原创 2015-11-02 季天鹤央行观察央行观察

作者：季天鹤，方正中期研究院研究员，央行观察专栏作家

引言：和挤提一样，暴跌不是随机分布的，不是概率的，而是必然的。状况一旦具备，便是迟早的事。

银行与金融市场常被看作两种平行的融资体系，但使用银行的分析框架，有助于更好理解金融市场，特别是股灾等暴跌。本文试图从准备金的角度，来讨论一下各市场中蕴含的准备金结构。

经典金本位下，储户存款带来黄金，银行用黄金放贷，使储户存款规模大于银行库存黄金，即准备金率小于100%。而银行为了赚取利差，往往使贷款期限长于存款，因而即使考虑每日归还贷款产生的黄金流入，银行也无法应对一日之间大量储户兑取黄金。而储户担心不及时取款将无款可取，于是蜂拥而至产生挤提。

客户挤提的原因，除了准备金率小于100%，还在于银行的兑付制度。银行对于活期存款保证随时兑付黄金，对于定期存款，有时也在罚息的条件下予以随时兑付。而兑付时，黄金的货币价值，与存款的货币价值相等。在银行持有的黄金价值小于存款价值时，先提款的储户可以获得与自身存款相等的黄金，而后提款的储户什么也得不到，因此储户必须努力“挤”到前面提款。

准备金率是为了减少挤提发生频率而采取的一种措施。对于存款规模相同的两家银行来说，拥有更多黄金作为准备金的银行，比起拥有较少黄金的银行来说，即使一段时间内面临的储户的提款情景完全一样，前者比后者更能承受严重的黄金流出。当然，准备金率不是万能的，毕竟银行持有的黄金依然小于存款规模，但准备金率高总比低安全。

提款使储户存款减少黄金增加，资产总额不变。银行在储户看来，便如同一个“存款做市商”，储户可以将黄金换成银行存款，也可以将银行存款换成黄金，兑换比率总是1:1，既不会变化，存取也没有价差。与之相对，典型的做市商报价分买价和卖价，两者既不断波动，又有价差。

典型的做市商持有证券和资金，并动态地不断买入和卖出证券，同时调整价格，使得其净持仓处在某个合适的水平。银行也希望其持有的黄金相对于存款处于某个合适水平，原因在于黄金不产生利息，而对客户存款要支付利息，黄金相对存款比例越大，在其它资产收益和负债成本相同的情况下，银行尽管更安全，但利润也越小，需要权衡取舍。

为什么银行要维持黄金与存款的1:1兑换率，而不能像典型做市商一样调整价格？问题的答案首先在于，做市商买卖的是两种自身资产，但银行吸收存款是同时扩张自身资产和负债，偿还负债当然要等量。有观点更认为，存款本来就不是借贷关系，而是存管关系，所以当然要保持1:1，本文不讨论这种观点。

另一个答案，在于最大限度地扩展银行负债的流通能力。通过维持黄金与存款的固定比例，所有银行的负债全都可以互相转换，表现为储户可以将在一家银行的存款汇到在另一家银行的账户。银行存款也就获得了与黄金相同的流通能力。毕竟，历史上货币长期是黄金白银等贵金属，银行存款并不是货币。

经历了漫长的过程，银行维系黄金兑换的能力越来越强，银行存款凭其优点战胜了人们的忧虑，这些优点包括有利息、不会像黄金一样被磨损，不会丢失被窃等。但黄金的历史遗迹依然存在，在中国央行的信贷收支表，以及国际收支平衡表中，都还保留了黄金占款与货币黄金的项目。

如果银行不维持固定的兑换价格又会如何？现实的例子是央行的外汇操作。银行持有人民币准备金，可以找央行兑换外汇。央行基本上保证银行兑换外汇的需要，但并不承诺一个固定价格。中国央行持有3.5万亿美元的外汇资产，但负债却有30万亿人民币。如果按照1美元兑换6.5人民币的水平卖出外汇，那么当所有的美元都卖出之后，负债减少了22.75亿人民币，但剩余的7.25亿人民币就无法兑换了。

央行被挤提了，一部分银行满意而归，另一部分亏得一干二净，原因是央行的准备金率在6.5的兑换率上不足100%。所以央行还有另外一手，即改变兑换比率，使将1美元兑换10人民币，那么央行在兑付了30万人民币之后，还能略有剩余。所有的银行都承担了汇率变化导致的损失，而不会像挤提的情况那样出现分化。央行通过调整兑换比率，使自己的准备金率超过了100%，当然就不会再遇到挤提的问题，只需要淡定兑付即可。这里我们看到，调整兑换率也就调整了准备金率。

