Abstract
This paper measures how securitized assets, including mortgage-backed securities and other asset-backed securities, have shifted across financial institutions during this crisis and how the availability of financing has accommodated such shifts. Sectors dependent on repo financing—in particular, the hedge fund sector—have reduced asset holdings, while the commercial banking sector, which has had access to more stable funding sources, has increased asset holdings. The banking sector also increased its leverage dramatically during this crisis. These findings are important to understand the role played by the government as well as the factors determining asset prices and liquidity during the crisis.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
A nonexhaustive list includes Gromb and Vayanos (2002), Allen and Gale (2005), Geanakoplos and Fostel (2008), He and Krishnamurthy (2008, 2009), Adrian and Shin (2010), and Brunnermeier and Pedersen (2009).
Bloomberg reports the aggregate repayment rate of 17 percent across a large (>$3 trillion) sample of ABS and MBS in 2008 (see Bloomberg CMO/ABS Market Profile; function mtge CAMP). They also report that the aggregate rate of new issuance is 10 percent. These numbers lead to our choice of 7 percent as a net repayment rate.
This estimate is based on the surviving funds, which lost $161 billion by redemption and lost $317 billion from asset trading, according to Barclay Hedge (2009).
Many of the other entities in Table 6 are owned by a bank holding company so that their balance sheet adjustments may have been influenced by the holding company with significant commercial banking operations such as Citigroup or JP Morgan Chase. Goldman Sachs and Morgan Stanley do become bank holding companies in the fall of 2008, so that there is a limit to how clean our pure broker/dealer measure can be. However, it is worth noting that even after converting to holding company status, commercial banking operation still represents a very small fraction of these entities and their main business remains to be in the broker/dealer industry. Separately, Merrill Lynch ceased to be a stand-alone broker/dealer and became part of the Bank of America as of January 2009. However, we do not observe major changes in Merrill Lynch's asset holdings in the first quarter of 2009.
The Federal Reserve's Flow of Funds is another data source for understanding the change in the broker/dealer sector. Although our computations result in a similar picture as painted by the flow of funds, the advantage of our computations is that the SEC filings allow a more detailed breakdown of asset holdings than is provided in the flow of funds.
There is another consideration that affects the interpretation of our computations in this section. We do not have information on derivatives positions. Thus, it is possible that some of these assets are hedged by derivatives so that the broker/dealers have a small exposure to the underlying asset risk. Nevertheless, our computation of asset sale is still correct, and we just have to modify our interpretation that the broker/dealers are unwinding positions as opposed to selling off risk.
Washington Mutual and Countrywide were also acquired by commercial banks in 2008. One point worth noting is that, because the Flow of Funds data is back-filled to reflect the effect of these mergers, Table 10 and Figure 1 are free of the data issue caused by these merger and acquisition activities. On the other hand, there may be a slow change in assets in the case of the broker/dealer acquisitions. Take for example, the JP Morgan Chase acquisition of Bear Stearns. Any MBS assets acquired in this merger will, at the time of the merger, be held in JP Morgan Chase's broker/dealer rather than the commercial bank. Thus the merger will not cause an immediate raise in commercial banking assets as computed from the call reports. However, suppose that over time, the securities held by Bear Stearns are transferred to JP Morgan Chase's commercial bank (perhaps because they can be financed more easily that way), then we would see a slow rise in banking assets.
The growth in AFS securities we document reflects growth in the holdings of JP Morgan Chase commercial bank and not the broker/dealer owned by the holding company. We can see this by comparing the AFS values reported in SEC filings to holdings data from call reports. The numbers are almost identical.
Note that ABCP outstanding shrinks from $1.25 trillion to $833 billion by December 2007. This suggests that the bulk of ABCP liquidation occurs in 2007 and not during 2008, and thus is likely not responsible for the 2008 asset growth at banks.
