原创 2015-12-31 季天鹤央行观察

季天鹤
方正中期研究院研究员、央行观察专栏作家

央行12月29日公布了宏观审慎评估体系MPA。从目前各界发布的信息来看，央行的MPA包含对广义信贷、委托贷款等指标，而监管标准包括广义信贷增速与目标M2增速的偏离、以及委托贷款增速与目标M2增速的偏离等。上述监管标准，或许延续了当初制定差别准备金率制度时的旧体，但在未来可能会面临挑战。

央行的监管脉络，在数量上是准备金率、公开市场操作规模、合意贷款数量等指标，其中合意贷款当然只针对贷款，不涉及银行负债侧；公开市场操作只涉及银行间，尽管涉及银行负债侧但不涉及货币量；准备金率同时涉及银行资产侧的准备金存款，以及银行负债侧的一般存款，则既影响银行间也影响货币量。

不过准备金率尽管横跨银行资产负债表两侧，但准备金规模是央行给定的，对银行而言是外生的。银行体系作为一个整体，无论自身怎样拆借买卖，只要央行不动作，银行在央行的准备金存款规模就不会变化。当然，法定准备金率通过外生给定的准备金规模约束了银行的一般存款规模，但银行资产负债的组合有很大自由变动的空间。对于准备金率在货币机制上的讨论，欢迎参考6月发表的《银行和存款准备金的那些事-《中央银行与货币供给》读后感之一》。

而广义信贷和货币量对银行来说是内生的量，两者具有内在的联系但也有背离。如果观察存款性公司概览，我们发现在2005年，国内信贷同比增长不足10%，但M2增速则在15%左右。但这显然不是银行信贷部门的问题，而是在外汇猛烈结汇的大背景下，货币增速的重要来源是结汇导致的。

结汇操作在增加货币存款的同时，并不会增加银行贷款，只会增加银行在央行的准备金存款或者银行的外币资产，因此国内信贷和货币量就会发生背离。这里最好的例子恰恰是央行自己：央行的信贷增速和自己的准备金存款增速显然就不一致，因为央行准备金存款的来源是巨大的外汇结汇，准备金存款增速更大程度上对应外汇增速而非信贷增速。

当然，在2009年的4万亿刺激时期，国内信贷和M2增速高度一致，都在30%附近，但这并不是必然。很快，在2015年，我们又看到了10月底国内信贷同比增长24%，但货币量增速仅在13%附近，两者又一次发生了较大背离。这次的原因，当然和外汇离开中国，人民币存款变为外币被汇出或者不记入货币量有关。

央行直接对非银行机构的信贷，本身也可能增加货币总量。例如央行借给证金公司2000亿，而证金公司用来买股票，则卖股票的人拿到钱存在银行，一样增加了货币量，而银行的信贷侧什么也没有发生。这种情况下，要求银行的信贷增长和货币量增长一致，对银行来说就有一定难处。而如果为了不让银行为难，而放弃对非银金融机构施加援手的工具，恐怕又削弱了维护金融稳定的能力。

当然，央行可以通过调整统计口径来让银行不那么难以达标，但我们从上面已经看到，信贷和货币增速相同不是必然，货币增速总是受到信贷之外的各个变量的影响。随着经济和货币环境的变化，信贷和货币量的增速也会不停变化，两者关系的常态是不稳定而非稳定。如果把同步增长看作常态，就会不停地被信贷以外的其他货币量影响因素挑战。

更仔细观察，央行要求信贷增速和货币量增速接近，与过去银监会要求信贷存量小于存款存量的75%，看似都是固化信贷和货币的关系，或者说固化了信贷和占据货币量绝大部分的存款的关系，但两者的假设是不同的。央行承认贷款创造存款，信贷和存款在机制上同步增减。银监会则认为存款创造贷款，本身不承认信贷和存款在机制上会同步增减，而认为存款是存款，信贷是信贷，银行在存款不变的情况下可以决定自己的信贷量。

所以信贷以外的货币量影响因素对于央行和银监会而言，就有着不同的角色。对央行而言，信贷以外的因素例如外汇是导致信贷和货币增速不同的干扰活动，如果没有这些扰动项目，信贷肯定和货币同步变动，正如4万亿刺激时期体现的那样。对银监会而言，信贷以外的因素恰恰是银行得以遵守存贷比的原因，没有大量的外汇流入，信贷和货币量的同规模增长必将使银行存贷比超过红线，存贷比的遵守乃是历史的偶然。

