#1Chinese Shadow Banking: Bank-Centric Misperceptions Tri Vi Dang Columbia University Honglin Wang Aidan Yao HK Institute for Monetary Research AXA Investment Managers HKMA August 2014#2Motivation China is second largest economy, biggest trading nation, has largest currency reserves and fastest growing middle class China will be one of the dominant players in global finance Its financial system is in fast transition Financial structure changed dramatically due to fast growth of non-bank finance Credit intermediation outside the formal banking system experienced rapid growth since the global financial crisis Non-bank finance 45% of total social finance (2013) versus Non-bank finance 40% of GDP (2013) versus ~ 0% (2003) ~ 0% (2003)#3Credit Intermediation in China % GDP 60 Banker's bill (Lhs) Entrust loans (Lhs) Trust loans (Lhs) Total credit* (Rhs) Bank loan (Rhs) 50 40 30 30 20 10 2004 2005 2007 2008 2010 2011 2013 * Total credit includes bank loans plus major shadow banking activities including trust loans, entrust loans and banker's bill % GDP 180 160 140 120 100 80#4Definition of Shadow Banking FSB: A system of credit intermediation that involves entities and activities outside the regulated banking system Bernanke: Shadow banking, as usually defined, comprises a diverse set of institutions and markets that, collectively, carry out traditional banking functions, but do so outside, or in ways only loosely linked to, the traditional system of regulated depository institutions Main Categories in US: Money market funds (MMF), loan securitization (ABS, MBS), repos#5International Comparison (2013) Size of shadow banking, as % of GDP and banking sector assets 400 350 ■% of GDP 300 ■% of bank assets 250 200 *Estimates taken from various media, brokerage and academic reports +Measure from Financial Stability Board (2013) 150 100 50 0 UK US Global Germany Average Japan China* China+#6Shadow Banking in China Wealth management products (WMPs) Trust products/loans Entrusted loans Informal lending Interbank market lending Repo Remark: Our paper focuses on WMPs and trust products#7WMPs Sold by banks as off balance sheets transactions and not subject to official oversight Funds raised from investors (high net worth individuals) are mainly invested in interbank lending and interbank bond markets and repo markets (with higher rates) These products are usually structured as short-term investment, typically less than 6 months. Majority is not guaranteed by banks#8Characteristics of Outstanding WMPs (2013) <1 month ■1-3 month 13-6 month 6-12 month 12-24 month >24 month Unspecified Principle and Interest guaranteed Principle guaranteed only No guarantee#9100 4600 Number of Issuance of WMPs 4100 3600 3100 2600 2100 1600 1100 600 Jan 2007 Jul 2007 Jan 2008 3 Jan 2010 Jul 2010 Jan 2011 Jul 2011 Jan 2012 Jul 2012 Jan 2013 Jul 2013#10Trust Products Structured by Trust Companies Sold through banks Funds raised from retail investors (high net worth individuals) channelled to more risky borrowers with restricted access to banks De facto corporate bonds labelled as trust products#11Example China Credit Trust company raised RMB 3 billion through a trust product called "Credit Equals Gold No.1" in 2011 sold to 700 hundreds of high net worth investors through the private banking arm of Industrial and Commercial Bank of China (ICBC). The fund channelled to Zhenfu Energy company for new projects in coal mining industry in Shanxi province and the product promised investors a yield of 10 cent in the next three years per Business model: Zhenfu pays 15% → 10% to investor, 2% to CCT, 3% to ICBC Remark: First high profile near default case in 2/2014 (later more on resolution)#12Research Objective Understanding the rapid growth of shadow banking in China in the context of its overall financial reforms Main Questions (1) What drives the rapid growth of Chinese shadow banking? (2) What is unique about Chinese shadow banking compared to US counterpart? (3) What is the reason for the Chinese system to evolve into different path? (4) What are the risks and the implications for regulation and reforms? → Focus is on the theoretical model of Chinese shadow banking#13Plan of Talk Drivers of Chinese shadow banking (3 common and 2 specific drivers) Chinese system is bank centric and different from US market based system A model of Chinese shadow banking using some new concepts Information sensitivity as a tail risk measure Micro-foundation for why Chinese shadow banking is bank centric System is built on the asymmetric perception of information sensitivity between banks and investors (~variant of "agreeing to disagree" in banking) Steps toward more transparency of tail risks and market based system#14Drivers of Chinese Shadow Banking Three common drivers Regulation on liabilities side (deposit rate ceiling) → Demand for save products with higher yields Regulation on assets side (loan quota, high reserve requirement, loan deposit ratio) Loan demand by risky borrowers that do not get bank finance → Off balance-sheet transactions can circumvent these restrictions → Our paper discusses how it works in China and what is special?