If the Fed bought three quarters of the new issuance of Treasury securities over an 8-month period, with a focus on longer maturities, the 10-yr Treasury yield would almost certainly fall, right? And if the Fed bought all the new issuance of MBS over an 8-month period, increasing its share of home mortgages by over 40% in the process, yields on MBS would almost certainly fall, right?
Wrong. The Fed has indeed been a huge buyer of Treasuries and MBS since last September, but Treasury yields and MBS yields have moved significantly higher, not lower.
What we've witnessed over the past 8 months—the duration so far of the Fed's Quantitative Easing Part 3—is almost a laboratory experiment designed to discover which is the more important determinant of longer-term interest rates: the market's willingness to hold the existing stock of bonds, or the actions of a very large purchaser of bonds on the margin (i.e., the stock vs. flow argument).
It's my impression that most market participants have been persuaded by the flow argument: namely, that the Fed's massive QE3 purchases have artificially depressed market interest rates. After all, that's been the Fed's stated intention: to buy lots of bonds in order to depress interest rates and thereby stimulate borrowing and economic activity. This line of reasoning says that the fact that 10-yr Treasury yields averaged an exceptionally low 1.75% over the past year has nothing to do with the market's view of inflation or economic growth; Treasury yields have in fact become meaningless inputs to valuation models and offer no insight into market and economic fundamentals, other than as a distorting influence.
I've argued to the contrary on many occasions over the years. I believe that interest rates are determined by the market's willingness to hold the existing stock of bonds, especially since Fed purchases on the margin represent only a small fraction of the existing stock. I think the Fed can only influence yields to the extent that the market's view of the economy is similar to the Fed's. If both expect the economy to be very weak, yields will be low, and prices will behave as if Fed purchases of bonds to stimulate the economy are in fact achieving their stated objective. But if the market thinks the economy is improving and/or inflation is rising, then no amount of Fed purchases will be able to keep yields from rising. That's the situation today, and it's been unfolding (in fits and starts) almost from the day QE3 began.
Since last September, when the Fed announced it would begin buying $40 billion per month of MBS, the Fed's holdings of MBS have increased by about $330 billion, and that means the Fed has essentially purchased all of the new MBS issuance since last September. As the chart above shows, the Fed now owns over 12% of all home mortgages, up from less than 9% at the end of last September. Yet despite these massive purchases, the yields on MBS have increased by over 100 bps!
When QE3 was first announced in late September, MBS spreads briefly plunged to almost zero as market participants attempted to "front-run" Fed purchases. But since then they have widened to almost 80 bps, not too far from their 30-year historical median of 114 bps. The MBS market was briefly distorted by QE3, but it is now behaving more normally.
Since last September, the Fed has been buying about $45 billion of Treasuries notes and bonds per month, and that adds up to almost 75% of the federal budget deficit over the same period. These purchases have boosted the Fed's holdings of marketable Treasuries from 15.3% of marketable Treasuries to about 16.5%, as the chart above shows, but that is still a relatively small fraction of the outstanding stock of Treasuries. Despite the Fed's massive purchases, 10-yr Treasury yields are up over 50 bps since last September. I would further note that if Fed purchases of Treasuries were a significant factor in driving yields, then the blue and red lines in the above chart should have a strong tendency to move in opposite directions (e.g., large purchases would increase the Fed's share of marketable Treasuries and would tend to depress yields). But for most of the past 10 years these lines have tended to move together; they only moved in opposite directions in the second quarter of 2008 and 2011.
As I explained a few weeks ago, when central bank purchases produce counter-intuitive results, it's a good indication that the economy is perking up. Yesterday I made that same case in reference to the even bigger increase we have seen in real yields on TIPS. Higher real and nominal yields are telling us that the market's outlook for economic growth is improving, while at the same time inflation expectations are moderating. If Fed purchases were artificially pumping up the economy, then we should have seen rising inflation expectations, lower real yields, and higher gold prices.
Of course, many observers will be quick to object that yields are up because the Fed has been actively dropping hints that it may taper its QE3 purchases somewhat earlier than expected, and this has caught the market by surprise. But the fact remains that the Fed is still a large purchaser of Treasuries and MBS, and their prices are still falling. My interpretation is that the market hears the Fed hints of an early QE3 tapering and concludes that it is reasonable for the Fed to end QE3 sooner than expected, because the economy outlook is improving, however marginal that improvement might be. As I've mentioned quite a few times before, avoiding a recession is all that matters for investors when short-term interest rates are virtually zero.
To sum up, I believe the history of QE3 shows us that the Fed cannot manipulate longer-term interest rates. Yields are fundamentally determined by the market's perception of the prospects for inflation and economic growth, not by Fed purchases of Treasuries and MBS.