Represents the estimated dollar change in net interest income expressed as a percent of net interest income based on asset yields and cost of funds as of such date. It includes the effect of periodic interest costs on our interest rate swaps, but excludes costs associated with our other supplemental hedges, such as swaptions and U.S. Treasury securities. Also excludes costs associated with our TBA position and TBA dollar roll income/loss. Base case scenario assumes interest rates and forecasted CPR of 8% and 7% as of September 30, 2014 and December 31, 2013, respectively. As of September 30, 2014, rate shock scenarios assume a forecasted CPR of 10%, 9%, 7% and 7% for the -100, -50, +50 and +100 basis points scenarios, respectively. As of December 31, 2013, rate shock scenarios assume a forecasted CPR of 10%, 8%, 7% and 6% for such scenarios, respectively. Estimated dollar change in net interest income does not include the one time impact of retroactive "catch-up" premium amortization benefit/cost due to a decrease/increase in the forecasted CPR and does not include dividend income from investments in other REITs. Down rate scenarios assume a floor of 0% for anticipated interest rates.
Includes the effect of derivatives and other securities used for hedging purposes.
Estimated dollar change in investment portfolio value expressed as a percent of the total fair value of our investment portfolio as of such date.
Estimated dollar change in portfolio value expressed as a percent of stockholders' equity, net of the Series A and Series B Preferred Stock liquidation preference, as of such date.
The change in our interest rate sensitivity as of September 30, 2014 compared to December 31, 2013 was a function of a flatter yield curve, partially mitigated by changes in the size and composition of our asset and hedge portfolio.
Because residential borrowers have the option to prepay their mortgage loans at par at any time, we face the risk that we will experience a return of principal on our investments faster than anticipated. Various factors affect the rate at which mortgage prepayments occur, including changes in the level of and directional trends in housing prices, interest rates, general economic conditions, loan age and size, loan-to-value ratio, the location of the property and social and demographic conditions. Additionally, changes to GSE underwriting practices or other governmental programs could also significantly impact prepayment rates or expectations. Also, the pace at which the loans underlying our securities become seriously delinquent or are modified and the timing of GSE repurchases of such loans from our securities can materially impact the rate of prepayments. Generally, prepayments on agency MBS increase during periods of falling mortgage interest rates and decrease during periods of rising mortgage interest rates. However, this may not always be the case.
We may reinvest principal repayments at a yield that is lower or higher than the yield on the repaid investment, thus affecting our net interest income by altering the average yield on our assets. We also amortize or accrete premiums and discounts associated with the purchase of agency MBS into interest income over the projected lives of the securities, including contractual payments and estimated prepayments using the interest method. Our policy for estimating prepayment speeds for calculating the effective yield is to evaluate published prepayment data for similar agency securities, market consensus and current market conditions. If the actual prepayment experienced differs from our estimate of prepayments, we will be required to make an adjustment to the amortization or accretion of premiums and discounts that would have an impact on future income.
When the market spread widens between the yield on our agency securities and benchmark interest rates, our net book value could decline if the value of our agency securities fall by more than the offsetting fair value increases on our hedging instruments tied to the underlying benchmark interest rates. We refer to this as "spread risk" or "basis risk." The spread risk associated with our mortgage assets and the resulting fluctuations in fair value of these securities can occur independent of changes in benchmark interest rates and may relate to other factors impacting the mortgage and fixed income markets, such as actual or anticipated monetary policy actions by the Federal Reserve, market liquidity, or changes in required rates of return on different assets. Consequently, while we use interest rate swaps and other supplemental hedges to attempt to protect against moves in interest rates, such instruments typically will not protect our net book value against spread risk.
The table below quantifies the estimated changes in the fair value of our investment portfolio (including derivatives and other securities used for hedging purposes) and in our net asset value as of September 30, 2014 and December 31, 2013 should spreads between our mortgage assets and benchmark interest rates go up or down by 10 and 25 basis points. These estimated impacts of spread changes are in addition to our sensitivity to interest rate shocks included in the interest rate shock table above. The table below assumes a spread duration of 5.3 years and 5.5 years based on interest rates and MBS prices as of September 30, 2014 and December 31, 2013, respectively. However, our portfolio's sensitivity of mortgage spread changes will vary with changes in interest rates and will generally increase as interest rates rise and prepayments slow. Additionally, we have limited transparency into the underlying investment and hedge portfolios of the other agency mortgage REITs in which we invest. Therefore, actual results could differ materially from our estimates.