#159- Jan 2000 -4 0 2 4 6 8 Real Deposit Rates 10 % Jul 2000 Jan 2001 Jul 2001 Jan 2002 Jul 2002 Jan 2003 Jul 2003 Jan 2004 Jul 2004 Jan 2005 Jul 2005 Jan 2006 Jul 2006 Jan 2007 Jul 2007 Jan 2008 Jul 2008 Real deposit rates Savings deposits Jan 2009 Jul2009 CPI Inflation Jan 2010 Jul 2010 Jan 2011 Jul 2011 Jan 2012 Jul 2012 Jan 2013 Jul 2013#16Two China specific drivers Economic stimulus package (RMB 4 trillion) after the global financial crisis Plus expansive monetary policy created many long-term projects, which demand funding from shadow banking after PBOC tightened monetary policy Endorsement by government Shadow banking is a mean to foster interest rate liberalization For high net worth individuals (WMPs), small savers (Alibaba YuEBao) Financial institutions to become more familiar with market force while keeping banks under relatively strict regulation (dual track framework) → Shadow banking is integral part of overall financial reform#17Comparison between Chinese and US Shadow Banking US shadow banking Regulation Q and high inflation led to creation of shadow banking Demand for save assets (MMF, Repos, Securitization) with higher yields#18° Jan 1961 Jan 1962 2 Jan 1963 Inflation 4 Jan 1964 60 Jan 1965 Jan 1966 00 8 Jan 1967 T-bill rate 10 Jan 1968 12 Jan 1969- % 14 Jan 1970 FFR Jan 1971 Jan 1972 Real Interest Rate in US Jan 1973 Jan 1974 Jan 1975- Jan 1976 Interest rate ceiling (Regulation Q) Jan 1977 Jan 1978 Jan 1979#19MMF in US -5 о 5 1970Q1 10 15 20 % 25 Trillions US$ 1973Q1 1976Q1 1988Q1 -MMMF % of total financial assets Inflation $AUM MMMFs (RHS) 4.3 3.3 2.3 1.3 0.3 -0.7#20Some Important Features of US Shadow Banking MMFS and securitization had been increasing even after interest rate liberalization Size is large: 20-25 trillion US$ in 2008 (MMFs: 4 Trillion, Repos: 8-10 Trillion, Loan securitization: 8-11 trillion) Funding mostly from wholesale capital market Shadow banking exists parallel to banks No deposit insurance De jure central bank is not the lender of last resort for shadow banking system#21Differences China US Products WMPs, trust products, trust loans MMF, securitization Product structure and characteristics Financial design Pooling and tranching (securitization) Backed my high quality loans (MMF) Simple Backed by risky loans (TP) Invests in interbank market (WMP) Investors Retail investors High net worth individuals, firms Institutional investors Financial institutions#22Key Differences Traditional banks China Selling Platform Capital markets Risk Transfers US No effective risk transfer Banks are not liable by contract Through security design ABCP: Pooling of commercial papers ABS: Senior tranche with loss absorption by junior claims#23Two Key Questions Why are commercial banks so dominant in Chinese shadow banking? How does Chinese shadow banking create safe assets out of risky loans and who bears risks? → Our model addresses these two and further questions#24Banking Theory Based on the Concept of Information Sensitivity Observations and central questions A number of financial products can be considered as privately-produced money like securities (certificates of deposits (CDs), senior MBS tranches, ABCP, repo) Agents accept those "private money” at par when transacting and expect to be able to redeem them at par Demand for higher yields Depositors expect their money back plus interests in future; otherwise, bank run To banks, deposit is their liability, which is supposed to be (almost) risk-free On asset side, loans are risky asset. So how can banks convince depositors their money is safe? Or, the face value of the private money issued by banks is stable?