而如果提供美元的一方，不再是单一的央行，而是许许多多的人民币买家，我们就进入了一个金融市场的情形。在这个市场上，人民币卖家能买到的最多的美元数量，就是美元卖家持有的全部美元。如果目前的价格使得人民币能够兑换的美元，多于美元卖家持有的全部美元，意味着美元对于人民币的准备金率是低于100%的，这个市场便存在着在目前价格上的挤提风险。

上述便是金融市场中的准备金结构。很多时候金融市场看起来是稳定和有效的，正如银行在没有被挤提的时候也是稳定和有效的。金融市场中的买单和卖单频繁成交，正如储户和银行间不停进行的存取款业务。金融市场的成交价格上上下下的温和波动，正如储户和银行间的存取款兑换比例稳定维持在1:1。

但金融市场和银行一样，都存在着挤提的因素。在金融市场里，虽然买方和卖方的地位是对等的，交易方向是对称的，但资金和证券的地位是不同的。大多数投资者和大多数人的目的，都是通过持有证券而获得更多的钱。类似“用存款换现金”的“用券换钱”需求，时时刻刻存在于金融市场。

金融市场中使用杠杆的市场主体从两个方面加剧了金融市场的挤提因素。一方面，这些主体借入资金买入证券，总是不时面临到期还钱的约束，因此比无杠杆的投资者更频繁地产生“用券换钱”的需求。另一方面，杠杆投资者对价格波动的容忍度小于无杠杆投资者，因此在同样波动分布下，杠杆投资者比无杠杆投资者更容易产生“用券换钱”的需求。

金融市场中的价格，影响了潜在的准备金率。如果持有3.5万亿美元的央行宣布，现在1美元兑换1人民币，那么他很容易遭到挤提，因为3.5万亿人民币的提款需求就会让央行违约。但如果央行宣布1美元兑换10人民币，则央行就要安全很多。股票市场也是如此，高股价很容易套走很多钱，使此后市场的买入能力一下子降低了，结果就是暴跌。

暴跌在固定价格的条件下表现为固定价格难以维持。外汇市场经常出现这种情况。而在股票等市场，暴跌表现为偏离正常的波动区间。暴跌的原因一方面可能和挤提无关，如市场基本面变化等，使得市场公认的股价（股票兑换银行存款的比率）迅速下降。但另一种则可能与挤提有关，即市场情绪变化（类似储户突然都对银行丧失信心），或买方没钱了（类似银行刚应付了一笔大额取现），卖方拼命想先卖券变现，导致挤提式暴跌。

与暴跌相对的是阴跌。与其一支股票的价格从100元迅速跌至70元，使股东甲持有的100股的市值，从10000元跌到了7000元，不如先让股票从100元跌到90元，让甲亏掉1000块钱卖给买家乙，然后股票继续跌，乙亏掉1000块钱卖给买家丙，买家丙亏掉1000块钱卖给买家丁。此时股价也是70元，但甲乙丙每人各只亏掉1000元。

银行挤提类似暴跌，而银行缩表类似阴跌。银行的利差利润，反应为贷款人归还贷款时缩减的存款较大，而银行发放存款利息增加的存款较小。例如银行发放了100元贷款给甲，甲并不提现而是存在银行。一段时间后，银行要给甲利息1元，但甲要偿还银行利息6元，于是银行在收回100元贷款时，社会整体存款减少了105元，即多减少了5元，而这有助准备金率的提升。

那么如何应对挤提和暴跌？对银行而言，如果央行和其他银行能够提供足量的黄金，供被挤提银行的储户提取，那么挤提便被顺利化解了。所谓足量，最简单的可以是100%的准备金率，即黄金和被挤提银行的存款等额，而对证券市场而言则是银行存款与市值等额。股票市场市值可以达到50万亿。如果想要所有市场主体都在这个市值上兑取银行存款，那么100%的准备金率意味着，救市当局要拿出50万亿的存款，把市场上的股票按现价照单全收。拿出几千亿应对几十万亿的挤提，用处非常微弱。

我们可以比对央行救汇市，8月份市场主体购汇近8000亿元人民币，央行就甩出了8000亿的等值外汇给银行，方保人民币汇率稳定不坠。如果只拿出几百亿人民币的等值外汇，仅数量上便应对不了8000亿的需求，结果要么是数量管制，人民币跌停，地量成交（例如2012年11月的人民币涨停），要么就是人民币跌到合适的水平。

如果市场是一个稳定的市场，没有大起大落，也没有戳泡沫和救市，那么金融市场的准备金率和稳定的兑换率，使得很多证券都可以货币化。前面提到，银行通过让存款和黄金自由等价兑换，成功地使存款上位，代替了黄金的地位，成为可以广泛购物的货币，那么金融市场的证券，也可以利用这一“现金-存款”的结构，具备货币的流通能力。

