Recently I’ve found myself interested in such a new option as investing in Case-Shiller 10 Index via leveraged trusts. Clearly, prior to taking any action of investment nature sane prospective investor certainly has to perform at least some sort of analysis on the matter. Not claiming ultimate truth, objectivity or significance, I’ll try to present my findings, hoping someone find them useful.
What is Case-Shiller index and how do I invest in it?
In this entry by “Case-Shiller index” I mean S&P/Case-Shiller Compisite-10 Home Price Index (not seasonally adjusted), designed to track changes in the value of the residential real estate market in 10 metropolitan regions across the United States. This index uses the repeat sales pricing technique to measure housing markets. The S&P/Case-Shiller Home Price Indices are calculated monthly and published with a two month lag.
Just to picture it:
Since June 30th 09 one can invest in national residential real estate market by purchasing shares of MacroMarkets’ UMM trust or go short with the real estate market via DMM trust. The whole idea is quite simple and is fairly visualized on the company’s website (www.macromarkets.com/index.shtml). In short, buying a UMM share at a given price you purchase a conditional right for a relevant portion of funds of the DMM trust and conditional obligation to the DMM trust – if for the date of final distribution (late Fall 2014 for shares circulating by now) Case-Shiller index doesn’t exceed/exceeds the value of index, that corresponds the price you paid, you’ll lose/gain an amount derived from the index value divergence. Obviously, in case you buy shares of DMM you will be entitled to a portion of UMM funds, if the value of Case-Shiller index published by the final distribution date is below the one supposed by the price you paid, and vice versa.
Relation between trusts shares prices and Case-Shiller index is shown in the table:
If CS-10 exceeds 216.23 or goes below 108.11 prior to the final distribution date, trusts will be terminated and either UMM or DMM investors will receive all the funds raised by both trusts. As proceeds from shares placing are held in form of T-securities, apart from money from final distribution investor may also be eligible for interest payment, provided the amount T-securities generate exceeds trusts management’s expenses. The amount of interest investor is eligible for depends not only on the settled interest rates for T-securities, but also on preceding changes in CS-10 index. For the sake of simplicity, I guess, it’s reasonable to consider interest payments from T’s always equal management’s expenses.
Why trade Case-Shiller via UMM/DMM?
> With average total trade volume for UMM and DMM of 30k a day market for these securities might be quite inefficient as of yet. Less then 2 months (see graph below) ago UMM/DMM shares were priced at the levels, that suppose, that real estate markets in 10 largest metropolitan areas were to experience another major prices decline just to return to something around -10% from current levels by late summer 2014 – clearly, it was not the likeliest scenario, but as trade volumes haven’t grown significantly ever since it’s feasible, that market will repeatedly deviate from common sense, thus, offering opportunities for promising bargains. In the same time, unlike most of stocks with comparable daily trade volume and volatility, residential real estate market is well covered in media and there can be found a plenty of relevant data for research.
> Such investment may be used as a sort of insurance against decline/increase in real estate prices.
First glance at CS-10 and particular real estate markets.
As otherwise CS-10 index would have been less adequate, metropolitan areas it features are given different weights, according to the sizes of their real estate markets. As a result, the index appears somewhat skewed towards the largest metro areas – almost half of the weight falls on NY and LA, making analysis of these two local markets crucially important for forecast of the index future value.
Although some similarity in price trends does persist among the areas, graph below suggests degrees of development of these trends differ significantly from area to area.
Inflation-adjusted CS indices:
Based on this graph, I guess it would be appropriate if these 10 areas were split in two groups: those with moderate bubble (peaking at +60% to +80% since 1987 – namely Boston, New York, Denver and Chicago – 47% of the index weight) and those with serious bubble (peaking at +120% to +180% since 1987). Some questions arise: what was that something, that pushed prices for the second group that high, but didn’t spread on other areas to the same extent? Of course the thesis about effect of lowered mortgage requirements can be recalled, but, as long as requirements were eased all around the nation, it doesn’t explain the difference observed between the areas.
