2014Cambridge Conference Business & EconomicsISBN : 9780974211428
Bubbles, Wobbles and Financial Instability: The Jenga Market[1]
Roger W. Clark[2]
Austin Peay State University
P.O. Box 4415
Clarksville, TN 37044
United States of America
Phone: 931-221-7574
Email:
Yu Zhang
Assistant Professor of Finance
Eastern Michigan University
420 Owen
Ypsilanti, MI, USA 48197
Phone: 734.487.1350
George C. Philippatos
University of Tennessee
Professor of Banking and Finance (Emeritus)
Knoxville, TN 37996-0540
United States of America
Voice: 865-690-9684
Email:
Bubbles, Wobbles and Financial Instability: The Jenga Market
Roger W. Clark, Austin Peay State University
Yu Chang Eastern Michigan University
George Philippatos, University of Tennessee Knoxville
A bubble is defined as a self-reinforcing increase in asset values to levels not justified by the fundamentals of the economy, as measured by some generally accepted paradigm (e.g. CAPM in its various forms).Bubbles can occur in financial markets without investors ever behaving irrationally, once the limitations of their knowledge is accounted for (Cooper, 2008). Financial asset price bubbles have been documented by many researchers, most notably by the recent Nobel Prize winner, Robert Schiller[3].
Such findings question the efficient markets notion that rational speculators always stabilize prices. In fact, according to research “…They are consistent with models in which rational investors may prefer to ride bubbles because of predictable investor sentiment and limits to arbitrage (Brunnermeier and Nagel, 2004). When rational investors prefer to ride the bubble for long periods of time, the self-reinforcing feedback effects may convert the bubbles to a wobble of instability, with serious consequences for the financial sector, and perhaps the entire economy.
While the results or M.K. Brunnermeier and S. Nagel focused on “Hedge Funds and the Technology Bubble” there is some recent evidence about the “Synchronization risk” where rational arbitrageurs “time the market” rather than correct the mispricing right away. This leads to “delayed arbitrage” which prolongs the life of the asset bubble (Abreu and Brunnermeier, 2002)
This paper proposes that what we commonly define as a stock bubble is actually a series of small bubbles as seen from the individual corporate basis. It is only when companies that are mispriced are placed on a sector of the economy do we see a classic “asset bubble”. While it would be very difficult for one speculator to attack an entire sector of the economy that is mispriced as it would involve a great deal of synchronization, two or more well-endowed speculators could attack and cause a panic in individual stocks and make a profit with the ensuing panic. Then they go to another, picking off the low hanging fruit until the entire sector collapses. This is similar to the board game “Jenga” where each player takes one part of a structure out. The loser is the one that collapses the structure with his/her pick.
In the crash of ’08 the blame primarily went primarily to the real estate industry. This bubble lasted several years. It would seem inconceivable that speculators did not realize that there was a bubble in this area. We would expect then, that the afore-mentioned self-reinforcing feedback effects converted this bubble into a wobble. We would also expect speculators would have been busily raiding individual companies before the wobble burst. This paper investigates the possibility that speculators behave in this manner to deflate the wobble before it bursts.
In the seminal paper of Abreu and Brunnermeier a key assumption was that one rational trader cannot alone correct a mispricing in the market. This paper can take as a given that while it is true that one trader cannot correct broad mispricing in the markets, traders do not trade “the market”. They trade individual company securities. Further, the actions of George Soros and Quantum Capital in Thailand seem to verify that under the proper circumstances one large arbitrageur can indeed bring down an entire monetary system (Stiglitz, 2002).
General Methodology and Databases
Should it be the case that traders are selecting individual companies to raid before a bubble bursts we should see evidence of it in a group of companies that have strayed in mispricing from their fundamentals. To test for this we will be taking companies by SIC code in the areas of construction and home sales and testing their mispricing against their individual fundamentals before the bubble burst and slightly before the bursting of the bubble. We should see some companies that return to their fundamental price before others in the same industry as they are raided by individual wealthy speculators. This could be a useful tool for judging the imminent bursting of a bubble.