Ivashina and Scharfstein (forthcoming) discuss another source of growth in bank assets. They document that many firms draw down credit lines during the turmoil of the fall of 2008, causing bank loans to rise. They stress that these loan increases are “involuntary” rather than voluntary. In the ABCP liquidations, banks involuntarily take on ABCP assets, but there decision to hold on to these assets is voluntary.
The data source is www.fdic.gov/regulations/resources/tlgp/reports.html.
There are finer patterns that match the dynamics of the crisis. As of December 31, 2007 the total outstanding advances rose to $875 billion. As of September 30, 2008, advances were at a peak of $1,011 trillion, before falling to $928 billion on December 31, 2008. The outstanding advance further falls to $817 billion on March 31, 2009.
Throughout the latter part of 2009 as financial conditions improved, banks have raised common equity from private sources, paying back the TARP money. It seems likely that leverage fell through this period.
As an example, news reports suggest that BlackRock Asset Management purchased asset-backed securities during the crisis. From their SEC filings, BlackRock's AUM in fixed-income funds decrease from $513 billion to $474 billion from 2007:Q4 to 2009:Q4. Similarly there are news accounts of private equity funds pursuing purchases of commercial banks (www.nytimes.com/2009/08/27/business/27bank.html). Note that this is not purchases of ABS, but purchases of banks. Moreover, it seems possible that the interest driving these purchases is the access to stable funding enjoyed by the banking sector.
Another possible sector we have left out of the analysis is long-only investors, such as private pension funds. The flow of funds reports total assets of pension funds of around $5 trillion. However the bulk of these assets are in corporate equities or mutual funds. The increase in holdings of GSE securities (which includes both MBS and straight agency debt) plus all corporate and foreign bonds over the relevant period is about $70 billion. Note that this figure likely includes a majority of debt securities which are not of interest for our analysis.
Bank-run explanations (Diamond and Dybvig, 1983; Allen and Gale, 2005, Gorton and Metrick, 2009; and He and Xiong, 2010) have similar predictions, although not stated explicitly in terms of debt constraints and haircuts. In these models, either the realization of a liquidity shock or deteriorating fundamentals trigger a bad equilibrium in which there is a disintermediation and asset sale.
Another exposition of the equity risk-capital theory focuses on risk-based regulatory capital considerations for commercial banks. Regulatory capital requirements penalize holdings of risky assets in favor riskless assets (for example, Kashyap and Stein, 2004). Thus, when losses erode capital levels, banks respond by shifting their portfolios to favor riskless assets. This in turn implies that banks require a higher risk premium to purchase risky assets, causing asset prices to fall. This theory shares the predictions of the equity-capital/risk-aversion theory, as the reasoning relies on the relation between asset demand, equity capital and asset riskiness.
In Brunnermeier and Sannikov (2009), intermediaries have linear preferences but are restricted to only have positive consumption. As a negative consumption implies a utility of minus infinity, this is isomorphic to assuming that intermediaries are risk averse.
According to TLGP, the maximum debt that can be issued by a bank is limited to 125 percent of the par value of the bank's senior unsecured debt that was outstanding as of the close of business September 30, 2008 and that was scheduled to mature on or before June 30, 2009. Banks only have used 43.7 percent of cap on March 31, 2009, and 43.7 percent on June 30, 2009 (www.fdic.gov/regulations/resources/tlgp/reports.html).
Allen and Gale (1994), Diamond and Rajan (2009), and Holmstrom and Tirole (1998) study a dynamic version of the leverage-constraints model. In their models, dynamic considerations lead agents to hold a buffer of liquidity at time 0, rather than saturating the maximum borrowing capacity immediately. In Allen and Gale and Diamond and Rajan, the behavior is because of the anticipation of future fire sales. In Holmstrom and Tirole, the behavior arises because the possibility of a binding constraint makes agents’ current value function concave. These models can rationalize low asset prices as well as banks’ ex ante decision not to saturate debt capacity. However, they require that banks expect that the U.S. Federal Reserve's lending and liquidity facilities will be insufficient to meet anticipated borrowing needs, which seems at odds with the unprecedented level of lending by the U.S. Federal Reserve. Also, this theory does not speak to high leverage directly.
