The widespread dispersion of credit risk and the unclear impact on nancial institutions have caused lenders to reduce lending activity or to make loans at higher interest rates. Similarly, the ability of corporations to obtain funds through the issuance of commercial paper was impacted. This aspect of the crisis is consistent with a credit crunch. The liquidity concerns drove central banks around the world to take action to provide funds to member banks to encourage the lending of funds to worthy borrowers and to re-invigorate the commercial paper markets. The subprime crisis also places downward pressure on economic growth, because fewer or more expensive loans decrease investment by businesses and consumer spending, which drive the economy. A separate but related dynamic is the downturn in the housing market, where a surplus inventory of homes has resulted in a signi cant decline in new home construction and housing prices in many areas. This also places downward pressure on growth. With interest rates on a large number of subprime and other Adjustable Rate Mortgages (ARM) due

Produced by The Berkeley Electronic Press, 2010 3 to adjust upward during the 2008 period, US Legislators and the US Treasury Department have been taking action. A systematic program to limit or defer interest rate adjustments was implemented to reduce the impact. In addition, lenders and borrowers facing defaults had been encouraged to cooperate to enable borrowers to stay in their homes. The risks to the broader economy created by the nancial market crisis and housing market downturn were primary factors in the January 22, 2008 decision by the US Federal reserve to cut interest rates and the economic stimulus package signed by President Bush on February 13, 2008. Both actions were designed to stimulate economic growth and inspire con dence in the nancial markets. The value of US sub-prime mortgages was estimated at $1.3 trillion as of March 2007, with over 7.5 million rst-lien sub-prime mortgages outstanding. Approximately 16% of subprime loans with ARM were 90 days delinquent or in foreclosure proceedings as of October 2007, roughly triple the rate of 2005. By January of 2008, the delinquency rate had risen to 21%. Subprime ARMs only represented 6.8% of the loans outstanding in the US, yet they represent 43.0% of the foreclosures started during the third quarter of 2007. A total of nearly 446,726 US household properties were subject to some sort of foreclosure action from July to September 2007, including those with prime, alt-A, and subprime loans. This was 2-fold of the 223,000 properties in the year earlier, and 34% higher than the 333,627 in the prior quarter. This increased to 527,740 during the fourth quarter of 2007, an 18% increase versus the prior quarter. For all of 2007, nearly 1.3 million properties were subject to foreclosure lings, up 79% versus 2006.

As a result extensive subprime loans, a few types of risk – credit risk, liquidity risk, asset valuation risk, and risk on relying on nancial institutions – came into being, and this 4-tuple risk shook the entire edi ce of the nancial world. Traditionally credit risk was assumed by the banks giving mortgage loans, but in recent years, owing to innovations in securitization, credit risk was being shared more broadly with investors, because the rights to these mortgage payments had been repackaged into a variety of complex investment vehicles, generally categorized as mortgage-backed securities (MBS) or collateralized debt obligations (CDO). The following diagram (Figure 1) depicts what transpired because of the securitization structure of borrowing, lending and investing. The small spikes within the circle represent the borrowers (home mortgage holders) from the banks, represented in the cluster (in the center of the circle) that created and sold mortgage-backed securities (MBS), and collateralized debt obligations (CDO). The spikes going beyond the circle represent the investors in the MBS and CDO, longer ones are the international investors and shorter ones are domestic investors. Originally, the home mortgagees with subprime loans were paying mortgages, but when the reset started, they could not pay their dues, and this reality came into being when the income in ows did not http://epublications.bond.edu.au/ijbf/vol7/iss2/2 4 get to the foreign and domestic investors. All spikes became insolvent and got crushed in the windmill of dire nancial distress. To manage their risk, mortgage originators (banks or mortgage lenders) created separate legal entities, called special-purpose entities (SPE), to both assume the risk of default and issue the MBS. The banks effectively sold all the mortgage assets (that is, banking account receivables, which were the rights to receive the mortgage payments) to these SPE. In turn, the SPE then sold the MBS to the investors. The mortgage assets in the SPE became the collateral. Another variant of SPE – known as Structured Investment Vehicles (SIV) – and other corporations soon realized the devastation in the markets on commercial papers, causing illiquidity an unmanageable situation. The amount of commercial paper issued as of October 18, 2007 dropped by 25%, to $888 billion, from the August 8 level. Along with or because of liquidity risk and credit risk, asset market prices fell sharply, and threw the nancial markets into chaos and disarray, where the domino effect was observed.

