## 监管
- 为什么要监管: externalities, market power, consumer protection.
- 为什么bank fail: fraction reserve prone to run, highly leverage, 监管opaque公司的容易solvency problem.
- deposit insurance 1. reduce bank runs partially, 2. bank's moral hazard, take excessive risk. Moral hazard解释: 增加 vola, 增加 option value.
- Capital regulation: 降低credit risk loss, 减少moral hazard by skin in the game.
- BS I II: 8% RWA, III: 3% leverage ratio. 因为: RWA扭曲投资目标, 不accurate可被操作。
- capital requirement的负面成本: 增加银行的cost 不能用利息税盾等, lending减少成为social cost.

## Run
- run的原因: business shocked can't repay loan, depositors fear bank insolvency, concern bank health. key concepts: information asymmetry, fractional reserve, sequential service.
- run的减少: deposit insurance, lender of last resort.
- panic-driven run: belief change, 2种均衡: efficient equilibrium(trust, late withdraw), sunspot equilibrium(coordination failure).
- Information-driven run: CJ88, more likely for weak bank(less repayment).
- Empirical: 1. SW96, 平常deposit从fail bank到solvent bank, panic的时候都取. 2.IP12, panic-driven, big run了，unrelated也run, 但没有informed坏消息. 当发生panic, 谁更容易run: uninsured 和关系不好的 depositors.
- Good balance也可能被run: Northern Rock的liquidity危机, 机构赎回, 因为highly leverage死掉了, 还影响whole market, macroprudential and fragility.

## 货币政策
- Monetary policy: 央行影响利率. 方法: open market operation, discount rate, reserve requirement.
- interest rate channel: short rate影响长期利率, cost of capital, investment, consumption. 但 Magnitude, timing, and composition puzzle.
- Balance sheet channel: firm balance sheet condition 影响 demand for loan: Interest -> future cash flow, asset, collateral -> net worth -> moral hazard and agency problem -> external finance premium, borrowing capacity -> investment, consumption. 2. weak or opaque firms suffer more, high cost of capital.
- Bank lending channel: Interest rate increase -> outflow of deposits, reserve decrease -> Bank loans decrease -> 不给公司借钱 -> investment decrease. Shocks on bank capital -> lending capacity impaired; opaque firms suffer more; Weak banks suffer more. 类比HT公司capital和投资的模型，银行capital下降导致投资(放贷)下降.2个条件: 1. 银行capital下降后必须减少贷款而不是其他方面补足. 2.公司不会依赖market finance而不受影响.
- Balance Sheet 模型: firm决定是否screen项目的成功概率, 是否投risk project. full information下只和project NPV有关, asymmetric information下和firm net worth有关. 越低, external finance premium越高, 并且不愿去screen, 只求稳定. 并且take too much risk, 有的负NPV项目被cross subsidize.
- Collateral的feedback effect揭示amplify: Feedback effect, collateral value $\leftrightarrows$ borrowing capacity. Knock-on effect on highly leveraged firms: temporary shock -> sell production tools -> lower productivity -> net value drops

## Real effects:
- Business cycle 解释expansion和recession: 1. real, RBC观点(exogenous productivity shock, bank credit does not matter), 2. credit观点: driven by credit cycles.
- Empirical evidence of BSC: rate导致了cost of capital上升, 尤其是小公司.
- BLC: Khwaja and Mian, 控制firm, 研究不同bank受liquidity shock不同 reduce lending.
- Real effects:
并不一定drop in loan就会导致output缩减, 公司还能转向其他融资方式。但是supply of credit 影响return to investment.
- Real effect on output: PR00: Japanese bank reduce lending -> real estate activity.
ACLW11: loan maturity before onset of 2007 crisis, refinance时降低了investment, 因为higher cost.
CST12: collateral 影响 invest
CR14:还有公司受影响，继而影响employment的。
H18:lending减少，影响GDP的。(Commerzbank)

## 长期关系:
- 机制: 银行先投资获得specific信息, 再利用这个在后来的长期中获取垄断利润. 好: good for small 因为考虑soft information. dark: monopoly rents.
- Benefits of banking relationship: reduce information asymmetry, improve loan terms(前期要得少, 可以根据soft制定条款), signal背书 for other sources of finance(因为知道受银行监管, 安全)
- Hold-up result: 第二期 Bank use monopoly power to extract informational rents. 第一期 Firms受到补贴(来自于银行的竞争).
- SBC: renegotiate to keep, get bigger loan. Relationship by reducing the threat of termination, 纵容了negative NPV projects.
- Market Finance 和 Bank Finance对比: Bank Finance 优点是monitoring effort, 但容易renegotiate.

