We maintain an allowance for loan losses for HFI loans held by Fannie Mae and by consolidated Fannie Mae MBS trusts, excluding loans for which we have elected the fair value option. When calculating our allowance for loan losses, we consider the unpaid principal balance, net of unamortized premiums and discounts, and other cost basis adjustments of HFI loans at the balance sheet date. We record write-offs as a reduction to our allowance for loan losses at the point of foreclosure, completion of a short sale, upon the redesignation of nonperforming loans from HFI to HFS or when a loan is determined to be uncollectible. See “Note 1, Summary of Significant Accounting Policies” for additional information and accounting policy changes resulting from our adoption of the CECL standard.


	For the Three Months Ended June 30, 2020		For the Six Months Ended June 30, 2020
	(Dollars in millions)
Single-family allowance for loan losses:
Beginning balance	$	(12,070)	$	(8,759)
Transition impact of the adoption of the CECL standard	—		(1,229)
Benefit (provision) for loan losses	389		(1,788)
Write-offs	46		116
Recoveries	(3)		(6)
Other	40		68
Ending Balance	$	(11,598)	$	(11,598)
Multifamily allowance for loan losses:
Beginning balance	$	(1,139)	$	(257)
Transition impact of the adoption of the CECL standard	—		(493)
Provision for loan losses	(228)		(635)
Write-offs	(1)		18
Recoveries	—		(1)

Ending Balance	$	(1,368)	$	(1,368)
Total allowance for loan losses:
Beginning balance	$	(13,209)	$	(9,016)
Transition impact of the adoption of the CECL standard	—		(1,722)
Benefit (provision) for loan losses	161		(2,423)
Write-offs	45		134
Recoveries	(3)		(7)
Other	40		68
Ending Balance	$	(12,966)	$	(12,966)

Our benefit (provision) for loan losses can vary substantially from period to period based on a number of factors, such as changes in actual and forecasted home prices or property valuations, fluctuations in actual and forecasted interest rates, borrower payment behavior, events such as natural disasters or pandemics, the types and volume of our loss mitigation activities, including forbearance and loan modifications, the volume of foreclosures completed, and the redesignation of loans from HFI to HFS.

Estimating the impact of the COVID-19 pandemic on our expected credit loss reserves required significant management judgment as we continue to observe uncertainty in our economic forecast caused by elevated levels of unemployment, the impact of fiscal stimulus, and variability surrounding the continued spread of COVID-19; thereby impacting projected borrower behavior. We applied management judgment in estimating the number of single-family borrowers who will receive forbearance and any resulting repayment plans, deferrals, or loan modifications that will be provided once the forbearance period ends. Under our CECL methodology, depending on the type of loan modification granted, loss severity estimates vary. Our provision can also be impacted by updates to the models, assumptions, and data used in determining our allowance for loan losses.

•Expected loan losses as a result of the COVID-19 pandemic, which includes adjustments to modeled results. In the first quarter of 2020, the rapidly changing conditions as a result of the unprecedented COVID-19 pandemic caused us to believe our model used to estimate single-family credit losses does not capture the entirety of losses we expect to incur relating to COVID-19, which includes our expectations surrounding loan forbearance and borrower behavior once the forbearance period ends. As a result, management used its judgment to significantly increase the loss projections developed by our credit loss model in the first quarter of 2020, contributing to the provision for loan losses for the six months ended June 30, 2020.

During the second quarter of 2020 our credit loss model consumed data from the initial months of the pandemic, including loan delinquencies, updated profile data for loans in forbearance, and an updated home price forecast. As more of this data was consumed by our credit loss model, we reduced our non-modeled adjustment. However, management continued to apply its judgment and supplement model results as of June 30, 2020, taking into account uncertainty regarding the type and extent of loss mitigation that may be needed when loans complete their forbearance period and the continued high degree of uncertainty regarding the future course of the pandemic and its effect on the economy, including the continued availability of fiscal stimulus to support borrowers.

For the second quarter of 2020, our estimate of expected losses due to the pandemic, which includes both modeled and non-modeled adjustments, remained relatively flat compared with the first quarter. A reduction in overall expected forbearance volumes was offset by a weaker credit profile for loans entering forbearance. Furthermore, the positive impact of higher-than-initially-expected borrower prepayment activity was offset by heightened uncertainty as noted above.

