[Update 1 clarifies recent weekly MBS purchases by the Fed.] The Federal Reserve continues this week to buy mortgage-related assets from government-sponsored enterprises (GSEs) – at a slower pace than recent weeks. The Fed’s forthcoming exit from the purchase program is stirring investor fears that mortgage spreads may widen out again, but one recent research paper dares to suggest the Fed’s purchases are not responsible for the recent tightening of spreads to near-historic lows. The Federal Reserve Bank of New York bought $12bn of mortgage-backed securities (MBS) from mortgage giants Freddie Mac (FRE), Fannie Mae (FNM) and Ginnie Mae in the week ending January 8. Gross purchases totaled $14.5bn — $1.8bn of Freddie MBS and $12.7bn of Fannie MBS — before $2.5bn of MBS sales during the same time frame, according to details released Thursday by the NY Fed. It’s another week of slowed purchases as the Fed winds down its purchase program, which is on track to buy up $1.25trn of agency MBS by the end of Q110. It’s a slight increase from the previous reporting week, but a marked decline from recent weeks, like the one before Christmas when the Fed bought $15bn of MBS, net of $2.43bn in sales (for a weekly gross of $17.43bn MBS). The Fed has considered extending and expanding asset-purchase programs, including the MBS program, if its exit this quarter is not replaced with private investor demand, causing MBS spreads to treasuries to blow out again. Global financial services firm Credit Suisse maintains spreads will keep tight or perhaps tighten further in the near-term, but investors remain nervous. HousingWire‘s Linda Lowell recently pointed out that net agency MBS issuance was around $384bn as of mid-December, while the Fed had bought about $1.1trn — nearly three times the net new pass-through supply coming into the market in 2009. This significant demand, Lowell writes, held mortgage spreads near historic tights for months. But a recent working paper from the National Bureau of Economic Research points out that demand generated by the Fed’s purchases is not enough to account for tightened spreads, due to simultaneous changes in prepayment and default risks. The paper notes the chief difficulty in assigning the Fed credit for narrowed spreads: A decline in either the prepayment or default risk could trigger the same tightening effect that coincided with the Fed’s purchases. “When we control for these risks, we find evidence of statistically insignificant or small effects of the program,” says the paper’s authors and Stanford University economists, Johannes Stroebel and John Taylor. Many investors and traders already take account of the prepayment risk with the use of the option-adjusted spread, so the authors essentially had to add the risk back in to see how spreads would behave. With default risk, the most effective measurement – the credit default swap series on GSE debt – has not been reported since the GSEs went into conservatorship, so the report’s authors used an alternative measure of default risk to model the behavior of spreads. In both cases, the result was telling, as Stroebel and Taylor note: “Movements in prepayment risk and default risk explain virtually all of the movements in mortgage spreads.” Write to Diana Golobay. The author holds no relevant investments.