原创 2016-08-09 季天鹤央行观察
文 \ 季天鹤央行观察专栏作家

在此前通过专栏沟通人民币定价机制和资产负债角度的流动性分析之后，央行在2季度的货币政策报告中又继续深化了上述两方面的内容，引人深思。

2016年2季度的《货币政策执行报告》近日出炉。当中，央行回应了市场关心的诸多问题，特别包括人民币汇率机制以及央行避不降准这两个问题。本文也就上述两个问题进行一些讨论，非常欢迎各位读者批评指正。

在人民币汇率机制方面，2016年1季度的《货币政策执行报告》已经介绍了目前“收盘汇率+一篮子货币汇率变化”的定价机制，而2季度的《货币政策执行报告》将收盘汇率与一篮子货币汇率变化变化这两个因素对于人民币中间价的影响进一步细化，可以看作是对1季度关于定价机制介绍的深化。

央行选择的考察区间起点2月15日并非随意挑选，而是2016年春节之后的第一个交易日，而在节前联储刚刚放出鸽派信号，逆转了市场关于加息走势的预期，再叠加长达一周的休市，人民币汇率市场如同经历了“格式化”的电脑，处于一个平衡而清新的起点，这一天成为央行眼中“收盘汇率+一篮子货币汇率变化”定价机制的开始也就不足为奇了。

此番央行指出，自2月15日至6月末，收盘汇率对中间价的累计影响为贬值2489 个基点，一篮子货币汇率变化对中间价的累计影响为升值1295 个基点。由此计算，两者合计的影响为贬值1194基点。而考察2月15日的中间价和6月30日的中间价，我们看到美元兑人民币从6.5118变化至6.6312，变化幅度恰为1194基点。

也就是说，央行其实是给出了它认为的中间价变化的影响来源，其中收盘汇率对中间价的影响是主要因素，主导了贬值，一篮子货币汇率变化对中间价的影响是次要因素，缓解了贬值。之所以一篮子货币汇率变化对人民币贬值起到了缓解作用，主要是非美货币对美元总体走强，为了维持人民币指数对美元的稳定，人民币对于美元需要升值。

但问题在于，人民币最主要的交易是对美元交易，这一货币对占到人民币对外币交易额的99%以上。非美货币对人民币的汇率，完全决定于欧元兑美元的汇率和美元兑人民币的汇率，而非交易的结果。一篮子货币汇率变化成为人民币中间价变化的影响因素，非美货币对美元的交易者，意外获得了对美元人民币中间价的影响力。

在美元兑人民币中间价变动和人民币指数变动方面，央行称“美元兑人民币中间价变动=今天美元兑人民币中间价-昨天美元兑人民币中间价=昨天美元兑人民币收盘价-昨天美元兑人民币中间价+今天维持篮子稳定需要的美元兑人民币汇率调整”，等式中的三个要素在确定今天美元兑人民币中间价时都是给定的。

而对于每天收盘时候人民币指数的变动，我们考虑一个以中间价作为计算基准的CFETS指数。今天的CFETS指数，首先因非美兑美元的波动和昨天产生变化，然后又因美元兑人民币为了维持指数稳定而修正，被改调节为昨天的CFETS指数，但此后又因为叠加了前一日美元兑人民币收盘价和中间价价差，从而和昨天的CFETS指数不同。也就是说，指数的变化应该只和收盘价差有关。

需要注意的是，美元兑人民币汇率只占CFETS指数中的部分权重，所以因价差带来的指数偏差波动幅度，只是价差自身波动的一部分，我的计算是57%的权重。看到这里，我们可以思考一个问题：央行认为“参考一篮子货币”有利于稳定人民币指数，但人民币篮子有没有可能比中间价更不稳定？

如果未来人民币收盘价和中间价进一步收敛，但收盘价仍处于更贬值的位置，而非美兑美元仍维持同样的升值变动，那么来自非美兑美元的对中间价的升值向修正，相对于收盘价贬值方向的价差就会更厉害，从而使得中间价波动率的下降快于价差收敛导致的指数波动率的下降，可能会发生指数波动大于中间价波动的情况。

换一个角度说明：如果非美兑美元市场休市一个月，汇率冻结，只有美元兑人民币市场开始，那么中间价就不再有来自非美的扰动，如果收盘价依然持续较中间价更为贬值，那么人民币指数每天按照价差*57%的水平下跌（人民币贬值），人民币中间价每天按照价差*100%走高（人民币贬值），人民币篮子的波动肯定是要小于人民币中间价的波动的。