Gatev and Strahan (2006) and Pennacchi (2006) document “reintermediation” during disruptions in the commercial paper market, and attribute the FDIC deposit insurance (which is only enjoyed by commercial banking sector) to this phenomenon.
There is another important theory linking government-backed financing and bank decisions that requires discussion. The classic risk-shifting theory (Jensen and Meckling, 1976) as applied to the banking sector is that banks exploit the government guarantee, turning risk-loving, and purchase the riskiest assets. On one hand, this theory is consistent with the fact the banks have increased asset holdings and have raised leverage. On the other hand, this theory seems at odds with a number of other stylized facts. First, even in their security purchases, banks have concentrated on buying the lower risk agency-backed MBS, rather than on seeking out the riskiest ABS to purchase. Second, the liquidity problems and apparent high market prices of risk seem most pronounced on the riskiest assets. Yet, if banks had strong reasons for buying the riskiest assets, these assets would have the lowest risk premia and the least liquidity problems. Finally, risk-shifting incentives would lead banks to saturate the debt guarantees, but the data suggest otherwise.
Note, however, that all of them exclude repurchase agreement transaction volumes.
It is possible that this treatment will exclude part of agency MBS holdings. However, by reading the notes of SEC filings, usually agency MBS are in the category of MBS, not in “U.S. government and agency securities.”
After being acquired by Bank of America the information on investment securities is in note 7.
Liberty mutual is not a publically listed company therefore does not have SEC filings. We obtain its annual reports from their website.
References
Acharya, Viral, Irvind Gujral, and Hyun Song Shin, 2009, “Dividends and Bank Capital during the Financial Crisis of 2007–2009,” New York University Stern School of Business Working Paper.
Acharya, Viral, Philipp, Schnabl, and Gustavo Suarez, 2010, “Securitization without Risk Transfer,” NBER Working Paper 15730 (Cambridge, MA, National Bureau of Economic Research).
Adrian, Tobias, and Hyun Song Shin, 2010, “Liquidity and Leverage,” Journal of Financial Intermediation, Vol. 19, No. 3, pp. 418–437.
Allen, Franklin, and Douglas Gale, 1994, “Liquidity Preference, Market Participation and Asset Price Volatility,” American Economic Review, Vol. 84, No. 4, pp. 933–955.
Allen, Franklin, and Gale, Douglas, 2005, “From Cash-in-the-Market Pricing to Financial Fragility,” Journal of the European Economic Association, Vol. 3, No. 2–3, pp. 535–546.
Ashcraft, Adam, Morten Bech, and Scott Frame, 2008, The Federal Home Loan Bank System: The Lender of Next-to-Last Resort? New York Federal Reserve Staff Report 357.
Barclay Hedge, 2009, The Hedge Fund Flow Report, (March). Available via the Internet: www.barclayhedge.com/research/hedge-fund-flow-report/hedge-fund-flow-report.html.
Beltran, Daniel, Laurie Pounder, and Charles Thomas, 2008, “Foreign Exposure to Asset-Backed Securities of U.S. Origin,” Board of Governors of the Federal Reserve System International Finance Discussion Paper Number 939.
Brunnermeier, Markus, 2009, “Deciphering the Liquidity and Credit Crunch 2007–08,” Journal of Economic Perspectives, Vol. 23, No. 1, pp. 77–100.
Brunnermeier, Markus, and Lasse Pedersen, 2009, “Market Liquidity and Funding Liquidity,” Review of Financial Studies, Vol. 22, No. 6, pp. 2201–2238.
Brunnermeier, Markus, and Yuliy Sannikov, 2009, “A Macroeconomic Model of the Financial Sector,” Princeton University Working Paper.
Caballero, Ricardo, and Pablo Kurlat, August, 2009, “The ‘Surprising’ Origin and Nature of Financial Crises: A Macroeconomic Policy Proposal,” Massachusetts Institute of Technology Working Paper.
