想不到,人民币盯住货币篮子结果竟是这样!Unexpectedly, pegging the Chinese yuan to a currency basket has led to this outcome!
2016年7月15日(微信公众号原文已不见,现在采用知乎上的时间)
李大鸟,扑克投资家作者,资深利率汇率研究员。
导语:“先天缺陷”的人民币汇率定价机制,该怎么改?
目前市场对于人民币中间价的定价机制越来越清楚了,无论是做市银行还是非做市机构,每天都可以对中间价做出基本正确的预测。中间价的波动也带动了人民币市场价的波动,同时带来了市场参与者持有的外汇头寸人民币估值的波动。但仔细思考这种波动,可以看到建立在人民币篮子基础上的波动,给市场带来了一些出乎意料的特点。
美元、人民币、欧元大三角汇率博弈
在人民币篮子设立之初,就已经有市场人士指出,按照贸易权重设立的篮子并不能反映交易的实际情况。在岸市场交易量最大的货币对是美元人民币,2015年达到29万亿元人民币,而第二位的欧元仅有4257亿元人民币,是美元兑人民币交易额的千分之一。交易量份额和贸易权重有着巨大的差别,反映出决策层没有考虑外汇的交易用途,而只考虑了贸易这个使用外汇的细分领域。不过,该不该考虑交易用途并不是本文要讨论的重点。
外汇市场报价往往用表格表示,两两货币之间有一个汇率。对于外汇市场之外的观察者来说,给出美元、欧元和人民币,他很容易看到三个汇率:美元兑欧元、欧元兑人民币、美元兑人民币。从数学上来看,知道欧元人民币汇率和美元人民币汇率,就可以自动计算出欧元美元汇率,比如1欧元兑换10人民币,1美元兑换6.25人民币,那么1欧元必然兑换1.6美元。
但从金融上来看,因果关系并不是任意的,也就是说并不是欧元人民币交易者和美元人民币交易者共同决定了欧元美元的汇率,1欧元兑1.6美元的汇率,并不是1欧元兑换10人民币和1美元兑换6.25人民币的结果。因为欧元兑美元、美元兑人民币和欧元兑人民币这三个市场完全不在一个量级上。欧元兑美元的交易量单日就可以超过万亿美元,是美元人民币的数十倍。
在欧元兑美元、美元兑人民币和欧元兑人民币这三个市场上,欧元兑美元汇率相对于中国投资者是给定的,中国交易者的欧元兑美元交易恐怕完全无法对价格产生影响,而是完全淹没在全球投资者的买卖单当中。在境内市场,美元兑人民币的交易量,又完全盖过欧元人民币的交易量,因此我们不能说欧元兑美元汇率和欧元兑人民币的汇率决定了美元兑人民币的汇率,而只能说欧元兑美元的汇率和美元兑人民币的汇率决定了欧元兑人民币的汇率。
CFETS人民币指数包括13个外币兑人民币的汇率,这种表达法没有反映人民币市场的真实情况,因为欧元等12个非美外币兑人民币市场和美元兑人民币市场根本就不平等。后者是中国交易者用自己的美元和人民币交易出来的结果,而前者并不是非美兑人民币交易者交易出来的结果,而是被动地在给定的非美兑美元以及美元兑人民币价格上交易的结果。
“交易出价格”和“在价格上交易”的差异,反映了交易者对市场的影响力以及影响本身,两者本没有一刀切的划分标准,但前面提到,非美对人民币的交易量是美元兑人民币交易量的千分之一,其影响简直可以忽略不计。同理,欧元兑美元在全球的交易量是中国外汇市场总交易量的数十倍,中国对欧元兑美元汇率的影响力也非常微弱。
真正反映人民币市场情况的表达形式,应该是美元兑人民币汇率+12个非美外币兑人民币汇率。这种拆分看似是数学上的形式等价变化,但反应的其实是实际交易的情况,即非美兑人民币汇率的走势,乃是非美兑美元交易和美元兑人民币交易的结果。非美兑美元的汇率相对中国市场而言先定,然后中国市场确定美元兑人民币汇率,最终得到一个人民币篮子的结果。
在13个外币兑人民币的表达法下,我们看到美元兑人民币占CFETS篮子的比重是1/4,但事实上每个非美货币兑人民币当中,都包含了非美兑美元和美元兑人民币两部分。如果考虑美元兑人民币+12个非美兑美元这一表达法,我们会意识到,这一表达法意味着把美元兑人民币的部分从每一个非美兑人民币汇率中剥离,而每一个剥离出来的美元兑人民币部分,都带着这个非美货币兑人民币的汇率在CFETS篮子里面的权重。
