There is an ongoing policy debate about when and at what speed the Federal Reserve Bank should reduce its portfolio of mortgage-backed securities (MBS). The MBS purchase programme was set up with an initial limit of $500 billion, was later expanded to $1.25 trillion, and is about to expire at the end of the first quarter of 2010. At its meeting in mid-December 2009, the Federal Open Market Committee announced that it was continuing its MBS purchases at a “gradually slowing pace.” However, the minutes from that meeting reveal a disagreement among the board members with respect to the need for continuing the programme:
“A few members […] observed that it might become desirable at some point in the future to provide more policy stimulus by expanding the planned scale of the Committee's large-scale asset purchases and continuing them beyond the first quarter, especially if the outlook for economic growth were to weaken or if mortgage market functioning were to deteriorate. One member thought that the improvement in financial market conditions and the economic outlook suggested that the quantity of planned asset purchases could be scaled back, and that it might become appropriate to begin reducing the Federal Reserve's holdings of longer-term assets if the recovery gains strength over time.”
An important input to this discussion involves an empirical assessment of the impact of the MBS purchase programme. The programme was introduced with the explicit aim of reducing mortgage interest rates (Board of Governors, 2008), so an empirical analysis of whether this has been achieved is required. Another longer-term question is whether such price-keeping operations (Fisher, 2009) should be a regular part of monetary policy in the future. Brian Sack, who runs the trading desk at the New York Fed, argued in a recent speech that they should be (Sack, 2009).
Conventional wisdom suggests that the MBS purchase programme has successfully contributed to the reduction in mortgage interest rates throughout 2009. Figure 1 shows that at the same time as the Fed (and the Treasury, which ran a smaller purchase programme) increased their holdings of mortgage-backed securities, the primary and secondary mortgage market spreads declined.1 Ben Bernanke consequently claimed as early as January 2009 that “[…] mortgage rates dropped significantly on the announcement of this programme and have fallen further since it went into operation.“ More recently, he concluded that: “The Federal Reserve’s agency debt and mortgage-backed securities purchase programmes stabilised the functioning of private secondary mortgage markets during the height of the financial turmoil. These actions also provided significant benefits to primary mortgage markets” (US Senate, 2009). A similar optimism about the success of the programme was expressed by Sack (2009).
Figure 1. Mortgage spreads and stock of MBS purchases
It is likely, however, that other factors – beyond the Fed’s MBS purchases – have also affected mortgage interest spreads. In particular, movements in prepayment risk and default risk should impact the yields on MBS, and consequently the primary mortgage rate. A question that needs to be considered is whether movements in prepayment risk and default risk that contributed to the increase in mortgage spreads in 2007 and 2008 can also explain their decline in 2009.
The first factor we consider that impacts the yields of MBS is the prepayment risk. Mortgage-backed securities are structured products that are collateralised by residential mortgages. Most of these mortgages entail a prepayment option, allowing the borrower to prepay the mortgage at any time prior to the maturity of the loan. Yields of mortgage-backed securities need to be adjusted upwards to compensate the holder of the MBS for this prepayment risk, which can vary over time and is driven by factors such as the implied volatility of interest rates. We control for the prepayment risk in the underlying mortgages by considering the effect of the MBS purchase programme on the swap option-adjusted spread, which is regularly used by MBS traders and investors. The option-adjusted spread is the yield-spread of the MBS over a term-structure of alternative interest rates after controlling for the value of the prepayment option. More details on the calculation of the option-adjusted spread can be found in Windas (1996).
The other factor that affects MBS pricing is the default risk. Since we are considering yields on MBS insured by Fannie Mae and Freddie Mac, the default risk is related to the default risk of the underlying mortgages as well as to the potential of the insuring Government-Sponsored Enterprise being unable to meet its guarantee obligations. We control for default risk in our regressions by using spreads on senior and subordinated agency debt. In an alternative specification, we also instrument for these spreads using housing market and credit market indicators. This is necessary to ensure that yields on Government-Sponsored Enterprise debt were not primarily driven by the simultaneous (albeit smaller) intervention of the Fed in the agency debt market. Since MBS guarantees rank pari passu to senior bonds, we believe that the spreads of agency senior and subordinated debt to Treasuries capture the default risk of agency-insured MBS well.