Diamond, Douglas, and Philip Dybvig, 1983, “Bank Runs, Deposit Insurance and Liquidity,” Journal of Political Economy, Vol. 99, No. 3, pp. 401–419.
Diamond, Douglas, and Raghuram Rajan, 2009, “Fear of Fire Sales and the Credit Freeze,” University of Chicago Working Paper (May).
Gatev, Evan, and Phillip Strahan, 2006, “Banks’ Advantage in Supplying Liquidity: Theory and Evidence from the Commercial Paper Market,” Journal of Finance, Vol. 61, No. 2, pp. 867–892.
Geanakoplos, John, and Ana Fostel, 2008, “Leverage Cycles and the Anxious Economy,” American Economic Review, Vol. 98, No. 4, pp. 1211–1244.
Gorton, Gary, and Andrew Metrick, 2009, “Securitized Banking and the Run on Repo,” Yale University Working Paper.
Gromb, Denis, and Dimitri Vayanos, 2002, “Equilibrium and Welfare in Markets with Financially Constrained Arbitrageurs,” Journal of Financial Economics, Vol. 66, No. 2–3, pp. 361–407.
He, Zhiguo, and Arvind Krishnamurthy, 2008, “Intermediary Asset Pricing,” NBER Working Paper 14517 (Cambridge, MA, National Bureau of Economic Research).
He, Zhiguo, and Krishnamurthy, Arvind, 2009, “A Model of Capital and Crises,” NBER Working Paper 14366 (Cambridge, MA, National Bureau of Economic Research).
He, Zhiguo, and Wei Xiong, 2010, “Dynamic Debt Runs,” NBER Working Paper 15653 (Cambridge, MA, National Bureau of Economic Research).
Holmstrom, Bengt, and Jean Tirole, 1998, “Private and Public Supply of Liquidity,” Journal of Political Economy, Vol. 106, No. 1, pp. 1–39.
International Monetary Fund (IMF), 2008, Global Financial Stability Report, World Economic and Financial Surveys (Washington, October).
International Monetary Fund (IMF), 2009, Global Financial Stability Report, World Economic and Financial Surveys (Washington, April).
Ivashina, Victoria, and David Scharfstein, forthcoming, “Bank Lending during the Financial Crisis of 2008,” Journal of Financial Economics.
Jensen, Michael, and William Meckling, 1976, “Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure,” Journal of Financial Economics, Vol. 3, No. 4, pp. 305–360.
Kashyap, Anil, and Jeremy Stein, 2004, “Cyclical Implications of the Basel II Capital Standards,” Economic Perspectives, Federal Reserve Bank of Chicago, First Quarter, Vol. 28, pp. 18–31.
Krishnamurthy, Arvind, 2010, “How Debt Markets Have Malfunctioned in the Crisis,” Journal of Economic Perspectives, Vol. 24, No. 1, pp. 3–28.
Laux, Christian, and Christian Leuz, 2010, “Did Fair-Value Accounting Contribute to the Financial Crisis? Journal of Economic Perspectives, Vol. 24, No. 1, pp. 93–118.
Lo, Andrew, 2008, “Hedge Funds, Systemic Risk, and the Financial Crisis of 2007–2008,” paper prepared for the U.S. House of Representatives Committee on Oversight and Government Reform.
Pennachi, George, 2006, “Deposit Insurance, bank regulation, and fiscal system risks,” Journal of Monetary Economics, Vol. 53, No. 1, pp. 1–30.
Shleifer, Andrei, and Robert W. Vishny, 1992, “Liquidation Values and Debt Capacity: A Market Equilibrium Approach,” Journal of Finance, Vol. 47, No. 4, pp. 1343–1366.
Vyas, Dushyantkumar, 2009, “The Timeliness of Write-downs by US Financial Institutions during the Financial Crisis of 2007–2008,” University of Toronto Working Paper.