ABC

In the wake of the almost free-falling economies all around the world in general, and the economy of the US in particular, President George W. Bush

Produced by The Berkeley Electronic Press, 2010 7 signed into law on February 13, 2008 an economic stimulus package of $168 billion, mainly in the form of income tax rebates, to help stimulate economic growth. The Troubled Asset Relief Program (TARP) which was a program of the US government to purchase assets and equity from nancial institutions to strengthen its nancial sector was created. It was the largest component of the government’s measures in 2008 to address the subprime mortgage crisis. TARP allowed the US Treasury Department to purchase or insure up to $700 billion of “troubled” assets. “Troubled assets” were de ned as “(A) residential or commercial mortgages and any securities, obligations, or other instruments that are based on or related to such mortgages, that in each case was originated or issued on or before March 14, 2008, the purchase of which the Secretary determines promotes nancial market stability; and (B) any other nancial instrument that the Secretary, after consultation with the Chairman of the Board of Governors of the Federal Reserve System, determines the purchase of which is necessary to promote nancial market stability.” Next, under the new Presidency of Barack Obama, the American Recovery and Reinvestment Act of 2009 (ARRA (Pub.L. 111-5) - an economic stimulus package enacted by the 111th United States Congress in February 2009 appeared to calm and soothe nancial nerve centers and had indeed controlled the turbulence. The Act followed other economic recovery legislation passed in the nal year of the Bush presidency including the Economic Stimulus Act of 2008 and the Emergency Economic Stabilization Act of 2008 which created the Troubled Assets Relief Program (TARP). These measures are nominally worth $787 billion. The Act includes federal tax cuts, expansion of unemployment bene ts and other social welfare provisions, and domestic spending in education, health care, and infrastructure, including the energy sector. The Act also included numerous non- economic recovery-related items that were either part of longer-term plans. The stock markets all around the world have rebounded with still bumps here and there. European governments and central banks have followed the US lead in the containing contagion. Japan has been on board. The Fed and other central banks have conducted open market operations to ensure member banks have access to liquidity. These are effectively short- term loans to member banks collateralized by government securities. Central banks have also lowered the interest rates charged to member banks (called the discount rate in the US) for short-term loans. Both measures effectively lubricate the nancial system, in two key ways. First, they help provide access to funds for those entities with illiquid mortgage-backed assets. This helps lenders, SPE, and SIV avoid selling mortgage-backed assets at a steep loss. Secondly, the available funds stimulate the commercial paper market and general economic activity. Speci c responses by central banks are included in the subprime crisis impact timeline. The Fed has been utilizing the Term Auction Facility (TAF) to provide short-term loans (liquidity) to banks. The Fed has increased the monthly amount of these auctions to $100 billion during March 2008, up from $60 billion in prior months. In addition, term repurchase agreements expected to cumulate to $100 billion were announced, which enhanced the ability of nancial institutions to http://epublications.bond.edu.au/ijbf/vol7/iss2/2 8 sell mortgage-backed and other debt. The Fed indicated that both the TAF and repurchase agreement amounts would continue and be increased as necessary. It appears that these measures have created a slow but steady rebound.

