THE QUESTION OF HOW RISING TREASURY YIELDS WILL IMPACT CAP RATES has been a major topic of discussion over the past six months. Although many investors are concerned by the increase in Treasury yields, it appears that cap rates can accommodate the recent rise through a tightening of real estate risk premiums, which previously had been well above their historical average. However, an additional 50 to 75 basis point expansion in Treasury yields will begin to put pressure on cap rates (assuming that debt availability and NOI growth projections remain constant). Above that level, cap rates will increase but not on a 1:1 basis due to compressing risk premiums. Furthermore, individual markets and property types will react differently. Markets which attracted substantial investment due to their perception of safety are at much greater risk for losses in value. The risk profile of core assets has increased significantly because low-yielding assets will need more excess NOI growth to compensate for higher cap rates. In this environment, the new safe havens will be higheryielding markets and property types that have favorable conditions for NOI growth. Cap rate spreads to Treasuries have compressed significantly but are not at atypical levels. From April to October 2013, the all-property cap rate to Treasury yield spread compressed from 505 bps to 416 bps. With the exception of early 2011, this is the lowest spread since the recession ended. However, it is still within levels seen from late 2009 to early 2011 before Treasury yields compressed significantly. The spread is also in line with late-2003 levels, indicating that it is within its natural historical range. The historical trend is that as Treasury yields rise, the spread between those yields and cap rates narrows, as was the case from 2003 to 2007. At the same time that yields rose from 3.3% in 2003 to 5.1% in 2007, cap rates declined from 8.1% to 6.6%, decreasing the spread from 480 bps to 150 bps. This significant compression occurred as a result of an improving economy, which led to more optimistic projections about future revenue growth while at the same time spurring the Federal Reserve to raise interest rates. From 2003 to 2006, GDP grew between 2.5% and 3.5% annually, CPI increased between 2.9% and 3.3% (above the 2.0% target) and most importantly, revenue (measured by market occupancy x rent) increased 4.6%, 8.7% and 10.5% in 2005, 2006 and 2007, respectively. Most investors expected such growth to continue and when it did not, a correction ensued. Given the current lackluster economy, spreads are unlikely to compress as much as they did from 2005 to 2007. The Gordon Dividend Growth Model can be used to illustrate the relationship between higher interest rates, NOI growth and valuations. The equation is as follows:
where P equals the valuation; D1 equals the current dividend; r equals the discount rate (or required return); and g equals the rate of growth This model can be extended to real estate where the valuation = NOI / cap rate. By definition, a cap rate should equal r g, the required return minus the growth rate. For cap rate compression to occur, the change in g must be greater than the change in r. Let s begin with a look at g, the NOI growth rate. Between 2003 and 2007, investors were likely using excessive assumptions for g (given the growth rates experienced during the period from 2005 to 2007). As the expected return component of r increased due to rising Treasury yields, it was outweighed by a larger g. Whether g will increase in the current cycle depends entirely upon whether investors begin to adjust their assumptions for stronger future NOI growth. If NOI surprises to the upside, g could begin to increase. However, the pace of recovery is slow and does not seem to be accelerating, as annual GDP growth over the past two quarters has been the weakest since the economy began expanding in the second half of 2009. As shown in the chart at right, nominal GDP and office NOI are correlated with a lag. Y/y nominal GDP growth is not only growing at an anemic 3.1% rate but has also decelerated over the past year with negative implications for NOI growth. Because of this, an imminently increasing g seems unlikely. Additionally, because Treasury yields have increased over such a short time period, simply not enough has changed in the economy for g to increase significantly. Now let s consider r, the discount rate or required return. The required return is expected to increase as a result of higher real Treasury yields, though it will have a different impact for different types of assets (discussed in the next section). Because the achievable risk-free return is now higher, it would seem logical that yields for other assets should increase as well. As mentioned previously, the rise will not be 1:1; many investors are incentivized to target certain absolute return hurdles regardless of the Treasury yield. The lack of a 1:1 rise in general is observed by the compressing BBB corporate bond to Treasury spread, which fell from 152 bps on May 1st to 126 bps on October 29th of this year. Between 2003 and 2007, cap rates continued to compress despite rising Treasury yields. As noted above, this compression was likely due to an improved growth outlook. In addition, aggressive debt markets may have also dampened a potential rise in r because easily accessible capital could have generated higher leveraged returns to compensate for pressure on unleveraged asset values. In the current cycle, it is possible that debt markets will continue to open up as the deleveraging process plays out, which would have positive implications for real estate valuations.