“现金-存款”结构包含两个阶段。第一个阶段是市场主体依然频繁地取现，即现金比存款用处大。90后可能已经离存折太远，但用过存折的储户们对此必定不陌生。第二个阶段是市场主体直接收付存款，现金的用途越来越少，现在我们正处于这个时代。取现和存现的操作相对过去少了很多，但依然重要，因为这是维系现金与存款1:1兑换关系的关键，正如期货交割量哪怕只是持仓量的小部分，也是使期货和现货价格收敛的关键。

达到第二阶段的方法，一是主观的，即市场相信取现是可以做到的，而这当然有自我实现的因素，即只要大家都相信，这事儿就值得相信，而一旦开始不相信，就不值得相信。二是客观的，即人们能够运用存款进行支付。例如就在几年以前，两个人之间付款一般还需要现金，均摊吃饭要计算找零，现在全都可以使用手机支付了，这部分现金需求就没有了。

一个已经成功使用这一结构上位的是货币市场基金份额，即余额宝。余额宝持有的银行存款规模小于其基金规模，即其份额的准备金率小于100%。这一方面意味着份额持有者赎回超过其在银行存款规模的余额时，余额宝需要卖出其他资产，在现有银行存款基础上获取更多银行存款，才能应对全部赎回，而另一方面也意味着余额宝把份额规模扩大到存款规模以上，供份额持有者来购买商品。

余额宝的策略是想搞出一个自己的货币王国，这样他就不会遭遇到用户取现的过程，以致使余额宝在银行的存款流向企业个人在银行的账户，而这会降低余额宝的准备金率。在这方面他已经取得了很大的成绩，淘宝和天猫都是他的。而只要降低了取现需求，准备金率也就可以稍低，余额宝就能放心地通过“蚂蚁花呗”来创造信贷和更多余额。

但毕竟余额宝的规模只有几千亿。所以他还要扩张，在和其他的商家合作。最好的结果是对方把银行存款账户关掉，而在余额宝开户。次好的结果是对方保留银行存款账户，但不每日结算，而是定期结算，这样余额宝就能更好地经营自己的流动性。

在金融市场中，较有上位可能的是债券，其特点在于价格比较稳定，并且给投资者两类选择，一是在金融市场上卖出变现，二是等待到期让发行人偿付而变现。目前还不能直接使用债券购买商品或金融资产，但市场普遍认可卖出短期优质债券换取银行存款的能力，如同认可将银行存款兑换为现金的能力一样，只不过债券兑换时还会面临少许价格的波动。

流动的债券市场如正常运转的银行。银行有人存款有人取款，实现了现金出入的动态平衡，出入的波动体现为准备金的增减，足够的准备金率使现金进出在一般情况下不会导致准备金枯竭，也就不会产生挤提。同样，债券市场里有买家有卖家，也存在现金出入的动态平衡。

如果市场主体一得到现金，就把现金用来买债，以供卖债一方获得现金，用于购买商品，而卖出商品一方又把现金用来买债，那么债券市场就像一个大银行一样，总有一些持有现金的人等待买券，如同银行持有准备金；而总有一些持券者卖券，如同银行储户提款用来购买商品。

因此正如“现金-存款”的体系中，两者目前都被看作是货币量的一部分一样，我们也可以把“存款-高流动债券”体系中的两者也都看作是某口径下货币量的一部分。存款因能够提取现金而流通，而高流动债券因能够换取存款而流通。如果我们坚持使用存款作为货币量，而不扩展至债券，那么高流动债券便加速了存款的流通速度。

现在中国的情况是债券市场流动性不好，购买债券很多只能持有到期。在这个情况下，能否作为抵押品就非常重要。如果能作为抵押品，那么质押债券虽然不如卖出债券那样，一劳永逸地得到银行存款的所有权和使用权，但至少可以在一段时间之内获得银行存款的所有权和使用权，这总比死等到期日偿付要好。

除了现货市场和抵押品，期货市场等衍生品市场的发展，也为债券的流动性增加了渠道。期货合约可以有非常多的种类和交割日，适用于很多债券，同一发行人发行的不同期限和票息率的债券都通过转换因子来标准化，用于到期交割。成为期货市场合约的可交割券，或者开发覆盖很多现券的期货合约，对于客观上增强债券的流动性，也可以起到正面的作用。

到期偿付是债券市场的一个重要特点，使债券市场兼具银行与金融市场的特色。企业发行债券获取银行存款，正如银行发行存款获取黄金。如果企业发行的债券是随时可售回的，则其与银行存款的差异将更加微弱。债券市场的投资者，也就同时面临了两个准备金体系，一个是发行企业的，一套是债券市场的。

这两个准备金体系有时会有矛盾。例如处于准备金率小于1:1的企业，大量发行债券获得等量银行存款后缓慢支出，这虽使企业的准备金率提高，但使金融市场上准备用于购买债券的资金减少，而债券总量由于企业的发行而增加，市场的准备金率就降低了。解决矛盾的办法恐怕是让企业将获得债券后的银行存款重新投入债券市场，而不是让债券市场遭遇“现金流出”。当然，企业支出存款之后，获得存款的人也要重新投入债券市场，不然也会使债券市场遭遇“现金流出”。

