Construction Spending Falls 0.3% in May after March & April Spending Revised Higher, Commenter RJS, Marketwatch 666
The Census Bureau report on construction spending for May (pdf) estimated that May’s seasonally adjusted construction spending would work out to $1,545.3 billion annually if extrapolated over an entire year, which was 0.3 percent (±1.0 percent)* below the revised annualized estimate of $1,549.5 billion of construction spending in April, but 7.5 percent (±1.3 percent) above the estimated annualized level of construction spending in May of last year…with this release, unadjusted construction spending data was revised back to January 2019 and seasonally adjusted data was revised back to January 2014…as a result of that and the usual monthly revision, the April spending estimate was revised 1.7% higher, from $1,524.2 billion to $1,386.1 billion, while the annual rate of construction spending for March was revised 1.8% higher, from $1,521.0 billion to $1,548.555 billion…we would normally suggest that a large upward revision to annualized March construction spending would have large positive impact on first quarter GDP when the annual revisions to GDP are released in late July, but with 4th quarter construction also being revised, the entire quarter over quarter change will need to be recomputed, a chore which we’re not about to undertake at this time . . . after revisions, construction spending tor the first 5 months of 2021 now amounts to $594.8 billion, 4.6 percent (±1.0 percent) more than the $568.5 billion in construction spending for the same 5 months of 2020.
A further breakdown of the different subsets of May construction spending is provided in a Census summary, which precedes the detailed spreadsheets:
- Private Construction: Spending on private construction was at a seasonally adjusted annual rate of $1,203.3 billion, 0.3 percent (±0.8 percent)* below the revised April estimate of $1,206.8 billion. Residential construction was at a seasonally adjusted annual rate of $751.7 billion in May, 0.2 percent (±1.3 percent)* above the revised April estimate of $750.3 billion. Nonresidential construction was at a seasonally adjusted annual rate of $451.6 billion in May, 1.1 percent (±0.8 percent) below the revised April estimate of $456.5 billion.
- Public Construction: In May, the estimated seasonally adjusted annual rate of public construction spending was $342.0 billion, 0.2 percent (±1.8 percent)* below the revised April estimate of $342.7 billion. Educational construction was at a seasonally adjusted annual rate of $82.0 billion, 1.9 percent (±1.8 percent) below the revised April estimate of $83.6 billion. Highway construction was at a seasonally adjusted annual rate of $98.6 billion, 1.4 percent (±6.1 percent)* above the revised April estimate of $97.2 billion.
This construction spending report will be used as source data for 3 subcomponents of GDP; investment in private non-residential structures, investment in residential structures, and government investment outlays, for both state and local governments….however, gauging the impact of the revised April and May construction spending as reported here on 2nd quarter GDP is difficult because all figures given in this report are in nominal dollars and as you know, data used to compute the change in GDP must be adjusted for changes in price…accurately adjusting construction spending for price changes is no easy matter, either, because the National Income and Product Accounts Handbook, Chapter 6 (pdf), lists a multitude of privately published deflators that are used by the BEA for the various components of non-residential investment, ie, such as using the Engineering News Record construction cost index for utilities’ construction . . . in lieu of trying to find and adjust for all of the obscure price indices the BEA uses, we’ve opted to just use the producer price index for final demand construction as an inexact shortcut to make an aggregate price adjustment sufficient to make an estimate…that index showed that aggregate construction costs rose 0.6% in May, after being up 1.1% from March to April, up 0.5% from February to March, and up 0.2% from January to February.
On that basis, we can estimate that May construction costs were roughly 1.7% greater than those of March, 2.2% greater than those of February and 2.4% greater than those of January, and obviously 0.6% more than those of April . . . we then use those percentages to inflate spending for each of the months of the first quarter which, for purposes of comparison, is arithmetically the same as deflating April and May construction spending vis-a vis the 1st quarter . . . annualized construction spending in millions of dollars for the five months in question is given here as 1,545,272 for May, 1,549,488 for April, 1,548,555 for March, 1,533,252 for February, and 1,549,793 for January….thus to figure the annual rate of change of May’s nominal construction spending figure of $1,293,872 and April’s figure of $1,304,007 from those of the ‘inflation adjusted’ figures of the first quarter, our calculation becomes (((1,545,272 + 1,549,488 * 1.006) / 2) / (((1,548,555 * 1.017) + (1,533,252 * 1.022) + (1,549,793 * 1.024)) / 3)) ^ 4 = 0.939855, which means that after adjusting for inflation, construction spending has been shrinking at a 6.01% annual rate over the first 2 months of the second quarter . . . put another way, that would be a construction spending contraction at a $24.26 billion annual rate, which means that if June construction shows no improvement, the 2nd quarter contraction in real construction would subtract a net of about 0.64 percentage points from 2nd quarter GDP across those components that it influences…