这样一来,我们会发现美元兑人民币这一汇率对CFETS篮子的影响,远大于12个非美兑美元对篮子的影响。如果用美元兑人民币+12个非美兑美元一起表达CFETS篮子,那么美元兑人民币在上述表达法下的权重,就不是美元兑人民币+12个非美兑人民币表达篮子时美元兑人民币的权重1/4,而是1/2强。
问题的关键,并不是在这1/2强上,而是在于另外的1/2弱上。对于美元兑人民币+12个非美兑美元表达的这个篮子,欧美、美日、澳美、镑美等货币对加总的权重也能达到1/4。而考虑了中间价定价机制中包括维持篮子稳定这一项,那么当12个非美兑美元汇率变化的时候,美元兑人民币首先要修正上述变化至篮子稳定的程度,然后考虑收盘价和中间价的价差。
美元兑人民币的汇率以及12个非美兑美元的汇率,就这样被稳定篮子这个目标在一定程度上联系了起来。非美兑美元的汇率变动不受中国交易者影响,中国交易者被动接受12个非美兑美元汇率变化。这些变化出来后,美元兑人民币中间价要变化,来修正12个非美兑美元汇率变动给篮子造成的影响。
人民币汇率篮子的“先天缺陷”
人民币篮子以及维持篮子稳定的定价机制,让非美兑美元的交易者对美元兑人民币汇率有了“意外”的影响力,也就是说这些交易者即使没有在美元兑人民币市场放哪怕一美元,也可以对美元兑人民币的中间价产生影响。最近市场人士也注意到了,英国退欧导致的英镑欧元暴跌,导致了人民币从机制上需要对美元贬值。如果没有这个机制,恐怕人民币对美元还不会贬值这么多,因为退欧后新兴市场国家货币包括卢布、林吉特、雷亚尔、等货币对美元都是升值的,市场在赌联储维持宽松和更宽松。
这样的定价机制,对于市场参与者而言也带来了“意外”的挑战。人民币汇率的研究者如果想研究美元兑人民币的汇率,首先要研究清楚欧美、美日、镑美、澳美等货币的走势,以及别忘了美元兑卢布,这个货币由于自己波动率大,篮子权重也不小,因而对美元人民币常常有不容忽视的影响。研究清楚非美兑美元走势之后才能研究人民币汇率,这相当于人民币汇率研究员要做一些全球汇率研究总监的工作。
而搞清楚了上面的情况,我们也能更好理解现在的状况。过去考虑人民币升值贬值,只需要关注美元兑人民币的汇率即可,因为这是人民币外汇市场的最主要交易。不过,即使在那个时候也存在两种可能,汇率从6.1变为6.2可能意味着人民币贬值,也可能意味着美元升值。很多人会觉得人民币贬值和美元升值是一回事,但其实不是。
这件事情在更广阔的外汇市场里面会更清楚。比如“退欧”导致英镑暴跌,恐怕不会有人认为是欧元升值。又比如“退欧”后日元走强,恐怕也不会有人说是美元疲软。之所以说是英镑贬值而不是欧元升值,是日元升值而不是美元贬值,其根本原因还是在于促使上述贬值和升值的,是人们广泛卖出英镑的操作和广泛买入日元的操作,而不是广泛买入欧元和广泛卖出美元的操作。所以我们也看到,上述的广泛卖出英镑体现为英镑对所有其他货币贬值,而广泛买入日元的操作体现为日元对所有其他货币升值。
那人民币篮子下跌是否意味着人民币的贬值呢?从2016年初到现在,人民币篮子从100.94下跌至94.25,美元兑人民币从6.5上升至6.7,那么人民币到底是贬值了6%还是3%呢?如果仔细观察13个外币兑人民币汇率兑CFETS指数的贡献,我们会发现导致CFETS指数下跌的罪魁祸首是日元兑人民币的上涨,从年初的5.4变为现在的6.6,幅度达到22%。次要的因素是欧元、卢布、美元和林吉特兑人民币的上涨。
看上去,人民币贬值很严重。但我们前面提到,日元兑人民币、欧元兑人民币、卢布兑人民币、林吉特兑人民币的交易量非常有限,我们必须把上述非美兑人民币的汇率拆成两部分,一部分是中国市场参与者被动接受的非美兑美元的交易导致的汇率,一部分是中国市场参与者主动影响的美元兑人民币的交易导致的汇率。非美兑人民币的交易者根本没有定价能力,他们都是被动接受者。