The data utilized for out tests will be monthly observations from the CRSP database, as well as the Case-Schiller House Pricing Index and other published data sources. For wealthy speculators we may resort to Hedge-Fund data. All told the data bases will encompass the period of the market crisis from 2001 to 2011, where available. In addition to the standard databases we have also utilized information about several classes of Real Estate Investment Trusts (REITs) for several categories and segments.
All told, our results support weakly the existence of stock bubbles and point to the formation of asset bubbles when the results are segmented three sequential periods: (a) the Accumulation period; (b) the peak period; and (c) the bubble pre-burst period.
Because of the vast databases the present research focuses on the sub period August 2005 to August 2006. This sub period is more convincing since housing prices were still increasing even though the bubble burst in December of 2007. The results and explanations are codified by two appendixes. Appendix(A) presents the real estate industry descriptions classified by SIC codes. Particular attention is given to the three major categories and the eight subcategories of Real Estate Investment Trusts (REITs). Appendix (B) is a timeline of news events that may have affected the financial markets.
In a future versions of our research we shall encompass the entire database from August 2005 to the present and compute both financial returns and cumulative returns for the entire period.
Specific Methodology and Results
For this paper we used the Case-Shiller 20-city composite home price index, Jul04 – Dec07. This measures sales price of existing single-family houses in 20 metropolitan areas in the U.S. The index reached its peak at 206.52 in July 2006. Based on the trend of the 20-city home index and the market news, the timeline is separated into three periods: accumulation period from July 2004 to July 2005; peak period from August 2005 to July 2006; and pre-burst period from August 2006 to February 2007.
To be conservative, we used the peak of the housing price as a bench mark, which is earlier than the stock market peak. Since housing prices were still climbing before July 2006, it will be harder to argue that the market burst before that time. So if we can find evidence that some stocks were attacked before the housing price peak, it is more convincing that rational investors have realized the potential bubble ahead of the bust.
We chose Feb, 2007 as the endpoint of pre-burst period for three reasons:
1)New Century Financial Corporation, the largest U.S. subprime lender, filed for chapter11 bankruptcy in April 2007. The company had problems in March but filed in April. To avoid arguments, we put the endpoint one month early, Feb. 2007
2)In the ABI database, when search “subprime” or “mortgage” or “crisis”, there were two news items in Jan 07 and nine in Feb 07 and 88 in March 07, 46 in April 07, 24 in May 07, 49 in June 07, 108 in July 07, 340 in August 07 and 319 in September 07. Therefore, Feb 2007 is chosen.
3)These two cut-off points match the local peak and trough in the stock market below, except Feb 07 is one month earlier than the local drop in S & P 500.
Figure 1 Price Change by Portfolios in Accumulation Period: Jul04 – Jul05.
Based on the geometric mean returns over the accumulation period from July 2004 to July 2005, five portfolios were formed with the highest return (P0) to the lowest (P4). The rank remains the same over the period. Figure 1 shows the monthly average price trends of all five portfolios. Price is the unadjusted price. If the price is not available then the mid-point of the highest bid and lowest ask is taken as the price of that month.
1)All portfolio prices remain stable or increasing over the first period. No significant price drop is observed. There is no obvious evidence that the market identified potential overpricing at this stage, or potential bubble is still in accumulation and stocks are not substantially overpriced.
2)The price of the highest return portfolio has much higher prices compared to the other portfolios and its price increases the most in the period.
P4 composition in three periods. The components of P4 in three time periods are investigated. Panel A demonstrates the number of firms in P4, percentage of P4, and percentage of that industry. Panel B shows the number of firms in P4, percentage of P4, and percentage of that segment in REITs. P4 is the portfolio with the highest geometric mean return in the period, which is the bottom 20% of the stocks in the real estate section. Period 1 is the accumulation period from July 2004 to July 2005; Period 2 is the peak period from August 2005 to July 2006; Period 3 is the pre-burst period from August 2006 to February 2007.
