2015-09-27 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

2013年的“钱荒”原因，已经成为银行业务的一大转折点。人们探讨了很多钱荒的原因，其中经常提到的一个原因是期限错配，本文试图从“钱荒”的原因出发，来讨论期限错配的丛林法则。

1，“钱荒”的原因
在包含所有银行的银行体系中，每一家银行都和另外一些银行有债权债务关系，规模当然可大可小，也可以为零。这些债权债务的形成，也就是银行间的拆借过程，伴随着这些银行在央行的存款换手。银行间完全不存在借贷，和银行间广泛存在借贷，所对应的央行存款的总规模，在央行不主动扩张的前提下，是完全一样的，但每发生一次借贷，就产生一笔债权债务关系，借贷越多，债权债务关系也越多。两个人之间借钱也是这样，借来借去，钱都是那么多。

从这个角度看，我们意识到，“钱荒”中缺乏的，并不是静态的资金规模，因为无论有没有债权债务，钱就在那里，不多不少，不会无端的“荒”。所谓“荒”，乃是描述一种动态过程，即还贷与借款过程中，人们感到还贷的风险很大，借款的难度很大。这就像一个湖里有很多鱼，夏天很容易打到鱼，但冬天由于湖面结冰，很难打到，但鱼还是在湖里。

在中国，有些银行总是需要向其他银行借钱，而有些银行总是很有钱从而往外出钱。这种现实的态势，有其客观的原因，比如存款的分布和经营的实力等等。由这些因素造成的银行间借贷，超出了本文讨论的范围。本文感兴趣的是，在银行间势均力敌的情况下，是不是还会出现“钱荒”。如果这种情况下依然还会出现“钱荒”的话，现实中出现“钱荒”也就不难理解了。

首先，我们考虑这样一群银行，它们之间的借贷关系都是同期限的，例如隔夜拆借，那么有没有可能有序地解除债权债务关系，而不是通过“钱荒”的方式呢？乍一看这个问题并不难回答，即同业债务是怎样产生的，那么让各银行的钱反向流动一遍，如同磁带倒带一样，即可解除所有债权债务关系，怎样来的就怎样回去。

但困难是，为确保这种解除债权债务关系的方法万无一失，需要完全严格按照债权债务关系产生的时间顺序反转回去，甚至日间的操作也要按照顺序来做，不然某家银行就可能出现没有在央行存款可供还款的情况。而这样严格的条件，在现实当中恐怕是不具备的。这有点像弄乱魔方容易，但还原魔方困难。

当然，有一种轻松的办法可以解决上述问题，即通过清算所。上述银行都去清算所登记自己的债权和债务情况，每天由清算所轧差清算净额，于是银行的“倒带”策略就简单了很多。如果每家银行自身的债权规模等于自身的债务规模，那么清算所会轻松地让每家银行的债权债务直接两清，而不用采取“倒带”的策略。但借贷期限的不同，则会使上述“倒带”和清算的策略遇到“梗阻”。

例如一家小行和一家大行，大行借给小行隔夜资金，小行又借给大行等量一年期资金，双方的资金量处在各自认为的理想水平。这样的两次操作后，大行和小行各自的资金量并无变化，但如果大行某天处于某种原因，决定停止借给小行隔夜资金，从而使自己的资金量达到新的理想值，由于小行借给大行的资金没有到期，小行单方面还款会使其资金量低于放款前的水平，只能在新情况下，被迫接受不理想的资金量。

推而广之，如果有很多家银行，借贷关系的期限多种多样，那么上述问题就更为普遍，在这个过程中，即使每家银行的债权等于其债务，但由于期限不同，“倒带”和清算便无法消除所有债务。这当然就是人们所说的“期限错配”导致的问题。