上面的情景是基于一个没有银行的债券市场。银行参与的债券市场中，银行存款会因银行的购债行为而增加，因此我们除了企业和债券市场的准备金率之外，又要考虑银行的准备金率。危机中，企业准备金率不足，实力不足以还债的前景引发市场抛债，银行此时也会卖出债券，卖给非银行机构会减少自己负债侧的存款，或者减少其它银行负债侧的存款，总之是缩小存款规模。如果是卖给另一家银行，那么会减少另一家银行的央行准备金，降低买家银行的准备金率。

而银行存款的减少又会降低债券市场的准备金率，让抛债进程成为恶性循环，即银行抛债缩表，导致非银行机构能通过抛债获得的钱也越来越少，结果就是债券价格越来越低。而债券的跌价会使市场对于银行的信心也受到打击，于是银行遭到挤提，现代的版本便是同业收回拆借而客户提取存款，银行的准备金率也将迅速降低。

拯救这一局面的办法，乃是央行直接从银行和非银行机构买债，这样一方面增加非银行机构在银行的存款，一方面增加了银行在央行的存款，同时还减少了银行和非银行机构持有的债券，最大限度增加了债券市场和银行两个体系的准备金率。给非银行机构放贷能增加银行存款和银行在央行存款，但并不能减少债券规模，效果弱于前者。给银行放贷或购买银行资产，只增加银行体系的准备金率，但银行仍然可以选择继续收缩资产和负债，银行存款规模未必得到有效扩张，债券市场的准备金率未能提升。

相对于债券市场，股票目前还未能通过“现金-存款”结构上位，因为兑换比例实在是太不确定，股价的变化也太大。但股价的大幅波动使价格因素可以对经济产生较大影响。我们可以考虑一家银行有很多储户，突然有一天这家银行的储户发现，自己在银行的存款余额，变成了前一天的10倍。储户的第一反应可能是取现，但也有可能是取现后再重新存入。例如一个储户拿膨胀了的存款取现买了一辆车，而卖车的一方又将钱存回了银行。对于银行而言似乎一切都没有变化，但对于实体经济而言，一辆车就这么卖出去了。

股票泡沫，增加的不是银行存款，而是兑换银行存款的能力，正所谓股价上涨不是印钞胜过印钞。每个股民持有的股票股价上涨，增加了卖出后能够得到的存款规模。股民卖出股票获得存款买入汽车，汽车卖家收款后用来买入股票，这和前面银行的例子是一样的。如果股票可以直接用来买车，则相当于汽车买家直接把股票卖给了汽车卖家，同时买了一辆车。如果股票处在较低的估值，这辆车就不会被卖出去。

但我们前面提到了，兑换率和准备金率是非常重要的。股票的高估值意味着高兑换率，高兑换率意味着对准备金率的更大扰动。高估值股票卖家买掉股票就会去买汽车，从股市中抽出买汽车那样多的钱，而股票估值较低时，卖价卖掉同样多的股票，只能买得起自行车，也就从股市中抽出买自行车那样多的钱。

而汽车卖家收款后也未必用于买股票，而完全可以用来在股票市场之外使用，这就是银行过去面临的现金流出问题。所以估值越高，维持高估值也就越难。泡沫越久越大，也就越容易破，破的时候的响声也越大。泡沫市场中的最优策略，当属圈钱，泡沫越大，圈钱越划算也越踊跃，但圈钱本身就是戳破泡沫的针。

最后来看外汇市场。由于是两种货币在交易，不存在上位成为货币的问题，但是否还存在挤提？我们需要考虑市场主体更想换得什么，或者市场主体被迫要换出什么，一个重要的因素就是债务。在外汇市场中，有些货币是融资货币，一般这些货币都是低息的，并且低息前景还非常稳定，典型的就是日元。市场一般不会借美元买日元，而会借日元买美元，因此市场上总是存在潜在的买日元的需求。

所以在美元日元这个市场里面，美元不会被挤提，因为大家都持有美元头寸，但日元会被挤提，因为大家都抢着购买日元还债。如果两个非银行主体借贷了日元，那么买入日元还债，意味着还债方把日元交给债权人，日元本身并不会减少，只是在流动中消除债务。而如果考虑银行发放的日元贷款，那么归还贷款意味着日元的消失和银行的缩表，日元的流动到此为止。

如同配资炒股当中，炒家疯狂卖股票还银行贷款后，这部分钱就没了一样，外汇市场里面，钱也会在归还银行贷款中消失，而不光是在现有价格上找不到卖家的问题。所以外汇市场出事儿一般会由央行出面，最有效的方法是央行货币互换，凭空创造出各自的本币借给对方国家，一方面让钱在流动中消除各类债务，同时通过自身的扩表来弥补银行的缩表。