因此,人民币相对日元、欧元、卢布、林吉特下跌这种表述虽然在数学上是正确的,但是在金融上是不反映实际情况的,因为市场上真实发生的情况是日元上涨、欧元上涨、卢布上涨、林吉特上涨。做多日元做空人民币的获利根源和风险敞口,不在于人民币,而在于日元。欧元显然体现的更明显,在脱欧当天欧元兑人民币从7.43暴跌至7.25,做多欧元做空人民币的投资者总不能将亏损归咎于人民币升值。
所以即使出现了人民币对所有外币贬值的情况,我们也不能认为人民币是在广泛地贬值,而只能说非美相对美元上涨,人民币相对美元贬值,因为这是市场的真实情况。关注人民币相对非美货币的走势当然有价值,但是我们时刻要知道非美兑人民币的汇率不是中国市场参与者直接交易出来的,而是被动接受非美兑美元以及美元兑人民币交易者的交易结果。考虑到目前美元和人民币的波动水平,非美兑人民币的主要波动来自非美兑美元的波动。
汇率定价机制应提升主动参与度
什么时候中国才能成为12个非美兑人民币的主动定价参与者而不是被动价格接受者?第一步是境内单只非美货币兑人民币的交易量和美元兑人民币的交易量旗鼓相当,在这个时候非美和人民币的汇率掌握在了中国交易者手里,乃至美元兑人民币的汇率也会受到非美兑人民币交易的影响。第二步是非美兑人民币的交易量和非美兑美元的交易量旗鼓相当,在这个阶段非美兑人民币的交易甚至可以影响非美兑美元的汇率,更不用说影响美元兑人民币的汇率。
第二步的实现当然就是人民币国际化,而第一步的实现,要靠国内外汇市场的进一步活跃,引入越来越多的市场参与者。一个特殊重要的要素是美元兑人民币汇率的双向波动,因为非美兑人民币汇率一直是双向波动的。但如果美元兑人民币维持某种单向波动的状态,那么考虑了波动之后,美元兑人民币交易的吸引力仍将远高于非美兑人民币的吸引力,这不利于市场参与者广泛进入非美兑人民币市场。
这里还隐含了另外一个问题,那就是2%的波动区间现在似乎已经丧失了意义,和去年这个时候人民币市场价“贴边”交易不同。这给人的感觉是:市场自己还是不知道怎么动才好,现在有个中间价在那里动,那就先跟着动,不去想这个动的办法是否符合自己的意愿,因为自己本来也没有想法。隔日的波动有了,日间的波动还很少。
一个婴儿被拉着来回走,看起来去了不少地方,但还不会自己迈腿,路也不是自己走的。放在推车里虽然实现了去各种地方的目的,但是他还是不会走路,腿上没有肌肉,神经系统也没有形成反射。如果让他自己走,他极有可能乱走一气,也可能摔倒在地,但为了早日学会走路,他自己的路,当然还是应该让他自己走。
The current market has become increasingly clear about the pricing mechanism of the Chinese yuan's central parity rate. Whether it's market-making banks or non-market-making institutions, they can make reasonably accurate predictions of the central parity rate every day. The fluctuations of the central parity rate have also influenced the fluctuations of the market-based exchange rate of the Chinese yuan, resulting in fluctuations in the valuation of foreign exchange positions held by market participants. However, upon careful consideration of these fluctuations, unexpected characteristics arising from the fluctuations based on the yuan currency basket become apparent.
