Diversified Local Economy... but Oil Still Plays a Large Role
Houston has experienced one of the strongest post-recession rebounds in private-sector employment, as shown in Figure 1. While strength in the oil industry has been responsible for a large part of the area’s growth, we note that the local economy is quite diverse. Houston has a particularly large medical establishment presence that includes the Texas Medical Center—the world’s largest medical complex. In addition, Houston’s port is quite active, with the Port of Houston handling about 70% of all the containerized cargo in the U.S. Gulf of Mexico; in 2012, it was the top-ranked U.S. port in terms of foreign tonnage.
(Click or tap images to enlarge)
Even though Houston has a relatively low number of office employees as a percentage of private sector workers (Figure 2) it still has a sizeable office market, particularly when compared to a city such as Dallas, which has 40% more office workers occupying only 10% more space (Figure 3).
Growth in the energy sector has made Houston a very active (and expanding) office market, with five of the top ten private sector employers in the oil and gas business (Figure 4).
Two of the four largest refineries in the U.S. are in Houston, and Houston is also home to 40 of the nation’s 145 publicly traded oil and gas exploration and production firms. Of the 24 Fortune 500 companies located in Houston, 20 of them are in the oil and gas business, and as shown in Figure 5, Houston is home to numerous others. In addition, while Texas produced almost 30% of the US’s natural gas output and more than 25% of the US total of crude oil in 2011, Houston-area oil and gas companies are responsible for an even greater share of output once international production is included; many international operations are still run out of a Houston headquarters.
Engineering services firms, particularly those that support the construction and maintenance of oil and gas pipelines, dominate the Harris County landscape. As shown in Figure 6, these firms comprise 40% of all oil and gas-related establishments, and roughly 20% of oil and gas-related employment.
Oil is Correlated with Office Employment…While Office Employment Is Correlated with Asking Rents
It should not be surprising, then, to see that private-sector employment has been tied closely to oil prices, as shown in Figure 7, suggesting a high degree of correlation between the two variables.
Similarly, office-using employment and asking prices on office rents appear linked (Figure 8), although perhaps not surprisingly, employment and rents show significantly less volatility than oil prices.
More generally, nominal oil prices and Houston-area rents have tended to move together (Figure 9), though the relationship has diverged on several occasions.
The following variables were tested to model Houston-area Class A and B office rents:
- Changes in lagged office rent growth
- Houston Metropolitan Statistical Area office-using employment
- West Texas Intermediate (WTI) spot oil prices
- Available square feet
- Net absorption
- Occupancy rate
- Total availability rate
- Total rentable building area (RBA)
- Amount of space under construction
- Total number of active rotary rigs in Texas and in the U.S. (quarterly average)
Both office rents and oil prices were deflated by the Consumer Price Index (CPI) in order to remove the general level of consumer price inflation from both series. For the sake of simplicity, the change in the change of office employment, real rents and occupancy rates are referred to as the “change in employment growth,” “change in rent growth,” and “change in occupancy growth,” respectively. The change in CPI-deflated oil prices is referred to as “change in oil” throughout the remainder of this paper.
Oil Prices Matter More Than Employment
Demand for office space is generally driven by employment.Given the long-term nature of leases and the costs associated with lease terminations, the quantity of space demanded is generally sticky. As a result, in an environment of unchanged employment there should be little net absorption of space (aside from moves driven to increase space efficiencies.) When changes in employment growth are regressed on changes in lagged oil prices, we find that oil price changes explain 20% of the variability in changes in Houston metro area employment growth. Knowing that lagged oil price changes help explain changes in employment growth, we test whether lagged oil price changes matter more than lagged changes in employment growth in determining changes in the future growth of rents.
Below, we show correlations between changes in rent growth and changes in employment growth (Figure 10) and correlations between changes in rent growth and changes inoil prices (Figure 11).
What Happens When We Try to Predict Changes in Rent Growth?
