2015-11-27 央行观察

汇金和中投的诞生，是央行和银行资产负债分析中最复杂的案例，目前来看还没有之一，比外汇缴存人民币准备金、四大资产管理公司置换不良资产、债券置换地方政府债务等，都要复杂，也精彩更多。其中潜藏了扩表和缩表、资产负债和股权的形成、外汇存放的形态等操作和问题，本身非常深刻。

汇金和中投是中国举足轻重的两家机构，对中国金融领域的发展做出了巨大的贡献。由于其规模的庞大，准确理解两者对于金融体系的影响，对实际的宏观以及流动性研究非常重要。本文试图运用现有的公开信息，回顾一下两家公司的诞生过程，来揭示人行与财政当年是如何注资银行的。

汇金的“汇”是什么形态？

汇金在2003年12月诞生，注册资本450亿美元（3724亿人民币），一下子就在2003年的央行资产负债表上留了个痕，表现为外汇占款的减少，以及其他资产的增加。当月，汇金就分别注资了中行和建行各225亿美元，其中对中行的出资包括196亿美元和部分黄金，看来汇金的“金”是黄金的“金”，但什么是“汇”呢？中行的年报透露了玄机。

2003年底，中行年报记载了拆放给外汇局1622.39亿元人民币。考虑到汇金注资中行发生在12月30日，而当年底中行的人民币拆放同业仅有127亿元人民币，因此这1622亿人民币应该是美元的人民币价值，而不是人民币，这和196亿美元也较接近。如果汇金给中行建行的，就是这个在外汇局的外汇存款，那么人行给汇金的外汇，应该也表现为在外汇局的存款。

除了前面提到的央行资产负债表之外，2003年12月外汇储备减少171亿美元，而当年前几个月，每个月新增的外汇储备规模都在100-200亿美元水平。而且，当月的外汇局口径银行代客结售汇顺差143亿美元，代客涉外收付款顺差129亿美元，不太应该导致外汇储备减少171亿美元的结果。唯一的答案就是，央行把汇金对应的外汇资产从外汇储备给剥离了。

这里对汇金持有的外汇局外汇存款的处理，和央行2007年开始对银行以外汇缴存人民币准备金的处理，有一些不同之处。（参考拙作《如何辨析“藏汇于民”与“资本外逃”》后者的情形下，外汇被通过“其他国外资产”项目持有，即央行认为这还是自己的国外资产，但不能算自己能任意使用的外汇储备。而汇金则被看作一种国内股权资产，因此相应地，央行国外资产项目干脆就被摘掉一部分，外汇储备也就随之减少了。

除了国外资产项目之外，央行用来注资汇金的资金来源，还可能会影响到外汇资产以外的项目。比如黄金，在央行的表上长期没有变化，但央行买入黄金时，必定会对资产负债表的其他项目造成影响，而为了保持平衡，央行或许会把不计入货币黄金的项目，先暂放在例如“其他资产”之类的项目里，使得注资汇金导致的“其他资产”的增幅，并没有注资规模那么大，因为有一部分已经在这个“其他资产”项目里了。

如果要总结一下，那么汇金的诞生过程如下。首先，央行在境外持有外汇资产或者在外汇局有外汇存款。然后，央行将外汇资产或外汇存款给汇金，汇金可能把外汇资产存在了外汇局。此后，汇金把这些资产给了银行，银行把资产存在了外汇局。之后，银行将资产又挪到境外，或者又存回央行。

汇金带来的几个问题

首先，人行注资汇金是不是扩了表？从银行年报反映的2003年底的情况来看并没有，因为汇金被设立之后，其在外汇局的存款规模变动，已经和央行资产负债表无关。汇金在外汇局的存款转为银行在外汇局持有，是汇金、外汇局和银行三者之间的事情。而汇金注资后银行在央行的存款，还是汇金注资之前那么多。

从黄金的角度或许更好理解这个过程。央行把黄金给了汇金，这部分黄金被谁持有和被运到哪里，都和央行没有关系。央行只知道自己持有了汇金的股权，但汇金怎么用这个黄金资产，是汇金自己的事情。汇金用黄金注资银行后，银行拿着黄金给外汇局也好，放在其他地方也罢，和央行无关，也和汇金无关。汇金知道的，就是自己持有了银行股权。

其次，中行如果将在外汇局的外汇存款回存给央行，会发生什么？这个交易会使央行重新在外汇局拥有外汇存款，而银行在央行的外币存款增加，也就是央行扩了表，不过这个扩表恐怕要叫作“被动扩表”，正如银行的“被动负债”一样。而对于外汇储备来说，参照外币缴存人民币准备金的操作来看，在外汇局的外汇存款回存央行，很可能并不影响外汇储备，而是放在央行资产负债表的“其他国外资产”项目下。

