It all seems so reasonable. US Treasury yields have fallen around 50 bp since the March rate hike. Market-based measures of inflation, like the 10-year breakeven and the five-year/five years forward, have fallen around 35 bp over the same period. That is to say that the decline in market-based measures of inflation expectations can account for nearly three-quarters of the decline in nominal yields.
It begs the question, why these inflation expectations have fallen. Commodity prices in general, and oil prices, in particular, have fallen, and this coincides with a decline in inflation expectations.
While this is clearly part of the story, many narratives are stopping at this point. It is fundamentally incomplete.
Let's begin with the basics. The nominal yield is a function of a real yield and a premium. In many high income countries, that premium is thought to be inflation. The conventional narrative uses market-based measures of inflation expectations. There are some problems with the market-based measures.
One widely recognized problem is that there is a difference in liquidity between conventional bonds and inflation-protected securities. This lends itself to breakevens falling an environment when bonds are rallying, increasing breakevens when bonds are being sold.
There are survey-based measures of inflation as well. These are considerably more stable than market-based measures. There are two surveys that have been cited by officials. The first, and arguably the more important, is the survey the Fed conducts itself, the Survey of Professional Forecasters. In May 2016, the average (2016-2020) was 2.1% and then 2.3% for a longer period (2016-2025). Last month's survey resulted in an average of 2.35% and 2.30% (2017-2021 and 2017-2026 respectively).
The other survey is the University of Michigan's consumer sentiment survey. The 5-10 year expectation was at 2.6% in June 2016. The preliminary estimate for June 2017 was reported last week. It was also at 2.6%. The final estimate will be reported next week (June 30). Taken together and individually, the surveys suggest inflation expectations are considerably more stable than the market-based measures.
The surveys encourage us not to be completely satisfied with the conventional narrative. It is missing something: Demand.
There are at least five sources of demand that a robust explanation would address. First, the post-crisis regulatory environment (national as well as international) requires large financial institutions to bolster their core capital. Government bonds meet the need. Second, the central bank continues to buy bonds as it recycles the proceeds from maturing issues.
We reckon that in September, provided there are no new shocks, the Fed will announce that starting in October, it will begin refraining from fully recycling the maturing funds, but this will be a very gradual process. The Fed has not announced the size of the balance sheet it will target. Many suspect something closer to half its current size may be appropriate.
Third, there is a demand for US debt instruments by foreign investors trying to escape the negative or low yields available in their domestic market. In the middle of the noughts, as the Fed under Greenspan was raising the Fed funds target, long-term rates fell. This was dubbed the Greenspan Conundrum. It was not such an outlier as commonly claimed. In any event, a Governor at the time, none other than Bernanke, provided what arguably was the most robust explanation of the time.
He suggested that developing Asian countries, many of whom began running current account surpluses after the 1997-1998 financial crisis had capital markets that were underdeveloped and could not absorb the local savings. Instead, the savings were exported to the US. Bernanke's explanation had a role for OPEC and recognized a demographic function. Bernanke's broad narrative still seems to be intact, but the players have changed.
Fourth, although the Federal Reserve has raised interest rates four times since late 2015, and three times since the US election last November, the higher rates have not been passed to consumers. In order to secure a better return, it appears that retail investors have moved into high-quality bond funds as an alternative to cash.
The fifth factor is not operational yet, but it is something that is another potential source of demand. Earlier this month, the US Treasury recommended revisions to financial regulation including some adjustment to the supplemental leverage ratio (SLR), which is one of the metric regulators look at in the US (and the UK). In particular, the recommendation is that holdings of US Treasuries, cash being held by the central bank, and the initial margin requirements for centrally cleared derivatives should be excluded from the calculation of the SLR.
The Bank of England has taken similar measures but have not excluded Gilts from its SLR calculation. This would be a US innovation. It effectively cuts the cost of funding US Treasuries significantly and would increase the capacity of bank balance sheets. The lower cost funding would make US Treasuries more appealing for US financial institutions.
Sometimes lowering funding costs can offset other constraints to make an asset look attractive. Remember in Q4 last year; the cross currency swap basis made it very cheap for dollar holders to swap for yen and buy a short-term Japanese debt instrument. By securing very cheap funding, they were able to buy negative yielding instruments and earn a better return that was available in the Treasury market itself.
In broad strokes, we have painted a picture that builds on what we argue is an incomplete conventional narrative that seeks to explain the drop in nominal yields as a decline in inflation expectations. We suggest that the market-based measures of inflation expectations may be compromised by liquidity differentials. Survey-based measures are considerably more stable. The conventional narrative does not integrate the strong demand for US Treasuries from both domestic and international sources. It stems from regulatory issues, competitive considerations and the lack of pass-through of Fed rate hikes to higher interest rates on household savings.