Xiong, Wei, 2001, “Convergence Trading with Wealth Effects: An Amplification Mechanism in Financial Markets,” Journal of Financial Economics, Vol. 62, No. 2, pp. 247–292.
Additional information
*Zhiguo He is an Assistant Professor at the Booth School of Business, University of Chicago; In Gu Khang is a Ph.D. student at the Kellogg School of Management, Northwestern University; and Arvind Krishnamurthy is a Professor at the Kellogg School of Management, Northwestern University. The authors thank participants in seminars at the Minneapolis Fed, Northwestern University, New York University, UC-Davis, University of Chicago, University of Illinois Urbana-Champaign, University of Waterloo, the National Bureau of Economic Research Securitization Group, and the IMF's 10th Annual Jacques Polak Research Conference for their comments. The authors also thank Viral Acharya, Tobias Adrian, Markus Brunnermeier, John Cochrane, Doug Diamond, Nicolae Garleanu, Pat O’Brien, Pierre-Olivier Gourinchas, Christian Leuz, David Lucca, Anil Kashyap, Ayhan Kose, Raghu Rajan, Amit Seru, Jacob Goldfield, and Hyun Shin for helpful comments and discussions.
Appendix
Appendix
Data Construction and Calculation
Hedge Funds
Asset under Management and Loss Estimates
We obtain AUM from the Hedge Fund Flow Report by Barclay Hedge (2009). Both redemptions and trading losses contribute to the drop of AUM of the hedge fund industry. A significant fraction of these redemptions and trading losses are because of the hedge funds that liquidated all of their positions completely and went out of business. However, only surviving funds report the breakdown between redemption and trading losses, which is a decrease of $161 billion from redemption and a loss of $317 billion from asset trading. Based on this information, we assign 66.3 percent of the drop in AUM to trading losses in all of our computations.
Leverage Information, both Strategy-Specific and Overall Average
For strategy-specific leverage information, we use the TASS hedge fund database which provides measures of leverage across different strategies as of 2006. We assume that this captures the leverage that hedge funds were using in 2007:Q4, that is, before the crisis affected the hedge fund industry.
We do not have strategy-specific leverage information for 2009:Q1. Instead, we rely on Lo (2008) who provides the annual leverage information across the entire hedge fund industry during 2007 and 2008. Although useful, these annual leverages are of limited use to us since it is well known that the credit markets tightened considerably toward the end of 2008 (as reflected in the significant increase of the haircuts in Table 5); this suggests that the leverage of the hedge fund industry in January 2008 must be quite different from the leverage in December 2008. Since we are primarily interested in finding out the leverage ratio for 2009:Q1, we ask if we can combine these two pieces of information (that average leverage ratio for 2008 is 2.3 and that hair-cuts on debt securities rose steadily throughout the year 2008) to arrive at a closer estimate of the leverage in 2009:Q1. Specifically, we ask the following question. What does the final 2008 year-end leverage have to be in order to match two facts: (1) haircuts double over the year 2008 (although we do not take a stand on what the level of the haircuts are); and (2) the average leverage ratio over the year 2008 is 2.3. The answer is a leverage ratio of 1.7 at the end of year 2008. We use 1.7 as the estimate of the leverage in 2009:Q1 for all hedge funds regardless of their investment strategies in all of our later computations.
Main Calculations
As mentioned in the main text, we consider the following three scenarios to estimate the net sale (or purchase) of credit/mortgage-related assets by the hedge fund industry.
Low scenario: only fixed-income strategies hold credit/mortgage-related assets. This strategy has an AUM drop of $91 billion. Given the estimation that 66.3 percent of AUM drop is owing to trading losses, the trading losses in credit/mortgage-related assets are $60.3 billion in this case.
