Retail sales for the May reporting period were very weak and caused even the most optimistic analysts to question the prospects for high economic growth. I have consistently been on the opposite side of this, forecasting growth would continue to come in below estimates and that so far is what has been happening.
The retail sales control group, the number that feeds into GDP, declined to 3.0% year over year growth from 3.7% last month and down from 5.4% in January of this year.
With consumption metrics continuing to disappoint, how did the Atlanta Fed raise their GDP estimate after the May retail sales report? They revised their forecast from 3.0% to 3.2% on weaker than expected data?
The Atlanta Fed GDP model has two inputs: actual reported data and estimates for the data that has not been reported yet for the quarter. Each week, the reported data comes in weaker than estimates and causes the GDP estimate to fall, but the Atlanta Fed has been increasing the estimate for consumption in the months to come to offset the recent declines. As the data comes in to replace estimates, it is my forecast that growth estimates will continue to decline and converge in the rage of 2.0%-2.4% for Q2 vs the current 3.2% estimate above. I will break down how I arrive in the low 2s at the end.
(All data from Census Bureau and BLS) (All control group)
Retail sales minus CPI (inflation), or real retail sales, have been falling quite rapidly since the beginning of 2016. The peak of the cycle is easy to see in the chart above, and based on the available data, retail sales growth will continue to decelerate in the coming months. Real retail sales growth is already at levels seen at the start of prior recessions.
This is another way to look at real retail sales that smoothes the effects of the volatile CPI number. The picture is however the same. The peak of growth in retail sales came in 2016 and has been decelerating at an increasing pace. Again, we are at levels that are the same as the start of the prior two recessions.
I adjust all data for inflation using the headline CPI but I also find it useful to use rent inflation vs. headline CPI as it is more stable (less volatile) and likely a more accurate representation of the inflation in the economy. Looking at retail sales in this way shows a negative real growth rate and at the risk of sounding alarmist, a very dire picture for the consumption side of the economy.
The Atlanta Fed Games
The data above is actually reported, hard data. It cannot be disputed. Real retail sales are falling quite dramatically. With this data in hand, how could the Atlanta Fed revise up their GDP estimates?
As you can see in the red box, the Atlanta Fed revised up the total GDP estimate from 3.0% to 3.2% and the same for PCE (Consumption).
We only have one month of consumption data that has been reported (April) and the current rate of PCE growth for Q2 is 2.2%. The other two months of data are estimates that the Atlanta Fed can essentially make up to achieve whatever number they would like. Of course, as the actual data comes in, they have to change the estimates with hard data.
It is for this reason, the Atlanta Fed has a tracking error of around 200 basis points from their first estimate to the final reading.
For the 3.2% GDP figure to be correct, consumption for the next two months must come in a full 1% higher than we currently are. This has nearly a 0% probability of occurring. The highest probability is that consumption remains where it has been for years, and where it currently sits as of April, at the 2.0% to 2.5% range. Using 2.25% (the middle) as the input for PCE rather than the ridiculous assumption for 3.2% generates a GDP growth rate for Q2 at 2.3%. This is where the Atlanta Fed will end up at the end of Q2 when they have to replace their made up assumptions with actual data.
The Atlanta Fed, for whatever reason, revised up their consumption numbers when the actual consumption data came in weaker than expected.
As the data is reported and their estimates are replaced with real data, their GDP tracker will fall another 100 basis points to the low 2% range.
As growth and inflation data continue to slow, bonds, or TLT, will continue to reap the benefits as they have been doing.
TLT is up over 6% YTD. Over 7% including dividends. The yield curve also reached the lowest level since the election dropping below 80 bps. These are two massive markets that do not believe the nonsense growth stories that are merely wishes and totally removed from the actual data that is being reported.
Growth will continue to slow and inflation will continue to disappoint. The bond market gets it.
Disclosure: I am/we are long TLT. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.