Over the past two weeks, treasury note yields have risen rather vigorously from a higher low. As a side effect of this, mortgage rates are now increasing in spite of Bernanke's truly massive purchases in that sector. Note that the Fed is not only adding $40 billion in mortgage related debt to its balance sheet per month, it is also replacing debt that matures. As a result, the purchases have exploded. In fact, the year-to-date purchases of agency debt amount to $345 billion so far, which is a monthly run rate of $69 billion. Together with its treasury debt purchases, the Fed has therefore bought securities amounting to $113 billion per month this year (of course the replacement purchases do not increase the money supply, unless the old purchases were from banks and the new ones are from non-banks).
In any case, there has been a strong increase in yields, in spite of the fact that market-based 'inflation expectations' continue to decline in both the US and Europe.
Inflation expectations in the US (orange line) and euro-land (measured by comparing nominal yields to yields on 'inflation protected' bonds): apart from a small recent uptick, they continue to fall in both regions, and yet, nominal yields are now heading higher anyway – click to enlarge.
We are not sure at this stage what this development portends, but it needs to be watched closely, due to the phenomenon of convexity selling. An in-depth explanation can be found here. In brief, what this means is that holders of mortgage backed bonds are forced to buy treasuries to protect themselves against prepayments when yields are falling, and are forced to short treasuries when yields are rising, to maintain duration and hedge against a fall in the value of their portfolios. Perversely this means they are buying high and selling low, but the point is that convexity related trading can exacerbate moves in treasury yields noticeably.
In short, there could be moves in treasury yields that are related to market structure and have little to do with changes in inflation expectations. According to what we have heard and read, the danger of increased convexity selling will become acute at a 10 year yield of about 2.25% (we can not confirm if this is actually the case, but it does sound plausible).
One point that is perhaps worth making is that a great many inter-market correlations have either broken down or become subject to large lags over the past year or so. It would therefore probably be dangerous to assume that rising bond yields will automatically coincide with rising stock prices – that may no longer be the case.
Charts by: Bloomberg, StockCharts