Understanding the Construction Revisions
Yesterday I commented Government "Processing Error" Sinks Housing Reports for Entire Year.
In that article, I stated 2015 GDP would be revised lower. Some disagree.
For example, MarketWatch reports IHS Global Insight US economist Patrick Newport wrote in a research note "The upward revision to spending in 2014 is enough to raise growth that year from 2.4% to 2.6%-2.7%. The revisions are likely to boost growth for 2015 as well."
Let's investigate that claim with a look at the actual revised construction data as posted by the Census Bureau.
Note: Don't study this table too long. Instead, skip to the analysis and tables that follow.
Initially Reported vs. Revised - Seasonally Adjusted Data
| Date | Total Construction | Total Residential | Total Private Construction | Private Residential | ||||
| Previously Published | Revised January 4, 2016 | Previously Published | Revised January 4, 2016 | Previously Published | Revised January 4, 2016 | Previously Published | Revised January 4, 2016 | |
| Oct-15 | 1,107,381 | 1,127,040 | 405,604 | 433,313 | 802,435 | 829,729 | 399,036 | 426,784 |
| Sep-15 | 1,096,637 | 1,123,892 | 401,658 | 432,381 | 795,841 | 824,201 | 395,021 | 425,703 |
| Aug-15 | 1,089,801 | 1,121,907 | 395,401 | 427,507 | 788,698 | 820,804 | 388,636 | 420,742 |
| Jul-15 | 1,080,358 | 1,114,716 | 388,681 | 423,039 | 781,249 | 815,607 | 382,063 | 416,421 |
| Jun-15 | 1,074,314 | 1,113,424 | 382,553 | 421,663 | 773,489 | 812,599 | 376,055 | 415,165 |
| May-15 | 1,068,424 | 1,107,569 | 379,735 | 418,881 | 776,452 | 815,598 | 373,063 | 412,208 |
| Apr-15 | 1,044,641 | 1,084,961 | 373,346 | 413,666 | 757,209 | 797,529 | 366,837 | 407,157 |
| Mar-15 | 1,006,351 | 1,052,899 | 363,879 | 410,428 | 729,713 | 776,261 | 357,512 | 404,061 |
| Feb-15 | 993,465 | 1,037,527 | 366,651 | 410,713 | 720,819 | 764,880 | 360,575 | 404,636 |
| Jan-15 | 990,051 | 1,033,261 | 364,833 | 408,043 | 716,185 | 759,396 | 358,940 | 402,151 |
| Dec-14 | 989,118 | 1,031,635 | 360,323 | 402,840 | 707,551 | 750,069 | 354,834 | 397,351 |
| Nov-14 | 976,888 | 1,016,054 | 352,509 | 391,674 | 699,344 | 738,510 | 347,159 | 386,325 |
| Oct-14 | 979,573 | 1,015,977 | 347,242 | 383,645 | 692,127 | 728,530 | 342,114 | 378,517 |
| Sep-14 | 959,182 | 992,349 | 342,529 | 375,695 | 684,862 | 718,028 | 337,109 | 370,275 |
| Aug-14 | 955,017 | 986,876 | 335,035 | 366,894 | 676,325 | 708,184 | 330,001 | 361,860 |
| Jul-14 | 952,468 | 984,990 | 334,543 | 367,065 | 673,787 | 706,309 | 329,531 | 362,054 |
| Jun-14 | 950,282 | 983,906 | 334,459 | 368,083 | 674,049 | 707,673 | 329,510 | 363,134 |
| May-14 | 957,641 | 990,737 | 338,268 | 371,364 | 681,986 | 715,082 | 333,522 | 366,618 |
| Apr-14 | 964,738 | 992,914 | 345,879 | 374,054 | 691,365 | 719,540 | 341,109 | 369,284 |
| Mar-14 | 964,995 | 989,496 | 347,470 | 371,970 | 696,050 | 720,551 | 342,607 | 367,107 |
| Feb-14 | 968,303 | 985,280 | 351,366 | 368,343 | 704,219 | 721,196 | 346,436 | 363,413 |
| Jan-14 | 961,002 | 973,655 | 354,047 | 366,700 | 696,718 | 709,370 | 349,142 | 361,795 |
Construction Spending Revised Lower 7 Consecutive Months!
For two years, construction spending went up vs. previous reported data. The net effect is GDP did indeed rise more than reported in 2014. The flip side is 2015 GDP will be lower than previous reported.
To understand why, we need to look at month over month differences as compared to previously reported numbers. Let's take a look.