第三，如果中行将回存给央行的外汇存款结汇，会发生什么？首先，这应该就和正常情况下，银行找央行结汇一样。央行的负债侧外币存款减少，人民币存款增加，而资产侧的“其他国外资产”减少，“外汇占款”增加。其次，虽然这个过程往往被称作所谓“二次结汇”，但“二次结汇”只是把外汇变为人民币，本身并不扩大央行的资产负债表。央行非常了解这一点，因此经常着重强调外汇资产的境外使用，因为一旦回存央行就会扩表，一旦结汇就是增发基础货币。

“二次结汇”和一般的流动性操作

“二次结汇”这个操作可以有多种理解，特别是考虑到与之相似的几种操作。既然有信贷资产抵押再贷款，那么可不可以有外汇资产抵押再贷款？既然有债券抵押的逆回购，可不可以有外汇抵押的逆回购？对银行来说，上面提到的抵押再贷款和抵押逆回购，和找央行结汇的有区别也有联系。

结汇中，银行外汇资产减少，人民币资产增加，总资产不变。而在前面虚拟的外汇抵押再贷款和外汇抵押逆回购中，银行依然持有外汇资产，但多了对央行的负债，以及在央行的存款，总资产扩大。但对于央行而言，此处的外汇是抵押品，本身不能使用，所以央行就是单纯的放贷扩表，和一般的再贷款和逆回购只是抵押品不同，创造流动性的方式是一样的，都是扩表。

从信贷抵押再贷款和债券抵押逆回购，到外汇抵押再贷款和外汇抵押逆回购，我们就能看出，“二次结汇”的扩表，一是前面提到的回存，二是要求这里的“结汇”是用央行之前给出的资产。如果是国外来了一箱美元纸币，存到央行换人民币，那么就是正常的结汇，没必要称其为“二次”了。

如果在汇金的例子里，真的发生了“二次结汇”，我们可以一眼看穿交易的本质。看上去，央行注资汇金，汇金注资银行，银行找央行结汇，钱在流动。但如果我们把汇金拿开，便可以看出，央行采取的乃是一种扩表股权投资，即通过扩大负债侧的人民币准备金让银行持有，同时资产侧增加了银行的股权，而不光央行资产扩大，银行的资产也扩大了。

我们还可以假设另一个过程：央行对中行建行进行人民币股权投资3724亿，使央行资产侧多了对银行股权，负债侧多了银行在央行存款。这时，央行又要求银行用新增的在央行人民币存款购买外汇，于是央行国外资产少了，刚增加的银行在央行存款也没了。如果央行又用两家银行的股权，作为新设立的汇金公司的出资，然后要求中行建行和汇金在上述3724亿元的股权上全部用美元计价，那么最后我们看到的就和真实发生的汇金注资一样，但大家都看得出央行是创造了3724亿进行注资。

汇金注资的银行有没有找央行“二次结汇”呢？我们不得而知，因为这一过程被夹杂在太多央行结汇的业务里。从年报中看，注资银行还是用汇金注资的外汇在境外做业务的，应该不会尽数找央行结汇。而“二次结汇”即使有，和总体的外汇流入相比规模也较小，何况与央行其他的流动性操作比起来，性质上也其实并无特殊之处。

中投诞生前传：2700亿特别国债

中投诞生在2007年，但其渊源要追溯到上个世纪末。那时候亚洲金融危机肆虐，中国银行业有很多坏帐。国家做了很多努力解决这个问题。坏帐减计会导致净资产由正转负，变成一个窟窿。填窟窿的一个办法，就是财政部发行2700亿特别国债给四大行，而四大行认购后，财政部用所得款项注资四大行，从而充实了资本，填上了窟窿。

这个交易名义上需要央行配合，即银行在央行备付金先转移到国库，作为国债认购款，然后财政部把这笔钱又挪回银行，作为出资。当时央行还出了个文件，表示可以提供再贷款支持。但明眼人一看就知道，财政部的出资，就是通过扩表的方式入股。前面的央行配合和备付金转移未必真的发生了。财政部自己印好了债券，派四辆车送给四家银行，注资恐怕就OK了。

这也是为什么后来人民银行通过汇金注资的时候，总说汇金在用“真金白银”注资，因为央行是拿出了自己的外汇资产，而财政部则是扩表发债，不是转移资产入股，而是发行负债入股。但前面也提到了，如果银行拿到外汇之后又回存到央行，央行的外汇储备就又补齐了，银行得到的，也不过是央行债务，即银行在央行的存款。

