Monday, April 24, 2017

US Treasury rates and curve: a neoFisherien perspective.


US Treasury Curve and Outright Trade Strategy Analysis

[Using a Fisher Rate Understanding of NGDP: Spot and Forward Expected]

This report is for market close immediately preceding:
## [1] "April 25 2017"
This analysis justifies or dismisses the use of long duration US Treasurys for a rebalanced equity and fixed income portfolio, with the fixed income almost always long duration US Treasurys and the weights constantly reweighted. The only “hedge” or offset for equity risk available to the American investor is a position in long duration US Treasurys.
However, the following will show that the Federal Reserve is significantly behind the market and long duration US Treasurys have jump risk in rates, with long US Treasurys potentially reaching 4% or higher yield.
Therefore long US Treasurys are currently unusable in a dynamically balanced equity portfolio strategy.

Current Points:

1. US Treasury rates are unusually low given intense pressure constantly applied by the US Federal Reserve in the form of deliberately delaying normalization well after the crisis setting that required “emergency” low rates. The US Federal Reserve is now over 3 years “late” in normalizing rates. The Taylor Rule currently solves for a 3% Federal Funds rate. The Taylor Rule is not a multi factor regression but a theory of the relationship of US growth and inflation to Fed Funds. US Treasury 5 years are likely to jump towards 4% and the US Treasury 20 year will jump towards 5%. The expected NGDP as derived from market rates, for this year to the longest date we derive of 7 years, are about 1/2 the rate level that would align with US Treasurys curve rate levels and about 1/3 a reasonable studied view of US potential GDP (Keynes, Fisher, Taylor). There are similarities between the Japanese economy from 1992 to date and the US economy now.

2. The curve will steepen - 5 years to 20 years - by 100 to 175 basis points or more, again likely in a jump.

3. Volatility is extraordinarily low so that all trades should be long vega and long convexity in structure. This “artificial” level of volatility is induced by the current FOMC policy.

