MFA FINANCIAL, INC. quarterly report, Q1 FY2013

MFA Financial, Inc. — Q1 2013 Form 10-Q

Reporting period: Three months ended March 31, 2013; balance-sheet comparisons are to December 31, 2012. MFA is a REIT investing primarily in residential Agency and Non-Agency mortgage-backed securities (MBS), financed principally with repurchase agreements. Figures below are in U.S. dollars; financial-statement amounts are generally in millions unless otherwise noted.

Financial performance and position

MetricQ1 2013 / March 31, 2013Comparable period / prior year-end
Interest income$122.7 million$123.5 million in Q1 2012
Interest expense$40.2 million$40.1 million in Q1 2012
Net interest income$82.6 million$83.4 million in Q1 2012
Net income$77.4 million; $0.21 per common share, basic and diluted$84.7 million; $0.23 per share in Q1 2012
Net interest spread / margin2.32% / 2.69%2.54% / 2.96% in Q1 2012
Total assets$13.653 billion$13.518 billion at December 31, 2012
MBS portfolio, fair value$12.542 billion: $7.154 billion Agency and $5.388 billion Non-Agency$12.608 billion at December 31, 2012
Cash and cash equivalents$601.6 million$401.3 million at December 31, 2012
Repurchase agreements$8.903 billion$8.752 billion at December 31, 2012
Securitized debt / Senior Notes$542.0 million / $100.0 million$646.8 million / $100.0 million at December 31, 2012
Stockholders’ equity / book value per common share$3.266 billion / $8.84$3.311 billion / $8.99 at December 31, 2012
Cash flow$71.2 million operating; $349.0 million investing; $219.9 million used in financing; cash increased $200.3 millionOperating cash flow was $79.9 million in Q1 2012

Comprehensive income available to common stock and participating securities was $205.7 million, including $130.4 million of other comprehensive income, primarily unrealized MBS and hedge gains. These valuation changes are distinct from net income.

Changes versus comparable periods

  • Net income available to common stock and participating securities declined to $75.3 million from $82.7 million in Q1 2012. Management attributed the decline mainly to lower gains and net interest income from Linked Transactions and a higher share count.
  • Net interest income fell $0.8 million. Lower Agency MBS yields contributed: Agency MBS yield was 2.42% versus 3.15% in Q1 2012. Non-Agency MBS interest income rose, but its yield edged down to 6.80% from 6.95%.
  • Other income declined to $3.2 million from $10.7 million, chiefly because Linked Transaction gains and income were lower. Operating and other expenses were broadly flat at $8.5 million.
  • Debt-to-equity leverage increased to 3.1x from 3.0x at year-end. Book value per share fell to $8.84; management said the $0.50 special dividend was the primary cause and that book value would have increased excluding it.
  • No credit-related OTTI losses were recognized in Q1 2013, compared with $0.9 million in Q1 2012. Non-Agency MBS net unrealized gains increased to $769.2 million from $623.9 million at year-end.

Outlook, commentary, risks and unusual items

  • Management intended to continue selectively acquiring Agency and Non-Agency MBS during the remainder of 2013, expecting Agency MBS to remain the majority of assets. It described Non-Agency MBS as offering attractive loss-adjusted returns, citing a discounted average amortized cost of 73% of face value. No specific earnings or return guidance was provided.
  • Management reported a $1.055 billion liquidity “Cushion” at quarter-end: $601.6 million cash, $410.7 million unpledged Agency MBS and $43.0 million excess Agency collateral. It also reported $182.1 million of unpledged and $342.1 million of excess-pledged Non-Agency MBS.
  • Key exposures include short-term, renewable repo funding against longer-duration assets; collateral margin calls and changing haircuts; rising rates and asset/liability repricing mismatch; prepayment and reinvestment risk; Non-Agency mortgage credit losses; and concentration in Southern California, Northern California and Florida. Management warned that falling collateral values or tighter funding markets could materially impair liquidity and leverage capacity.
  • At quarter-end, $2.514 billion notional of interest-rate swaps hedged funding exposure; swaps had a $50.5 million net unrealized liability. The filing’s rate-shock analysis estimated that an immediate 100-basis-point rise would reduce projected 12-month net interest income by 7.43% and portfolio value by 1.08%; these are model estimates, not guidance.
  • A tax-calculation review identified previously undistributed REIT taxable income for 2010–2011 and a potential shortfall for 2012. MFA declared a $0.50-per-share special dividend, paid April 10, 2013, and expected additional 2013 dividends may be needed. The 2012 tax return was not yet final.
  • After quarter-end, MFA issued 8.0 million Series B preferred shares, with estimated net proceeds of $193.3 million, and announced redemption of all Series A preferred shares for approximately $97.0 million on May 16, 2013.
  • Disclosure controls were deemed ineffective as of March 31, 2013 because remediation of a material weakness in controls over MBS taxable-income calculations remained ongoing. Management gave no assurance remediation would succeed. The filing reported no material pending legal proceedings.

Important facts for investors to verify

  • Progress and effectiveness of remediation for the material weakness, and the resulting reliability of taxable-income calculations and future REIT distributions.
  • Final 2012 REIT taxable income, any additional dividends needed for prior-year undistributed income, and the effect of the preferred-stock issuance and Series A redemption on funding costs and capital.
  • Repo renewals, counterparty concentration, collateral haircuts and margin-call capacity, particularly under adverse market conditions.
  • Non-Agency MBS cash-flow assumptions, credit-reserve releases, realized losses and regional mortgage performance; management’s estimates are judgmental and may change.
  • Interest-rate, prepayment and reinvestment outcomes versus the filing’s modeled sensitivities; actual results may differ materially.