New evidence on the impact of the MBS purchase programme
When we control for the impact of prepayment risk and default risk on the option-adjusted spread, we find that that the MBS purchase programme has not had a consistently large or statistically significant effect on mortgage spreads. Our results can be summarised as follows (see Stroebel and Taylor 2009 for more details):
- Using conventional option-adjusted spreads constructed by using a LIBOR swaps term-structure to control for prepayment risk, we find that the MBS programme has had no significant effect on mortgage spreads. Movements in prepayment risk and default risk explain virtually all of the movements in the option-adjusted spreads. In particular, the decline in the option-adjusted spreads that occurred during the period of the MBS programme can be better explained by a general decline in default risk. We use Figure 2 to illustrate this finding. It shows the development of the Swap option-adjusted spreads in blue. The red series shows the predicted Swap option-adjusted spreads using the agency debt spread (a measure of default risk). Movements in default risk explain the movement in Swap option-adjusted spreads very well. The residual between the actual and the predicted Swap option-adjusted spreads series, shown in green at the bottom of the graph, indicates that there is little left for the Fed’s MBS portfolio to explain.
Figure 2. Residual analysis of swap option-adjusted spreads
- If one uses an alternative measure of the option-adjusted spreads based on the Treasury yield curve, a somewhat more significant effect on mortgage spreads—about 30 basis points—can be detected. This is illustrated in Figure 3 by the downward-shift in the residuals in Q3/Q4 2009, around time of the start of the Treasury’s MBS purchase programme and the announcement of the Fed’s programme. However, even with this measure, the volume of purchases appears to have no effect over and above the mere existence of the programme. It is interesting to observe that as the volume of MBS held by the Fed and Treasury grew to $1000 billion, the residual did not become more negative – the expansion of the MBS purchase programme does not appear to have lowered mortgage spreads, beyond the initial downward shift.
Figure 3. Residual Analysis of Treasury-OAS
- We also estimate the impact using indirect methods to control for prepayment risk (by controlling for the implied volatility of interest rates). The results generally confirm the analysis using option-adjusted spreads, but shows somewhat larger effects in the secondary market. The impact of the programme on primary market spreads ranges from the wrong sign and insignificant to around 30 basis points. For secondary market rates the estimated impact is in the 30 to 60 basis point range. This corresponds to a less than full pass through of any impact to primary mortgage spreads, the rate paid by the home-buyer. This suggests that a portion of any reduction in funding costs contributed to the profit margin for lenders.
If our estimates hold up to scrutiny, they raise doubts about price-keeping operations such as the MBS purchase programme and suggest that the Fed could gradually reduce the size of its portfolio without a significant impact on the mortgage market.
1 The primary market mortgage rate series comes from Freddie Mac’s Primary Mortgage Market Survey. The secondary market mortgage is the Fannie Mae MBS 30 Year Current Coupon. The spreads are created by subtracting the yield on 10-year Treasuries from both series. The maturity difference between these series captures the fact that most 30-year mortgages are paid-off or refinanced before their maturity.
Bernanke, Ben S. (2009), “The Crisis and the Policy Response,” Stamp Lecture, London School of Economics, 13 January.
Board of Governors of the Federal Reserve System (2008), Press Release, 25 November.
Federal Open Markets Committee (2009), Minutes of December 15/16 meeting.
Fisher, Peter (2009), “The Market View: Incentives Matter” in Ciorciari, John D. and Taylor, John B, eds. (2009), The Road Ahead for the Fed, Hoover Press
Sack, Brian (2009), “The Fed's Expanded Balance Sheet”, Remarks at the Money Marketeers of NYU, 2 December.
Stroebel, Johannes C. and John B. Taylor (2009), “Estimated Impact of the Fed’s Mortgage-Backed Securities Purchase program”, NBER Working Paper 15626
Windas, Thomas (1996), An introduction to option-adjusted spread analysis, Revised Edition, Bloomberg Press