Taking the AUM under fixed income strategies at 2007:Q4 (which is $160 billion) and the leverage ratio of the fixed income strategy (which is 4.5 according to TASS hedge fund database in 2006), we estimate the entire holdings of credit/mortgage-related assets for the hedge fund industry to be about $720 billion=$160 billion × 4.5 in 2007:Q4. We then calculate the asset holdings in 2009:Q1 to be $117.3 billion (we multiply the $69 billion AUM in 2009:Q1 by the leverage ratio 1.7). Taking into account the net repayment of 7 percent on existing assets in Equation (1) and applying the loss estimate of $60.3 billion, we arrive at the estimate sale of $492 billion=$720 billion × (100 percent−7 percent) −$117.3 billion−$60.3 billion by the hedge fund industry in this scenario.
Medium scenario: only fixed-income and macro strategies hold credit/mortgage-related assets. The total drop of AUM is $121 billion under this scenario and the trading loss estimate is $80.2 billion. Our calculation is performed in the same way as above and is omitted here; the only difference is that we do the same exercise as above with macro strategy and combine the resulting net sale with the net sale of fixed income strategy. The estimate sale is $546 billion in this case.
Upper scenario: credit/mortgage-related assets are held by a broad class of hedge funds which includes the following strategies—distressed securities, fixed-income, and macro as well as a fraction of the multistrategy and sector specific strategy funds. To determine the fraction of multistrategy, we assume constant proportionality and assign the proportion of the combined capital of distressed securities, fixed-income, and macro in relation to the industry total capital excluding multi-strategy, other, and sector specific strategies for both times, 2007:Q4 and 2009:Q1. To determine the fraction of sector specific strategy, we assume that it is proportional to the share of two industries in GDP, real estate and finance. Since this broad class of funds is close to the entire hedge fund sector (by the size of AUM), we use the average leverage, instead of sector specific leverage, of the entire hedge fund sector for 2007:Q4: this is 2.8 according to Figure 3 in Lo (2008).
Under this scenario, the total drop of AUM is $514 billion and the trading loss estimate is $170 billion. We then follow the same steps as in the low/medium scenarios to reach the estimated sale of $754 billion of credit/mortgage-related assets by the hedge fund sector.
Brokers and Dealers
The Federal Reserve's Flow of Funds does not provide detailed enough information on the breakdown of assets so that we are not able construct an accurate measure of the credit/mortgage-related assets by this sector. So, we rely on the individual SEC filings of major broker/dealers in the United States instead.
We take the top eight broker/dealers in Table 6 as the entire broker/dealers sector and compute their trading assets based on information from their individual SEC filings. Our goal is to estimate what fraction of the trading assets in the broker/dealer sector can be counted as credit/mortgage-related assets. To this end, we first restrict our attention to the top three broker/dealers (Goldman Sachs, Morgan Stanley, and Merrill Lynch) and calculate the fraction of credit/mortgage-related assets in their trading assets. Then we extrapolate this fraction to the other five firms in Table 6 based on the assumption that Goldman Sachs, Morgan Stanley, and Merrill Lynch are representative of the broker/dealer sector. We take this approach because of three reasons: (1) Goldman Sachs, Morgan Stanley, and Merrill Lynch are the only meaningful “pure” broker/dealers remaining in this sector since Lehman Brothers and Bear Stearns disappear out of the industry during 2008; (2) the other three broker/dealers in Table 6 are subsidiaries of the three largest bank holding companies in the United States, which also own the three largest commercial banks in the United States, and are likely to be under nonnegligible influence by the considerations of the commercial banking operation (which behooves us to focus solely on Goldman Sachs, Morgan Stanley, and Merrill Lynch to obtain a relatively clean estimate for the holdings of credit/mortgage-related assets; and (3) the broker/dealer subsidiaries mentioned in (2) do not report detailed enough information on their holdings of credit/mortgage-related assets because their holding companies report on a consolidated level, which does not provide sufficient information for analysis.
Most of the firms in Table 6 have an item called “financial Instruments owned” on their balance sheet, which includes derivative contracts, U.S. government and agency securities, sovereign debt, corporate equity, MBS and ABS, and so on.Footnote 23 We label this category as “trading assets.” For the top three firms (Goldman Sachs, Morgan Stanley, and Merrill Lynch), whose credit/mortgage-related holdings are our focus, we try to exclude derivative contracts, sovereign debt, U.S. government and agency securities,Footnote 24 and corporate equity to obtain an estimate of “credit/mortgage-related assets.” We include corporate debt because this category includes “other debt securities” such as private MBS, which we are mostly interested in.