The USD, CNY, EUR Triangle Exchange Rate Game
Since the inception of the yuan currency basket, market participants have pointed out that the basket constructed based on trade weights does not reflect the actual trading situation. Among the currency pairs traded in the onshore market, the largest is the USD/CNY pair, which reached ¥29 trillion in 2015, while the second-largest, EUR/CNY, only accounted for ¥425.7 billion, just a thousandth of the trading volume of USD/CNY. The difference between trading volume shares and trade weights reflects that the decision-makers did not consider the trading use of foreign exchange but focused only on the trade's specific sector. However, whether or not to consider trading use is not the focus of this article.
Exchange rate quotations in the foreign exchange market are often represented in tables with exchange rates between currency pairs. For observers outside the foreign exchange market, given the USD, EUR, and CNY, they can easily observe three exchange rates: USD/EUR, EUR/CNY, and USD/CNY. Mathematically, knowing the EUR/CNY exchange rate and the USD/CNY exchange rate allows for the automatic calculation of the EUR/USD exchange rate. For example, if 1 EUR equals 10 CNY and 1 USD equals 6.25 CNY, then 1 EUR must equal 1.6 USD.
However, causality in finance is not arbitrary, meaning that the EUR/USD exchange rate of 1.6 is not a result of the EUR/CNY exchange rate of 10 and the USD/CNY exchange rate of 6.25. This is because the EUR/USD, USD/CNY, and EUR/CNY markets are not on the same scale. The daily trading volume of EUR/USD can exceed trillions of USD, which is tens of times larger than the USD/CNY trading volume.
In the EUR/USD, USD/CNY, and EUR/CNY markets, the EUR/USD exchange rate is given relative to Chinese investors; Chinese traders' EUR/USD trades likely have minimal impact on prices and are submerged within the buy and sell orders of global investors. In the domestic market, the trading volume of USD/CNY completely overshadows that of EUR/CNY. Therefore, it cannot be said that the EUR/USD and EUR/CNY exchange rates determine the USD/CNY exchange rate; it can only be said that the EUR/USD and USD/CNY exchange rates determine the EUR/CNY exchange rate.
The CFETS CNY Index includes the exchange rates of 13 foreign currencies against the CNY. This representation does not reflect the real situation of the CNY market, as the 12 non-USD foreign currency exchange rates against the CNY and the USD/CNY market are fundamentally unequal. The latter is the result of Chinese traders' transactions using their own USD and CNY, while the former is the result of trading on given non-USD/USD and USD/CNY prices.
The difference between "trading out a price" and "trading on a price" reflects the influence and impact of traders on the market. There is no strict criterion for categorizing them, but as mentioned earlier, the trading volume of non-USD against the CNY is just a thousandth of the USD/CNY trading volume, making their influence negligible. Similarly, the trading volume of EUR/USD globally is tens of times larger than China's total foreign exchange market volume, so China's influence on the EUR/USD exchange rate is also very weak.
The expression that truly reflects the CNY market situation should be the USD/CNY exchange rate plus the exchange rates of 12 non-USD currencies against the CNY. Although this decomposition appears to be a mathematically equivalent change in form, it actually reflects the situation of actual trading. The trend of non-USD against the CNY is the result of trading non-USD/USD and USD/CNY. The EUR/USD exchange rate relative to the CNY market is predetermined, and then the USD/CNY exchange rate is determined by the Chinese market, ultimately resulting in a CNY basket.
Under the expression of the 13 foreign currencies' exchange rates against the CNY, we see that the weight of the USD/CNY exchange rate in the CFETS basket is 1/4. However, each non-USD currency exchange rate against the CNY includes two parts: the non-USD/USD and USD/CNY parts. If we consider the expression of USD/CNY plus 12 non-USD/USD, we will realize that this expression means removing the USD/CNY part from each non-USD exchange rate against the CNY, and each removed USD/CNY part carries the weight of the non-USD currency exchange rate against the CNY in the CFETS basket.