## 银行竞争:
- 对定价的影响: oligopoly竞争, intermediaiton margin 受银行个数和弹性影响.
- Cournot-Nash equilibrium: take the other's strategy as given, maximize profit + Market clearing condition + symmetric equilibrium.
- 对金融稳定性：competition-fragility观点: 银行投资，竞争让银行利润减少, 只好增加risk. competition-stability观点: investor投资，银行竞争降低了cost of capital.
- Hump-shaped: 小银行太多竞争压力大, 大银行大而不倒, moral hazard.
- 衡量竞争的指标: Lerner index: (MP-MC)/MP, k largest concentration, HHI index: $\sum_{all } s_i^2$
- Shadow bank: credit intermediation outside banking system, 如Lehman. **特点**：short-term, high leverage, 没有监管也没有deposit insurance.
- Alternative finance: direct lending between lenders and borrowers online. 如peer-to-peer, crowdfund, marketplace lending. 特点：没有中间商所以**interest rate low**, access to credit便利, unsecured lending.

## Credit Rationing
- Adversion selection产生的条件: heterogeneous borrower, can't observe.
- Interest rates 对 bank profit 的双重影响: 1. increase returns, 2. impair average quality of loans.
- Stiglitz and Weiss(1981): same NPV, heterogeneous risk.
Interest rate 上升, moral hazard: borrowers choose riskier projects; Adverse selection: good borrowers are discouraged.
- Williamson 87', Costly State Verification.
repayment increase, Default probability increases, audit area enlarges, expected audit cost increases.
$\rho'(R)$, when R large the derivative is negative, thus credit rationing.
- **Solutions to mitigate credit rationing**: 1. collateral, 2. credit registers reduce information asymmetries, 3. microcredit, decrease monitoring costs.
- Microcredit: Group lending(several borrower), joint liability(if one default, others exclude), peer monitoring, frequent repayment(weekly repayment, reduce Information Asymmetry).
critic: better access to durable goods but not business creation(Banerjee 14')

## 合约设计
- Cash flow manipulation may arise: 1. incentives not aligned, 2. can't directly observe private infomation.
- 证明incentive compatible得说两个方向: profit小不会多报, profit大不会少报.
- Coventant: 条款, a clause in a loan contract that restrict the borrower's actions, such as restrictions on firms' capital expenditures(Nini 09').
- Roles of collateral: 1. screening device, 2. incentive to monitor 3. monitoring device(控制工具), 不会misbehave.
- Collateral + HT 模型: 有collateral reduce credit rationing，可以借到更多. Collateral relaxes credit rationing, pledgeable income $R\leq$ increase, threshold $A\geq$ decrease.
- Subprime: 以房子做collateral, 房价上升，借得更多，房价下跌，还不上.

## 代理监管
- 三种information assymetry: Ex ante, interim, ex post: adverse selection, moral hazard, cashflow manipulation. Act(Monitoring in broad sense): screening by the lender or signaling by the borrower; preventing opportunistic behavior and auditing.
- small investors monitoring: fixed cost, inefficient replication, free-riding.
- 银行监管有优势的必要条件: 1.小投资者 2.监管规模优势 3.low costs of delegation: costs of monitoring the monitor is low.
- NPV: expected future cash flows from an investment minus investment put in.
- HT model: 银行监管减轻了 information asymmetry 导致的credit rationing.
- Borrowers' incentive compatibility constraint(ICC):(选Good project 不选 Bad的原因)
- Pledgeable income: 愿意给的最大R.
- Implications: 1.access to external funds 对没资本的公司难！ 2.银行监管减轻了credit rationing, 能从银行借到钱是好signal. 结束关系是坏信号.
- 为什么银行选择存款finance自己？1.economies of scope，要是depositor也贷款, transaction account. 2.被regulator监管容易. 3. funding stability, 有能力管理illiquid assets.