2，期限错配和违约

由于短期的利率低于长期的利率，因此，借入短期资金，放出长期资金，本身是一种赚钱的生意。回到前面提到的小行和大行，如果没有信用风险，小行从大行借入隔夜资金，并放给大行长期资金，而大行则从小行借入等量长期资金，并放给小行隔夜资金。这样的交易不会改变两家银行各自的资金量，唯一不同的是，小行每天都赚一点利差，大行则亏一点利差。

小行的动机很容易理解，但人们可能会很疑惑，为什么大行会愿意做这样的交易呢？答案其实在前面的例子里已经提到：大行可以通过停止借钱给小行而使小行陷入不理想的状态当中，而小行如果想恢复临近理想的状态，就不得不给与大行一些好处。这些好处在现实中，随着市场条件的变化而有多有少，而大行的目标，则是在这一“逼债”的过程中，获得尽可能多的好处，从而覆盖掉之前“长借入短借出”造成的利息亏损。

更现实一点的情况下，市场中有两类银行，一类银行是稳健型银行，而另一类则是激进型银行。稳健型银行由于平时赚钱较少，因而很可能撑不到挣大钱的那一天，就被股东抛弃，而被很挣钱的、股价高估的激进型银行收购。但稳健型银行等待的，就是激进型银行陷入危机，此时稳健型银行可以通过高息担保贷款或者入股的方式，大捞一笔。

而激进型银行则是一般情况下赚钱，但在危机当中会亏到难以想象。他们的生存之道，一是想办法在危机来临之前全身而退，变身稳健型银行，而二是在危机之中，尽量让其他激进型银行当替死鬼，而努力让自身存活。这两个办法甚至衍生出第三个办法，就是自己及时退出，让别的激进者死去，然后再把激进者低价收入，赚取平时的激进利润，以及危机时的捡便宜利润。

上述债权债务的特点，在于没有二级市场，债权人永远是同一个。而会计上也用历史价值入帐，因为没有二级市场也就不会有价格的变化。市场参与者的唯一挣钱方式就是收利息。真实情况下，整个市场是由可流通和不可流通的债权债务组成的，是历史价值和市场价值并用的。除了期限错配之外，稳健型银行还可以迫使激进型银行低价卖出资产还债，而在资产价格的回升中获取收益，而激进型银行借钱买入资产，以扩大收益率的行为，一方面能够获得巨大的账面利润，但同时也给了稳健型银行以可乘之机。

上面是期限错配游戏的基本法则。一方面，短债务长债权的期限错配主体是可以挣钱的，这一点人们都知道。但另一方面，人们较少提及的是，和这些主体作对手的另一方稳健派，也是可以挣钱的。所以这个游戏才能够不停地在动态中玩下去，因为两方面都有挣钱的可能性，都愿意玩。双方的精髓都是一方面等待时机吃掉对方，另一方面在面临危机时让同类比自己先死去。央行不主动扩表，在央行的存款就那么多，怎么分也不会更多，不吃对手和队友，自己的一份怎么会多？

而之所以是一个“丛林法则”，一方面是想说明，这个游戏是十分残酷的，需要吃掉对手和牺牲队友。但另一方面也想说明，这个丛林是动态平衡的，正如丛林中有大型的捕食者，也有很多小动物。期限错配起来，两类参与者才能有动态的实力和地位变化。当然，由于考虑的是同业的情况，这里不考虑银行存款之类的负债的期限问题。

最后我们考察一个常见的表达，也就是“货币空转”。很多人对于这个概念感到十分迷茫，但上面例子里银行之间的期限错配操作，显然和实体经济没有任何关系，连一根铅笔都生产不了，目的就是在于吃掉对手和牺牲队友来获得利润。银行在央行的存款从一家银行转移到另一家银行，但这并不是由于实体经济里企业和个人之间的转帐需要，而是以银行系统的风险为代价为自己谋取利益，也不难理解央行想要出手打压了。

The cause of the "money shortage" in 2013 has become a major turning point in the banking business. Many reasons have been discussed for the "money shortage," among which the mismatched maturities are often mentioned. This article attempts to discuss the jungle law of mismatched maturities starting from the cause of the "money shortage."