Total Construction Spending vs. Previous Reports
| Date | Total Construction Spending | ||
| Previous M/M Increase | Revised M/M Increase | Difference | |
| Oct-15 | 0.98% | 0.28% | -0.70% |
| Sep-15 | 0.63% | 0.18% | -0.45% |
| Aug-15 | 0.87% | 0.65% | -0.23% |
| Jul-15 | 0.56% | 0.12% | -0.45% |
| Jun-15 | 0.55% | 0.53% | -0.02% |
| May-15 | 2.28% | 2.08% | -0.19% |
| Apr-15 | 3.80% | 3.05% | -0.76% |
| Mar-15 | 1.30% | 1.48% | 0.18% |
| Feb-15 | 0.34% | 0.41% | 0.07% |
| Jan-15 | 0.09% | 0.16% | 0.06% |
| Dec-14 | 1.25% | 1.53% | 0.28% |
| Nov-14 | -0.27% | 0.01% | 0.28% |
| Oct-14 | 2.13% | 2.38% | 0.26% |
| Sep-14 | 0.44% | 0.55% | 0.12% |
| Aug-14 | 0.27% | 0.19% | -0.08% |
| Jul-14 | 0.23% | 0.11% | -0.12% |
| Jun-14 | -0.77% | -0.69% | 0.08% |
| May-14 | -0.74% | -0.22% | 0.52% |
| Apr-14 | -0.03% | 0.35% | 0.37% |
| Mar-14 | -0.34% | 0.43% | 0.77% |
| Feb-14 | 0.76% | 1.19% | 0.43% |
| Jan-14 | 0.11% | 1.21% | 1.10% |
Total Residential Construction Spending vs. Previous Reports
| Date | Total Residential Construction Spending | ||
| Previous M/M Increase | Revised M/M Increase | Difference | |
| Oct-15 | 0.98% | 0.22% | -0.77% |
| Sep-15 | 1.58% | 1.14% | -0.44% |
| Aug-15 | 1.73% | 1.06% | -0.67% |
| Jul-15 | 1.60% | 0.33% | -1.28% |
| Jun-15 | 0.74% | 0.66% | -0.08% |
| May-15 | 1.71% | 1.26% | -0.45% |
| Apr-15 | 2.60% | 0.79% | -1.81% |
| Mar-15 | -0.76% | -0.07% | 0.69% |
| Feb-15 | 0.50% | 0.65% | 0.16% |
| Jan-15 | 1.25% | 1.29% | 0.04% |
| Dec-14 | 2.22% | 2.85% | 0.63% |
| Nov-14 | 1.52% | 2.09% | 0.58% |
| Oct-14 | 1.38% | 2.12% | 0.74% |
| Sep-14 | 2.24% | 2.40% | 0.16% |
| Aug-14 | 0.15% | -0.05% | -0.19% |
| Jul-14 | 0.03% | -0.28% | -0.30% |
| Jun-14 | -1.13% | -0.88% | 0.24% |
| May-14 | -2.20% | -0.72% | 1.48% |
| Apr-14 | -0.46% | 0.56% | 1.02% |
| Mar-14 | -1.11% | 0.98% | 2.09% |
| Feb-14 | -0.76% | 0.45% | 1.21% |
| Jan-14 | -1.25% | 1.69% | 2.94% |
4th Quarter GDPNow Forecast
The GDPNow forecast for 4th quarter is now down to 0.7% from 1.3% on December 23.
2015 GDP Will Decline vs. Previous Estimates: How Much?
The question is not whether 2015 GDP will rise vs. previous estimates, but rather by how much it will sink.
Let's start with the GDPNow forecast.
Of the decline since December 23, 0.5 percentage points came following the Census Bureau's release on construction spending and the ISM report, both on January 4.
I do not know how to separate ISM from construction, but the net result was as follows:
- Nonresidential structures declined by 0.10
- Residential declined by 0.14
- PCE declined by 0.13.
That's a total of decline for those three components of 0.37 percentage points.
Part of that decline was based not only on revisions, but also on a month-over month decline in November construction spending of 0.4 percentage points.
If 3/4 of the decline in those components is due to construction spending, then I estimate third quarter GDP will be revised about .57 percentage points lower, second quarter .56 percentage points lower, and first quarter .21 percentage points higher.
Those are very crude calculations that may be wildly off the mark.
If accurate, first quarter GDP would be 0.0%, second quarter GDP 3.3%, third quarter GDP would be 1.4%. And if the Atlanta Fed model holds with no changes from here, fourth quarter GDP would be 0.7%.
In that case, 2015 GDP would be about 1.35% with the Fed hiking and GDP decelerating rapidly.
