Real Gross Domestic Product expanded by 2.54% in Q2 2017, below most estimates including the final one from the Atlanta Fed’s GDPNow model. That latter method was close once again in its final days (+2.8%), but earlier in the quarter was predicting GDP growth of 4.3%. That would have been like what many people were thinking after another awful first quarter, a meaningful rebound in the second that would give “reflation” a statistical boost suggesting at last some real economy action behind it.

With the July release of GDP estimates the BEA has as it does every year updated its benchmarks. As a result, Q1 2017 GDP was revised lower to 1.22% growth, meaning that for the first half of this year the US economy (as measured this way) expanded by just 1.89%; the same awful trend rather than “reflation.”

The most revealing aspects of these revisions was in how they treated the two parts (so far) of the “rising dollar” period. The first part up to Q3 2015 was revised higher, and therefore due to compounding GDP overall was revised up. The quarterly figure for the highly contentious Q1 2015, the one that gave the world “residual seasonality”, is now imagined to have been 3.19% rather than -0.20% for the so-called final estimate.

While that may give proponents of “residual seasonality” some mild comfort, it actually proves instead the negative interpretations of the economy drawn from when the quarterly estimate was near-zero and less. The economy was, in fact, facing grave difficulties, as what followed later in 2015 skirted just barely the boundaries of recession.

Under the July 2015 benchmarks, Q4 2015 GDP was thought (final estimate) to have been 1.38%, growth followed by 1.07% in Q1 2016 (final estimate). The July 2016 update reworked those rates both less than 1%, to 0.86% and 0.83%, respectively. Now the BEA figures those crucial quarters were together barely positive at all, 0.48% and 0.58%, respectively.

Though that was over a year in the past, the economy especially in the manufacturing sector still has an inventory problem. In Q2, inventory neither added nor subtracted from GDP. As in Q1, private inventories were (after revisions) believed to have declined but again only slightly (in GDP accounting, second derivatives are what get including in quarterly GDP growth, meaning that the -$2.6 billion contraction in the second quarter was essentially the same rate as the -$100 million in Q1).

Those small reductions, however, aren’t anything like the drastic inventory liquidations we find at the end of every business cycle. The post-Great “Recession” period is itself atypical and unlike any cycle, but there are still cyclical processes at work in the economy especially in inventory and its effects on manufacturing where the marginal economic direction is clearly being set.

Benchmark revisions now suggest that the inventory imbalance gathered largely in late 2014 and early 2015 was larger than thought, even though it was already estimated to have been the greatest inventory gain since the middle 1990’s. That only makes the lack of liquidation during and after the near-recession all the more confounding, and therefore a factor in why US and foreign industry (through US purchases of goods produced anywhere) can’t gain any rebound momentum beyond the absence of further economic contraction.

Despite those upward revisions for several years ago, the change in GDP levels didn’t alter the true picture of the overall economic situation. Slightly better in 2014 and early 2015 does not make up for the ongoing non-linear contraction, where had the Great “Recession” been a recession Real GDP in Q2 2017 would have been around $21.7 trillion rather than slightly more than $17 trillion.

That almost $5 trillion difference is everything, an enormous gap in output that explains the lack of demand for labor input. It became years ago, especially after 2011 (“dollar”) and the 2012 slowdown, self-reinforcing upon real economic conditions as well as future expectations. Businesses don’t hire because revenue (and cash flow) has remained subdued, meaning fewer Americans in proportion working (or even looking for work) therefore buying far less goods and services, confirming business expectations for little or no revenue growth.

In that situation, as noted before, liquidity preferences rule all considerations including the over-management of costs. In a normal business cycle/growth period, firms would consider first their ability to meet future growth rather than overemphasizing caution and liquidity.

Translated into GDP, that’s the roller coaster of Non-Residential Fixed Investment shown above. Under prior economic cycles, it tends to grow in steady rather than uneven fashion, reinforcing the overall uneven economy.

Because the economy keeps alternating between near-recessions (2012, 2015-16) and shallow rebounds this is no economic expansion. Together with the initial recovery period that only partially offset the enormous contraction of 2008-09, the economy has gone nowhere despite the passage of a decade since trouble first erupted monetarily.

Unfortunately, the latest GDP estimates and updates all merely confirm that through the middle of 2017 it isn’t likely to get somewhere anytime soon. The only aspect that seems to have changed is that in GDP terms the rebound this time after the near-recession quarters is considerably weaker and less convincing than the one in 2013-14. That would certainly fit curve comparisons from the UST yield curve to eurodollar futures and even WTI.