•Changes in our expectations for home price growth. In the first quarter, we significantly reduced our expectations for home price growth to near-zero for 2020, which contributed to our provision for loan losses for the period. However, the negative impact from the first quarter was partially reduced in the second quarter as we revised our home price forecast to reflect an increase in home price appreciation on a national basis for 2020 based on strong home sales data for the period. This improvement in the 2020 home price forecast was partially offset by our updated long-term projected home price growth estimate, which was lowered as a result of a longer projected economic recovery period. Higher home prices decrease the likelihood that loans will default and decrease the amount of credit loss on loans that do default, which impacts our estimate of losses and ultimately decreases our loss reserves and provision for loan losses. Conversely, lower home prices increase the likelihood that loans will default, thereby increasing loss reserves and provision for loan losses.

•Credit benefit from lower actual and projected mortgage interest rates. As mortgage interest rates decline, we expect an increase in future prepayments on single-family loans, including modified loans. Higher expected prepayments shorten the expected lives of modified loans, which decreases the expected impairment relating to term and interest-rate concessions provided on these loans and results in a benefit for loan losses.

The primary factor that contributed to a decrease in single-family write-offs for the three and six months ended June 30, 2020 compared with the three and six months ended June 30, 2019 was a reduction in the number of reperforming loans redesignated from HFI to HFS.

•Actual and projected economic data and expected loan losses as a result of the COVID-19 pandemic. Similar to the single-family provision for loan losses discussed above, we believe our model used to estimate multifamily loan losses as of June 30, 2020 does not capture the entirety of losses we expect to incur relating to COVID-19. As a result, management used its judgment to increase the loss projections developed by our credit loss model to reflect our current expectations relating to the impact of the pandemic. Accordingly, our multifamily provision for loan losses was primarily driven by elevated unemployment rates compared with pre-COVID-19 levels. We expect higher unemployment rates will reduce the operating income of multifamily properties in the near term, resulting in an increase in the number of loans in forbearance. Additionally, property values are expected to decrease, increasing the probability of loan defaults.

In the second quarter, we increased our expected loan losses as a result of significant economic uncertainty and worsened forecasted capitalization rates. We also increased our expected loan losses on senior housing loans to reflect that these properties have been disproportionately impacted by the pandemic, which has resulted in increased operating expenses and has limited these borrowers’ ability to attract new tenants. These increases in expected loan losses were partially offset by a reduction in our overall estimated forbearance volumes and lower actual and projected interest rates. In developing these adjustments, management considered the current credit risk profile of our multifamily loan book of business, as well as relevant historical credit loss experience during rare or stressful economic environments.

The following tables display prior period changes in single-family and multifamily allowance for loan losses and the prior period amortized cost in our HFI loans by impairment or allowance methodology and portfolio segment prior to the adoption of the CECL standard. For a description of our previous allowance and impairment methodology refer to “Note 1, Summary of Significant Accounting Policies” in our 2019 Form 10-K.


	For the Three Months Ended June 30, 2019		For the Six Months Ended June 30, 2019
	(Dollars in millions)
Single-family allowance for loan losses:
Beginning balance	$	(12,985)	$	(13,969)
Benefit (provision) for loan losses	1,239		1,886
Write-offs	558		939
Recoveries	(15)		(60)
Other	(7)		(6)
Ending Balance	$	(11,210)	$	(11,210)
Multifamily allowance for loan losses:
Beginning balance	$	(247)	$	(234)
Benefit (provision) for loan losses	(27)		(40)
Write-offs	4		4
Recoveries	(2)		(2)
Ending Balance	$	(272)	$	(272)
Total allowance for loan losses:
Beginning balance	$	(13,232)	$	(14,203)
Benefit (provision) for loan losses	1,212		1,846
Write-offs	562		943
Recoveries	(17)		(62)
Other	(7)		(6)
Ending Balance	$	(11,482)	$	(11,482)


	As of December 31, 2019
	Single-Family		Multifamily					Total
	(Dollars in millions)
Allowance for loan losses by segment:
Individually impaired loans	$	(8,175)		$	(45)	$	(8,220)
Collectively reserved loans	(584)			(212)		(796)
Total allowance for loan losses	$	(8,759)		$	(257)	$	(9,016)
Amortized cost in loans by segment:
Individually impaired loans	$	97,196		$	680	$	97,876
Collectively reserved loans	2,909,115			329,938		3,239,053
Total amortized cost in loans	$	3,006,311		$	330,618	$	3,336,929