#25Remark An important function of banking is the creation of safe (stable, information insensitive) liabilities under regulatory oversight An important function of shadow banking is the creation of safe (stable, information insensitive) liabilities without regulatory oversight#26Information Sensitivity (IS) as a Tail Risk Measure Need an economic measure that links assets (loans to firms) and liabilities (demand deposits) so as to model balance sheet dynamics Dang, Gorton and Holmstrom (2013) define IS as π L = = XH XL max[p-s(x),0]. f(x)dx where x is the asset (loans) with distribution f(x) that backs s(x) a security (liability, demand deposit) p is the price of the security or the amount invested#27Example: s(x) is debt P s(x) IL " P D WMP (x)=min√x.D] P f(x) fi(x) AA f(x) fs(x) X P D X XH π₁ = [max[p—s(x),0] · f(x)dx L XL measures expected loss in low payoff states (tail risk measures)#28Example: s(x) is WMP P s(x) IL " WMP (x)=min√x.D] P f(x) fi(x) A f(x) fs(x) P D P D X D=p*(1+r) Blue distribution: Principal and interest are safe (л=0) Green distribution: Principal is safe Red distribution: Neither principal nor interest is guaranteed X#29Remark We use this measure as a unifying concept to address: What drives the growth of Chinese shadow banking? Why are commercial banks so dominant in Chinese shadow banking? How does Chinese shadow banking create safe assets out of risky loans and who bears risks?#30Commercial banks are reluctant to finance projects with high IS Proposition 1 A commercial bank holds a portfolio of assets (i.e. finance projects) such that its information sensitivity лL≤у where y measures information costs of depositors.#31Proof (Sketch) T=0: T=1: Bank that obtains deposit of w and gives a loan to a firm that invests in a long term project which pays off x at T=2. The first depositor withdraws w from the bank. Since the fund is lent out, the bank needs to attract a new depositor. A bank will only be able to obtain new deposits if new depositors will deposit. This depends on the information sensitivity of the asset of the bank. The project has information sensitivity of л and depositors can learn about the bank before depositing at costs y.#32If the second depositor learns that the payoff of the asset is low he does not deposit which means the first deposit cannot withdraw. Anticipating this, the first depositor either demands for a higher interest rate or does not deposit. In order to avoid information production by late depositors the loan must be information insensitive. We can also interpret л as a measure of “suspicion". If лL is larger than a threshold value y then depositors have more reason to become concerned about how safe their deposits would be. Remark: See Dang, Gorton, Holmstrom and Ordonez (2014) for full model.#33Implications Information sensitive projects (that leads to лL>Y) are not financed by commercial banks. Chinese commercial banks prefer to lend to state owned companies since these loans have a low information sensitivity because state owned companies are implicitly backed by the government. Since loans to small and medium size enterprises and developers have higher information sensitivity per unit capital, they do not get bank loans.#34Corollary 1.1 Lending quota magnifies the shortage of funding for risky projects.#35Numerical Example Consider an economy with two dates (t=0,1,2) and three equally likely possible states (S1, S2, S3) at t=2. The bank has a safe asset and two potential projects with the following cash flow at t=2. S1 S2 S3 Investment amount Project 1 0.5 1 2 1 Project 2 0.4 1 2.1 1 π(Project 1) = (1-0.5) = 1/1/ πL (Project 2) = (1-0.4)= 6 30 π (Project1+ Project2) = 1/3 (2-0.9) = 110 Π#36Illustration (Proposition 1) Suppose Y=6 5.6 Bank will finance project 1 but not project 2. Illustration (Corollary 1.1) Y 11 Suppose > 30, bank will finance both projects and lend L=2. If Q=1, then project 2 will not get funding.#37Demand for information insensitive products and shadow banking as a bank-centric phenomenon Proposition 2 Suppose investors are looking for information insensitive financial products and the information sensitivity of a financial product that is backed by a risky project is π >0. A sufficient condition for "/" Investor 0 is that the seller or distributor L provides credit guarantee. =#38Proof f(x) fi(x) A P D f(x) fs(x) P X Bank has asset y such that y+s(x)≥D for all x.#39Corollary 2.1 Since state owned commercial banks are the few entities that investors trust, the involvement of commercial banks is needed. Proof Investors have expectation about f Bank (y) or that yMin for bank is sufficiently large. In contrast, investors have little information about fTrust (y) or yMin of a trust company. If it is costly to learn about fTrust (y) then investors do not buy from trust company.#40Shadow banking based on the asymmetric perception of information sensitivity Most WMPs and trust products are not guaranteed by banks (by contract). Investors think trust products are safe. Who bears tail risk?#41Proposition 3 Trust >0. In an L Suppose the information sensitivity of a trust product is π/ Bank =πL "asymmetric tail risk perception equilibrium", "π/ Investor 0 and trust = product is sold.