Overall, it all comes back to whether the change in g will outweigh the change in r. We do not believe this will happen. As fundamentals currently stand, the dramatic compression in spreads during 2006 and 2007 is unlikely to be repeated due to weaker NOI growth and less relative debt availability (see chart at right). Therefore, even though there is some cushion in the spreads, there is not enough to take them to 2006 levels. Certain markets and property types are more exposed to rising interest rates. The impact on r (the discount rate, or the required rate of return) from rising Treasury yields is not necessarily equal among markets and property types. Instead, the types of assets which are priced relative to bonds are more likely to be impacted by rising rates. The required return on these assets is expected to more closely mirror the increase in Treasury yields; by contrast, for riskier assets from which investors require certain return hurdles regardless of the yield level, higher interest rates may have only a negligible effect. The narrowing spread is most pronounced in markets that have been perceived as perpetually safe. Manhattan s current office cap rate to Treasury spread is 168 bps, down from 422 bps in early-2012 and equal to early-2006 levels for that market. Washington, D.C. s office cap rate spread has fallen to 267 bps, down from 428 bps in mid-2012 and equal to early-2005 levels. These spreads are much more compressed than they are for real estate in general, where spreads are at mid-2003 levels of approximately 400 bps. assets do not have a lot of capacity for spreads to compress further; therefore these cap rates will be the first to increase should rates continue to rise. For the non-core office, industrial, retail and apartment property types, spreads remain at healthy late-2003 levels and for this reason there still appears to be room for cap rates to absorb another 50 bp increase in Treasury yields. Even for gateway apartments, the spread remains at late-2003/early-2004 levels. Real vs. nominal interest rate increases will have different effects. It is possible that these spreads have compressed for a reason. Perhaps NOI growth projections are being revised upwards for these markets. However, at least in Washington, D.C. s case, market NOI actually decreased in Q2-2013 so it is unlikely that this theory holds (unless the market was initially pricing in a much larger decrease in NOI). In summary, gateway office
can absorb more percentage points in cap rate increases than those with lower yields. The break-even point across going-in cap rates is depicted in the chart at top right. Looking at the same situation from a different perspective, assets with lower going-in yields require more excess NOI growth in order to equalize IRRs compared to a scenario in which cap rates do not rise. Indicated by the chart at bottom right of the last page, if cap rates increase by 100 basis points then an asset with a going-in cap rate of 4.5% needs 384 basis points of The recent spike in bond yields has been attributable to a rise in real yields as inflation has remained benign. Going forward, if inflation accelerates and yields rise as a result, spreads will be set to further compress because replacement costs and rents tend to increase with inflation (g will increase). In this case, property types with shorter leases (hotels and apartments) will outperform the other property types because their NOIs will be able to quickly adjust to rising rents. However, a scenario where real yields continue to increase will have a more negative effect on real estate (as well as other assets like equities and bonds) because g will not compensate for much of the difference. Core is not low risk; going-in cap rates matter. additional annual NOI growth to achieve its underwritten return. In comparison, an asset with a 7.5% going-in cap rate only requires 224 basis points. Notably, the change in the excess NOI growth requirement trends down as yields increase; the difference between a 4.5% and a 5.0% yielding asset is greater than that between a 7.5% and an 8.0%. This is because a 50 basis point increase from 4.5% is a greater percentage than a 50 basis point increase from 7.5%. All else equal, higher-yielding assets are better positioned to accommodate a rise in cap rates. This is because the same NOI growth rate will produce a greater overall change in NOI for assets with higher yields. For example, assuming a five year hold and 3% NOI growth, an asset with a going-in yield of 4.50% will end its hold period at a yield-on-cost of 5.22%, good for a 72 basis point increase. In comparison, an asset with a going-in yield of 8.00% will end at a yield-on-cost of 9.27%, a 127 basis point increase. Simply put, higher-yielding assets In an efficient market scenario, it could be argued that loweryielding property types and markets will exhibit higher NOI growth due to barriers to supply. However, stronger rent growth and low cap rates can actually result in higher supply by making more projects viable. In the current cycle, high-barrier office markets like San Francisco, New York and Washington, D.C. are actually witnessing healthier construction pipelines as a percent of inventory compared to lower-barrier markets with similar vacancy rates like Denver, San Antonio, Raleigh and
Charlotte (see chart below). Over the long run, these trends may work themselves out but it should not be automatically assumed that a lower-yielding asset translates to short-term safety. Finally, the scenario below depicts how much excess NOI growth would be required to compensate for rising interest rates at various correlations between interest rates and cap rates. The chart illustrates that a modest rise in interest rates could be easily absorbed by slightly above-average NOI growth. Conclusion Changes to debt availability and/or the NOI growth outlook can alter the effect of rising yields The recent rise in bond yields can be absorbed by general real estate cap rates although core gateway office may experience a slight correction If real yields continue to rise (assuming debt availability and the NOI growth outlook remain constant): Cap rates will also increase but not on a 1:1 basis (due to tightening risk premiums) Core office properties in gateway markets will experience a more significant correction Non-core, higher-yielding properties are best positioned to absorb rising rates, and thus will be the least affected If yields rise as a result of inflation: Cap rate increases, if any, would be muted Hotels and apartments will outperform other property types because of their shorter-term leases Modest rises in cap rates can be overcome by slightly above average NOI growth