All the market indicators point to clear sailing for commercial real estate investment. What could possibly go wrong? Hopefully nothing, but the next couple of years will be a good time to closely watch government regulatory agencies for actions that could trigger a downturn.

From a standard supply-and-demand perspective, the CRE market is in good shape.  An Urban Land Institute study in 4Q 2016 found that CRE leaders see moderate industry growth through the end of 2019 and beyond.  The economy is expected to continue a moderate growth pace, fueling space demand that will allow property values to rise by an average of 4 percent per year over the next three years.  The biggest challenge is in the area of rent growth, as vacancy rates are expected to increase in every property type except for industrial.  The study estimates that office vacancies estimated at 13.4 percent and retail at 10.3 percent by the end of 2019.

ULI’s views on vacancy rates, absorption levels and cap rates are shared by most real estate economists and investment market experts.   But there’s another data set that’s worth watching: the level of bank exposure to CRE, which has triggered sudden downturns in the past.

A March report from the General Accountability Office (GAO) focuses on increased CRE lending by banks over the past several years.  According to the report, the number of banks with intense concentrations of commercial real estate loans is large and getting larger. Hundreds of banks now hold CRE loans that add up to more than 300 percent of their total assets.  “Property prices have been increasing rapidly, and property valuations also have risen,” GAO reported.  “Risk has increased, based largely on the simultaneous increase in bank CRE lending and CRE prices.”

Should we be alarmed? Not yet.  GAO notes that the over CRE risk level is well below that of the latest peak in 2006-2007, so there appears to be no near-term danger of government action to force banks to back off.  But we should be alert– when government agencies have acted in the past, markets that appeared to be in equilibrium have crashed hard.

That’s exactly what happened in the 1990 commercial real estate crash that contributed to the 1990-1991 recession. CRE was definitely a bubble economy, as lender and investors with too much cash had pushed prices beyond their sensible limits. But what burst the bubble was a directive by the Office of the Controller of the Currency (OCC) that banks and insurance companies would have to dramatically increase their capital reserves on real estate loans, which had grown too large as a percentage of total assets.  The OCC directive took effect immediately, turning every institution into a portfolio liquidator on the same day, with no group of buyers to take their place. Crash!

There’s also a credible argument that government accounting contributed to the Global Financial Crisis, albeit not in a way that’s specifically related to real estate.  Former Federal Deposit Insurance Corporation Chair William Isaac argues that FAS 157, which mandated mark-to-market accounting, was partly responsible for the sub-prime mortgage collapse. Also known as fair value accounting, the mark-to-market system doesn’t reflect accurate asset values in times of illiquidity or volatility and may have fueled a vicious circle of repeated write-downs.

FAS 157 went into effect in November 2007, just before the market downturn, and the SEC agreed to relax the mark-to-market rules in April 2009, after which time the economy began its recovery. Isaac and other experts believe this correlation was more than a coincidence.  To the extent that they’re right, it illustrates the potential impact of government regulatory action on the economy.

Again, there’s nothing in the recent GAO report to suggest that action on bank lending is imminent.  However, it’s important to keep an eye on these reports and consider the possible impact they could have on the health of our industry.  History shows that we should be paying attention not only to market data, but also to agency interpretations of that data.