Continuing with yesterday's commentary on "residual seasonality" being applied to GDP measures, there are a few more observations I will make before discussing the potential capital market response to these actions, which I will discuss tomorrow.
The immediate catalyst for residual seasonality to become an issue was not the fact that the professional economists so badly missed the mark for first-quarter GDP, offering public estimations that it would come in at about 2% before the Bureau of Economic Analysis (BEA) estimated it to be 0.2%.
It was because the Federal Reserve Bank of Atlanta's GDPNow model much more accurately forecast, well in advance, that the BEA would announce a 0.1% annualized rate of growth for Q1.
The importance of this is that the GDPNow model is designed specifically to mirror the BEA model using publicly available data. The reason for structuring the model this way and to provide the estimates and methodology publicly is to prove that it is possible to accurately forecast the BEA's GDP estimate without the BEA's proprietary data and that there is no reason for the private-sector economists and Federal Open Market Committee (FOMC) members to be so consistently wrong.
I wrote about this issue last July in the column, "The Rise of 'Nowcasting' Economic Activity," following the Atlanta Fed's announcement that it would start to release the GDPNow forecasts publicly.
The GDPNow model, in other words, is not a system for estimating GDP. It's a system for estimating what the BEA will state the GDP is.
That makes it very different from the GDPplus model created by the Federal Reserve Bank of Philadelphia.
That model enhances the BEA model in an attempt to more accurately forecast what GDP is, rather than what the BEA will announce it to be.
This distinction is important because although the GDPplus model implies a Q1 2015 annualized rate of growth in GDP of 1.8%, which is very close to the consensus of 2% offered by the private-sector economists, the two groups perform different functions.
The purpose of private-sector publicly offered forecasts is to anticipate the BEA's announcement, not to independently offer a competing or alternative narrative.
The bottom line is that the private-sector forecasts are wrong and have been consistently wrong for many years. Applying a residual seasonality adjustment to the BEA's data and conclusions doesn't change that. In fact, it makes the BEA model less public and more proprietary, and thus less predictable.
We'll have to wait until July 30 to find out how much of the new system for calculating GDP by the BEA will be made public.
There is, however, another big problem with this issue for monetary policy and the potential for rate hikes by the Fed.
The residual seasonality adjustment doesn't just cause Q1 figures to be revised upward, it causes the figures for the other three quarters to be revised downward. It is a means of smoothing the cyclicality of quarter-to-quarter differences.
The reason this is important now is that the second-quarter 2015 GDPNow forecast, at 0.8% annualized, is well below the historical average of the past 30 years of about 3.3% and of the past 15 years of about 2.85%.
That rebound is absent this year and is almost certainly independent of any residual seasonality.
The second quarter on average has the highest GDP measure of all four quarters, and yet this quarter the estimated rate of growth is not only substantially below the historical average, it's also below the estimate for this year's Q1 growth at the same point in the quarter.
If the current GDPNow forecast of 0.8% holds, although applying a residual seasonality adjustment to Q1 will cause the figure to increase, applying it to Q2 could cause it to show a contraction.
However, even if it doesn't show a contraction as having started, it will clearly indicate, at the very least, that a substantial deceleration in economic activity is under way at the same time the Federal Reserve is trying desperately to create an economic logic narrative that justifies raising interest rates.
That justification, for the past six months, has been grounded in the Fed's repeatedly offered assessment that the lack of growth in economic activity is transitory and the unemployment rate has been falling.
If Q2 GDP growth comes in near Q1's and well below the historical average for Q2, the Fed will be hard pressed to continue to pursue that line of reasoning.
