The Delayed Recovery of Investment in Nonresidential Structures
While real GDP has long passed its pre-recession peak, business fixed investment is still 4 percent below its previous high. This is mainly due to the delayed recovery of one of its components, investment in nonresidential structures (factories, plants, office buildings, stores, hospitals etc.). This investment category dropped by 35 percent in the years 2008-2009 and didn't begin to recover until mid-2011, two years after the recession ended. Since then, it has been growing fast, but it is still 23 percent below its peak. In contrast, investment in equipment and software, the other component of business fixed investment, dropped by 20 percent during the recession, began to pick up right when the recovery started, rapidly bounced back, and is now 4.8 percent above its previous peak.
Investment activity across industries followed a similar pattern. Investment in equipment and software tended to reach bottom in 2009, the year the recession ended, while investment in structures tended to remain depressed throughout 2011 (the most recent year for which industry data are currently available). This was true both for industries that performed relatively well during the business cycle, like information and health care, and for industries that were hit harder by the recession, like manufacturing (See The Recession and Recovery from an Industry Perspective).
One reason why investment in structures recovered later was that it was held down by the overhang of structures that had been built before the recession. Structures are very long-lived productive assets, with an average age of approximately 24 years, so investment in these assets crucially depends on forecasts of long-term growth. Forecasts of long-term growth were revised down around the beginning of the Great Recession (see Behind the Slowdown of Potential GDP). Suddenly, the stock of structures that had been built based on pre-recession forecasts became excessive, and firms had to reduce their investment activity, absorb the overhang, and bring the stock back in line with the new forecasts. This process took especially long because these assets last so long. Equipment and software, in contrast, are shorter-lived assets, with an average age of approximately 7 years, so the overhang of equipment and software was smaller and quicker to absorb.
Data on investment activity by type confirm our previous observations - Investment in equipment and software tended to behave more in sync with economic activity, dropping during the recession and bouncing back during the recovery, while investment in structures tended to lag. Within each category of investment, however, different types behaved differently. In 2011, investments in industrial and transportation equipment were on their way to recovery, but still well below their peaks. In contrast, investment in information processing equipment and software didn't decline much during the recession, and in 2011 it was already above its previous peak, due in part to its stronger underlying trend growth. Investment in structures tended to decline later, as investing in these long-lived assets is planned more in advance and is more difficult to reverse. In 2011, investments in most types of structures were still below their peaks. Investment in commercial structures, which include office buildings and multi-merchandise shopping structures, was especially depressed, 50 percent less than its 2007 level.