Cause of the "Money Shortage"

In the banking system encompassing all banks, each bank has creditor-debtor relationships with other banks, with scales ranging from large to small or even zero. The formation of these creditor-debtor relationships, which is the process of interbank lending, is accompanied by these banks' deposit turnover at the central bank. Whether interbank lending is completely absent or widely present corresponds to the total scale of central bank deposits, provided that the central bank does not actively expand, is exactly the same. However, with each occurrence of interbank lending, a creditor-debtor relationship is created, and the more lending occurs, the more such relationships arise. The same principle applies to individuals lending money to each other.

From this perspective, we realize that what is lacking in the "money shortage" is not a static fund size, because regardless of whether there are creditor-debtor relationships, money is there, neither more nor less, and there won't be an arbitrary "shortage." The so-called "shortage" describes a dynamic process where people feel a high risk in repaying and difficulty in borrowing during the process of repayment and borrowing. This is like a lake full of fish; it's easy to catch fish in the summer but difficult in the winter when the lake freezes over, yet the fish are still there.

From this standpoint, it's clear that what's lacking in the "money shortage" is not a static shortage of funds, because money is always there, regardless of creditor-debtor relationships. Instead, it's a description of a dynamic situation where risks are high in repayment and borrowing is challenging. Just as in a lake with abundant fish, there are times when fish are easy to catch, but there are also times when they are difficult to catch due to external factors.

In China, some banks always need to borrow money from other banks, while others always have money to lend. This situation has objective reasons, such as the distribution of deposits and business strength. The interbank borrowing and lending resulting from these factors are beyond the scope of this article. The focus here is whether a "money shortage" would still occur when banks are evenly matched. If such a situation still results in a "money shortage," it's easier to understand why it happens in reality.

First, consider a group of banks where their borrowing and lending relationships have the same maturity, for example, overnight interbank lending. In this case, is it possible to orderly dissolve the creditor-debtor relationships without resorting to a "money shortage"? At first glance, this question might seem easy to answer: if interbank debts are created this way, then by reversing the flow of money among the banks, much like rewinding a tape, all creditor-debtor relationships could be eliminated. However, the difficulty lies in ensuring the foolproof nature of this debt resolution method, which requires an exact reversal of the chronological order of creditor-debtor relationships, even the intraday operations need to be executed in order, otherwise, a bank might find itself without central bank deposits for repayment. Yet, in practice, such strict conditions are likely not achievable. This is somewhat analogous to how scrambling a Rubik's Cube is easy, but solving it is difficult.

Of course, there is an easier way to solve the above problem, and that is through a clearinghouse. If the above-mentioned banks all register their creditor and debtor information with the clearinghouse, which then calculates the net amount to be cleared daily, the "rewind" strategy becomes simpler. If each bank's creditor scale equals its debtor scale, the clearinghouse can easily clear the creditor-debtor relationships for each bank directly without resorting to a "rewind" strategy. However, differing maturities of debts create an obstacle to both the "rewind" and clearing strategies.

For instance, a small bank and a large bank: the large bank lends overnight funds to the small bank, and the small bank lends the same amount of one-year funds to the large bank. After these two transactions, neither bank's funds have changed, but if the large bank decides to stop lending overnight funds to the small bank for some reason, thereby achieving its new desired fund level, the small bank, as its funds lent to the large bank haven't matured yet, will have to repay unilaterally, leading to a lower fund level than before lending. Thus, in the new situation, it is forced to accept an undesirable fund level.

Broadening this scenario, if many banks are involved, each with varying maturity terms for their borrowing and lending relationships, the same problem becomes more widespread. In this process, even if each bank's creditor matches its debtor, the differing maturities make it impossible for the "rewind" and clearing strategies to eliminate all debts. This, of course, is the problem caused by mismatched maturities.