但是，财政的扩表发债入股，和银行回存央行后的扩表发准备金存款入股，还有个重大的区别。债券没有提前返售条款，银行拿着国债，到期前什么也得不到。但银行拿着在央行存款，可以找央行购汇，央行的负债和央行的资产还存在着联系，正如储户可以用银行存款兑取纸币一样。不过如果央行表示，银行不能用在央行的存款换任何央行资产，那么央行和财政的操作差异就更小一些。

财政部发债注资的操作很容易产生误解，比如描述国有银行改革的神作《红色资本：中国的非凡崛起与脆弱的金融基础》中，作者认为“财政部将2700亿元人民币从储户和企业名下悄无声息地转移到了银行的资本金中”。这一观点是完全错误的，作者没能理解财政部当时在创造着国债资产，这2700亿和储户和企业完全没关系。这里的国债不是用来补充食物的鼻饲管，它就是食物本身。

财政部用国债入股补窟窿，满足银行的资本充足率要求，而央行也在1998年降低了准备金率，银行此时在理论上已经有了足够的超额准备金率和资本充足率。不过没想到，四大行没过几年又搞了一堆坏账出来，让财政部的投资打了水漂。比如2004年底的建行股东，分别是汇金、建银、国家电网、宝钢和长江电力，没有财政部的影子了。

中投的诞生：1.55万亿特别国债

财政发行了2700亿特别国债之后，又发行1.55万亿特别国债组建中投。上述特别国债一部分发行给商业银行，另一部分1.35万亿通过农行发行给央行。其中6000亿是8月发行10年期，利率4.3%，7500亿是12月发行15年期，利率4.45%。两次发行在央行资产负债表上的“对政府债权”一项都有记载。

上述交易在一般的叙述中，便是财政发债，农行买债，财政拿到钱找央行买外汇，而央行拿财政给的钱从农行买债。但事实上，新闻报道提到，农行一买入国债，便被央行买走，而这个时候钱刚被财政圈走，央行哪里来的钱用来买农行的国债呢？所以真正发生的情节，并不同于一般叙述那样简单，不过下面两段纯粹是过程的推导，读者可以直接跳过。

首先，财政定向给农行发债，但农行未必有6000亿的在央行存款，因此可能先要有个临时的央行再贷款，使央行资产负债各增6000亿，分别是对农行贷款和农行在央行存款；而农行资产负债也增加6000亿，分别是对央行负债和在央行存款。然后农行买债，在央行存款减少6000亿，债券资产增加6000亿；而国家资产负债双增，分别是在央行国库存款增加6000亿，以及债券发行6000亿。

下面，央行再扩表购买国债，方法是资产侧增加6000亿国债，负债侧增加农行在央行存款6000亿，而农行多了6000亿在央行存款，少了6000亿国债。此时农行可以归还再贷款，使自己在央行存款和对央行负债各减少6000亿，而央行的再贷款资产和农行在央行存款负债也减少6000亿。然后，财政用国库存款找央行购买外汇资产，使央行外汇资产和国库存款负债各减少6000亿，而财政则是减少6000亿国库存款，增加6000亿外汇资产。

上述交易伴随着频繁的扩表缩表，远非简单的钱的过手，普通的表述完全扭曲了交易的过程。但本质上，这个交易和财政部发国债当股东类似，只是这次换到的不是央行的所有者权益，而是央行的外汇资产。整个交易可简化为财政部通过发债，一手交债，一手拿外汇。这其实和央行在结汇时，一手给银行人民币准备金，一手拿外汇是类似的。财政部和央行都非常清楚，自己的负债就是别人的资产，凭空创造负债就是凭空创造资产。

发债换资产的手法，在银行改革中普遍被用来置换坏账，这方面的主力是四大资产管理公司（AMC）。他们发了很多自己的债券用来换银行的坏账，100亿的资本金能支持上千亿的债券发行，这当然是因为财政部和央行的支持。而央行也有时参加进来，发行一些特别央票来置换资产，有时候也通过给AMC再贷款的方式让AMC去购买。需要注意，AMC的各类操作也不是新闻报道那样钱的转手，里面也藏着很多扩表缩表的动作，本文不展开讨论了。

不过换一个思路来看，发债置换资产其实就是资产证券化。银行把资产打包给AMC，而AMC用这些资产作为基础资产发抵押债券。只不过这些抵押债券又全都被银行持有，因此造成了AMC扩表发债，用自己的债来置换不良贷款的效果。同理，财政部用自己的债来置换央行外汇和汇金，恐怕特别国债要被叫作“外汇资产抵押债券”了，事实上也差不多如此，从新闻报道来看，当时的楼总曾经说每天醒来都有几个亿的利息要还，看来特别国债简直就是个ABS。