从这个角度看，银行直接炒股会造成更大的影响。除了外汇市场，前面提到的债券市场中也体现出，银行在公允价值的逼迫下抛出债券造成的存款紧缩，减少了非银行债券投资者的市场的资金，降低了债券市场的准备金率。如果银行参与股票市场，股票的波动比债券更大更频繁，银行缩表的影响恐怕更大。

除了债务的币种之外，所有者权益的币种也非常重要。对于没有负债的投资者而言，有些用美元计利润和所有者权益，有些用欧元计利润和所有者权益。发生危机时，美元计利润的投资者希望减少美元计价损失，便会减少欧元敞口，而欧元计利润的投资者希望减少欧元损失，便会减少欧元敞口。如果两类投资者都投资人民币的话，那么两方恐怕都会抛出人民币了。

文章至此该作小结了。如前所述，银行和金融市场中均存在着弹性的自由与刚性的约束。弹性的自由在金融市场表现为价格的温和波动，但刚性的约束意味着买家再也无法支持原有的卖出流，而市场一发现这一点便会开始暴跌。刚性的约束在市场稳定时隐藏于幕后，正如银行不需要100%准备金也可以运行。

一旦出现事件，例如大的证券卖出或者大的存款提款，哪怕事件的行为主体并不是出于对市场信心或银行信心的崩溃，而仅仅是由于自己的其他需要，也会导致市场的暴跌和银行的挤提。金融市场的准备金率当然无法计算，因为证券买方的买入意愿总是不确定的，愿意买入的价格也是时时波动的，但知道市场中存在这一刚性约束是非常重要的。有准备金不是万能的，没有准备金是万万不能的。

科技的发展使金融市场中可流通的工具种类以及流通性都大幅度增加，但随着流通性的增加，金融市场中出现了越来越多的准备金结构。这种准备金结构可能是银行式的，即建立在一家机构的资产负债表上，也可能是金融市场式的，有很多买家和卖家，但两种方式最后均形成了某种“现金-存款”的准备金结构。这一结构既导致了挤提不停的发生，又使得很多金融工具可以借机上位而成为货币。

由此我们可以重新审视中国和美国。中国银行存款规模较大，伴以较小的金融市场和金融工具规模，即较多的“银行存款”和较少的“银行存款兑换能力”；美国银行存款较少，但对应了巨大的金融市场和庞大的金融市场规模，即较少的“银行存款”和较多的“银行存款兑换能力”。中国的经济运行是存款持有者直接支出存款购买商品劳务，而美国的经济运行则是金融工具持有者先卖出金融工具，获取买者的银行存款，然后购买商品和劳务。

这一结构也把交易者分为了两类。一类交易者如同在推土机前捡钱，满足于“存款”带来的收益，承担“挤提”带来的血本无归风险。另一类交易者开着推土机，随时准备发动一个“挤提”，或在偶然到来的“挤提”中趁火打劫，把推土机前的捡钱者一举铲掉。关注金融市场中的准备金结构，有助于理解暴跌的产生，如何避免暴跌造成的亏损，以及利用暴跌获利。暴跌不是随机分布的，它是到达某一状况后的必然结果，如同一个拿着杯子的人，如果杯子的水很少，那么即使人跑步水也不会撒，但如果水很多很满，那么即使人慢慢走路，水也可能会溢出，更别说背后突然跑过来一个人，一把猛推。

Introduction: Similar to a bank run, a sudden stock market crash is not a random distribution, nor a matter of probability, but rather a certainty. Once the conditions are in place, it's only a matter of time.

Banks and financial markets are often seen as two parallel financing systems. However, using a banking analysis framework can help better understand financial markets, especially events like stock market crashes. This article attempts to discuss the reserve structure inherent in various markets from the perspective of reserves.

Under the classical gold standard, depositors bring in gold, and banks lend using that gold, making the total deposits greater than the bank's gold reserves, resulting in a reserve ratio of less than 100%. Banks often extend loan maturities beyond deposit terms to earn interest differentials. Even considering the daily inflow of gold from loan repayments, banks cannot handle a sudden rush of gold withdrawals in a single day. Depositors fear being unable to withdraw in time, leading to a rush for withdrawals, similar to a bank run.

The reasons for customer withdrawals are not only due to reserve ratios less than 100%, but also the redemption system of banks. Banks guarantee immediate redemption of gold for demand deposits and sometimes for time deposits under penalty conditions. During redemption, the monetary value of gold is equivalent to the monetary value of the deposit. When the value of gold held by the bank is less than the value of deposits, those who withdraw first receive an equivalent amount of gold to their deposits, leaving nothing for those who withdraw later. Therefore, depositors must rush to withdraw early.