另一个极端情况下，如果收盘价和中间价没有价差，但是非美兑美元不断变化，那么人民币指数每天维持不动，波动为零，但人民币收盘价每天按照非美的变动而调整，波动为正。此时，人民币指数的波动也小于人民币中间价的波动。但如果又有价差又有非美扰动，人民币中间价和人民币指数的波动孰大孰小，就涉及到价差和非美变动的相关关系了，而不是确定无疑的。

而讨论了定价机制之后，我们自然就过渡到了第二个问题，也就是降准能否成为人民币贬值的推手，或者说人民币贬值是不是降准与否的真正考虑因素。2015年4季度的《货币政策执行报告》中的专栏5，已经从资产负债表的角度讨论了准备金率和流动性以及人民币贬值预期的问题。

央行对于降准的担忧有两个，一是通过利率渠道影响人民币汇率，二是通过情绪渠道影响人民币汇率。情绪渠道当然是非常有可能的，但情绪渠道其实是无视了降准是否真的意味着宽松或者导致宽松，不容易讨论。而从利率渠道看，目前市场讨论较多的是委外资金配置债券压低债券收益率，这一市场行为恐怕无关央行。

而且，这一现象本身就反映了一个问题：如果人们真的很担心人民币下跌，那为什么还会有这样多的委外资金愿意投资收益率如此之低的人民币债券？这是否说明，人们对于人民币资产其实是相当有信心的，或者说人们出于很多原因还是愿意或者被迫持有人民币？如果降准会引发人们抛弃人民币资产购买美元资产，会不会在抛弃委外资产的过程中抬高收益率？

对于买汇炒汇的“投机者”而言，在中国境内这些投机者恐怕主要以非银行机构或者个人为主。非银机构和个人的结售汇并不直接影响银行的资产侧，而非银机构和个人向境外汇出外汇，银行在合法合规的情况下只能执行指令，因此央行通过对银行资产这边避不降准的方式，来隔山打牛，作用于银行负债那边的投机者，效果恐怕也比较间接。

在人民币的定价机制方面，目前包含维持指数稳定这一因素，而维持指数稳定的动机是非美货币兑美元的汇率变化，而非美货币兑美元的汇率变化恐怕不受中国在降准这一级别的政策影响。也就是说，美元兑人民币汇率中有一块儿是不受降准影响的。如果日本央行和欧洲央行都很强硬，把美元指数推到90，尽管美元的利率汇率环境可能依旧，美元和人民币交易者遇到的市场情况几无变化，但人民币会遭受强大的升值推力。

因此，考虑避不降准的原因，可能还是需要回归到降准本身。央行在降准的同时应该会大幅削减OMO/MLF的规模，6月份新增的7000亿对银行信贷恐怕都要收回，才能维持目前的利率水平。从央行的角度看，原本可以用来制约银行的“MLF到期停止叙做”这一工具就没有了，银行也不用考虑在央行操作到期时准备人民币流动性的问题。在这一情况下，银行可能更有动力配置外币资产，不利于央行保守外汇储备。

另外一个原因，可能还是和MLF不能随意停做有关。如果央行的流动性支持本身意味着政策性定向投放的需要，那么在政策不变的情况下，MLF的量就不会减少而会继续叙做甚至增加下去，而这就意味着央行无法实现伴随着回笼流动性的降准。除非出现大规模的购汇，导致银行除了因为负债侧存款不断扩大而遭遇越来越高的法定缴存规模外，还因为从央行购汇而面临在央行存款的流失，从而满足缴存规模的能力也遭到削弱，那么在MLF自身就能够满足银行体系的缴准需要的前提下，央行也就确实没有必要缴准了。

After previously communicating liquidity analysis from the perspectives of RMB pricing mechanism and asset-liability, the central bank further delved into these aspects in its second-quarter monetary policy report, leaving one with much to ponder.

The Monetary Policy Implementation Report for the second quarter of 2016 has recently been released. In it, the central bank addressed numerous market concerns, particularly concerning the RMB exchange rate mechanism and the central bank's decision to refrain from reducing the reserve requirement ratio (RRR). This article also discusses the aforementioned two issues and welcomes readers' critique and correction.