中投购买汇金

上述1.55万亿元特别国债换了2000亿美元，全都是中投的资本金。2008年底的中投年报显示，其持有的长期股权投资为1710亿美元，而这部分主要对应了汇金的权益。通过汇金的资产负债表倒推，2007年底汇金自身的净资产估值应该约为1300亿美元。而2007年12月，我们看到央行的其他资产减少了4881亿，而这更接近于2007年底，汇金卖给中投后，新的实收资本670亿美元（5521亿人民币），而不是1300亿美元。

而且，央行多了1.55万亿债券资产，为了资产负债的平衡，央行要么相应增加1.55万亿负债，要么相应减少1.55万亿别的资产。可能的情况是，财政部一方面换到了几百亿美元外汇，不然的话汇金也没有200亿美元能够注资国开行了。另一方面，财政部又买了整个汇金。

这个时候我们看到，央行可能是通过多个统计口径持有的汇金。一部分是在“其他资产”里，大概670亿美元水平。而另一部分，可能对应了“对其他金融性公司债权”的项目，这一项目在当年8月和12月都出现了锐减，恰好和特别国债的发行重合。这或许和央行在汇金的权益如何计算有关，即一部分看作是“其他资产”中的股权，另一部分看作是对汇金的某种债权，不过这都只是推测了。

无论如何划分，央行出售汇金都不影响央行的负债侧，即不影响银行在央行的资产。但此前汇金的增值，需要资产负债表的其他项目来平衡。如果央行把汇金带来的增值全都上缴给了财政，财政又给支出了，那么央行其实是在发行了新的基础货币。当然，如果央行创造一个盈余公积类的项目，把对应增值部分的负债/所有者权益放在单独的科目里，就不会有这样的流动性被动创造了。

后记

汇金和中投的诞生，是央行和银行资产负债分析中最复杂的案例，目前来看还没有之一，比外汇缴存人民币准备金、四大资产管理公司置换不良资产、债券置换地方政府债务等，都要复杂，也精彩更多。其中潜藏了扩表和缩表、资产负债和股权的形成、外汇存放的形态等操作和问题，本身非常深刻。作者资历浅薄，没有亲身经历那个时代，对细节的了解有限，但对已有的框架手法已经叹为观止。

财政和央行的扩表注资方式，使中国的货币量虚高。注资银行在正常情况下，应该是银行存款转变为银行股权的过程，无论这笔存款是财政通过征税从纳税人的银行存款里征来的，还是企业和储户乃至外国人认购形成的，都应该造成银行存款的减少。如果银行放贷创造了很多存款，但贷款收不回来形成坏账，而贷款人贷到的存款已经支出，所有权已经转移，怎么办？

银行无法自己选择负债扣除，便只能在减计坏账资产的同时减计所有者权益。而银行补充所有者权益，应该是通过公开募股，让对应坏账那样多的存款消失掉变成股权，这样的操作前后，银行存款乃至货币量是不变的。但银行改革中，国家并没有让企业和个人为银行买单。无论是央行还是财政，都在自己的资产负债表上想办法，而并没有强迫企业和个人出钱，所以坏账创造的存款就依然留下了。

当然，会有人说外汇储备不注资银行可以干别的，财政发国债不注资银行可以干别的，注资就是对社会的掠夺。但银行的资本金总要充实，任务是要完成的，而完成总是有代价的。不可能大家都不出钱，银行的资本充足率就自动OK了。何况，企业找银行要“包饺子贷款”的时候，有想过这笔钱是谁在出，自己的饺子是由谁买单么？

或许又会有人说，财政凭空就能创造成千上万亿的负债用来注资银行，那银行就不用担心坏账的问题了，资本出了窟窿就让国家发债注资好了。央行也能凭空创造成千上万亿的负债，那银行出了问题让央行注资就行了。这种观点，忽视了金融之外还有实体经济的分析。尽管金融上银行永远可以通过上述方法满足资本充足率，但整天发生坏账还不停创造存款的银行，必将导致通货膨胀以及不平等：谁能贷款后拿着存款成功跑路，谁就是富翁。

同时我们看到，国有银行当时面临的问题，并不只是资本充足率的问题。亏钱的背后肯定是一些实际的原因，而不只是财务的原因。很多诸如制度、产权关系、金融技术等等问题，才是当时银行不能挣钱、总有坏账的原因。财政已经不惜负债来补充银行资本了，最后还是大亏一场，也说明这一问题。

此外，这种观点还忽视了另外一点。天上没有免费的午餐，我们必须看到，发行负债的方式必将导致一个结果，那就是准备金率的下降。在财政的角度看，就是其在央行的存款与整个国债余额比变小了，而从央行的角度看来，就是外汇储备（外汇占款）和央行债务的比例变小了。或许财政部不能发新债偿旧债了，还可以通过央行的支持来避免债券违约，但央行维持汇率的能力削弱，并不是财政能轻松补充的。