In this section how much the spill-over effect is evaluated. How did sub-prime loans spread the virus across nations? Numerous studies and serious research have shown different conclusions and this paper is more in disagreement on the question as to whether international markets are integrated or segmented. Black (1974), Stulz (1981a, 1981b), Eun and Janakiramanan (1986), Hietala (1989) have viewed that international capital markets are segmented, and Adler (1974), Adler and Dumas (1983), and Stapleton and Subramanyam (1977), ignoring exchange risk, examined the implication of market segmentation. French, Schwert, and Stambaugh (1987) have presented evidence that there is a positive correlation between the conditional expected excess return on the market portfolio and the conditional variance of its return, using generalized autoregressive conditional heteroscedasticity representation, as exposited by Bollerslev (1986) and Engle (1982) for the excess return on the market with the Standard and Poor’s 500 index (S&P 500) for the market portfolio. In their sample, it appeared that foreign capital markets are signi cantly segmented from the US capital markets. The studies of Cho, Eun and Senbet (1986), Wheatley (1988), Korajczyk and Vialett (1990), and Harvey (1991) have provided evidence of integration by employing different asset pricing models and using monthly data. By making use of the arbitrage pricing theory (APT), Gultekin, Gultekin and Penati (1989) have established that for fteen years the US and it the Japanese markets had been more and more integrated. Chan, Karolyi, and Stulz (1991) have attempted to reexamine and generalize the work of French, it Schwert, and Stambaugh (1987) by assuming that (i) markets are internationally integrated, (ii) investors are optimizers in the mean-variance framework in a common numeraire currency such as the US dollar, and (iii) the aggregate relative n n risk aversion, Z ¦ Z / ¦W i 1 it i 1 it , is constant, where Z it is the i-th individual’s risk aversion, and Wit is the wealth at time t.

Under these assumptions, they have begun with the following equations:

E t ( Rmt R ft ) Z 1 vart ( Rmt R ft ) (1)

Et (rdt ) Z 1 cov(rdt mmt ) (2)

Produced by The Berkeley Electronic Press, 2010 9 where Et is the expectation operator, Rmt the return on market portfolio, Rft the risk-free rate, and vart (Rmt Rf ) the variance of the excess return on the market portfolio, rdt and mmt the excess return on the US domestic market portfolio and that of the world market portfolio, while cov stands for the covariance. Since the return on the world market portfolio can be expressed as:

mmt wdt rdt (1 wdt )rnt (3)

where wdt is the weight on the excess return on the domestic market portfolio and is rnt the excess return on the non-domestic (foreign) market portfolio, one can rewrite (2) by plugging (3) into it as follows:

Et (rdt ) Z 1 [ wdt var(rdt ) (1 wdt ) cov t (rdt , rnt )] (4)

It is obvious now that if rdt equation (4) is reduced to French et al. relation. However, if the covariance term is zero in (4), it is not the same case that French, Schwert, and Stambaugh (1989) had examined. Chan, Karolyi, and Stulz (1991) parameterize equation (4) as follows:

rdt D d E dv wdt hdt E dc (1 wdt )hct T d 1H dt 1 T d 2 H dt 2 H dt (5)

rnt D n E nv wn1 hn1 E nc (1 wnt )hct T n1H nt 1 T n 2 H dt 2 I n H dt 1 H nt (6)

where hdt and hnt are the variances and hct is the covariance, which depends upon past returns, lag structure of stochastic terms were introduced through H , and I captured the effects of infrequent trading and non-synchronous trading hours. Making use of the general process, modeled after Baba, Engle, Kraft, and Kroner (1989), Chan et al. reported that their “tests are … supportive of the hypothesis that markets are internationally integrated over sample period”they consider.” Blenman (1991), in an interesting theoretical model, consisting of two countries in which one country’s market is perfectly competitive for everyone, including domestic and foreign investors, but the other country’s market is barrier-free for its domestic investors, but quite partitioned for foreign investors, derives a number of useful result in a uni ed exchange rate regime. It was shown that all investors’ risky assets hold a common factor share, appropriately weighted by their tastes. When they face further restrictions, their asset demands are modi ed to re ect those restrictions unless the variance-covariance matrix is block diagonal, and/or the restrictions mutually cancel each other, and/or the total risk tolerance factors are all zero. Along these traditional routes some more works are noteworthy. Stulz (1995) discussed globalization of capital market and the cost of capital, which under a different fold extend the issue of international connectivity of markets. Portes, and Rey (1999) examined the determinants of cross-border equity, and Pagano, Roell, and Zechner (2001) looked at the geography of equity listing and http://epublications.bond.edu.au/ijbf/vol7/iss2/2 10 asked the question: Why do companies list abroad? In his Presidential Address to American Finance Association, published in The Journal of Finance (2005), Stulz re-examined the limits of nancial globalization, and threw the issue of market integration or segmentation into further focus.