## Liquidity
- 一些假设: shocks independent across individual, no stochastic withdraw pattern.
- autarky: no trade; market allocation: bonds; optimal allocation.
- autarky的解: not on efficient line. 证明efficient: $C1 \pi_1 + \pi_2 C2/R = 1$
- market allocation的解: pR=1, C1=1, C2=R.
- Optimal: $R\rho = \frac{U'_1}{U'_2}$, 在risk-averse情况, C1>1, C2<R.
- Why can't markets achieve optimal allocation? Markets can't open at t=0.
- Implications: under independent shocks, financial intermediaries improve assets allocation, econ growth increase with banks.
- Commercial paper: short-term debt instrument issued by large corporations, 不从bank借钱了.
- Credit line: loan from bank, gives a firm an option to borrow up to a limit at a fixed spread, alternative to CP.
- Why banks well suited to provide liquidity insurance?
synergies between deposit taking and lending(Kashyap 02),存款管理和贷款管理需要的技能一样, stress的时候 CP 市场 dry, 要安全，都去银行 deposit, Credit line use 正向 correlated，银行甚至不用动自己的liquid assets. (Gatev and Strahan 06).

------
### 一句话说清
Ch1的故事: 为什么bank improve outcome: because individuals invest little in long run projects, 因为他们害怕liquidity shock。有了bank，每个人发生liquidity shock的时间不一样，银行能aggregate，又提供liquidity insurance, 能投到长期项目，赚取更高的回报。

Ch2的故事: 银行channel相比于direct finance的好处(monitoring方面): 如果每个人都去监管，inefficient replication.

(如果太麻烦都不监管?)，银行监管，可以降低融资门坎threshold, 更多公司借到钱 relax credit rationing.

两类项目说明moral hazard: 一个好项目，一个坏项目，坏项目能得private benefit, 如果这个private benefit 足够高，我就不管它成不成功，去赚这个private benefit就行，所以要是利率R太高，我就去做坏项目了。 银行的监管成果体现在减少了这个private benefit, private benefit没那么有吸引力，我可以容忍更高的利率R，对银行来说，那你R高一点，我门槛可以放低一点，所以更多公司能借到钱了。

Ch3的故事: 博弈的开始：怎么设计贷款合约:
1.一个让公司不strategically default/manipulate cashflow 的贷款合约(incentive compatible contract): 引入audit.
但audit是有成本的，在ch4中讲Williamson.
2.i)通过抵押物筛选，应对逆向选择。
高风险的人容易玩脱，所以他们不愿意有抵押物被没收。
ii)有抵押物可以借到更多的钱，对应上面即使失败都能保障的private benefit, 抵押物是在另一个方向，更鼓励你去成功，否则就将惩罚。你相当有自信，你愿意回给银行的钱更多，银行的贷款门槛就降低了。(relax credit rationing)

Ch4的故事: 银行将利率设定在低于市场均衡水平的原因:

利率上升对银行的双重影响: 1.增加repayment, 2.影响贷款合约的成分, 降低average loan quality. 会经常遇到违约，银行利润受影响, 并且要执行audit, 增加成本。所以银行不会设利率不会到无穷大。

怎样影响贷款合约组成：
- Adverse Selection: 利率过高，只有那些谋取private benefit 或者Bold的人才愿意借.
- Moral Hazard: 对borrower来说更勇敢, return无上限, limited liability有下限保障, 但对银行刚好相反, 如果你违约了，银行只能被迫拿个collateral, 拿不到它想要的(1+r)B.

银行作为贷款的供给方，会出现backwards bending supply (profit - r plot).
有可能供给需求曲线不相交了，就银行说了算，定到max profit的利率，但是低于市场均衡，所以出现了credit rationing.(有想借的借不到, supply < total demand)

Ch5的故事:
银行怎么赚钱: 银行的竞争是imperfect，在savings market和loan market上都能赚取intermediation margin。他们还能有一些local monopoly power.(因为走路费腿，大家不会跑大老远去).
银行竞争体现在利率报价上。

银行的竞争对金融稳定性产生影响，而且是hump-shaped的。小银行太多，为了在激烈的竞争脱颖而出，银行被迫risk-taking. 如果竞争太少，大银行大而不倒，他们会攫取这种救火队的好处，moral hazard, risk-taking.