This analysis uses a Fisher Rate spot and forward derivation. James Bullard of the St Louis Fed provides the relationship of the Fisher Rate with a rules based FOMC policy. And we believe significant errors by the Fed (either by political commission or happenstance) have been made such that great inefficiencies are now in UST pricing as well as the USTC. The recent work on developing the “Neo Fisherism”" thesis correctly identifies why and to what degree UST have been distorted. The once in a 1/2 a century problem of QE distortion has large inefficiencies in UST and the curve, with flattening the yield curve out of proportion to how, indicated by via the Fisher Rate, to the normal relationship to the US economy. The long intermediate US Treasurys rate is approximate of NGDP expected over time adjusted for the risk premium per year, and the yield curve for longer intermediate US Treasurys to long duration US Treasurys is the growth of the NGDP.
These inefficiencies allow good opportunity in trading UST for at least another two years or more. If the stress and use of FR is misplaced, then positions are established at levels which are, in fact, efficient so while inordinate return will not occur, inordinate loss will not occur either.
However the use of a neoFisherien ideas in trading and pricing the yeild curve and US Treasury rate levels is rare, we are the only ones we know if who do so, that there will not be a discrete market move to effecient levels but large sudden jumps, making risk control difficult.
The Fisher Rate, spot being Fed Funds, for every 3 month forward out to 7 years is plotted.
The process is unique. Most US Treasury and curve traders or portfolio managers use an approach like what Yellen applies where the monetary policy Fed Funds rate is set to a level in relationship to a Wicksell “natural rate”. The Wicksell natural rate is set by intuition as it cannot be observed. This “natural rate” approach does not consider Fed Funds in terms of the relationship of the Fed Funds to Fisher Rate to NGDP; with the Fisher Rate being the equivalent, in the end, to inflation + GDP, or NGDP. Furthermore, the use pf a Wicksell “natural rate” can easily allow the Fed to pursue a political agenda as it is cloaked under the unobservable nature of the Wicksell rate.
The following analysis is based on the spot and future NGDP/Fisher Rate/Fed Funds rate which in the end syncs with the spot pricing of US Treasurys rate and the yield curve - the relationship between the curve and spot US Treasury prices is co-dependent where if the Fed strives to a certain curve and certain US Treasury yield, that will for the short and intermediate run define NGDP based on the power of the Federal Reserve, which if course is considerable, and thereby the curve and rates, the NGDP will in turn define the curve and US Treasurys as the two are o to the limits of the Fed power - cp-dependents. This is the “Neo Fisherism” effect of the Fed Funds defining NGDP and NGDP in the end defining Fed Funds, and thereby having managed rates defining on NGDP or if the Fed is truly “rules based” the Fed Funds rate is defined by NGDP.
A Federal Reserve insisting upon extraordinary low Fed Funds rate can over time end with a low NGDP well below the potential NGDP. Japan has shown this for the last 16 years and was the Japanese experience from 2000 to date.
The process to derive the forward Fisher Rate is to first derive the risk premium, how much more yield is required to extend a year in UST maturity.
To do this is counter- intuitive but simply put: 1) the curve is monitored over the 7 years term structure to get a read - 7 years being a point in time when a business cycle will occur, this is to get past the immediate trading of anticipating a business cycle; 2) the curve is deconvexed by calculating what the positive convexity of the curve is worth using bond volatility; 3) then a one year 6 years forward read is taken of the forward deconvexed curve. That is the risk premium. The risk premium, considering the riskiness of normalization and the risk pricing as to the effect of the new POTUS on fiscal policy, is abnormally low, moving below 10 basis points. We feel this indicates how behind the market the FOMC is, how it is leaning in to maintain low rates. The risk premium should be around 20 basis points, not the current sub 10 basis points. And the Fed continues to pressure rates down such the risk premium is decreasing. The additional 10 to 14 basis points in risk premia if the Fed had or was normalizing would result in curve 100 to 150 basis points steeper than the current curve.
We feel this will occur in jumps and not discreetly as Yellen is now stating. In fact we find it silly (if not arrogant) to propose r depend upon the Federal Reserve to table a one or two year plan on how many increases of Fed Funds will occur and at what rate of increase.
The risk premium above is used to build a term structure of the instantaneous Fed Funds rate in the future, or the Fisher Rate, which in turn is considered to be NGDP. In this case a 7 years forward instantaneous NGDP. Keep in mind this is not a forward forward but the instantaneous NGDP (Fed Funds) expected in market prices. NOte the risk premia and expected NGDP are moving in opposite directions. The risk risk premium and expected NGDP should move in the same direction. The ethusiasm for Trump’s expected economic policy is opposite to the risk premum. Large jumps in the curve and in rates should materialize.
Note as well that as soon as the Yellen Fed started, the RP fell from 30 basis points to now about 7 basis points, despite a brief rise towards 20 basis points foe the first Fed Fund raise in December 2015. But as it became clear that Yellen would not normlize - the RP reversed and fell to current lows. This drop in risk premium is a stochastic measure, similar to vega, and ends with pricing a abnormally flat yield curve.
The above shows forward NGDP expectation in 7 years over time. We now derive the forward Fisher Rate, the NGDP, in quarterly periods up to 7 years, a term structure, and look at this term structure of expected NGDP at certain times and consider how it is changing.
Trump potential policy is raising expected NGDP, though for the last few weeks this has reversed and expected NGDP is lower. This indicates less confidence in the pro growth policies of Trump. This reems to have peaked at 2 1/2% and is now towards 2 1/4%. The Trump raise in expected NGDP combined with the Federal Reserve refusal to deal with this potential by the “go slow”Fed Funds raise also explains the drop in risk premium (above), which is making for unprecedented explosive pressure on the yield curve and US Treasury 20 year rate as the curve flattens and the 20 year rate stays at low rates. The curve and the 20 year US Treasury will jump by about 100 to 175 basis points in a short time frame - week or month - when the correct risk premium and a forward NGDP matches the Trump realized NGDP. Or the recent drop in expected NGDP and risk premium indicates that Trump will fail.