This way, we find that the impact of the USD/CNY exchange rate on the CFETS basket is much greater than the impact of the 12 non-USD/USD exchange rates on the basket. If we express the CFETS basket with USD/CNY plus 12 non-USD/USD, the weight of USD/CNY in the above expression is not the 1/4 weight when expressed as USD/CNY plus 12 non-USD/USD in the basket. Instead, it is a stronger 1/2.
However, the key issue lies in the weaker half, not the stronger half. Concerning the expression of USD/CNY plus 12 non-USD/USD, the weights of currency pairs such as EUR/USD, GBP/USD, AUD/USD, etc., when summed up, also reach 1/4 of the CFETS basket. When considering the mechanism of maintaining the stability of the basket in the central parity pricing mechanism, when the exchange rates of the 12 non-USD/USD pairs change, the USD/CNY rate must first be adjusted to stabilize the basket, and then the difference between the closing price and the central parity rate is considered.
Thus, the exchange rate of USD/CNY and the exchange rates of the 12 non-USD/USD pairs are to some extent connected by the goal of stabilizing the basket. The fluctuation of the non-USD/USD exchange rate is not influenced by Chinese traders; they passively accept the changes resulting from non-USD/USD and USD/CNY trading. After these changes, the central parity rate of USD/CNY must change to correct the impact of the changes in the 12 non-USD/USD exchange rates on the basket.
The "Inherent Defects" of the CNY Exchange Rate Basket
The CNY basket and the mechanism to maintain the stability of the basket allow non-USD/USD traders to have an "unexpected" influence on the USD/CNY exchange rate. This means that even if these traders have not traded a single dollar in the USD/CNY market, they can still impact the central parity rate of USD/CNY. Recently, market observers have also noted that the pound sterling's sharp drop due to Brexit led to the need for the CNY to depreciate against the USD structurally. Without this mechanism, the CNY might not have depreciated as much against the USD because post-Brexit, emerging market currencies including the ruble, ringgit, and real have all appreciated, and the market is betting on the Fed maintaining loose and looser policies.
This pricing mechanism also poses an "unexpected" challenge for market participants. Researchers studying CNY exchange rates who wish to study the USD/CNY exchange rate must first understand the trends of EUR/USD, GBP/USD, AUD/USD, etc., not to mention USD/RUB, a currency that, due to its high volatility and significant basket weight, has a non-negligible impact on the USD/CNY exchange rate. Only after understanding the trends of the 12 non-USD/USD pairs can researchers study the CNY exchange rate. This is equivalent to CNY exchange rate researchers needing to perform the role of a global exchange rate research director.
Understanding the above circumstances also allows us to better understand the current situation. In the past, considering whether the CNY appreciated or depreciated only required focusing on the USD/CNY exchange rate since it was the primary trade in the CNY foreign exchange market. However, even at that time, two possibilities existed: a change from 6.1 to 6.2 could mean CNY depreciation or USD appreciation. Many people might think that CNY depreciation and USD appreciation are the same, but they are not.
This issue is clearer in the broader foreign exchange market. For instance, the "Brexit" led to a sharp drop in the British pound, but no one would likely claim the euro appreciated. Similarly, the yen's strength after "Brexit" wouldn't be attributed to USD weakness. The reason why it's described as British pound depreciation rather than euro appreciation, or Japanese yen appreciation rather than USD weakness, is because the actions that prompted these depreciations and appreciations were the widespread selling of the pound and the widespread buying of the yen, not widespread buying of the euro or widespread selling of the USD. This is also why we see that the widespread selling of the pound led to its depreciation against all other currencies, and the widespread buying of the yen led to its appreciation against all other currencies.