Period 1 / Period 2 / Period 3Firm Number / % in Portfolio / % in Industry / Firm Number / % in Portfolio / % in Industry / Firm Number / % in Portfolio / % in Industry
Panel A: By Industry
Operative Builders / 0 / 0.00% / 0.00% / 15 / 26.79% / 93.75% / 7 / 14.29% / 35.00%
General Contractors / 0 / 0.00% / 0.00% / 6 / 10.71% / 60.00% / 1 / 2.04% / 33.33%
Mortgage Bankers and Loan Correspondents / 11 / 20.75% / 78.57% / 5 / 8.93% / 33.33% / 4 / 8.16% / 40.00%
Real Estate Agents and Managers / 3 / 5.66% / 42.86% / 5 / 8.93% / 50.00% / 2 / 4.08% / 20.00%
Title Insurance / 0 / 0.00% / 0.00% / 1 / 1.79% / 16.67% / 0 / 0.00% / 0.00%
All REITs / 39 / 73.58% / 18.14% / 24 / 42.86% / 10.81% / 35 / 71.43% / 17.50%
Panel B: Segments in REITs
Mortgage_Home Financing / 13 / 33.33% / 56.52% / 8 / 33.33% / 32.00% / 11 / 31.43% / 45.83%
Mortgage_Commercial Financing / 3 / 7.69% / 25.00% / 2 / 8.33% / 14.29% / 3 / 8.57% / 27.27%
Residential_Apartment / 0 / 0.00% / 0.00% / 1 / 4.17% / 4.55% / 6 / 17.14% / 33.33%
Residential_Manufactured Home / 2 / 5.13% / 40.00% / 1 / 4.17% / 20.00% / 2 / 5.71% / 40.00%
Other REITs / 21 / 53.85% / 13.91% / 12 / 50.00% / 7.69% / 13 / 37.14% / 9.15%
Table 1
This table is quite interesting as it shows the composition of the stocks in the real estate industry before the bubble burst. Here there is found evidence of attacks within several sectors of the industry:
- No stocks in operative builders and general contractors industries performed in the bottom 20% of real estate sector,but, they did in the second period.
- Almost all Mortgage Bankers and Loan Correspondents (78.57%) performed in the bottom 20% in the real estate sector. Even in periods 2 and 3, they still had a big proportion in the worst performing portfolio.
- More than half of the REITs invested in the home financing mortgage (Mortgage_home financing) belong to the bottom performing stocks. The number alsoremained higher than other segments in REITs through period 2 and 3.
Figure 2
Price and Short Interest Ratio Co-movement of P4 (Low) in the Peak Period: Aug05 – Jul06. Thisfigure 2 shows the co-movement of the monthly average price and short interest ratio (SI) of portfolio 4 from August 2005 to July 2005. Portfolio 4 is the portfolio with lowest geometric mean return in the period.
- A clear mirror image shows strongly that short sellers did affect the market. When short interest ratio increases, price of P4 responds by decreasing. Alternatively, when short interest ratio decreases, price of P4 increases. During this peak period, short interest of P4 increased over the whole period. After Apr06, the short interest increased substantially. Responding to this increase, portfolio 4 price decreased.
- It appears that a group of investors realized that there was potential overpricing and attacked some of the stocks in the market (P4 in our sample), well before the bubble burst in 2007. Remember this is the period when housing prices were climbing to their peak. Also, for some portfolios, the short sellers jumped in too early, the increasing of the short interest did not get a response of prices dropping. The price continued to increase as positive correlation indicates. However, it does not mean the short sellers were not attacking them. They were just unsuccessful, because too many investors did not believe the market was over heated or the stock prices were overpriced and they kept buying and pushing out the price. Some of these investors are momentum traders; some are informed traders (like hedge funds) trying to catch the bubble and profit from the momentum traders (or rational bubble in some literature in this case).
Review and tentative Conclusions
In this paper, based on the Housing index price of 10 cities, we found the price peaked in August 2006. Therefore, we tested whether from Aug 05- Aug 06 stocks were attacked. This period is more convincing since during that period, housing price is still increasing, even though the bubble burst in Dec. 2007. We have found different sub categories for real estate industry, the majority of them being REITs. We then separated REITs into Mortgage REITs and non-Mortgage REITs. Based on the period from August 2005-August 2006, returns were calculated for all stocks in the sample. We ranked the returns and categorized the sample into 5 groups (P0, P1...P4). Based on the subgroups, we check whether their prices drop or increase. The results supports that P4 (low) has dramatic drops even when the housing price is still increasing in that period. P4 also has the highest short interest ratio. When SIR and price are graphed P4, P3 clearly show that when SIR is higher, price is lower. However, P0 (high) does not have that pattern. The P0 graph shows the SIR is not enough and when SIR increases, price also increases.