Regarding the RMB exchange rate mechanism, the first-quarter Monetary Policy Implementation Report of 2016 introduced the current pricing mechanism of "closing exchange rate + a basket of currency exchange rate changes" for determining the central parity rate. The second-quarter report further refined the impact of both the closing exchange rate and the basket of currency exchange rate changes on the central parity rate, which could be seen as a deepening of the introduction of the pricing mechanism in the first quarter.

The central bank's chosen examination period starting on February 15th was not randomly selected, but rather the first trading day after the 2016 Spring Festival. Prior to the holiday, the central bank had just sent a dovish signal, reversing market expectations of an interest rate hike. Coupled with a market closure lasting a full week, the RMB exchange rate market resembled a "formatted" computer and started from a balanced and fresh point. As such, this day being the starting point for the central bank's "closing exchange rate + a basket of currency exchange rate changes" pricing mechanism is not surprising.

In this instance, the central bank pointed out that from February 15th to the end of June, the cumulative impact of the closing exchange rate on the central parity rate resulted in a depreciation of 2489 basis points, while the cumulative impact of the basket of currency exchange rate changes led to an appreciation of 1295 basis points. Thus, calculating the combined effect of the two results in a depreciation of 1194 basis points. Examining the central parity rate on February 15th and June 30th, it's evident that the USD to RMB exchange rate changed from 6.5118 to 6.6312, precisely a change of 1194 basis points.

In other words, the central bank is essentially providing the sources of influence it believes are behind changes in the central parity rate. The impact of the closing exchange rate is the primary factor leading to depreciation, acting as the main driver, while the impact of the basket of currency exchange rate changes serves as a secondary factor alleviating depreciation. The reason the basket of currency exchange rate changes alleviates RMB depreciation is mainly due to the overall strength of non-USD currencies against the USD. To maintain the RMB index's stability against the USD, the RMB needs to appreciate against the USD.

However, the challenge lies in the fact that the most significant currency pair traded for RMB is the USD, accounting for over 99% of RMB's foreign exchange transactions. The exchange rate of non-USD currencies against RMB is entirely determined by the EUR to USD exchange rate and the USD to RMB exchange rate, outcomes unrelated to trade. The basket of currency exchange rate changes becomes a factor influencing changes in the RMB central parity rate. Unexpectedly, non-USD currency traders gain influence over the USD to RMB central parity rate.

Regarding changes in the USD to RMB central parity rate and RMB index, the central bank states that "USD to RMB central parity rate change = today's USD to RMB central parity rate - yesterday's USD to RMB central parity rate = yesterday's USD to RMB closing rate - yesterday's USD to RMB central parity rate + today's USD to RMB exchange rate adjustment required to maintain basket stability." The three components of this equation are given when determining today's USD to RMB central parity rate.

As for changes in the RMB index at the end of each day, the article considers CFETS index based on the central parity rate as a benchmark. Today's CFETS index is initially affected by non-USD to USD fluctuations and the changes from yesterday. It is then adjusted to maintain stability in the index due to USD to RMB changes, becoming the adjusted CFETS index from yesterday. However, this adjusted index further differs due to the addition of the USD to RMB closing rate and the central parity rate difference from the previous day, meaning the index changes are solely related to the closing rate difference.

It's worth noting that the USD to RMB exchange rate only carries a partial weight in the CFETS index, so the deviation in the index caused by the closing rate difference's fluctuations is merely a fraction of the rate difference itself, specifically 57% according to my calculations. At this point, one can contemplate a question: while the central bank believes that referencing the basket of currencies benefits RMB index stability, could the RMB basket of currencies be even less stable than the central parity rate?

If in the future the RMB closing rate and central parity rate converge further, but the closing rate still remains more depreciated and non-USD to USD changes continue to appreciate at the same pace, then the upward adjustment from non-USD to USD towards the central parity rate, in relation to the downward closing rate direction, will lead to a more significant rate difference. Consequently, the reduction in central parity rate volatility due to rate difference convergence might surpass the decline in index volatility, possibly resulting in a scenario where index volatility exceeds central parity rate volatility.

Looking at it from a different angle: if the non-USD to USD market is closed for a month with exchange rates frozen, leaving only the USD to RMB market active, then the central parity rate will no longer be influenced by non-USD currencies. If the closing rate continues to depreciate relative to the central parity rate, the RMB index will daily decrease based on a level of rate difference multiplied by 57% (RMB depreciation), while the central parity rate will daily rise based on the same rate difference multiplied by 100% (RMB depreciation). Consequently, the RMB basket's volatility will undoubtedly be smaller than that of the central parity rate.