#42Proof Use "Agreeing to Disagree" argument. Since banks only distribute the trust product it is not liable. A default of trust product does not affect the information sensitivity of the assets on the balance sheet of the banks. The trust product contributes zero information sensitivity to the bank's portfolio, i.e. π/ L Bank =0. πί Investor L = 0 if they believe banks are liable. From Proposition 2, π/ Bank Since ==πt L Investor =0, investors buy and banks sells trust products.#43Corollary 3.1 If banks and investors have consistent beliefs, then Пі Investor = (1-K)л" Trust share tail risks. π Bank L Trust Πι = KπL and where [0,1] denotes how banks and investors and investors s(x) Bank P IL " Investor WMP (x)=min[x.D] P D X#44Remark Conceptually our notion is similar to dogmatic beliefs in stock market trading E.g. Harris and Raviv (1993), Pearson and Kendal (1995), Geanakoplos (2009) Traders have mutually inconsistent priors and also posteriors High profile example: Herbalife Ackman Pershing Square: "pyramid scheme" Icahn: "very undervalued" → Ex post only one can be right → Ex ante disagreement generates trade#45Examples of asymmetric risk perception equilibrium Example 1 (Trust products) China Credit Trust company raised RMB 3 billion through a trust product called "Credit Equals Gold No.1" in 2011, which was sold to hundreds of high net worth investors through ICBC. Investors believed they were buying something with an implicit guarantee from the bank. There are anecdotal evidences that local bank branch managers told investors that the product is safe. The fund raised by the trust product was channelled to Zhenfu Energy company for new projects in coal mining industry in Shanxi province and the product promised investors a yield of 10 per cent in the next three years. In the end of 2013, it became clear that the Zhenfu cannot pay 3 billion back to the trust company due to deteriorating profits in the coal mining industry. The market became more nervous when ICBC refused to bail out. Under this intense glare, China Credit Trust announced in the last minute that it had reached an agreement with an unnamed third party to sell the shares it held in the Zhenfu so that the investors is offered a deal to recoup their principle and only three percent of interest.#46Example 2 (Yu'E Bao) Before the internet giant Alibaba entered the money market funds (MMF) business in June 2013, the MMF sector was small and did not attract many retail investors. After Alibaba acquired about 50% of the MMF provider Tian Hong and offered MMF types of products through YuE Bao, these investment products sold online gained huge popularity: AUM of RMB500 billion by the end of February 2014. Since Chinese consumers and investors are very familiar with Alibaba and its online market place, they might implicitly assume that in case of default Alibaba will bail out the failed investments products because of reputational concerns. Investors have information about the financial strength of Alibaba that it is able to rescue any failed product although legally Alibaba does not provide any credit guarantee.#47Example 3 (Agency MBS) Ginnie Mae is the only mortgage-backed securities (MBS) issuer with explicit government guarantee. Although there were no such guarantees for Fannie Mae and Freddie Mac before the financial crisis, MBS investors seemed to have implicitly assumed this. As long as the market is functioning well and there were no defaults of the AAA rated Agency MBS tranches, investors may have no reason to question that MBSs were information insensitive. When the losses of Fannie and Freddie accelerated as housing prices continued to decline, the US government took both enterprises into conservatorship in early September 2008 and provided explicit guarantee so as to avoid a potential collapse of the primary and secondary Agency MBS markets (FHFA, 2008). ABCPs also exhibit such features. Despite their off-balanced sheet characteristics banks provide credit guarantees. (Acharya, Schnabl and Gustavo (2013))#48Remark Gennaioli, Shleifer and Vishny (2012) propose the notion of neglected risks by both investors and financial intermediaries. Chinese shadow banking is different. Banks are aware of the risks so they are not surprised since these products are not complicated financial products (see Example 1). Also, banks do not face additional risks since they are not liable by contractual design. Rather than neglecting risks, investors overlook or (intentionally) neglect the contractual clause that banks are not liable. Financial institutions are not the buyers of shadow banking products and they are not traded in secondary markets so the implication for systemic risks is different.