不过没有了财政和央行，银行依然有捷径可循。完全可以想象银行和保险公司联手，前者给后者放同业拆借，而后者向前者注资，相当于银行放贷给自己注资，而保险公司的负债率尽管增加（或者说准备金率尽管降低），也还在可控的范围，说不定这笔贷款利率可以很低。但还是争取让投资者用自己的钱来说话吧。希望当年央行和财政部亲自主演的注资大戏不再上演，而是定格在那个年代，成为永远的例题，而不是反复做的习题。

Huijin and CIC are two pivotal institutions in China that have made significant contributions to the development of the country's financial sector. Due to their substantial scale, accurately understanding their impact on the financial system is crucial for macroeconomic and liquidity research. This article attempts to use publicly available information to review the birth processes of both companies, revealing how the central bank and the finance ministry capitalized banks.

What Form is "Hui" in Huijin?

Huijin was established in December 2003 with a registered capital of $45 billion (372.4 billion RMB), leaving a mark on the central bank's balance sheet for the year. This is reflected in the reduction of foreign exchange deposits and an increase in other assets. In that month, Huijin injected $22.5 billion into both Bank of China (BOC) and China Construction Bank (CCB), with BOC's contribution including $19.6 billion and some gold. It appears that the "gold" in Huijin is indeed related to physical gold. But what about the "Hui" (exchange)? Clues can be found in BOC's annual report.

At the end of 2003, BOC's annual report recorded RMB 162.239 billion placed with the State Administration of Foreign Exchange (SAFE). Since Huijin's capital injection into BOC occurred on December 30th, the RMB 162.2 billion should likely represent the RMB value of dollars, rather than RMB itself. This is closer to $19.6 billion. If Huijin's injection into BOC and CCB corresponds to these foreign exchange deposits with SAFE, the foreign exchange provided by the central bank to Huijin should also be reflected as deposits with SAFE.

Apart from the central bank's balance sheet, there was a decrease of $17.1 billion in foreign exchange reserves in December 2003. In the preceding months, the monthly increment of foreign exchange reserves had been in the range of $10-20 billion. The surplus from bank-client foreign exchange transactions was $14.3 billion, and the surplus from cross-border payments and receipts was $12.9 billion for that month. These figures shouldn't lead to a $17.1 billion decrease in foreign exchange reserves. The only plausible answer is that the central bank separated the foreign exchange assets corresponding to Huijin from its foreign exchange reserves.

The treatment of foreign exchange deposits held by Huijin outside the foreign exchange administration, and the central bank's handling of banks' RMB reserve requirements in foreign exchange deposits starting from 2007, have some differences. In the latter case, foreign exchange is held as "other foreign assets" by the central bank, as it is considered foreign assets owned by the central bank but not usable as foreign exchange reserves. In contrast, Huijin is treated as a domestic equity asset, which led the central bank to adjust the "other foreign assets" category, thereby reducing foreign exchange reserves.

Apart from foreign assets, the source of funds used by the central bank to capitalize Huijin might also impact items beyond foreign exchange assets. For instance, consider gold. While the central bank's balance sheet didn't show significant changes in gold holdings, buying gold would impact other items on the balance sheet. To maintain balance, the central bank could temporarily include non-monetary gold assets under categories like "other assets," dampening the increase in the "other assets" category caused by Huijin's injection.

In summary, the birth process of Huijin unfolds as follows: The central bank holds foreign exchange assets overseas or maintains foreign exchange deposits with SAFE. The central bank then transfers foreign exchange assets or deposits to Huijin, which might further deposit foreign exchange assets with SAFE. Subsequently, Huijin invests these assets in banks, which deposit them back with SAFE. The banks might then move these assets overseas or deposit them back with the central bank.

Questions Arising from Huijin

Firstly, did the central bank expand its balance sheet with the Huijin injection? Based on the year-end 2003 data reflected in bank annual reports, it seems not. The changes in the size of deposits held by Huijin with SAFE no longer had any relation to the central bank's balance sheet. The transformation of Huijin's deposits at SAFE into banks' deposits at the central bank is a matter involving Huijin, SAFE, and the banks themselves. The deposits held by banks at the central bank after Huijin's injection remained unchanged.

Considering the gold aspect might provide a better understanding of this process. When the central bank provided gold to Huijin, the subsequent ownership and location of that gold aren't relevant to the central bank. The central bank is only aware of its equity stake in Huijin, but how Huijin uses those gold assets is its own affair. After Huijin's gold injection into banks, whether the banks give that gold back to SAFE or hold it elsewhere doesn't involve the central bank or Huijin. All Huijin knows is that it now holds equity in the banks.