Reserve ratios are a measure taken to reduce the frequency of runs. For two banks with the same deposit size, the bank with more gold as reserves can withstand more severe outflows compared to the one with less gold, even if facing the same withdrawal scenario over a period. Of course, reserve ratios are not foolproof, as bank-held gold is still smaller than total deposits, but higher ratios are safer than lower ones.

Withdrawals reduce deposits but increase gold holdings, keeping the total assets unchanged. From the perspective of depositors, banks act as "deposit market makers." Depositors can exchange gold for bank deposits and vice versa at a fixed exchange rate of 1:1 without any price difference. In contrast, typical market makers have bid and ask prices with fluctuating rates and spreads.

Market participants using leverage exacerbate withdrawal factors. They frequently face constraints due to repaying borrowed funds to buy securities. Moreover, leveraged investors have lower tolerance for price fluctuations compared to non-leveraged ones. This makes them more likely to demand the conversion of securities to cash during similar price volatility.

Market prices affect potential reserve ratios. If a central bank holding $3.5 trillion in reserves announces a fixed exchange rate of 1 USD to 1 CNY, it's vulnerable to runs as the demand for withdrawing $3.5 trillion in CNY exceeds its reserves. However, if the central bank announces a rate of 1 USD to 10 CNY, it becomes much safer. Similarly, in stock markets, high stock prices can lead to siphoning off funds, reducing buying power and potentially causing a crash.

Crashes occur when fixed prices are difficult to maintain. This often happens in foreign exchange markets. In stock markets, crashes deviate from the normal range of fluctuations. Crashes can result from fundamental changes in the market or emotional shifts (similar to a loss of confidence in banks), causing a run-like scenario leading to a crash.

In contrast to a steep crash, there's a gradual decline known as a "shadow drop." Instead of a stock's price swiftly falling from 100 yuan to 70 yuan, causing shareholder A's 100 shares' value to drop from 10,000 yuan to 7,000 yuan, it might be preferable to let the stock decrease from 100 yuan to 90 yuan. This way, shareholder A incurs a loss of 1,000 yuan when selling to buyer B. Then, the stock continues to drop, causing B to lose 1,000 yuan when selling to buyer C, and so on. Eventually, the stock price reaches 70 yuan, but each of the participants, A, B, and C, only loses 1,000 yuan.

Bank runs are similar to steep crashes, while balance sheet reduction resembles shadow drops for banks. The bank's net interest margin reflects a greater decrease in deposits from borrowers repaying loans than the increase in deposits due to interest payments. For instance, if a bank lends 100 yuan to individual A, who keeps it in the bank, the bank has to pay A 1 yuan in interest, but A owes the bank 6 yuan in interest. When the bank reclaims the 100 yuan loan, the overall deposits in society decrease by 105 yuan, i.e., 5 yuan less, helping raise the reserve ratio.

So, how can bank runs and crashes be managed? For banks, if the central bank and other banks can provide sufficient gold to meet the withdrawal demands of the bank's depositors, the run can be successfully defused. The simplest way is with a 100% reserve ratio, where gold matches the bank's deposits or the market value for the securities market. If the market value of the stock market reaches 50 trillion yuan and everyone wants to convert their bank deposits at this value, a 100% reserve ratio means the rescue authorities need to put up 50 trillion yuan in deposits to buy all the stocks at the current price. Handling trillions with just billions is insufficient to counteract the trillions, making it ineffective.

To compare, in the foreign exchange market, when market participants purchased almost 800 billion yuan in August, the central bank provided equivalent foreign exchange to banks, stabilizing the yuan exchange rate. If the central bank had provided only a few hundred billion yuan in foreign exchange, it would not have matched the 800 billion demand, resulting in either quantity controls or a depreciation to a suitable level.

In a stable market with no significant fluctuations or market manipulation, the reserve ratio and stable exchange rates allow many securities to be monetized. Just as banks successfully replaced gold with deposits, securities in financial markets can use this "cash-deposit" structure for liquidity.

The "cash-deposit" structure consists of two phases. In the first phase, cash is more valuable than deposits, and in the second phase, direct deposit transactions take precedence over cash transactions. Currently, we're in the second phase. While cash transactions are less frequent, they're still important to maintain the 1:1 cash-deposit exchange relationship, similar to how futures contracts' delivery volume, even if only a small portion of open interest, is crucial for convergence of futures and spot prices.

To achieve the second phase, it's subjective as the market must believe in the feasibility of cash transactions. This self-fulfilling element means as long as everyone believes in it, it's worth believing in. Objective factors involve people using deposits for payments. Several years ago, cash was commonly needed for peer-to-peer payments, but now, mobile payments are more common, reducing the demand for cash.

An example of successful implementation of this structure is the currency market fund, such as Alipay's Yu'e Bao. The fund's deposits are smaller than its size, resulting in a reserve ratio below 100%. When investors redeem more than the fund's deposit holdings, the fund must sell other assets to generate more deposits, to meet the full redemption demand. However, the fund expands its size beyond its deposit holdings, allowing investors to buy goods with fund shares.

