions about the indices</span><span style="font-style&#58;normal;">.</span><span style="font-style&#58;normal;">&#160;P</span><span style="font-style&#58;normal;">lease cite this working paper when using the supertract&#160;HPIs. The data files can be downloaded in a&#160;compressed format&#160;saved in&#160;t​he following file types&#58;</span><br></p><ul><li><span style="font-style&#58;normal;"><span style="font-style&#58;normal;"></span><span style="font-style&#58;normal;"><a href="/PolicyProgramsResearch/Research/PaperDocuments/wp2101-data-tract-csv.zip" target="_blank">CSV</a> [29MB compressed, 229MB extracted]</span></span></li><li><span style="font-style&#58;normal;"><span style="font-style&#58;normal;"><a href="/PolicyProgramsResearch/Research/PaperDocuments/wp2101-data-tract-dta.zip" target="_blank">DTA</a> (</span><span style="font-style&#58;normal;">Stata)</span>​ [62MB compressed, 254MB extracted]</span></li></ul>6/23/2022 2:55:15 AMHome / Policy, Programs & Research / Research / Working Paper 21-01: A Flexible Method of House Price Index Construction using Repeat-Sales Aggregates In this vein, we make two 2844https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
Working Paper 20-02: Nowcasting Unemployment Insurance Claims in the Time of COVID-1928208<font color="#000000" face="Times New Roman" size="3"> </font><h4>Abstract&#58; </h4><p>Near term forecasts, also called nowcasts, are most challenging but also most important when the economy experiences an abrupt change. In this paper, we explore the performance of models with different information sets and data structures in order to best nowcast US initial unemployment claims in spring of 2020 in the midst of the COVID-19 pandemic. We show that the best model, particularly near the structural break in claims, is a state-level panel model that includes dummy variables to capture the variation in timing of state-of-emergency declarations. Autoregressive models perform poorly at first but catch up relatively quickly. Models including Google Trends are outperformed by alternative models in nearly all periods. Our results suggest that in times of structural change there is a bias-variance tradeoff. Early on, simple approaches to exploit relevant information in the cross sectional dimension improve forecasts, but in later periods the efficiency of autoregressive models dominates.</p><font color="#000000" face="Times New Roman" size="3"> </font>8/6/2020 8:12:46 PMHome / Policy, Programs & Research / Research / Working Paper 20-02: Nowcasting Unemployment Insurance Claims in the Time of COVID-19 Near term forecasts, also called nowcasts 2580https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
FHFA MORTGAGE ANALYTICS PLATFORM (Version 2.0)27705<p><strong>Release notes&#160;&#160;</strong></p><p>This version of the white paper incorporates the following two major updates to the FHFA Mortgage Analytics Platform&#58;</p><ol><li>Re-estimated performing loan equations for 30-year fixed rate, 15-year fixed rate, and 5/1 adjustable-rate mortgages.&#160; The re-estimated equations incorporated new explanatory variables important to predicting single-family mortgage delinquency.</li><li>Replaced the constant mortgage insurance (MI) haircuts with rating-based MI haircuts. </li></ol>5/21/2020 5:54:30 PMHome / Policy, Programs & Research / Research / FHFA MORTGAGE ANALYTICS PLATFORM (Version 2.0) White 3502https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
Working Paper 20-01: Land Valuation using Public Records and Kriging: Implications for Land versus Property Taxation in Cities27657<h4>Abstract&#58;</h4><font color="#000000" face="Times New Roman" size="3"> </font><p>We construct land values for each parcel in Maricopa County (Phoenix), Arizona, from 2000 through 2018 using a novel administrative dataset containing the universe of land sales and parcel records in the county. We then compare residential land values constructed using two classes of source data, vacant land sales and land under existing structures. Between 2012 and 2018, estimated land values for developed parcels are, on average, 14% higher when estimated using vacant land due to plattage effects and other unobserved factors. Growth rates are similar, facilitating the use of vacant land price indices to trace valuations over time from an accurate base year valuation. Dynamics between prices of Maricopa County land and housing suggest hypothetical land value tax revenues are more pro-cyclical than property tax revenues, with Betas with respect to national house prices of 3.3 and 2.3, respectively. By 2018, houses had recovered 96% of pre-crisis (2007) values, but land had only recovered 66%. These findings demonstrate a source of risk of dependence on public revenues from land value taxes versus a base-period revenue-neutral property tax.</p><font color="#000000" face="Times New Roman" size="3"> </font><font color="#000000" face="Times New Roman" size="3"> </font><p>Note&#58; An earlier version of this paper was posted in March 2020. The May 2021 update includes a refresh of external data sources used in the paper and refinements to some of the underlying methods.</p><font color="#000000" face="Times New Roman" size="3"> </font> <font color="#000000" face="Times New Roman" size="3"></font> 6/23/2022 8:07:35 PMHome / Policy, Programs & Research / Research / Working Paper 20-01: Land Valuation using Public Records and Kriging: Implications for Land versus Property Taxation in Cities 2542https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
Working Paper 19-02: A Quarter Century of Mortgage Risk26282<h4>Abstract&#58;</h4> <font color="#000000" face="Times New Roman" size="3"></font> <p>This paper provides a comprehensive account of the evolution of default risk for newly originated home mortgages over the past quarter century. We bring together several data sources to produce this history, including loan-level data for the entire Enterprise (Fannie Mae and Freddie Mac) book. We use these data to track a large number of loan characteristics and a summary measure of risk, the stressed default rate. Among the many results in the paper, we show that mortgage risk had already risen in the 1990s, planting seeds of the financial crisis well before the actual event. Our results also cast doubt on explanations of the crisis that focus on borrowers with low credit scores.</p><p>Note&#58; These indices are works in progress and all data, tables, figures, and other results in this working paper are subject to change. Earlier versions of this paper were posted in January, March, and October 2019, and May 2021. The January and March 2019 versions were posted under the title “Mortgage Risk Since 1990.&quot; The May 2021 version adds refinance mortgages to the prior analysis which focused exclusively on purchase-money mortgages.&#160; Other improvements to the imputation and validation methods are also included.</p>5/25/2022 5:13:34 PMHome / Policy, Programs & Research / Research / Working Paper 19-02: A Quarter Century of Mortgage Risk William Larson (FHFA), Morris Davis (Rutgers), Stephen Oliner (American 14344https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx
Working Paper 19-01: The Price of Residential Land for Counties, ZIP codes, and Census Tracts in the United States26065<h4>Abstract&#58;</h4><p>Data from millions of appraisals in 2012-2019 are used to estimate residential land prices, the share of house value attributable to land, and related statistics down to the census-tract level for areas that include the vast majority of U.S. population and single-family housing. The results confirm predictions about land prices from canonical urban models. Over 2012-2019, we show that land prices rose faster than house prices in large metro areas, boosting the land share of house value, while the land share fell in smaller metros. The data are available for download at <a href="/PolicyProgramsResearch/Research/Pages/wp1901.aspx">https&#58;//www.fhfa.gov/papers/wp1901.aspx</a>.</p><p>Note&#58; These indices are works in progress and all data, tables, figures, and other results in this working paper are subject to change.</p><p>&#160;</p>11/9/2020 3:27:28 PMWilliam Larson (FHFA), Jessica Shui (FHFA), Morris Davis (Rutgers), Stephen Oliner (AEI Data from millions of appraisals in 2012-2019 are used to estimate residential land prices 32632https://www.fhfa.gov/PolicyProgramsResearch/Research/Pages/Forms/AllItems.aspxhtmlFalseaspx

© 2022 Federal Housing Finance Agency