此外，对于委托贷款而言，银行在这个业务里，扮演的应该是委托人存款搬运工而非融资人存款创造者的角色。如果委托贷款只是银行把委托方的银行存款转给了指定融资方，从而对货币量增速没有贡献的话，把委托贷款增速和货币量增速绑在一起，感觉两者关系较为间接。

不过银行想要完成央行的要求也很容易，只需要抓住一点，即目标M2增速不是实际的M2增速。最简单的方法当然是资产证券化，或者说间接融资直接化，把信贷卖给保险和基金等非银行金融机构，甚至通过互联网卖给个人投资者。这样非银和个人的存款减少了，银行的贷款也减少了，货币量和信贷双双减少。银行以货币量增速不达标为代价，使信贷增速达到目标区间。同时，证券化还能够帮助央行完成提高直接融资比重的政策目标。

阅读了“央行观察”10月24日发布的《利率市场化了，下一站货币量》的读者，或许会指出，在本文中，作者认为贷款和货币增速是不绑定的，但在《利率市场化了，下一站货币量》中为什么多次提到绑定？原因在于《利率市场化了，下一站货币量》想指出的，是在信贷以外的因素给定时，比如外汇流入减少时，央行为了达到某个预设货币增速目标，就只能通过信贷这一途径创造货币，信贷和货币因此而绑定。如果不想实现某个货币增速目标，仅看信贷与货币实际增速，那么两者显然又会受到信贷以外的货币影响因素所影响。

同样，正如《利率市场化了，下一站货币量》文末指出的，当银行自身无法完成目标货币量增速，或者不想通过信贷方式完成时，央行要么放弃目标货币量增速，要么就是想别的办法完成。比如，海外人民币如果回流境内但不用于归还银行贷款或跑到央行，就会体现为货币量的增加。而境内金融市场开放吸引境外投资者的外币在境内结汇，也可以使货币量增加。不过海外人民币存量总体有限，并且自身还在迅速通过购汇而削减，而境外主体进入国内肯定也不会形成外汇内流的狂潮，用来填补货币增速的缺口或许不够迅速。

还有一种可能是央行自己，因为央行自身可以贷款、股权投资或结汇，因此只要找到一个非银行主体，通过它在银行资产侧创造出准备金存款，在银行负债侧创造出货币然后支出，就行了。如果央行只在银行间努力宽松，但银行不进行足够的信贷扩张，那么央行的政策就无法传导到货币侧。比如如果2015年没有股市崩盘，13%的M2增速恐怕很难完成，因为2015年6月底的M2增速还不到10%，而7月一下子就到了13.3%。

在目前通缩的压力下，央行通过上述方法可以让这个非银行主体成为一个池子，里面可以装股票、债券甚至商品等，一旦通胀或者资产泡沫抬头，就可以抛出而打压价格，正如证金和汇金收入了大量股票之后，股市的任何疯长企图都会面临证金汇金的打压。除了货币层面，银行在央行的准备金规模提升，也有利于MPA的执行。如果准备金规模和准备金率一起下降，谁还会在乎MPA给好银行在法定存款准备金利率上的一点点上浮？

不过说到法定准备金利率上浮的好处，现行的法定准备金率是1.62%，如果每家银行都奖励上浮30%，如果20万亿在央行存款全都是准备金存款，那么央行每年最多要给出972亿元。考虑到不可能每家银行都有奖励上浮，幅度也不会都是30%，因此央行每年多给的利息应该非常少。相比之下，现在1个季度全国的银行不良贷款就能新增1000亿。

让银行面临“多放贷”与“多从央行领利息”这一权衡取舍是不是一个好主意？央行多给的一点利息恐怕都会被计提拨备的费用给抵消掉。但银行缩减信贷或者要求市场主体给与多一点好处以弥补不能从央行多取的利息，似乎又有推高融资成本的可能。最好的情况，还是直接融资的大发展，人们转向银行担保的较为正规的融资标的，把参与庞氏骗局的钱拿出来用于支持实体融资需求，在直接融资的大发展中降低融资成本，方为更有前途更可持续的模式。

On December 29th, the central bank released the Macro-Prudential Assessment (MPA) framework. From the information released by various sectors so far, the central bank's MPA includes indicators such as broad credit and entrusted loans, while regulatory standards include deviations between the growth rates of broad credit and the target M2 growth rate, as well as deviations between the growth rates of entrusted loans and the target M2 growth rate. These regulatory standards may continue the old framework from when the differential reserve requirement system was initially formulated, but they might face challenges in the future.

