Thank you to all my readers for making this construction economics blog worthwhile. Here’s ten of my most visited articles in 2016.
Just a few thoughts on forecasting construction spending
If we look at actual data through September, with ~75% of the actual annual construction spending already in the year-to-date and only 25% of annual spending in the estimate to complete in the 4th quarter, if forecasts for year-end totals are not already within 2.5% of the final, that means the projections for the final quarter are off by more than 10%.
Cash flows from construction starts are the strongest predictor of total spending next month. Projects that already started are in backlog. Backlog is used to develop projections for spending. For next month we already have 96% of all work in backlog. The remaining 4% will be new starts next month. For two months out we have 92% and for three months out we have 88% already in backlog. So, the uncertainty on spending in the near term is pretty low. Even averaged over a quarter rather than being exact every month, it takes a major event to have a big miss on projections for the next quarter.
Although exact monthly values in spending may be difficult to predict, trends for monthly changes in spending are easier to predict. Often these trends have very small changes month to month. Monthly changes in spending of +/- 5% are uncommon and changes of +/- 10% are quite rare.
Some forecasts I’ve seen for construction spending would demand that month to month changes in the final quarter exceed 20% to 30%. It’s just not going to happen folks.
For all of 2014-2015-2016
- Total Construction jobs increased from 5,950,000 to 6,700,000, +12.6%
- Total Construction spending increased from $960bil to $1,170bil, +22%
BUT, much of that spending increase is inflation. We need to compare to constant dollars which = Volume. Inflation was particularly high in residential work, over 15% for the 3 years, which means $1.00bil in spending 3 years ago would be worth more than $1.15bil today. Converting everything to constant Oct 2016 dollars, after inflation we get:
- Total Construction volume increased from $1,070bil to $1,170bil, +9.4%
So, for the 3 years, jobs increased by 12.6% while real work volume increased by 9.4%.
A more thorough analysis, which takes hours worked into consideration, shows from the Jan 2011 bottom of the recession in construction to current, both jobs and volume have increased equally by 28%. But jobs growth is often out-of-balance with real volume growth. In the beginning of the recession years of 2008-2011 firms let go of people much faster than work volume declined. 2009 showed a big gain in productivity. By 2010-2011 firms hadn’t let go enough to match the loss in spending. Then from 2012-2014 workload grew faster than firms filled jobs. Since 2011 we are sort of on an even keel.
For the last 3 years, jobs increased more than real construction volume. I pointed this out in my last detailed jobs report. The 6 months from Oct 2015 – March 2016 encompassed the fastest new construction jobs growth period in a decade. It’s no surprise to me that jobs growth has been slow since March this year. Frankly, I wouldn’t be the least bit surprised if it remains slow for awhile. When we look at jobs growth vs. volume growth, there is reason to believe that slow jobs growth is not entirely due to labor shortages. Part of the blame is due to recent over-hiring.
Just a few important facts here (that you won’t read in the headlines).
- Jobs have increased by 23% since the recession bottom January 2011. Construction spending has increased 52% in same period.
- Adjusting jobs for hours worked and spending for inflation, both work output and constant volume construction grew at 28% since recession bottom.
- Residential construction jobs are down 22% from the 2006 peak. In constant $ after inflation, real residential volume of work is down 39%.
- Nonresidential buildings construction jobs are down 8% from the 2008 peak. In constant $ after inflation, real nonres bldgs volume of work is down 17%.
- Construction spending average for the last 3 months is at a 10 year high. Construction volume in constant $ after inflation is still 18% below 10 years ago.
- Census construction data is ALWAYS revised in the following two months after initial release. Census updates all the values for the previous year, usually with the May data release (on July 1) the following year.
- In 2016, 7 times the first release of spending showed a decline vs the previous month. After revisions, the values show no declines vs the previous month.
- In the last 36 months, there were 16 spending releases that originally showed a decline vs the previous month. After revisions there were no mo/mo declines.
- It’s hard to add $100 billion in new construction spending in one year. Since 1993 it’s happened only 3 times; 2004, 2005 and 2015.
- It’s real damn hard to add $100 billion in new construction volume in a year. After adjusting for inflation, construction volume has never increased by $100 billion. It has increased by $75 billion 4 times and 3 more times by $50 billion.
- It takes about 6,000 jobs to put-in-place $1 billion of construction in a year. $100 billion in new work would require 600,000 new jobs in a year. The largest construction jobs growth ever recorded is about 700,000 jobs in 2 years.
- Now think of #9, #10 and #11 in terms of ONLY the Infrastructure sector. Infrastructure, about 25% of total construction spending, added spending more than $25 billion in a single year only once. The average annual growth for the past 20 years is less than $10 billion/year. The average growth in jobs (excluding all recessionary years because they would make the result approach zero) is about 25,000/year.
- The Aug-Sep-Oct 3mo average of construction starts for Nonresidential Bldgs (by Dodge Data) is the best 3mo since Q1 2008. Q1 2008 was the PEAK of the nonresidential buildings construction boom.
- More infrastucture projects started construction in the 1st 6mo of 2015 than any time in history. This will boost infrastructure spending through 2017. Infrastructure spending is low in 2016 due to a low volume of starts in 2014.
October construction spending put-in-place was released today by U.S. Census. This report includes the first revision to September data and the 2nd revision to August.