The drop in risk premium and now a easing of expected NGDP could also indicate a neoFisherien world is starting to make the US econommy sceloric and a Japanese “lost decade”, or a secular stagnation economy will occur.
The slope of the term structure of expected NGDP will over time be in synch with the US Treasury curve.
If it is not, that it is a trade opportunity. However the slope is flattening consideably since the year end 2016.
This slope in expected NGDP is equivalent to the FOMC “dots”, but the dots are not derived as expected NGDP is derived in this report, but a “best guess” of the FOMC which is is assumed has better information than.
NGDP expected should make sense in terms of the schedule for NGDP change, and the level of NGDP at any time. The yellow dotted line ins NGDP before the crisis, indicating that after A Fed tightening was digested, the US NGDP would be consistently over 4%. The blue line shows that the nadir of the crisis as far as US Treasurys and Fed policy goes was not in 2009, but in 2012, Then the expectations indicated expectations for a Japanese like slow burn depression. The red line is the current terms structure of expected NGDP which still shows an abnormal low expectation for NGDP, though higher since the time when pre election polls showed most thought HRC would win the election; not in sync with a 4% plus regular NGDP and certainly not at all in sync with a Trump strong fiscal stimulus plan.
The current pricing of expected NGDP shows that long duration US Treasurys are unusually low and though already backing about 50 to 75 basis points. And the 2% level of expected NGDP is not in sync with record high SP500 as it indicates sub-par Japanese “lost decade” GDP is expected. This will resolve moving to the long term levels of March 2007 (yellow dotted), in a series of large sudden jumps.
Plotting the expected NGDP in 7 years over time versus long duration UST shows whether the UST is currently rich or cheap versus the expected NGDP. While US Treasury rates have risen since the election, they will rise another 150 basis points, assuming NGDP expected in 7 years does not rise. Since we feel expected NGDP will rise by 150 basis points to 200 basis points, that will be an additional 100 basis points in the US Treasury 20-year yield to 4%
By locating certain key dates, the effect of Neo Fisherism and how out of whack UST has been since the Yellen Fed began are obvious. The recent jump in rates with the Trump election victory rise in NGDP expectations is obvious, as is the distance the 20 year UST has still to go as NGDP (monetary rate policy) normalizes. There is significant return in various short US Treasury rate trades until the UST 20 year is 4% or higher.
While the expected NGDP has dropped by about 25 basis points from the highs of the post election surge in expected NGDP, the 20 year US Treasury has dropped by almost 1/2 %.
The relationship of long term expected NGDP to the UST 20 year is regressed and a rich/cheap pricing of the 20 year US Treasury is arrived at, graphed in terms of the 20 year US Treasury given expected NGDP compared to the current trading of the US Treasury 20 year. After the election to Feb 2017, extpected NGDP indicated a fair value for the US Treasury 20 year of 3%. Since February the expected NGDP has dropped and the US Treasury 20 year is now fair value of about 2 5/8%.
Taking a closer look (smaller rainge in “x” axis - or US Treasury 20 year yield) shows the US Treasury 20 year is retracing the path from the beginning of the Yellen Fed to the Dec 2015 rate rise. But this has reversed since February 2017 with both expected US Treasury 20 year, the US Treasury 20 year and the expected BGDP dropping. Our read is this is starting to forewarn that the Trump will fail in their stated goals of a 4% NGDP.
Yet the US Treasury 20 year is still rich by about 25 basis points even if Trump only succeeds with a 2 3/4% NGDP.
The same regression of expected NGDP to the UST 5 year is done using both NGDP expected in 7 years and NGDP expected in 4 years.
If NGDP stays at the current level or drops, the UST 5 year is about 20 basis points cheap. Our read is that expected NGDP is lagging , and the US Treasury 5 year prices and NGDP expected will jump 100 basis points as a 2% Fed Funds is reached and then exceeded.
The curve from intermediates to long duration in UST is the forward pricing of NGDP growth. Just as for UST 20 years and UST 5 years, we regress the curve to expected NGDP over time and compare that to the 20 to 5 curve. While the curve is not as flat based on expected NGDP as it was in February 2017, it is still too flat by about 20 basis points, even if this substandard growth is all Trump succeeds with.
Given the analysis for 20 year and 5 year above, and considering the curve graphed to expected NGDP, the curve is very flat now versus where it has been in the past for the last 50 years. A rise in expected NGDP will remove some of this very large cheapness in the curve (i.e. it is too flat and will steepen), but even then the curve will steepen by at least 50 basis points. A technical effect from QE is also capping long rates, not from supply and demand but from the “lack of safe assets” repo.
The 5:20 curve will steepen in a jump by at least 100 basis points.
Spot curve versus the 5 year is graphed, starting with 5 year and 20 year yield and the curve over time.
The curve from 5 years and 20 years can be considered the forward expectations of NGDP growth. The UST 5 years is a good forward expectation of expected NGDP.
Current levels show that the UST 5 year may jump suddenly by 150 to 200 basis points to be in sync with both a normalized NGDP expectations and in terms of the level in 5 years being in sync with forward views of “normal”" growth.
The SP500 lagging return is plotted against the forward expected NGDP.
The SP500 is reflected in the forward expected NGDP and the forward growth of NGDP. Or to sustain current levels of SP500 the NGDP expected must achieve certain levels. Currently the NGDP of around 2% will not sustain SP500 over 1800 level.
The curve 5:20 year UST is graphed versus expected NGDP. It is obvious the curve will steepen dramatically unless the US enters a serious recession, if not a slow burn depression of secular stagnation.