Secondly, what happens if banks deposit their foreign exchange deposits with SAFE back to the central bank? This would cause the central bank to have foreign exchange deposits with SAFE once more, and the foreign currency deposits held by banks at the central bank would increase. This could be seen as passive balance sheet expansion for the central bank, similar to passive liabilities for banks. As for foreign exchange reserves, according to the operations related to foreign currency reserve requirements in RMB starting from 2007, returning foreign exchange deposits to the central bank wouldn't likely affect foreign exchange reserves. Instead, those deposits might be placed under the "other foreign assets" category on the central bank's balance sheet.

Thirdly, what happens if banks conduct foreign exchange settlement for the deposits they return to the central bank? This process resembles the banks conducting normal foreign exchange settlement with the central bank. On the liability side of the central bank's balance sheet, foreign currency deposits held by banks decrease, and RMB deposits increase. On the asset side, "other foreign assets" decrease, and "foreign exchange deposits" increase. Despite the term "second-time foreign exchange settlement," this doesn't necessarily expand the central bank's balance sheet. The central bank well understands this and often emphasizes the overseas usage of foreign exchange assets. Once foreign exchange is returned to the central bank, its balance sheet expands, and if foreign exchange is settled, it effectively leads to base money issuance.

"Second-time foreign exchange settlement" and Ordinary Liquidity Operations

Various interpretations can be derived from the operation of "second-time foreign exchange settlement," particularly when considering similar operations. If there's credit asset-backed lending, can there be foreign exchange asset-backed lending? If there are reverse repurchases with bond collateral, can there be reverse repurchases with foreign exchange collateral? For banks, these operations exhibit differences and connections to ordinary liquidity operations, including foreign exchange settlement with the central bank.

In a foreign exchange settlement, the bank's foreign exchange assets decrease while RMB assets increase, leaving total assets unchanged. In the previously mentioned virtual scenarios of foreign exchange asset-backed lending and reverse repurchases, the bank still holds foreign exchange assets while incurring liabilities to the central bank, thus expanding total assets. However, for the central bank, the foreign exchange assets here serve as collateral and cannot be used. Consequently, the central bank simply extends a loan and expands its balance sheet, akin to general lending and reverse repurchases. Both involve collateral, with the former involving ordinary assets while the latter involves foreign exchange assets.

From credit asset-backed lending and bond collateral reverse repurchases to foreign exchange asset-backed lending and reverse repurchases, it becomes apparent that "second-time foreign exchange settlement" expansion is realized in two ways. The first involves the previously mentioned deposit return, while the second stipulates that the "settlement" uses assets previously provided by the central bank. If a box of US dollar bills arrived from abroad and was deposited with the central bank in exchange for RMB, it would be considered a standard foreign exchange settlement, rendering the term "second-time" unnecessary.

If indeed "second-time foreign exchange settlement" occurred in the case of Huijin, the nature of the transaction can be discerned. At first glance, the central bank capitalized Huijin, Huijin capitalized banks, banks conducted foreign exchange settlement with the central bank, and money flowed. However, if we remove Huijin from the picture, we can see that the central bank employed a form of balance sheet expansion through equity investment. By enlarging the liability side's RMB reserve requirements, the central bank allows banks to hold equity while also increasing assets through bank equity. Both the central bank's assets and the banks' assets expand.

Let's assume another scenario: The central bank makes a RMB 372.4 billion equity investment in BOC and CCB, resulting in additional equity for the central bank and deposits for the banks. Now, the central bank asks the banks to use their newly deposited RMB funds to purchase foreign exchange. This leads to a reduction in the central bank's foreign assets and the disappearance of the banks' deposits. If the central bank subsequently uses these banks' equity to contribute to a newly established company like CIC, and mandates that BOC, CCB, and CIC all denominate their contributions in US dollars, the outcome would resemble the actual birth of CIC. Yet, everyone could discern that the central bank created RMB 372.4 billion to capitalize the company.

Did the banks conduct "second-time foreign exchange settlement" for the capital injection from Huijin? We cannot be certain, as this process is intertwined with numerous foreign exchange operations conducted by the central bank. From the annual reports, it seems that banks used the foreign exchange injected by Huijin for overseas operations, likely avoiding extensive foreign exchange settlement with the central bank. Even if "second-time foreign exchange settlement" occurred, its scale is relatively small compared to the overall foreign exchange inflow, especially when contrasted with other liquidity operations of the central bank. In essence, it doesn't differ significantly from standard operations.

The Antecedent of CIC: The RMB 270 billion Special Bonds

CIC's birth in 2007 has its roots in the late 20th century. During the Asian financial crisis at that time, China's banking industry had many bad loans. The nation made significant efforts to tackle this problem. Writing off bad loans could lead to a shift from positive to negative net assets, creating a hole. To plug this hole, the Ministry of Finance issued RMB 270 billion in special bonds to the four major banks. After the banks subscribed, the Ministry of Finance used the proceeds to capitalize the banks, bolstering their capital and filling the hole.