The strategy of funds like Yu'e Bao is to create a "currency kingdom" where users don't redeem their funds, allowing the fund to accumulate deposits without cash outflows. This way, funds flow from banks to individual accounts, reducing the reserve ratio. Funds like Yu'e Bao have achieved considerable success, as seen with the popularity of Taobao and Tmall.

In financial markets, bonds might have a better chance of adopting this structure. Bonds have stable prices and offer investors the option to sell in the financial market or wait for maturity. Although bonds can't currently directly purchase goods or financial assets, they are widely recognized as having the ability to be sold for bank deposits, similar to the ability to convert bank deposits into cash. The exchange process might involve minor price fluctuations.

A liquid bond market functions similarly to a bank. Just as banks balance cash inflows and outflows, bond markets balance buyer and seller activities.

In China's current situation, the bond market liquidity is poor, and many bonds can only be held until maturity. Collateralization becomes crucial in this context. If bonds can be used as collateral, holders can at least gain ownership and usage rights to bank deposits for a certain period. This is better than waiting for the maturity date.

Additionally, the development of derivative markets like futures markets provides more channels for bond liquidity enhancement. Futures contracts come in various types and delivery dates, catering to a variety of bonds. For example, bonds issued by the same entity with varying maturities and coupon rates can be standardized using conversion factors for delivery. Developing futures contracts covering multiple bonds or making deliverable bonds into futures contracts can objectively enhance bond liquidity.

Maturity payments are a key feature of the bond market, giving it a blend of bank and financial market characteristics. Issuers obtain bank deposits by issuing bonds, similar to banks obtaining deposits by issuing liabilities. If bonds can be redeemed at any time, the difference between bonds and bank deposits will be minimal. Bond market investors deal with two reserve systems - the issuer's and the bond market's.

However, there can be conflicts between these two reserve systems. For example, when the reserve requirement ratio is less than 1:1 for a company, issuing a large number of bonds to gain an equivalent amount of bank deposits and slowly spending them can increase the company's reserve ratio. Yet, this reduces the funds available for buying bonds in the financial market. As the total bond supply increases due to the company's issuances, the market's reserve requirement ratio decreases. A solution might involve encouraging companies to reinvest their bond-secured bank deposits into the bond market, avoiding a "cash outflow." Similarly, after companies spend their deposits, deposit recipients also need to reinvest in the bond market to prevent a "cash outflow."

The scenario described above pertains to a bond market without banks. In a bond market involving banks, actions such as banks purchasing bonds increase bank deposits. Thus, aside from considering the reserve requirement ratios of companies and the bond market, the reserve requirement ratio of banks also requires attention. During crises, companies with insufficient reserves to repay debt can trigger market bond sales. Banks might also sell bonds in this situation, reducing their own deposits on the liability side or other banks' deposits, thereby shrinking the overall deposit base. Selling bonds to another bank would reduce the buying bank's reserve ratio if reserves aren't increased.

However, reduced bank deposits can also lower the bond market's reserve requirement ratio, creating a vicious cycle of bond sales, market contraction, and lowered bond prices. The decreasing bond prices affect market confidence in banks, leading to withdrawals and a reduction in the reserve requirement ratio of banks. It becomes a cycle, and the modern version involves banks calling back interbank loans and clients withdrawing deposits, causing rapid declines in banks' reserve requirement ratios.

A way to mitigate this situation is for the central bank to directly purchase bonds from banks and non-bank institutions. This simultaneously increases non-bank institutions' deposits in banks and banks' deposits with the central bank, while also decreasing the overall bond supply due to the central bank's purchases. This approach maximizes the reserve requirement ratios of both the bond market and banks.

Lending to non-bank institutions could increase bank deposits and central bank deposits, but this might not reduce the bond supply, yielding weaker results. Lending to banks or purchasing bank assets solely increases banks' reserve requirement ratios. Banks might choose to continue asset and liability contraction, limiting the expansion of their deposit base, and might not effectively raise the bond market's reserve requirement ratio.

Unlike the bond market, stocks have not yet advanced through the "cash-deposit" structure, primarily due to uncertain conversion ratios and substantial price fluctuations. However, significant stock price fluctuations can impact the economy. For instance, imagine a bank with many depositors where, suddenly, each depositor's account balance increases tenfold. Depositors might opt to withdraw, or they might decide to withdraw and reinvest. An individual might withdraw an inflated deposit to purchase a car, while the car seller who receives the payment might deposit it back into the bank. For the bank, nothing much seems to change, but for the real economy, a car is sold.