Details on how we construct the measure of credit/mortgage-related assets are provided below for each of the top three firms.
Goldman Sachs: Trading assets are “total financial instruments owned, at fair value” in the balance sheet. Detailed break-down of these assets are in Note 3 in Goldman Sachs’ 10-K filing. We compute the sum of “mortgage and other asset-backed loans and securities” and “corporate debt securities and other debt obligations” to be “credit/mortgage-related assets.”
Morgan Stanley: Trading assets are “Total financial instruments owned, at fair value” in the balance sheet with detailed decomposition. We take “corporate and other debt” to be “credit/mortgage-related assets.”
Merrill Lynch: Merrill Lynch has a slightly different reporting system than the first two. From the balance sheet, we sum up “trading assets, at fair value” and “Investment securities” (with detailed decomposition in note 5)Footnote 25 to reach the estimate of “trading assets.” To calculate “credit/mortgage-related assets,” we take “corporate debt and preferred stock” and “mortgage, mortgage-backed, and asset backed securities” from the “Trading assets, at fair value” and add “available-for-sale,” “trading,” and “held-to-maturity” securities from the “investment securities” (with data in note 5). Interestingly, Merrill Lynch reports that 93 percent of these securities are nonagency and agency MBS in the 10-Q filing of 2009:Q1.
As reported in Table 7, these top three broker/dealers have a total sale of $156 billion=$363 billion × (100 percent−7 percent) −$182 billion, before adjusting for the losses of the broker/dealer sector.
Main Calculations
The total trading assets of the industry are $2.601 trillion in November 2007, implying that the rest of the broker/dealer sector is holding trading assets of $1456 billion at that time.
Low scenario (extrapolation based on Goldman Sachs’ net sale): Goldman Sachs had 20.5 percent of trading assets as credit/mortgage-related assets and these assets dropped by 11.7 percent (as percentage of trading assets in 2007:Q4) from 2007:Q4 to 2009:Q1. Thus we estimate that the other five firms must be holding $317 billion ($147 billion) of credit/mortgage-related assets in 2007:Q4 (2009:Q1) under this scenario. Given the 7 percent rate of repayment on existing assets, the sale by the other five firms (before accounting for losses) is $148 billion. Therefore the total sale for the broker/dealer sector is $304 billion=$148 billion+$156 billion under this low scenario. Finally, from Table 3, we note that the sector lost $100 billion on mortgage/credit assets, implying a net sale of $204 billion.
Medium scenario (extrapolation based on the average of the three firms’ net sale): As a group, the three firms had 31.7 percent of trading assets as credit/mortgage-related assets and these assets dropped by 15.8 percent (as percentage of the trading assets in 2007:Q4) from 2007:Q4 to 2009:Q1. We carry out the same calculation as in Low Scenario and find that the net sale is $254 billion in this case.
Upper scenario (extrapolation based on Merrill Lynch's net sale): Merrill Lynch had 38.5 percent of trading assets as credit/mortgage-related assets and these assets dropped by 20 percent (as percentage of trading assets in 2007:Q4) from 2007:Q4 to 2009:Q1. We carry out the same calculation as above to find that the net sale is $307 billion.
Insurance Companies
The data source for the insurance sector is their SEC filings. Similar to the broker/dealer sector, the Flow of Funds data do not give detailed enough breakdown on the assets so that we are not able construct a reliable estimate of the credit/mortgage-related assets by the insurance industry.