In another extreme scenario, if there's no rate difference between the closing rate and the central parity rate, but non-USD to USD continues to change, then the RMB index will remain unchanged daily, with zero volatility. However, the RMB closing rate will adjust daily based on non-USD changes, leading to positive volatility. At this point, the RMB index's volatility will be lower than that of the central parity rate. Yet, if both rate difference and non-USD disturbance exist, the relative volatility of the RMB central parity rate and index depends on the relationship between rate difference and non-USD changes, rather than being unequivocal.

After discussing the pricing mechanism, we naturally transition to the second question, which is whether the reserve requirement ratio (RRR) reduction can act as a catalyst for RMB depreciation, or whether RMB depreciation is the genuine factor for considering whether to implement an RRR reduction. Column 5 of the fourth-quarter 2015 Monetary Policy Implementation Report has already discussed issues concerning reserve ratios, liquidity, and expectations of RMB depreciation from the perspective of the balance sheet.

The central bank has two concerns regarding RRR reduction. The first is the impact on the RMB exchange rate through the interest rate channel, and the second is the influence on the RMB exchange rate through the sentiment channel. The sentiment channel is certainly possible, but it disregards whether RRR reduction truly implies easing or leads to easing, making it difficult to discuss. Looking from the perspective of the interest rate channel, the topic that the market currently discusses most is external funds allocating bonds, which suppresses bond yields. However, this market behavior is likely unrelated to the central bank.

Furthermore, this phenomenon itself raises a question: If people are genuinely concerned about RMB depreciation, why would there be so many external funds willing to invest in RMB bonds with such low yields? Does this not indicate that people have significant confidence in RMB assets, or that for various reasons, they are still willing or compelled to hold RMB? If RRR reduction would prompt people to abandon RMB assets in favor of purchasing USD assets, could this process of relinquishing external assets lead to higher yields?

For the "speculators" engaged in forex trading, these speculators are mainly non-bank institutions or individuals within China. The forex transactions of non-bank institutions and individuals do not directly impact banks' asset sides. When non-bank institutions and individuals send forex abroad, banks can only execute orders in a legal and compliant manner. Therefore, the central bank's approach to refrain from reducing RRR on the asset side of banks acts as an indirect influence on speculators on the liability side of banks, attempting to reach a goal from afar.

Regarding the pricing mechanism of the RMB, it currently includes the factor of maintaining index stability. The motivation for maintaining index stability lies in the changes in non-USD currency against the USD. However, these changes in non-USD currency against the USD are probably not affected by policy actions such as RRR reduction at China's level. In other words, a portion of the USD to RMB exchange rate remains unaffected by RRR reduction. If the Bank of Japan and the European Central Bank maintain a strong stance, pushing the USD index to 90, even if the USD interest rate and exchange rate environment remains unchanged, the market conditions faced by USD and RMB traders may remain unaltered, yet the RMB would be subjected to a strong appreciation pressure.

Therefore, considering the reasons for not reducing RRR, one might need to revert to the essence of RRR reduction itself. While reducing RRR, the central bank is likely to significantly decrease the scale of open market operations (OMO) and medium-term lending facility (MLF). The added 700 billion in June for bank credit might have to be withdrawn in order to maintain the current interest rate level. From the central bank's perspective, the tool that was once available to restrict banks, "stopping MLF from rolling over upon maturity," would be lost. Banks would also no longer need to consider preparing RMB liquidity when central bank operations mature. In this case, banks might be more motivated to allocate foreign currency assets, which is unfavorable to the central bank's conservative foreign exchange reserves.

Another reason might be related to the inability to cease MLF arbitrarily. If the central bank's liquidity support implies the need for policy-driven targeted injections, then assuming policy remains unchanged, the quantity of MLF would not decrease but rather continue to roll over or even increase. This would mean that the central bank cannot achieve an RRR reduction accompanied by liquidity withdrawal. Unless there is massive forex purchasing, causing banks not only to encounter continuously expanding deposits on the liability side due to expanded liabilities but also to face outflows of deposits at the central bank due to forex purchasing, which in turn weakens their ability to meet deposit requirements. Under the precondition that MLF can meet the banking system's reserve requirement needs, the central bank indeed has no necessity to implement RRR reduction.