#49Towards more transparency of tail risks and market based shadow banking System is built on the asymmetric perception of information sensitivity (tail risks) and thus not sustainable. Since the underlying projects and loans that back WMPs and trust products are intrinsically risky, default risks do not vanish despite the asymmetric perception of tail risks. If banks and investors have mutually consistent perception of tail risks then they have to share it among themselves.#50Proposition 4 Suppose investors are aware that WMPs and trust product are information sensitive and they bear default risks. The more information sensitive the product the higher the required expected return. Proof See Dang, Gorton and Holmstrom (2013a): The Information Sensitivity of a Security#51Corollary 4.1 Market participants need (credible and independent) institutions to determine the information sensitivity of shadow banking products. Proof The argument is similar to the proof of Corollary 2.1. In order to determine the information sensitivity of a financial product s(x) which payoff is backed by project x investors need to determine distribution f(x). Investors typically do not have enough financial sophistication and knowhow to do that. So they need third party institutions to provide information about f(x) and thus the information sensitivity of s(x).#52Corollary 4.2 Tranching can provide a better information sensitivity-return profile and more investment products to investors. Proof See Dang, Gorton and Holmstrom (2013a)#53The European Central Bank and Bank of England (2014) A market for prudently designed asset backed securities (ABS) has the potential to improve the efficiency of resource allocation in the economy and to allow for better risk sharing. It does so by transforming relatively illiquid assets into more liquid securities. These can then be sold to investors thereby allowing originators to obtain funding and, potentially, transfer part of the underlying risk, while investors in such securities can diversify their portfolios in terms of risk and return. This can lead to lower costs of capital, higher economic growth and a broader distribution of risk. → Call for revival of ABS markets#54Remark We started a new project on loan securitization in China and its implications for the effectiveness of monetary policy.#55Concluding Remarks Shadow banking in both US and China was motivated by regulations on interest rates and developed rapidly. However, the systems have evolved into different paths because of different existing financial infrastructure and legal system. Chinese shadow banking is bank centric and invests mainly in information sensitive products, while US shadow banking is market system and mainly invests in information insensitive products. Chinese shadow banking is partly driven by misperceptions, which could add tail risks to Chinese financial system#56Short Term Risks Default risks of shadow banking products can trigger contagious panic among investors Possible collapse of issuance of these products Dry up of funding for risky borrowers and affect economic growth Since shadow banking products are not bought by institutional investors and not used as collateral in wholesale banking there is no immediate direct effects on banking system Remark Regulator can ask banks to provide funding (assets of RMB 150 trillion)#57Some Policy Implications: Correction of Misperception (1) Make "implicit guarantee" by banks "explicit” by requiring banks to bring the information-sensitive assets back on their balance sheets Consequences Increase the risks of the banking sector Bank will reduce funding for risky borrowers#58Some Policy Implications: Correction of Misperception (2) Educate investors about risks by allowing for some defaults WMPs backed by more risky projects should be labelled as having higher default risks. Supportive actions Promote third party institutions (rating agencies, market analysts) Securitization to create information sensitive assets Some wealthy investors should be willing to bear tail risks for higher returns → In our opinion this is more desirable#59Some Policy Implications: Full Interest Rate Liberalization Shadow banking (or other financial innovations) is likely to remain Commercial banks are still reluctant to fund (too) risky projects#60Some Policy Implications: Market Psychology Market psychology is an important determinant of financial stability. China has more than RMB24 trillion currency reserves. The credibility of PBOC is key for maintaining stability. The announcement to save the system whatever it takes can suffice to maintain stability of the financial system.