We find that changes in employment growth lagged from one to 12 quarters have no significant correlation with changes in rent growth across the sample time period (Figure 10) whereas changes in oil lead changes in rent growth by one quarter with a correlation of 0.27 (Figure 11). This finding is not altogether surprising; asking rents today should be correlated with anticipated changes in employment. Put another way, asking office prices are a function of future changes in demand, and future changes in employment are most correlated with recent changes in oil prices. From a modeling perspective, using changes in lagged oil prices versus changes in employment is advantageous given reporting delays; data on quarterly employment from the Bureau of Labor Statistics are not available until at least three weeks after the end of the quarter, whereas oil prices are published on a daily basis.
Even though there is a significant correlation between the current change in the growth of rents and the current change in the growth of employment, contemporaneous changes are useless in making future forecasts. As such, there is no lagged value of employment growth that is significant in forecasting future changes in rent growth.
Given oil’s importance as a predictive factor in Houston rents, a model to determine whether the active total rig count (in both Texas and the U.S.) provides any additional information over and above that provided by spot oil prices is tested. However, as Figure 12 suggests, the trajectory of rigs follows (and slightly lags) the path of oil prices, making
Explaining Rents via Lags of Oil, Net Absorption and Prior Rents
Using the complete data set from 1999, changes in rent growth are modeled using the variables previously mentioned on page 4; the most robust equation is presented in Figure13. (Appendix III shows cross-correlograms of rents against each of the variables tested.) The in-sample results show that 49% of the variation in changes in rent growth is explained by the four variables shown and lagged prior period errors (the AR term).
In order to test the above model’s ability to accurately forecast future rent trends, a re-estimation of the above equation is made using data from Q3 2001 through Q4 2011. The corresponding out-of sample forecasts (Figure 14) show a root mean square error (RMSE) of $0.22, just 1.2% of the actual average inflation-adjusted rents of $17.95 over the Q1 2012 - Q4 2013 period.
The model above highlights the sensitivity of future price changes to not only changes in the price of oil, but also changes in occupancy and construction trends. Note that changes in oil prices affect changes in rent growth only one quarter ahead; in contrast, the changes in the other variables affect rent growth changes 6-7 quarters ahead. Oil price changes affect rents only in the short term.
What Does It All Mean?
As of late, office space under construction in Houston has been climbing even as occupancy rates have risen from 85.3% seven quarters ago to 86.3% as of Q4 2013. A tremendous office construction boom has been underway: Houston currently has 35% more new office space under construction than Manhattan (Figure 15).
While a significant percentage of Houston’s new office construction is pre-leased, the roughly 25% of space that hasn’t been spoken for has the potential to slow or reverse the increase in occupancy rates in the near term, particularly if tenants trade a larger existing lease for a smaller, but more efficient, space in a new building. Similarly, given the negative coefficient on the “space under construction” variable in the regression equation,we can also surmise that the increase in construction activity will be a source of downward pressure on rents.
In the short term, we believe that local area rents are at or near a peak. The outlook for the next six months is for a modest increase in real rents on the order of $0.20/sf before we see rents returning to current price levels by the end of the year. Given the relatively low level of inflation at present, the forecasted rent change on a nominal basis is similar to a few pennies higher.
One caveat: new space tends to have higher prices than existing space; moreover, new space is more apt to have a listed asking price (versus existing space, where a significant percentage of asking prices are withheld from databases such as CoStar.) With “under construction” space in Houston at a record high as a fraction of total RBA, the higher prices associated with new space could offset some of the downward price pressure exerted by a rising occupancy rate and an increase in space under construction, potentially mitigating some of the price decline we forecast toward year end.
- Commercial rents in the Houston office market have reached an all-time high, with asking prices on combined Class A and B space up 12% since the end of 2010.
- Not surprisingly, the growth in the oil industry has spurred area employment to record levels, particularly in light of the recent boom in shale oil and gas, which has driven expansion in transportation and logistics areas as well.
- An analysis of the relationship among employment, crude oil prices and office rents finds that lagged changes in oil prices exert a more direct impact on short-term changes in Houston rent growth than do changes in office employment growth.
- A model of future rents based on lagged changes in rent growth, lagged changes in occupancy growth, lagged changes in oil prices and lagged changes in the degree of new space under construction suggests that rents should peak over the next one to two quarters before falling back to current levels.