This process nominally required the cooperation of the central bank, with banks' reserve deposits transferred to the national treasury for bond subscription, after which the Ministry of Finance moved the funds back to the banks as capital injection. The central bank even issued a statement indicating its readiness to provide refinancing support. However, it's evident that the Ministry of Finance's capital injection was essentially a means of balance sheet expansion through equity investment. The central bank's cooperation and reserve deposit transfer might not have actually occurred. The Ministry of Finance had already prepared the bonds, which were handed over to the banks directly, making the capital injection a straightforward process.

This is why later, when the People's Bank of China (PBOC) capitalized Huijin, it emphasized the use of "real assets" for capital injection, as the central bank employed its own foreign exchange assets. In contrast, the Ministry of Finance employed balance sheet expansion through bond issuance, not asset transfer for equity injection. However, it's also noted that if the banks return the deposits they received from the central bank, the central bank's foreign exchange reserves would be replenished. The banks' outcome would simply be a debt owed by the central bank—an increase in deposits.

There's a significant distinction between the Ministry of Finance's bond issuance for capital injection and banks' return of deposits to the central bank. Bonds lack early redemption clauses, and banks won't receive anything until maturity. In contrast, when banks return deposits to the central bank, they can conduct foreign exchange settlement, as the central bank's liabilities and assets are still connected. This mirrors depositors using bank deposits to obtain paper currency. However, if the central bank states that banks cannot exchange their deposits at the central bank for any central bank assets, then the difference between the Ministry of Finance and the central bank's operations is minimized.

The Ministry of Finance's issuance of bonds for capital injection is prone to misunderstanding. For instance, in the book "Red Capitalism: The Fragile Financial Foundation of China's Extraordinary Rise," the author wrongly assumes that "the Ministry of Finance secretly transferred RMB 27 billion from households and enterprises to banks' capital without making a sound." This perspective is entirely incorrect, as the author failed to understand that the Ministry of Finance was creating assets with the special bonds, which had no connection to households and enterprises. These bonds aren't a feeding tube to satiate hunger but the food itself.

The Ministry of Finance used special bonds for equity injection to fill the capital hole, meeting banks' capital adequacy requirements. Meanwhile, the PBOC had lowered reserve requirement ratios in 1998, meaning that banks theoretically already had sufficient excess reserve ratios and capital adequacy ratios. Unexpectedly, the four major banks accumulated a slew of bad loans a few years later, causing the Ministry of Finance's investment to go to waste. For instance, at the end of 2004, BOC's shareholders included Huijin, ICBC, State Grid, Baosteel, and China Yangtze Power, with no trace of the Ministry of Finance.

The Birth of Central Huijin: CNY 15.5 Trillion Special Treasury Bonds

After issuing CNY 2700 billion special treasury bonds, an additional CNY 15.5 trillion special treasury bonds were issued to establish Central Huijin. Part of these special treasury bonds were issued to commercial banks, and the remaining CNY 13.5 trillion was issued through the Agricultural Bank of China (ABC) to the central bank. Among these, CNY 600 billion was issued in August for a 10-year term with an interest rate of 4.3%, and CNY 750 billion was issued in December for a 15-year term with an interest rate of 4.45%. Both issuances were recorded in the central bank's balance sheet under the category "Claims on Government."

In the general narrative, the transactions could be summarized as follows: the government issued bonds, ABC bought the bonds, the government obtained money and used it to buy foreign exchange from the central bank, and the central bank used the money received from the government to buy bonds from ABC. However, the actual process is more intricate than this simple description.

Firstly, the government issued bonds specifically to ABC. However, ABC might not have CNY 600 billion in deposits with the central bank, so there might need to be a temporary interbank loan from the central bank, resulting in an increase of CNY 600 billion on both the central bank's assets and liabilities, representing loans to ABC and deposits by ABC. Subsequently, ABC purchased the bonds, leading to a reduction of CNY 600 billion in its deposits with the central bank and an increase of CNY 600 billion in bond assets. The national assets and liabilities increased by CNY 600 billion each: an increase in deposits at the central bank's treasury and bond issuance.

Next, the central bank expanded its balance sheet to purchase the bonds. This involved an increase of CNY 600 billion in bond assets and an increase of CNY 600 billion in ABC's deposits with the central bank, leading to a decrease in ABC's deposits and an increase in the central bank's liabilities. Subsequently, the government used the treasury deposits to purchase foreign exchange assets from the central bank, resulting in reductions of CNY 600 billion each in foreign exchange assets and treasury deposits at the central bank.