还有种相对的观点认为银行竞争降低利率，firm容易活下来，就不会去从事危险项目.

银行也受到圈外的竞争，shadow banks, alternative finance(peer to peer).

怎么越说越长....

Ch6的故事:
和银行建立长期关系的好处: 作为signal背书, 自己向银行提供的soft information, 降低了IA, 降低信贷约束，对小公司尤其好！

长期关系的陷阱:
对于公司的tricky: "Hold-up problem", 套牢。建立关系之初，银行需要先投入，取得你的信息，你要转到其他银行，其他银行也有这个cost，所以他们不会给你更好的deal. 现在的银行知道这一点，它能靠信息垄断赚取informational rent, 甚至它预期到这种长期的hold-up，可以在一开始用低利率吸引你，再在长期中recoup回来。

(My confusion: 可每个新银行都可以提供这样的初始低利率啊？而不是模型中用zero-profit推出的不利情况)

对于银行的tricky: soft budget constraint, 遇上了一个中道崩殂的项目，renegotiate被迫接手烂摊子。

Ch7的故事: 银行挤兑
run: 是指run out of money!

Bank天生脆弱: 因为fractional reserve system. 挤兑取决于公众的belief! 公众对于银行的健康状况也是有信息不对称的，一旦公众太敏感，就去排队！coordination failure导致的sunspot equilibrium, panic-driven bank run.

另一个情况是有内幕消息(知道银行健康状况)的人, 对于uninformed late consumer, 他们根据排队的长队判断，当队伍长的时候，可能要么early consumer太多，要么有内幕消息的late consumer也去排队, 他们权衡一下，也可能去排队。 如果是前者(即银行状况良好)，uninformed late consumer却去排了队，informed late consumer被迫博弈也去排队，否则就啥也拿不到，inefficient run.

Ch8的故事: 货币政策传导
中央银行怎么通过货币政策实现mandate(如price stability, employment, economic developments)?

利率线: 央行调控short-term rate, 自动影响 long-term rate,
cost of capital -> investment, consumption
调控手段: open market operations, discount window, reserve requirements, unconventional tools: forward guidance, quantitative easing.

由于利率线的力度不够，不及时，还有
信贷线:
信贷1线：firm side:
interest rate -> asset value and collateral value -> net worth -> agency problem and moral hazard -> borrowing capacity -> investment

这里collateral还会echo, 对highly leveraged firms,已经融不了新资了，他们被迫卖掉生产工具，降低生产力，再次降低collateral value.

信贷2线：bank side:
interest rate increase -> outflow of deposits, researve 下降(业务收缩，不需要那么多保证金了) -> Bank loans decrease -> 不给公司借钱 -> investment decrease
更简洁的解释是看银行的资产负债表，loan是银行的资产而deposit是银行负债，deposit减少，银行吸收贷款减少，因此能贷出的loan就少了.关键就在于利率提升导致deposit的减少.
*这里的interest rate是指央行的指导利率而不是银行给出的利率。

counterintuitive的地方是： outflow of deposits 是因为存款利率不会按比例增加，大家可以选择去买Bond或者货币基金。

(my confusion: deposit 影响 loan 的道理是？这里解释说公司的投资需求能否满足取决于初始资本，银行的投资是它的loan,取决于它的capital)

Ch9的故事:
Business cycle的形成：
real side 认为是 productivity 受到正的(比如innovation)或者负的冲击。
credit side 如Ch8所讲的.

Ch10的故事:
microprudential regulation的例子:
deposit insurance: uninsured depositors更容易去挤兑，有了
deposit insurance, 人们不会去挤兑银行了，但银行会借助这种力量take more risk, moral hazard.
capital requirements: 为什么要设定资本充足: 如果银行拿的都不是自己的钱，作为投资方，难免 moral hazard 干坏事, 得让他"skin in the game".