Following Leontief’s celebrated structure, ij Ghosh and Khaksari (187) taken us to a new line of analytical framework of ow structure. Bringing that ow structure alive, one can examine the investment ows of investors of different countries into different capital markets existing across several nations under the assumption that individual investors are rational and they face no political risks with their investment funds. Let aij be the amount of the investment funds in market i needed to buy one unit investment in market j, Xj be the total investment units in market j, and Yi be the amount of funds in market i kept exogenously for other purposes. Then the available funds balance can be stated as follows:

To put (7) in matrix notation, we have where I is an n u n identity matrix, A { [ a ij ] nun matrix, X and Y are n element column vectors. The solution of (7c) is given by:

Here Z ij is the ij-th element of the (I-A) inverse matrix.

Consider now that A matrix assumes one of the following forms:

Produced by The Berkeley Electronic Press, 2010 11

If A { A1 , X 1 is independent of other markets, but X 2 depends on X 1 , depends on X 1 , and X 2 , and so on. This sort decomposability of and X 3 investment ow structure de nes the rst market as the “key” market that drives the rest of the world market. Mortgage meltdown in the US with its tentacles on other mortgage-backed securities appeared to have driven world markets down. If, on the other hand, A { A2 , then n-th market is obviously the key generator of the global market movements. If A assumes the form of A3, we then have a number of subsets of the global markets, each subset being independent of the others. Here we nd that (1, 2), (3,4), ……., (n-1, n) markets are intra- subset integrated, and inter-subset segmented. However, if the ow matrix is not triangular (upper or lower), the key market can still be captured through the following indices of power of dispersion and sensitivity of dispersion, originally noted by Rasmussen (1956):

The index d j de nes the extent to which an increase in investment ow in market j is dispersed throughout the global economy. If d j ! 1 , it signi es http://epublications.bond.edu.au/ijbf/vol7/iss2/2 12 that market j draws heavily on the system of markets (and conversely, if d j 1 ). The index d i , on the other hand, measures the extent to which the system of markets draws upon market i . If d i ! 1 , it means that the i - th market is required to accommodate more investment funds than other markets corresponding to an exogenous increase in investment. These two indices serve well, but they are not without some weaknesses as they are un-weighted. To correct this de ciency, one can modify, for instance, the sensitivity of dispersion as follows:

Assume that markets are indexed h and i (where h 1,2,...., m; i 1,2,...., n ) such that {h} {i} , the balance holds for a nation, and for markets i balance holds for the world as a whole. That is,

For the world as a whole, and

for a nation, d (1 d d d n) , where

I NN { >G hk @mum G hk is the Kronecker delta that is,

G hk 1 h k ; G hk 1 hzk

A NN { >a hk @mum for h, k 1,2,......., m

A NW { >a hs @mu( n m ) X W ( d ) { { X h( d ) }, X N ( d ) { { X s( d ) },

Y N ( d ) { {Yh( d ) for d (1 d d d n) :: national markets. national markets.

Produced by The Berkeley Electronic Press, 2010 13

From (11) one can get the following:

X N (d ) ( I NN A NN ) 1 (Y N ( d ) A NW X N ( d ) ) (13)

Since ( n m) equations in (11) yield X V N Y , where N

X N { { X s }, [V N V W ] { ( I A) 1 , m 1 d s d n, equation (13) can be re-expressed as follows:

X N (d ) ( I NN A NN ) 1 (Y N ( d ) A NW ( E r( d )G rs ).V W Y ) (14)

provided the following proportionality holds:

X N (d ) ( E r( d )G rs ) X N , r , s m 1,........, n , and X N { { X s }, m 1 d s d n.

Following the procedure outlined earlier, one can compute the indices of power and sensitivity of dispersions of investment ows within this structure of global market network, and determine if markets are integrated or segmented. To support the above theoretical segments with the applications that are undertaken by other researchers, some parallel information may be given. For example, a variant of the above is empirically applied by Parhizgari (2002). Using equity data on the stock exchanges of seven regions and 18 countries for January 1982 October 2002, Parhizgari (2002) shows that substantial degrees of convergence have been achieved across the global stock markets over the noted period.