Does the drop in the CNY basket imply CNY depreciation? From the beginning of 2016 until now, the CNY basket has dropped from 100.94 to 94.25, and the USD/CNY exchange rate has risen from 6.5 to 6.7. So, did the CNY depreciate by 6% or 3%? If we carefully examine the contributions of the 13 foreign currency exchange rates against the CNY to the CFETS Index, we'll find that the primary culprit for the drop in the CFETS Index is the rise of the JPY/CNY exchange rate, which increased from 5.4 at the beginning of the year to 6.6 now, a 22% increase. The secondary factors are the rises of the EUR, RUB, USD, and MYR exchange rates against the CNY.
At first glance, it appears that the CNY has depreciated significantly. However, as mentioned earlier, the trading volumes of JPY/CNY, EUR/CNY, RUB/CNY, and MYR/CNY are very limited. We must separate the non-USD/CNY exchange rates into two parts: the part resulting from non-USD/USD trading that is accepted passively by Chinese traders, and the part resulting from USD/CNY trading that is actively influenced by Chinese traders. Non-USD/CNY traders have no pricing power; they are passive acceptors.
Therefore, while the description of the CNY depreciating against all foreign currencies is mathematically correct, it does not reflect the actual situation in finance. The actual situation in the market is that the JPY has appreciated, the EUR has appreciated, the RUB has appreciated, and the MYR has appreciated. The origin of profits and risk exposure when going long JPY and short CNY lies not with CNY but with JPY. The appreciation of the EUR is even more evident. On the day of the Brexit, the EUR/CNY plummeted from 7.43 to 7.25. Investors who went long the EUR and short the CNY certainly cannot blame their losses on CNY appreciation.
So, even if there is a situation where the CNY depreciates against all foreign currencies, we cannot conclude that the CNY is depreciating extensively. Instead, we can only say that non-USD currencies are appreciating relative to the USD, resulting in CNY depreciation, because this is the market's reality. While it's valuable to focus on the trend of the CNY relative to non-USD currencies, we must always be aware that the non-USD/CNY exchange rates are not directly traded by Chinese market participants; they are the result of the trading of non-USD/USD and USD/CNY traders. Given the current levels of volatility of the USD and CNY, the primary volatility of non-USD/CNY comes from the volatility of non-USD/USD.
Exchange Rate Pricing Mechanism Should Enhance Active Participation
When will China become an active pricing participant for the 12 non-USD/CNY exchange rates rather than a passive price acceptor? The first step is for the trading volumes of non-USD/CNY and USD/CNY to be roughly equivalent within the domestic and foreign exchange markets. At this stage, non-USD and CNY exchange rates are in the hands of Chinese traders, including even the ability to impact the non-USD/CNY exchange rate and, to some extent, the USD/CNY exchange rate. The second step is for the trading volumes of non-USD/CNY and non-USD/USD to be roughly equivalent. At this stage, the trading of non-USD/CNY could potentially impact the non-USD/USD exchange rate and, not to mention, the USD/CNY exchange rate.
The realization of the second step naturally involves CNY internationalization, while the realization of the first step requires further activity in both domestic and foreign exchange markets, attracting more and more market participants. An especially important factor is the two-way fluctuations of the USD/CNY exchange rate because non-USD/CNY exchange rates have always been two-way. However, if the USD/CNY exchange rate maintains a certain one-way fluctuation, then considering the fluctuations, the attractiveness of USD/CNY trading will still be much higher than that of non-USD/CNY. This is not conducive to widespread entry into the non-USD/CNY market by market participants.
This also implies another issue: the 2% fluctuation range now seems to have lost its meaning, different from the "sticking to the edge" trading of the market price from last year. The feeling this gives is that the market itself still doesn't know how to move, and there's a central parity rate there to move, so everyone follows it, without considering whether this movement is in line with their own desires, as they don't have ideas themselves. There have been fluctuations overnight, but intraday fluctuations are still rare.
A baby being pulled back and forth, although it seems to have gone to many places, it still doesn't know how to walk on its own, and its legs are not yet strong enough. Although achieving various destinations in a stroller accomplishes the goal, it still doesn't know how to walk; its legs lack muscles, and its nervous system hasn't formed reflexes. Even if you allow it to walk on its own, it might wander aimlessly or fall down; however, to learn how to walk sooner, it's essential to let it walk on its own.