- Next Friday’s Q2 GDP report is likely to come in better than expected – perhaps by a significant degree.
- While backward-looking in nature, any such “beat” will likely affirm the recent optimism in the domestic equity and credit markets – optimism that we’ve admittedly been on the wrong side of.
- That being said, however, we’re inclined to fade such optimism given our expressly dour outlook for domestic economic growth from here.
With the advent of the JUN Retail Sales, JUN Industrial Production and JUN CPI data, Friday was an important day for U.S. macroeconomic data and, more importantly, the predictive tracking algorithm we employ to forecast Real GDP growth on an intra-quarter basis.
The key read-through from these data points is that YoY consumption and investment growth are both tracking higher on a quarterly average basis (i.e. Q2 vs. Q1). With key metrics of inflation flat QoQ and the net export matrix largely unchanged as well, the probability that Real GDP growth decelerates in Q2 is extremely low. The more likely outcome is that growth accelerates modestly; our model is now anticipating a sequential bump up to +2.3% YoY in Q2.
On a headline (i.e. QoQ SAAR) basis this figure translates to +4.8%; no that is not a typo. Recall that our latest update had been calling for Real GDP to decelerate from Q1’s +2.1% YoY growth rate to +1.8% YoY, which translates to +2.7% on a QoQ SAAR basis. That +50bps of positive revision to the YoY growth rate equates to nearly a doubling of the headline growth rate.
For a variety of reasons detailed in our 9/2 Early Look titled, “Do You QoQ?”, we don’t lend too much credence to the headline growth rate – which itself is more “noise” than “signal” – but it’s important to track nonetheless – if only because Macro Consensus queues off of it when formulating opinions about the state of the macroeconomy. Indeed, formulating differentiated [and profitable] opinions about growth and inflation requires a differentiated modeling process, which we detail below:
- Stochastic (us) vs. Econometric (them): Our GIP Model employs a top-down approach that uses trailing momentum and volatility in the GDP series itself as inputs to determine a range of probable outcomes. Contrast this with traditional econometric models that anchor on reported high-frequency data to “build” a singular GDP estimate from the bottom-up. In conjunction with the Bayesian inference process highlighted below, we are able to have reasonable estimates for GDP growth up to four quarters out, as opposed to econometric models whose accuracy is severely limited on an out-quarter basis due to a lack of reported high-frequency data. As a result, we tend to spot critical inflections in the trending momentum of the economy 3-6 months ahead of Macro Consensus.
- Bayesian (us) vs. Frequentist (them): Our GIP Model employs a Bayesian inference process that uses base effects as inputs to determine the direction and magnitude of adjustments from the base rate (i.e. the prior reported growth rate) with the output being a singular growth estimate that falls within the aforementioned range of probable outcomes. Our Bayes factor is the incorporation of a predictive tracking algorithm that guides our estimate to the most appropriate level within (or sometimes outside of) the aforementioned range on an intra-quarter basis. Contrast this with the Frequentist approach employed by Macro Consensus which typically defaults stock estimates of +2.5%, +3.0% or, worse, +4.0% and adjusts from there according to reported high-frequency data.
To the extent that +4.8% headline growth rate is proven in the area code of accurate when Q2 GDP is reported next Friday morning, there are three key considerations for investors to consider:
1) After starting the quarter with Wall Street’s lowest estimate for Q2 GDP growth (+1% YoY; +0.3% QoQ SAAR), we are now the Street high estimate. For comparison’s sake, the Atlanta Fed and Bloomberg Consensus are currently at +2.4% and +2.5%, respectively, on a QoQ SAAR basis.
We pride ourselves on the accuracy of growth and inflation forecasts; our predictive tracking algorithm has an intra-quarter standard error of 28bps vs. 252bps for the Atlanta Fed’s tracker. As such, an intra-quarter revision of that magnitude is embarrassing to say the least and represents one of the few major mistakes we’ve made in recent years. While unachievable, perfection remains our goal and we’ll continue to evolve our models as needed in pursuit of that objective.
2) On their own, absolute growth and inflation numbers mean nothing to our investment process. Rather, our analysis has shown that analyzing such figures on a second derivative basis leads to more impactful conclusions from the perspective of predicting the performance of key factor exposures across asset classes.
We triangulate second derivative trends across Real GDP growth and Headline CPI in our four-quadrant GIP analysis (shown below) and the positive revision to our Q2 GDP estimate suggests the U.S. economy likely moved into #Quad1 last quarter; this represents a delta from our initial forecast of #Quad4 and can explain why the equity and credit markets were so resilient heading into and throughout last quarter.
3) Even if we’re too high by ~100bps on the headline GDP print next Friday, a growth rate in the mid-to-high +3% range will be interpreted very positively throughout the investment community. By reinforcing what we’ve shown to be misguided expectations of residual seasonality, it will also impact the views of policymakers as well.
Recall that the FOMC effectively shrugged off the initially weak Q1 GDP print as largely a function of poor seasonal adjustment techniques. A trend growth rate of +2.5% to +3% is likely to allow Yellen to look past trending weakness in her proprietary Labor Market Conditions Index and adjust policy guidance in the hawkish direction.
This means the time between the FOMC’s July 27th meeting and its September 21st meeting is likely to contain a meaningful degree of hawkish rhetoric out of the various Fed Heads. That may serve to catalyze incremental convergence in the factor exposures we’ve been bullish and bearish on in the YTD; last week may have been a preview to the extent all of this wasn’t immediately priced in. Conversely, toned-down expectations of fiscal stimulus might prove to be rather bearish for risk assets for a TRADE.
Looking ahead, there are three very important factors to consider as we progress throughout the back half of 2016 and into 2017:
- Relative to other key metrics of inflation, the GDP Deflator is being understated by a significant degree. Specifically, the spread between the GDP Deflator and the Fed’s preferred metric for inflation (i.e. the PCE Core Price Index) was -160bps in Q1; that 0.8th percentile reading represents a z-score of -2.2x on a trailing 30Y basis. Headline Real GDP growth would have been negative in Q1 at -0.5% had the two inflation figures been equal. #electionyearmath
- Our model has Headline CPI accelerating by a substantial amount throughout the balance of the year. To the extent the GDP Deflator and Core PCE Inflation remain co-integrated and the former mean reverts to where other key metrics of inflation are tracking, we could see a sizeable “accounting” hit to Real GDP growth over the next 2-3 quarters.
- The confluence of the aforementioned accounting drag, decelerating employment and consumer credit growth, as well as consumer confidence and bond yields concomitantly trending lower into peak base effects for consumer spending leads us to have a negative bias on Real GDP growth through at least 1Q17. Specifically, our model is calling for YoY Real GDP growth to decelerate from +2.3% in Q2 to +2.0%, +1.8% and +1.4% in Q3, Q4 and Q1, respectively. Those figures translate to +0.8%, +0.6% and -0.5%, respectively, on a headline basis.
Source: Bloomberg L.P.
All told, next Friday’s Q2 GDP report is likely to come in better than expected – perhaps by a significant degree. While backward-looking in nature, any such “beat” will likely affirm the recent optimism in the domestic equity and credit markets – optimism that we’ve admittedly been on the wrong side of.
That being said, however, we’re inclined to fade such optimism given our expressly dour outlook for domestic economic growth from here. As we detail on slides 43-45 in our Q3 Macro Themes presentation, the biggest downside surprise risk remaining in the economy is a pending pickup in the pace of labor market deterioration that should commence within the next 2-3 quarters.
Best of luck out there risk-managing this whiplash in U.S. growth expectations. As always, feel free to email or call with questions.