U.S. Treasuries are widely considered the benchmark of REIT performance. This frequent comparison of REITs to Treasuries can easily lead investors to the illusion of a much stronger causality effect by Treasuries on REITs than is warranted. There is evidence of solid directional correlation between Treasuries and REITs (see Pearson Coefficient work in my spreadsheet), but strong correlation does not indicate strong causation. Treasury performance certainly impacts REITs. However, using Treasuries as a central factor in predicting future REIT yields and prices will often result in unpleasant surprises. With Treasury yields in a historically low range, an analysis of the interaction between these two securities is crucial. My point should be established through the following four questions and answers:
(The modern REIT era began in the early 1990s, so my analysis is limited from 1990 to present. I used monthly data to focus on long-term trends. Data source: NAREIT, FRB St. Louis. Google Docs spreadsheet shows my work.)
Question 1: In periods of rising Treasury yields, should REIT stock performance fall because investors are exiting REITs for safer, government-backed alternatives?
I identified 16 periods between 1990 and present in which the 10-year Treasury ((CMT)) was in a growth trend. During 11 of those 16 periods, the REIT price index grew as well, and in the remaining five periods, the REIT price index fell (three of those five falling periods occurred from 1990-1994, while modern REITs were still in their infancy). One likely explanation for why REITs often grow when Treasury yields grow is that periods of rising Treasuries often mean a healthier, growing economy. REITs enjoyed increases in valuation as part of a more macro-economic movement.
(click images to enlarge)
Question 2: In periods of declining Treasury yields, should REIT stock performance improve, since investors will seek alternative securities with higher yields?
I identified 17 periods between 1990 and present in which the 10-year Treasury was in a declining trend. During 11 of those periods, the REIT price index grew, and in six periods, the REIT price index fell. These results were closer to expectations than the findings in question 1, but I still failed to make any conclusions. This data needs to be picked apart in more detail than a Seeking Alpha article allows. For example, in six consecutive periods of falling Treasury yields (Sept '96 - Nov '96; Apr '97 - Oct '98; May '00 - Mar '01; Apr '03 - Jun '03; and Jun '04 - Sep '04), I found that the change in the REIT price index was part of a larger REIT price movement trend. Meaning that leading up to the period of dropping Treasury yields, REITs were already moving up or down, continued in that direction during the period of dropping Treasury yields, and after the Treasury yields stopped declining, REITs continued that direction trend for a significant period of time.
Question 3: The September 2012 yield spread between the 10-year Treasury and the REIT index dividend yield was 1.63%, and the historical average yield spread is 1.03%. Should we expect a mean reversion soon?
The standard deviation for this spread is 1.27%, so the difference in yields is well within 1 standard deviation. Furthermore, very rarely is the average yield spread close to the 20-year 1.03% average. Although the spread appears to have been tightening up since 2010, historically, the spread can stay well above or below the mean for years at a time. This can be demonstrated through the following chart:
Question 4: What would be the effect on REIT prices if the 10-year Treasury yield surpassed the REIT Index yield?
Since Treasuries are in effect risk-free, representing only a combination of liquidity and inflation risk, it would seem logical that REIT investors would always demand a premium for the risk. However, since 1990, there have been 8 periods where the 10-year Treasury monthly average yield has exceeded the REIT Index monthly average yield.
These periods serve as a reminder that REITs are first stocks, and only second income producing vehicles. Each of these eight periods of negative yield spreads had a unique duration, ranging from days to 22 months, and there were six or seven different REIT price reactions:
- April '90: REIT Index fell after April '90, but this was in the middle of a severe REIT price decline that had started in August '89 and only ended 6 months later.
- April - July '91: The result was REIT price volatility with no clear direction: April +, May +, June -, July +, August -, September +, October -, November -.
- March '93: REIT prices increased from October '92, peaking in March '93. After the 10-year surpassed the REIT yield in March '93, REIT prices slid from April to July '93. This is what should happen, theoretically.
- April - December '94: April to December was a period in which Treasury and REIT yields were very similar, averaging 7.40% and 7.42%, respectively. During five of those months, Treasuries had higher yields, and during four of those months, REITs had higher yields. During this nine-month period, the REIT index dropped slightly from 233.38 to 218.55, and then declined until April '95 to 211.97, before proceeding on a nice rally.
- December '96 - March '98: This was a 16-month period where bond yields were higher than REIT prices each month. During this period, REITs increased from 294.24 to 327.61. REITs then dove between April and August '98, falling to 257.43.
- January '06 - October '07: This represented 22 straight months where the 10-year Treasury yield was higher than the REIT index yield. Twelve months into this period, the REIT index reached an all-time high before beginning to decline. The drop from this peak led straight into the Great Recession.
- April '10: REIT Index increased from January to April '10, and then slid for the next two months until June '10. This is similar to what happened in March '93.
- February '11 and April '11: REIT Index had a three-month rally into February '11, before subsequently dropping in March, only to again rally in April. In May, the REIT Index climbed to a 33-month high.
Some of the price growth in REITs over the past 13 months absolutely can be attributed to investors seeking steady and reliable dividend income at a higher rate than what is obtainable through Treasury yields. But Treasuries, like REITs, reflect the economic environment in which they occur. We are in an uncertain, low-growth environment for the foreseeable future, and both REITs and Treasuries reflect that reality.
If the yield spread tightens because, for example, inflation expectations rise, I do not believe that would cause a blow to REIT prices. As demonstrated in Question 1, REIT prices often appreciate in periods of Treasury growth.
When REIT dividend yields are significantly above or below the 10-year Treasury yield, it is rarely a clear suggestion of cheap or expensive valuations. Investors should try to figure out the reasons for the discrepancy; for example, perhaps Treasuries are priced to reflect unusual conditions, like QE3, or that REIT dividends are unsustainable. I am long REITs, but it's hardly because I believe Treasury yields will remain ultra-low for the next 12 months.