We choose the eight largest insurance companies listed in Table 8 and examine their holdings of mortgage and other ABS as reported in their SEC filings.Footnote 26 These eight insurance companies collectively have total assets of $2.136 trillion as of 2007:Q4, which accounts for about 34 percent of the insurance sector with asset size of $6.365 trillion (from the Flow of Funds, the total assets of the insurance sector are $6.365 trillion as of 2007:Q4 (including both property-casualty insurance companies, L116, and life insurance companies, L117)). Our methodology is to extrapolate these eight firms to the rest of the insurance sector.
Main Calculations
From Table 8, the sale of credit/mortgage-related assets, before accounting for losses, is $152 billion=$279 billion × (100 percent−7 percent) −$107 billion. The rest of the insurance sector has a total asset of $4.229. trillion.
Low scenario: extrapolation based on three firms: Berkshire, Travelers, and Liberty Mutual, which have the smallest shrinkage of toxic assets: These three companies have 5.7 percent of assets as credit/mortgage-related assets and the credit/mortgage-related assets drop by 0.8 percent (as percentage of assets in 2007:Q4) from 2007:Q4 to 2009:Q1. Thus we estimate that the rest of the insurance sector holds $241 billion ($207 billion) of credit/mortgage-related assets in 2007:Q4 (2009:Q4). Given the 7 percent rate of repayment on assets, the sale of credit/mortgage-related assets by the other firms in the insurance sector (before accounting for losses) is $17 billion. Therefore the total sale for the insurance sector is $169 billion=$152 billion+$17 billion. Finally, as reported in Table 3, we note that they have lost $207 billion on mortgage/credit assets, implying a net purchase of roughly $38 billion.
Medium scenario: extrapolation based all seven firms in Table 7 excluding AIG: These seven insurers have 8.9 percent of assets as credit/mortgage-related assets these assets drop by 3.1 percent (as percentage of assets in 2007:Q4) from 2007:Q4 to 2009:Q4. We carry out the same computation as in the Low Scenario above and find that the net sale is about $50 billion.
Upper scenario: extrapolation based on all eight firms including AIG: Including AIG, these eight insurers have 13.1 percent of assets as credit/mortgage-related assets and these assets drop by 8.1 percent (as percentage of assets in 2007:Q4) from 2007:Q4 to 2009:Q1. We carry out the same computations as in the two scenarios above and find that the net sale is about $247 billion.
Commercial Banks
The data are from the Federal Reserve's Flow of Funds and Call Reports. We do not have to rely on the commercial banks’ SEC filings as these two data sources provide us a fairly accurate read on the holdings of credit/mortgage-related assets in contrast to the broker/dealer and insurance sectors.
Main Calculations
From Table 10, the purchase of credit/mortgage-related assets, before accounting for losses, is $231 billion=$1.774 trillion−$1.659. trillion × (100 percent−7 percent).
Upper scenario: assign the entire total write-downs and losses of $500 billion to the credit/mortgage-related assets: Then the net purchase is $731 billion.
Medium scenario: assign a fraction of the losses to the security portfolio, based on the IMF's Global Financial Stability Report: The report estimates that the banking sector would eventually suffer losses/writedowns of $600 billion on loans and $1.002 trillion on security holdings. Using this ratio of 1002/1602, we assign $313 billion of losses to the security holdings. The net purchase is $544 billion in this case.
Low scenario: assign losses based on assumptions about loss rates on the specific assets in banks’ portfolios: Banks hold $1,192 trillion in agency-backed MBS and $467 billion in privately issued securitized assets. Most of the future losses likely arise from the privately issued securities. The IMF's Global Financial Stability Report of October 2008 estimates losses on the outstanding stock of ABS and ABS CDOs to be about 33 percent (IMF, 2008). They report loss rates on CMBS of 17 percent. Taking these numbers as representative of losses on private securitized assets, we assume that these securities fall in value by 25 percent between 2007:Q4 and 2009:Q1. Then, the losses on the private sector assets total $117 billion. We further assume that agency-backed MBS also fall in value by 5 percent, as spreads in this market widen by about 1 percent over the period we are interested in (see Krishnamurthy, 2010). Taken together, the total loss estimate is $176 billion. Therefore the net purchase is $407 billion.