Mismatched Maturities and Default

Due to short-term interest rates being lower than long-term interest rates, borrowing short-term funds and lending long-term funds is inherently profitable. Returning to the example of the small bank and the large bank mentioned earlier, assuming no credit risk, the small bank borrows overnight funds from the large bank and lends long-term funds to the large bank, and the large bank borrows the same amount of long-term funds from the small bank and lends overnight funds to the small bank. This transaction doesn't change the fund levels of both banks; the only difference is that the small bank earns a slight interest rate difference daily, while the large bank incurs a slight loss.

The motive of the small bank is easily understood, but people might wonder why the large bank would engage in such a transaction. The answer has already been mentioned in the example: the large bank can force the small bank into an undesirable situation by stopping overnight lending, making it necessary for the small bank to provide benefits to the large bank in order to restore a closer-to-ideal state. The amount of these benefits may vary depending on market conditions, and the large bank's goal is to obtain as many benefits as possible during this process of "forcing repayment," covering the interest losses caused by the previous "borrowing short, lending long" practice.

In a more realistic context, there are two types of banks in the market: conservative banks and aggressive banks. Conservative banks, due to lower earnings, are likely to be abandoned by shareholders before they manage to accumulate significant profits. Aggressive banks, on the other hand, are more profitable in normal times but face unimaginable losses during crises. Their survival strategies are twofold: to transform into conservative banks before the crisis hits, effectively becoming stable, and to let other aggressive banks take the fall during crises, ensuring their own survival. In fact, an even more realistic scenario might involve two classes of banks: one conservative and one aggressive. The conservative bank lends to the aggressive bank at a high interest rate, while the aggressive bank borrows from the conservative bank and invests in assets, effectively expanding its returns. The conservative bank earns interest income, while the aggressive bank seeks higher returns on investment.

The above features of creditor-debtor relationships lack a secondary market; the creditor-debtor relationships always involve the same entities. Accounting-wise, historical values are used because without a secondary market, there are no price fluctuations. The only way market participants make money is through interest income. In reality, the market consists of tradable and non-tradable creditor-debtor relationships, utilizing both historical and market values. Apart from mismatched maturities, conservative banks can also force aggressive banks to sell assets at a low price to repay debts. They can then profit from the subsequent asset price recovery, whereas aggressive banks borrow to invest in assets to expand their returns. While this behavior yields significant book profits, it also presents an opportunity for conservative banks.

The fundamental principles of the mismatched maturities game are as follows: On the one hand, it is profitable for entities with mismatched short-term debts and long-term credits; this is well-known. On the other hand, what people rarely mention is that the conservative players opposing these entities can also make money. This is why the game can continue dynamically—both sides have the potential to profit and are willing to play. The essence of both sides is waiting for the right time to take advantage of the opponent and at the same time, allowing the opponent to die off first during a crisis. As the central bank does not actively expand its balance sheet and its deposits in the central bank remain constant, there is no room to distribute more, whether to compete against opponents or allies.

The reason it is termed a "jungle law" is that it illustrates the brutality of this game, where opponents are devoured and allies are sacrificed for personal gain. However, it also aims to highlight the dynamic equilibrium of this jungle, just as there are large predators and many small animals in a real jungle. Mismatched maturities allow for dynamic shifts in power and status for the two types of participants. Since this analysis is focused on interbank scenarios, it doesn't consider the maturity of bank deposits or liabilities.

Finally, let's examine a commonly used term, "currency churn." Many people find this concept confusing, but the example of interbank mismatched maturity operations mentioned earlier is clearly unrelated to the real economy; it doesn't even produce a single pencil. Its purpose is to derive benefits by devouring opponents and sacrificing allies, incurring risks within the banking system while seeking personal gains. Although deposits move from one bank to another within the central bank, it's not due to transfers between enterprises and individuals in the real economy; rather, it's a strategy employed by banks to maximize their interests at the cost of system risks. It's not difficult to understand why the central bank would want to intervene in this process.