The regulatory context of the central bank involves indicators such as the reserve requirement ratio, the scale of open market operations, and the quantity of agreed-upon loans. Agreed-upon loans specifically pertain to loans and do not involve the liability side of banks. Open market operations only involve interbank activities, although they affect bank liabilities but not the money supply. The reserve requirement ratio simultaneously affects the reserve deposits on the asset side of banks and the general deposits on the liability side, influencing both the interbank sector and the money supply.

However, even though the reserve requirement ratio spans both sides of the bank's balance sheet, the reserve amount is determined by the central bank, making it exogenous to the banks. As a collective banking system, regardless of how banks engage in transactions, as long as the central bank remains inactive, the reserve deposit size in banks at the central bank will not change. Certainly, the statutory reserve requirement ratio constrains the size of general deposits relative to exogenous reserve amounts, but there is significant flexibility in the composition of bank assets and liabilities. For discussions regarding the monetary mechanism of the reserve requirement ratio, you are welcome to refer to the feelings shared in the June publication titled "Those Things About Banks and Deposit Reserves - Impressions on 'Central Banks and Monetary Supply'."

In contrast, broad credit and the money supply are endogenous quantities for banks, with inherent connections but also divergences between the two. If we examine an overview of deposit-taking companies, we can observe that in 2005, the domestic credit growth was less than 10% year-on-year, while the M2 growth rate was around 15%. However, this is evidently not an issue with the bank's credit sector but rather stems from the significant inflow of foreign exchange due to exchange settlements in the background, resulting in the main source of money growth being from the exchange settlements.

While exchange operations increase currency deposits, they do not increase bank loans. They only increase the central bank's reserve deposits or the bank's foreign currency assets. Therefore, a divergence between domestic credit and the money supply can arise. A prime example of this is the central bank itself: the credit growth of the central bank and the growth of its reserve deposits clearly do not align, as the source of the central bank's reserve deposit growth is primarily from significant foreign exchange settlements, which correspond more to foreign exchange growth than credit growth.

Of course, during the stimulus period of the 4 trillion yuan in 2009, domestic credit and M2 growth rates were highly consistent, both around 30%. However, this is not a given. Soon after, in 2015, we saw that by the end of October, domestic credit had grown by 24% year-on-year, while the growth rate of the money supply was only around 13%, marking another significant deviation between the two. The reason for this was linked to the outflow of foreign exchange from China, where RMB deposits were converted into foreign currency and not included in the money supply.

Direct credit by the central bank to non-banking institutions itself can increase the total money supply. For instance, if the central bank lends 200 billion yuan to a securities company, and the securities company uses it to buy stocks, the individuals who sell stocks will receive money in their banks, thus increasing the money supply, while no change occurs on the credit side of the bank. In such cases, aligning credit growth with money supply growth poses a challenge for banks. However, if the tools to assist non-banking financial institutions are abandoned in order to avoid burdening banks, it might weaken the ability to maintain financial stability.

Naturally, the central bank can make it easier for banks to meet the requirements by adjusting statistical methods, but as seen above, synchronous growth of credit and the money supply is not guaranteed. Money supply growth is always influenced by various variables outside of credit. As economic and monetary conditions change, the growth rates of credit and the money supply will also change continuously; their relationship is unstable rather than stable. Considering synchronized growth as the norm would constantly expose it to challenges from factors affecting the money supply other than credit.

Looking more closely, the central bank's requirement for credit growth and money supply growth to be close mirrors the China Banking Regulatory Commission's (CBRC) past requirement that credit stock should be less than 75% of deposit stock. While both seem to fix the relationship between credit and money or between credit and the majority portion of the money supply held in deposits, their assumptions are different. The central bank acknowledges that loans create deposits, and credit and deposits increase and decrease synchronously in terms of the mechanism. The CBRC, on the other hand, believes that deposits create loans, and it doesn't recognize that credit and deposits synchronously increase and decrease in terms of the mechanism. Instead, it views deposits as deposits and credit as credit, and banks can decide their credit level while keeping deposits unchanged.

Therefore, factors outside of credit that affect the money supply have different roles for the central bank and the CBRC. For the central bank, factors outside of credit, such as foreign exchange, are disruptive activities that cause differences between credit and money supply growth rates. If these disruptive factors are absent, credit and money would undoubtedly change synchronously, as demonstrated during the 4 trillion yuan stimulus period. For the CBRC, factors outside of credit are precisely the reasons that allow banks to adhere to the loan-to-deposit ratio. Without significant inflows of foreign exchange, synchronous growth of credit and the money supply in the same proportion would inevitably lead banks to surpass the red line of the loan-to-deposit ratio. The adherence to the loan-to-deposit ratio is therefore a historical accident.

