Implementing Long-term Security Purchases
During slowdowns in economic activity and periods of inflation, the monetary authority’s optimal response is to lower the real rate (to negative values if economic conditions require it). Traditionally, the Federal Reserve achieved this by reducing the target fed funds rate—which implies a reduction in short-term real rates because inflation expectations are practically fixed in the very short run. In general (but with notable exceptions), this reduction has an effect also on yields of longer maturity, which can be thought of as a combination of current and future expected short-term rates, thus stimulating the economy.
When short-term rates are close to zero the traditional tool is no longer feasible. However, with longer-term rates substantially above zero, since November 2008, the Fed has expressed its intention of purchasing long-term securities to affect their rates. This is part of a quantitative easing policy which aims to stimulate economic activity when the usual fed funds target is bounded by zero.
The strategy of intervening in long-term treasury markets is not designed to provide liquidity to those markets (usually they do not need it) but to change the relative supply of securities in order to affect their yields. This, in turn, should have spillover effects on other assets’ long-term rates. However, changing the supply of a security that is traded in a very liquid market does not necessary affect its price, which should be determined only by its “fundamentals.” For long-term U.S. treasuries, those fundamentals are inflation and output growth expectations (during the corresponding maturity period). However, not only are those fundamentals not observable but they are also a function of the policy strategy itself; moreover, the longer the maturity, the more those elements can interact with one another. As a result, it is difficult to make an accurate evaluation of the effects of the chosen policy.
For example, in order to stimulate the economy and fight the risk of deflation, the Fed has recently intervened in the open market to reduce yields on long-term treasuries. If that strategy is successful, inflation and output growth expectations may also be positively affected, which will put upward pressure on the yields.
In November 2008, the Fed decided to intervene in the long-term markets by announcing the buying of up to $500 billion mortgage-backed securities (MBS) and announced the TALF; in March 2009, the Fed specified the size of its intervention ($300 billion of long-term treasuries and some additional $750 billion in MBS during the year); this strategy was confirmed in April FOMC meetings.
Figures 1–3 show the daily rates of treasuries of different maturities combined with Fed interventions. The Fed purchases started during the spring of 2009; as the bars show, they concentrate mainly on medium-term maturities and are smoothed throughout. The biggest interventions are still relatively small compared to the volumes usually traded. For example, the thinner market for treasuries with maturities longer than 11 years has an average daily volume above $22 billion, whereas the biggest Fed intervention was below $3.5 billion, or less than 15 percent of the average trading volume. Hence, interventions per se do not seem to add unwarranted volatility to the market. In fact, we have not found any significant correlation between the size of the intervention and either the treasury yields or the daily returns. However, the announcement dates show an immediate effect on prices. This is when markets incorporate the information of the policy change. On November 25, when the TALF and $500 billion MBS purchases were announced, medium- and long-term treasuries lost 18 basis points and 15 basis points, respectively (notice that the rate increases on Monday, November 24 are related to the release of a joint statement on Citigroup by the Treasury, Federal Reserve, and FDIC). On December 15, when the Fed announced its intention of purchasing long-term treasuries, there was another decrease of 16 basis points for medium-term and 12 basis points for long-term treasuries. Finally, on March 18, when the size of the intervention was specified, there were substantial reductions of 46 basis points and 26 basis points, respectively; even the short-term maturity showed a reduction of 9 basis points.
The figures presented are aggregated at a daily frequency. Figure 4 shows the same data as figure 2 (medium-term purchases) but they are aggregated at a weekly frequency. The interpretation corroborates the one given earlier for the daily frequencies: Weeks with sizeable Fed interventions seem to have no statistically significant impact on prices.
It is worth noting that, except for the initial impact on announcement days, long- and medium-term treasuries not only have not been affected by Fed purchases at daily or weekly frequency but also have trended up—especially in the case of treasury bonds with maturities of 10 years and over; this has been true at least since the end of December 2008. Unlike medium- and long-term securities, the 1-year rate has been trending down, at least since the end of February 2009; rates are now below 0.5 percent. This might suggest that the policy has been successful and that the medium- and long-term rates have been trending up because the risk of a prolonged recession and deflation has faded—whereas, given monetary policy lags, shorter-term expectations have not changed so dramatically. To corroborate this idea, we have found that the correlation at a weekly frequency between 10-year treasury bonds and Standard & Poor’s 500 (which should mainly reflect output growth expectations) is strongly positive and significant at 46 percent—while on the March 18 announcement day they moved in opposite directions (see figure 5). Moreover, most measures of inflation expectations have also been trending up to values more consistent with the Fed’s long-run inflation target. For example, figure 5 shows the 10-year inflation expectation derived from TIPS (Treasury Inflation-Protected Securities). Even if part of the series’ volatility can be attributed to swings in the liquidity premium for the thinner market of TIPS, the trend is clearly upward.
The Fed has also targeted the market for mortgage-backed securities. In contrast with the treasury market, the MBS market has been particularly hard hit during the crisis. In fact, almost two-thirds of the Fed’s purchases of long-term securities have been concentrated in the MBS market. In figure 6, we show the weekly yield on a 30-year Fannie Mae MBS (5 percent coupon) and its spread with the 10-year treasury (the average maturity of a mortgage is 10 years). In a market that is not perfectly liquid, the Fed’s purchases are supposed to have a stronger effect. The impact of those purchases on MBS yields is beyond the scope of this page, but it is reasonable to believe that those purchases should have played a role in reducing the spread at least since January 2009, when they actually started.