The transactions mentioned above involved frequent expansions and contractions of the balance sheet, which go beyond simple money transactions. The common description overlooks the complexity of the process. However, at its core, this transaction is similar to the government acting as a shareholder when issuing bonds, but instead of acquiring ownership in the central bank, the government acquires foreign exchange assets. The entire transaction can be simplified as the government issuing bonds, simultaneously obtaining bonds and foreign exchange. This is analogous to the central bank's actions in foreign exchange transactions, where it provides banks with reserve money in exchange for foreign exchange.

The creation of bonds to replace assets is a form of securitization. Banks package their assets for asset management companies (AMCs), and these assets serve as the basis for issuing mortgage-backed securities. In this case, the mortgage-backed securities are entirely held by banks, allowing AMCs to expand their balance sheets and exchange their own bonds for non-performing loans. Similarly, the government uses its bonds to replace central bank foreign exchange and Central Huijin's assets. This process is akin to "foreign exchange asset-backed securities." From the available news reports, it appears that Central Huijin's value was often compared to asset-backed securities due to its significance.

Central Huijin's Acquisition by China Investment Corporation (CIC)

The CNY 15.5 trillion special treasury bonds were exchanged for USD 200 billion, which constituted the entire capital of CIC. The end-of-2008 annual report of CIC indicated that its long-term equity investments amounted to USD 171 billion, mainly corresponding to Central Huijin's equity. Extrapolating from Central Huijin's balance sheet, its net assets' valuation around the end of 2007 should have been approximately USD 130 billion. In December 2007, we observed that the central bank's other assets decreased by CNY 4881 billion, which aligns more with the issuance of the CNY 15.5 trillion special treasury bonds.

Furthermore, the central bank acquired CNY 15.5 trillion in bond assets. To balance its assets and liabilities, the central bank would either need to increase its liabilities by CNY 15.5 trillion or decrease other assets by the same amount. It is possible that the treasury gained several billion USD in foreign exchange or capital from Central Huijin's sale to CIC, enabling the government's investment in China Development Bank (CDB) of around USD 20 billion. Additionally, the government made the entire acquisition of Central Huijin.

At this point, it becomes apparent that the central bank might hold Central Huijin's equity under different reporting methods. Part of it might be categorized under "other assets," totaling approximately USD 67 billion. Another portion could correspond to the item "Claims on Other Financial Institutions," which experienced a sharp reduction in August and December of that year, coinciding with the issuance of the special treasury bonds. This arrangement possibly results from varying interpretations of how the central bank accounts for its equity in Central Huijin.

Regardless of how it's categorized, selling Central Huijin would not impact the central bank's liabilities, thus not affecting bank assets held at the central bank. Nevertheless, the appreciation of Central Huijin required adjustments in other balance sheet items. If the central bank transferred the entire appreciation gained from Central Huijin's sale to the government, which subsequently spent the funds, the central bank would effectively create new base money. However, if the central bank established a reserve-like account to record the appreciation portion separately, this liquidity creation would be avoided.

Afterword

The birth of Central Huijin and China Investment Corporation (CIC) presents one of the most intricate cases in central bank and bank balance sheet analysis. It is unparalleled, more complex, and captivating than scenarios like foreign exchange reserve requirements, the exchange of non-performing assets with asset management companies (AMCs), or the bond exchange for local government debt. This case involves actions and issues related to balance sheet expansion and contraction, the formation of assets and liabilities, and the nature of foreign exchange reserves.

The perspectives offered by this viewpoint could overlook the practical issues that state-owned banks faced at the time, which were not solely due to capital adequacy concerns. Many underlying factors such as institutional, property rights, and financial technology issues were responsible for the banks' inability to generate profits. While financial measures could temporarily improve capital adequacy, these measures did not address the broader problems that led to persistent losses and non-performing loans.

Furthermore, the issuance of debt to capitalize banks would inevitably result in a reduction of reserve requirements. From the perspective of the government, this reduction signifies a decrease in the government's deposits with the central bank relative to the overall balance of government debt. Conversely, from the central bank's perspective, it indicates a reduction in the proportion of foreign exchange reserves (foreign exchange deposits) to central bank liabilities. This lower reserve coverage could weaken the central bank's ability to maintain currency stability.

Nevertheless, even without direct interventions from the government and central bank, banks still have alternatives. Banks and insurance companies could collaborate, with banks extending interbank loans to insurance companies, and insurance companies reciprocating by injecting capital into banks. This arrangement could mimic banks lending to themselves for capital injection, and while insurance companies' liability ratios might increase (or reserve ratios decrease), they could still remain within manageable bounds. However, the aim should be to encourage investors to contribute their funds. Ideally, the dramatic capital injections carried out by the central bank and government in that era should remain in history, serving as timeless examples rather than repetitive exercises.