Furthermore, for entrusted loans, banks likely act more as carriers of deposits from trustees rather than creators of deposits for financing. If entrusted loans involve banks transferring the deposits of trustees to designated borrowers and thus do not contribute to the money supply growth rate, then linking entrusted loan growth with money supply growth seems to suggest an indirect relationship between the two.

However, it is relatively easy for banks to fulfill the central bank's requirements. They only need to grasp one point: the target M2 growth rate is not the actual M2 growth rate. The simplest way is undoubtedly through asset securitization or indirect financing directness. Banks can sell credit to non-bank financial institutions such as insurance companies and funds, or even sell to individual investors through the internet. This reduces non-bank and individual deposits, also reducing bank loans and consequently both money supply and credit. Banks can achieve the target credit growth rate at the cost of not achieving the target money supply growth rate. At the same time, securitization helps the central bank achieve its policy goal of increasing the proportion of direct financing.

Readers who have perused the "Central Bank Observation" article titled "Marketization of Interest Rates, Next Stop: Money Supply," published on October 24, might point out that the author in this text believes that loan and money supply growth rates are not bound together. However, the same author repeatedly mentioned binding in "Marketization of Interest Rates, Next Stop: Money Supply." This is because "Marketization of Interest Rates, Next Stop: Money Supply" intends to illustrate that, with non-credit factors given, such as reduced foreign exchange inflows, in order to achieve a predetermined money supply growth target, the central bank can only create money through credit. Therefore, credit and money supply are bound in this context. If one does not aim to achieve a certain money supply growth target and looks only at the actual growth rates of credit and money, it is evident that they will also be affected by factors affecting the money supply other than credit.

Likewise, as pointed out in the conclusion of "Marketization of Interest Rates, Next Stop: Money Supply," if banks themselves cannot achieve the target money supply growth rate or do not want to achieve it through credit, the central bank either abandons the target money supply growth rate or explores other ways to achieve it. For instance, if RMB abroad returns to the country but is not used to repay bank loans or deposited in the central bank, it will lead to an increase in the money supply. Similarly, the inflow of foreign currency from overseas investors attracted by the opening of the domestic financial market can increase the money supply. However, overseas RMB deposits are generally limited and are rapidly decreasing through exchange purchases, and foreign entities entering the country will certainly not result in a frenzy of inward foreign exchange. Thus, using them to fill the gap in money supply growth may not be swift enough.

Another possibility is the central bank itself. Since the central bank can lend, make equity investments, or engage in foreign exchange operations, it only needs to find a non-bank entity that creates reserve deposits on the asset side of the bank and creates money on the liability side, then spend that money. If the central bank only focuses on loosening in the interbank market but banks do not engage in sufficient credit expansion, then the central bank's policy cannot be transmitted to the monetary side. For example, if there hadn't been a stock market crash in 2015, achieving a 13% M2 growth rate would likely have been challenging because the M2 growth rate at the end of June 2015 was still less than 10%, but it suddenly jumped to 13.3% in July.

Given the current deflationary pressure, the central bank can make this easier by letting this non-bank entity act as a pool to hold stocks, bonds, and even commodities. Once inflation or asset bubbles arise, these can be sold to suppress prices. Similar to how China Securities Finance Corp. and China Investment Corp. held a large number of stocks and suppressed any attempt at rampant stock growth. Besides the monetary aspect, an increase in the central bank's reserve deposit size in banks can also facilitate the implementation of the MPA. If the reserve deposit size and the reserve requirement ratio both decrease, who would care about the slight premium that MPA gives good banks on the statutory deposit reserve interest rate?

However, when it comes to the benefits of the premium on the statutory deposit reserve interest rate, the current statutory deposit reserve rate is 1.62%. If every bank awards a 30% premium, and if all the 20 trillion yuan in central bank deposits are reserve deposits, then the central bank would need to provide a maximum of 97.2 billion yuan annually. Considering it's unlikely that every bank will have a premium and the magnitude won't all be 30%, the extra interest the central bank pays out each year should be very limited. In comparison, at present, bad loans of 100 billion yuan can be added every quarter nationwide.

Is it a good idea to make banks choose between "lending more" and "receiving more interest from the central bank"? The extra interest the central bank pays out might be offset by provisioning expenses. However, reducing credit expansion or requiring market participants to provide more benefits to compensate for the inability to take more interest from the central bank might increase financing costs. The best scenario is still the significant development of direct financing. People should turn to more formal financing targets guaranteed by banks, bringing the money involved in Ponzi schemes to support real financing needs. In the significant development of direct financing, reducing financing costs can pave the way for a more promising and sustainable model.