Bibliography
Abreu, Dilip and Markus K. Brunnermeier, “Synchronization Risk and Delayed Arbitrage”, Journal of Financial Economics 66 (2002) pp. 341-360
Brunnermeier, Markus K. and Stefan Nagel, The Journal of Finance, Vol LIX, No. 5, October 2004, pp2013-40
Cooper, George, The Origin of Financial Crises: Central Banks, Credit Bubbles, and the Efficient Market Fallacy. Vintage Press, 2008, pp.121-125)
Schiller, Robert, Irrational Exuberance, Princeton University Press, Princeton, N.J. 2000
Sorkin, Andrew Ross, Too Big to Fail: The Inside Story of How Wall Street and Washington Fought to Save the FinancialSystem--and Themselves, Penguin Books, Updated edition (September 7, 2010)
Stiglitz, Joseph E., Globalization and its Discontents, W.W. Norton, 2002, pp 94-98
Appendix A
Real Estate Industry Description
Classified by SIC Codes
Real Estate Sub-Industry Description
“General Contractors--Single-Family Houses, SIC:1521” covers general contractors primarily engaged in construction activities of single-family houses. General contractors primarily engage in building, remodeling, and repairing houses. Included in this industry are prefabricated housing units assembled on-site and townhouse construction. The industry is lack of concentration. A typical General Contractors firm may produce fewer than twenty-five houses each year, and employ fewer than twenty-five workers. Large General Contractor firms that do compete nationally are often relatively decentralized - consisting of generally autonomous regional operating companies. The few contractors that compete overseas usually do so through foreign-owned subsidiaries.
“Operative Builders, SIC:1531” covers builders primarily engaged in the construction of single-family houses and other buildings they sell on their own rather than as contractors. Unlike general contractors who perform construction work on a for-hire basis for the owner or owners of a development, operative builders are the owners of the structures they build and act as their own general contractors. In addition to construction, operative builders engage in land acquisition, sales, and a variety of other non-construction activities associated with developing and selling properties. Although operative builders were primarily involved in construction, principal industry activities also included subdividing and site development work, real estate management activities, land sales, construction-related activities, and miscellaneous operations.
“Mortgage Bankers and Loan Correspondents, SIC:6162” covers establishments primarily engaged in originating mortgage loans, selling mortgage loans to permanent investors, and servicing these loans. They may also provide real estate construction loans.
Mortgage bankers specialize in the origination or production of mortgage loans for sale to the secondary mortgage market. Many mortgage lenders make or buy loans, some sell loans, and others service loans. Mortgage bankers link theses three functions.
“Title Insurance, SIC:6361” covers establishments primarily engaged in underwriting insurance to protect the owner of real estate, or lenders of money thereon, against loss sustained by reason of any defect of title. The concept of title insurance is primarily used in the United States. Title insurance protects borrowers and lenders against legal complications often having to do with fraud, liens, taxes, and other unexpected issues that can arise during and after real estate transactions. Typical title defects result from liens and encumbrances on a property related to unpaid taxes, land use and zoning restrictions, unsettled contractor disputes for work done on the property, and unrecorded deeds. Title insurance also protects the buyer and lender from fraud perpetrated by the seller.
“Real Estate Agents and Managers, SIC:6531” covers establishments primarily engaged in renting, buying, selling, managing, and appraising real estate for others. The industry also includes agents and managers who work on commercial real estate, including high-rise office buildings, shopping centers, industrial plants, medical centers, and so forth. It is a service business involving a variety of professionals.
“REITs, Real Estate Investment Trusts, SIC:6798” covers establishments primarily engaged in closed-end investments in real estate or related mortgage assets operating so that they could meet the requirements of the Real Estate Investment Trust Act of 1960 as amended. This act exempts trusts from corporate income and capital gains taxation, provided they invest primarily in specified assets, pay out most of their income to shareholders, and meet certain requirements regarding the dispersion of trust ownership.