The rise in stock prices doesn't necessarily increase bank deposits, but it does enhance the ability to convert stock holdings into bank deposits. Just as depositors can withdraw and spend, stockholders can sell stocks and spend. Stockholders selling stocks obtain deposits that can purchase goods, and those deposits might end up back in the stock market, similar to the earlier example with the bank.

Stock market bubbles increase the ability to convert stock holdings into deposits, which isn't equivalent to printing money. As stock prices increase, the ability to acquire deposits after selling stocks grows. A stockholder selling stocks to purchase a car effectively removes money from the stock market. If stock prices are low, the car might not sell, preventing money from exiting the market.

However, this perspective considers exchange rates and reserve requirement ratios. High stock valuations mean high conversion rates, causing more significant disruptions to reserve requirement ratios. High valuation stocks result in higher deposit withdrawals, channeling funds away from the stock market. On the other hand, when stock valuations are low, although stocks are sold for the same amount, the proceeds might only afford a bicycle. Thus, it's important to manage both valuations and reserve requirement ratios.

Furthermore, the money received by car sellers might not necessarily be reinvested into stocks. Instead, it could be used outside the stock market. This mirrors the historical cash outflow issue faced by banks. Therefore, maintaining a high valuation for an extended period becomes more difficult. The longer a bubble lasts, the bigger and louder the burst. The best strategy within a bubble is to capitalize on the situation, as larger bubbles offer more lucrative opportunities. However, the act of capitalizing on the situation itself becomes the needle that bursts the bubble.

Finally, let's consider the foreign exchange market. Since it involves trading two currencies, there's no concern about a currency overtaking another. However, the question remains if runs or withdrawals could still occur. Debt plays a crucial role here. In the foreign exchange market, certain currencies serve as funding currencies, typically with low interest rates and stable prospects, such as the Japanese yen. People usually borrow in yen to buy other currencies, not the other way around. Therefore, there's an inherent demand for the funding currency.

Hence, in the USD/JPY market, for example, the dollar wouldn't be subject to runs as much, as people usually hold dollar positions. However, the yen might face runs since there's demand to buy it to repay debt. If two non-bank entities borrow yen, for instance, and one repays its debt by buying yen, it essentially transfers the yen to the creditor. The amount of yen doesn't decrease, but the debt is eliminated through the transfer. For bank-issued loans in yen, repaying the loan leads to the disappearance of yen and a contraction of the bank's balance sheet.

Similar to stock market leverage trading, after speculators sell stocks and repay bank loans, that money effectively disappears. In the foreign exchange market, money also vanishes when repaying bank loans, and it's not solely a matter of finding sellers at existing prices. Therefore, central banks often step in during foreign exchange market crises. The most effective method is currency swaps between central banks. This essentially creates their respective domestic currencies and loans them to each other's countries. This eliminates debt within the flow of money, simultaneously expanding their balance sheets while counteracting bank contraction.

From this perspective, banks engaging in stock trading would cause more significant impacts. Apart from the foreign exchange market, the bond market also demonstrates that bank-initiated sales due to fair value pressure lead to deposit contraction, reducing funds available for non-bank bond investors and lowering the bond market's reserve requirement ratio. If banks enter the stock market, the impact might be more substantial due to stock price volatility surpassing that of bonds.

In addition to debt currencies, the currency in which owners' equity is denominated is crucial. For investors with no debt, some calculate profits and equity in USD, while others use EUR. During a crisis, investors calculating profits in USD might aim to reduce USD losses and decrease EUR exposure. Similarly, those calculating profits in EUR might aim to reduce EUR losses and decrease USD exposure. If both categories of investors had CNY exposure, they might sell CNY.

To summarize, both banks and financial markets exhibit a balance between flexibility and rigidity. The financial market's flexibility is visible in moderate price fluctuations, while its rigidity is demonstrated by the inability of buyers to support a continuous selling flow and the subsequent market crash. Though concealed in stable market conditions, similar to how banks operate without a 100% reserve requirement ratio, rigidity becomes apparent during events. A financial market's reserve requirement ratio is impossible to calculate precisely due to fluctuating buyer willingness and prices. However, recognizing the presence of this rigid constraint is crucial. While reserves have their role, not having them is far more problematic.

Technological advancements have broadened the types of tradable instruments and their liquidity in the financial market. However, as liquidity increases, various reserve structures emerge. These can be bank-based, situated within an institution's balance sheet, or market-based, featuring numerous buyers and sellers. Regardless, both structures lead to a "cash-deposit" reserve system. This can create ongoing withdrawal pressures while enabling financial tools to gain currency-like status.

Understanding this structure helps comprehend the emergence of market crashes, strategies to mitigate their losses, and how to capitalize on them. Market crashes are not random; they are inevitable outcomes under specific circumstances. It's akin to a person holding a cup: if the cup has little water, it won't spill even if the person runs, but if it's full, the water might spill even if the person walks slowly, let alone if someone comes from behind and pushes them.