October spending 1st release came in at $1.172 billion, 0.5% higher than September which was revised up 1.4% to $1.166 billion from the 1st release of $1.150. August was revised up 2.1% to $1.166 billion from the 1st release of $1.142.
I predicted October spending would come in at $1.190 billion. Once revisions to October data are posted in Nov and Dec, we may reach that $1.190 billion forecast. Revisions have averaged over 1.4%/mo this year and 1.5%/mo for the last 4 months.
Average spending for the last 3 months is $1.169 billion, the highest three-month average since May-Jun-Jul 2006.
Year-to-date (YTD) spending is up 4.5% over last year, but this may go even higher once the revisions are in. There is now no doubt that we’ve clearly passed a previously forecast dip in spending that bottomed in Apr-May at $1.142 billion. The last 5 months of spending are all up from the low point and the trend is pointing higher.
My forecast for total spending in 2016 is $1,168 billion, up 5% from 2015. I expect 7.6% growth in 2017.
Residential spending is up 5.5% ytd and is on track to reach a 2016 total of $468 billion, +6.4% over 2015. Last year, peak spending was in September, then residential spending dropped slightly in Q4 2015. This year I expect 2016 spending to peak in Q4, so we should see ytd performance get better as we approach year end. Cash flow from new starts indicates growth of 9% in 2017 spending.
Total Nonresidential spending is up 3.8% ytd, on track to finish 2016 with total spending at $700 bil, up 4.2% over 2015. Almost all the 2016 growth is in nonresidential buildings, not infrastructure. For the 4mo period Jul-Oct 2016, compared to the same 4mo in 2015, all nonresidential spending is up only 1.7%, but the spending trends are not apparent unless we separate nonresidential buildings from non-building infrastructure. For Jul-Oct 2016, compared to the same period a year ago, nonresidential buildings spending is up 7.6% and non-building infrastructure is down 5.4%.
Nonresidential Buildings spending is up 8.2% ytd through October, led by Office, Lodging and Commercial Retail markets. We should finish 2016 up 8.1% with a total at $409 billion vs. $379 billion in 2015. Total sector growth for the last three years is 35%. I’m predicting 2017 spending for Nonresidential Buildings will increase 7.5%, led by Educational and Office spending.
We are currently at what may be 2016 peak nonresidential buildings spending. I’m expecting nonresidential buildings spending to stall or drop 1.5% to 2% over the next few months before resuming growth. This drop may be in large part due to uneven starts from the end of 2014 and beginning of 2015, a period when starts were abnormally high, that are now finishing and dropping out of the monthly spending values. Usual normal growth patterns in starts do not fill the void left when abnormally high volume of projects finish.
Non-building Infrastructure spending is down 1.2% ytd. Infrastructure spending in 2016 will total $291 billion, down less than 1% from 2015. Spending predicted from Dodge Data Starts predicted this drop. Negative drivers are Transportation, contributing -0.9% to overall decline, Sewage/Waste Disposal -1.0% and Water Supply -0.4%. Power, the largest infrastructure market at 33% of total, is up 1.4% ytd so adds about +0.5% to offset some of the declines. Highway/Street, 31% of infrastructure, is up only slightly. Growth resumes in Q1 2017. Although new starts in 2016 will finish down 10%, starts in 2015 were so high that 2016 will still be a good volume of new starts. Predicted spending from starts is indicating 2017 will be a record year for spending on infrastructure, up 7% from 2016.
Public spending is down 1.5% ytd, on track to finish 2016 with total spending at $285 billion, down 1.4% from 2015. Public spending will rebound in 2017, up 6.5%.
Educational spending is 80% public and 20% private. Education accounts for 25% of public work. Educational is by far the largest building type in public work. All the remaining building types contribute only 2% to 4% each.
60% of all public work is infrastructure. Highway/Street accounts for 31% of all public work. Transportation facilities is 11% of public work, Sewage and Waste Water 9% and Water 4.5%.
The biggest drivers of performance in public markets by far are Highway/Street and Educational spending. Highway/Street spending reached all-time highs from Dec 2015 to March 2016 but is currently 10% below that level and will end 2016 down 1% from 2015. In public markets educational is only up 5% ytd, but in October experienced the largest monthly increase in the public sector.
REVISIONS AND YEAR/YEAR COMPARISONS
Census construction data is always revised in the following two months after initial release. Census revises data and incorporates more data from additional sources to update spending values. Census updates all the values for the previous year, usually with the May data release (on July 1) the following year.
For the 1st nine months of 2016, seven of nine times the first release of spending showed a decline vs the previous month. After revisions, the values show no declines vs the previous month. The last 36 months of data shows there were 16 Census releases that originally showed a decline vs the previous month. After revisions there were no mo/mo declines in the last 36 months. Revisions in 2016 have averaged 1.4%/mo and 1.5%/mo for the last 4 months.
In 2015, spending peaked in the months of July, August and September, then dropped slightly and remained flat for the last quarter of the year. This 2015 pattern, along with the issue of revisions noted above, is one of the reasons comparisons of 2016 to same month last year was low for August and September. A growth trend is now in place. Expect this month vs same month last year for the remainder of 2016 to come in near or above +5%.
10-20-16 Starts Point to Robust 2017 Spending