One may assume, that the difference in rate of price increase was caused by difference in pace of socioeconomic parameters improvement. Grounds for such assumption are obvious: the faster number of employed and/or per capita income increase – the more money is absorbed by the area to fuel additional demand for the real estate. It this case, once again, we need to learn, whether there were some specific patterns of socioeconomic improvement among areas of the second group not observed among other areas.
First, let’s look at such an important indicator as unemployment rate:
Although Washington D.C. metropolitan area shows some remarkable dynamics, I guess, no sound conclusion is apparent in case of this indicator, as long as there appears no essential divergence between groups’ averages.
Another way to gauge employment dynamics is that by using relative change in number of employed. Grounds: even with growing/stable unemployment rate and stable per capita income local real estate market still may experience demand increase, if absolute number of employed and, thus, aggregate metropolitan income increase.
And yet again, despite remarkable dynamics on some markets, particularly on that of Las Vegas, Denver, Miami, San Diego and D.C., divergence between groups’ averages is not significant enough to draw a feasible conclusion.
Next socioeconomic indicator is average income growth (measured as nominal per capita gross metropolitan product growth since 2001; for years 2007 and 2008 nation average GDP growth rate was used). Here we can see some quite intriguing trends:
As seen, not only average rate of per capita GDP growth was higher for the second group, but also no area of the first group ever exceeds any area of the second one, no exemptions. Preliminary conclusion seems obvious: personal income growth was a necessary, if not determining, prerequisite for bubble inflation in a given large metropolitan area, whilst unemployment dynamics didn’t affect demand significantly.
As per capita GDP data has shown such interesting results, it might be worth comparing it with paces of bubble inflation. The graph below shows cumulated excess of nominal per capita metro GDP growth over local CS index growth (for years 2007 and 2008 nation average GDP growth rate was used), i.e. if above 100 – per capita GDP was growing faster than local real estate prices, and vice versa.
Although, most of areas have already returned to levels of 2001, the major ones – Los Angeles and New York – are still around 20% below the 2001st’s income-price ratio. Are they necessarily going to 2001 levels? It depends, clearly, on whether U.S. economic system is to restore pre-bubble regime for real estate market or bubble levels will be considered to remain a normal by market participants. In case, when future state of nature largely depends on hardly predictable actions of a relatively small group of people (fiscal and monetary policy makers in regard to bubble prevention, names having essential power over investors’ views in regard to definition of “normal”, etc.), no strict deterministic approach can be employed, i.e. there’s no data that allows you to predict with high accuracy, e.g., when, for how long and to what extent lending standards will be tightened and how market’s common sense will develop. I doubt, that in such case there is an approach leading to a forecast more accurate than plain comparison with previous bubble developments would suppose.
Japanese assets bubble and dot-com bubble usually cross investors’ minds as the most relevant recent examples real estate bubble may be compared to. These examples share some important peculiarities with recent U.S. real estate bubble inflation and burst: they largely affected the whole national economy and their developments involved direct participation of broad ranges of investors, thus, drawing media’s active attention. The bottom line of the comparison is that investors do not tend to re-inflate massive bubbles – the ones market considers to be the prime cause of economic downturn or slowdown – once they burst.
At this point my conclusion would be:
There is no reason to expect a rapid uptick in real estate prices towards bubble levels, unless slow recovery is replaced by economic boom, accompanied by fast paces of income growth, unemployment rate decline and credit standards conservation. Considering poor effects fiscal and monetary stimulus programs have had so far in the context of consumer spending acceleration and concerns about credit standards in regard to assets rapid inflation prevention monetary policy-makers seem to share, economic boom and demand for real estate soaring are doubtfully around the corner. Meanwhile, taking into consideration essentially slowed paces of economic contraction and relatively stable investors’ and consumers’ sentiment, another major decline in real estate prices doesn’t seem likely as well, thus, UMM/DMM shares do not appear interesting as long-term investments, unless they at least leave $20-30 price range.
Aspects to analyze later: plausible effects of delayed demand emergence and inventory disposal.