提要：美联储开始实施量化宽松的货币政策，市场担心这种非常规政策会影响到美元。本文认为，一方面，量化宽松措施仅会影响美元的名义价值，其发挥影响力的途径是美国相对通胀率的变化。鉴于欧洲央行和英格兰银行也将实行量化宽松举措，美元未必会受到不利影响。另一方面，迫使美联储采取量化宽松举措的动因在于美国金融基本面的恶化，该因素将损害美元的实际价值。考虑两方面因素，未来6个月之中，在全球去杠杆化进程的推动下，美元将继续对主要货币升值，但幅度不大，而后将有所回落，同时对新兴市场货币仍将大幅升值。
（外脑精华·北京）美联储已开始实施“量化宽松”的货币政策。本文分析了定量宽松的概念，以及这种非常规政策对美元的影响。我们的结论是，包括量化宽松在内，货币政策操作仅会影响名义变量，而不会对美元的实际价值产生明显影响。一方面，如果量化宽松改变了美国的中期通胀前景，就会影响美元的名义价值；另一方面，美联储的量化宽松是否会导致美元贬值还取决于其他央行的举措。鉴于欧洲央行和英格兰银行也将采取量化宽松举措，美元未必会受到不利影响。
虽然量化宽松自身仅会通过相对通胀率和通胀预期影响美元的名义价值，但迫使美联储实施量化宽松的内在结构性问题却会改变美元相对于其他货币的实际价值。从理论上说，美国金融体系的危险状况本应通过美元贬值表现出来。近期美元之所以持续升值，原因在于，在当前的去杠杆化时期，主要储备货币的地位支持了美元。美国财政赤字的扩大将对美元的内在价值构成进一步的压力。
总之，在名义汇率层面上，美联储的量化宽松对美元的影响取决于美国的相对通胀前景及其对通胀预期的影响。但就实际汇率而言，美元的内在价值确实会因其金融业受到的重创而下跌。我们仍然认为，在世界经济减速的过程中（这个阶段至少将持续到明年夏季），美元将继续升值，但在去杠杆化过程结束、而且事实证明美国的经济复苏将比其他经济体更艰难之后，美元将出现一定幅度的贬值。由于美国经济基本面的恶化，未来几个月之中，美元升值对主要货币升值的幅度将小于我们先前的预期。另一方面，我们仍然认为，美元仍将对新兴市场货币大幅升值。
量化宽松的经验与现实
面对通胀率下滑、失业率攀升的局面，在联邦基金利率逼近零点的情况下，美联储很可能会依靠量化宽松政策来维持刺激性货币政策。总体而言，央行可以通过两种方式放松银根：改变货币价格（即利率）或改变货币数量。多年以来，正统的货币政策一直以前一个政策杠杆为中心。然而，随着通胀率回落、短期名义利率逼近零点，从原则上说，央行可以后一种方式、即数量杠杆来实施扩张性货币政策。影响经济活动的是实际利率而非名义利率。如果经济处于通缩状态，那么即使名义利率为零，实际利率也会保持正值。2000年日本面临的情况就是如此－－名义利率已降至零点，但在实际利率为正值的情况下，低迷的货币需求仍不足以令货币政策发挥效力。这就是过去所说的“流动性陷阱”。
然而，目前美国、欧元区和英国的情况却有所不同，它们的名义利率仍未降至零点。这种情况下，货币政策面临的主要问题并不是通缩和货币需求不足、即总需求问题，尽管如果需求进一步减速，就可能出现这个问题。当前的货币政策之所以缺乏效力，主要原因在于短期政策利率对货币总量、信贷和总需求的传导机制已经失效。因此，日本是在零利率的情况下实行了量化宽松政策，而美国、欧元区和英国在没有达到零利率的情况下就实行了量化宽松。
央行放松银根的非常规方式主要有三种。第一，央行可以通过与外界沟通或量化宽松等方式，培养短期利率将长期保持低位的预期。事实上，2001年3月－2006年3月，日本央行实行量化宽松政策的主要目的就在于此。再如，2003年8月，美联储公开市场委员会在公报中称“适应性政策将维持相当长的时间”也是此类放松银根承诺的事例。第二，央行可以扩大其资产负债表的规模，以左右通胀预期。目前，美联储的资产负债表规模已由8月的9000亿美元升至2万亿美元左右。从这个角度看，美联储通过“印钞票”实行量化宽松的做法是显而易见的。第三，央行可以改变其资产负债表的结构。如果投资者将不同资产视为非完全替代品，央行买进特定资产的操作就会对资产价格产生明显影响。就此而言，最好的例子就是长期美国国债。从理论上说，美联储可以大规模买进美国国债，以抑制收益率上升；日本央行在实行量化宽松政策期间，就曾进行这种操作。上述三种量化宽松的方式虽然概念不同，但在操作上可以相互替代。随着美国经济进一步减速，我们预计美联储将运用所有这三种非常规方式。
2008年9月，美联储开始扩大其资产负债表。联储采取量化宽松的目的有两方面。一是买进市场上的证券，以图“启动”银行体系；二是承担私人部门所不愿承担的一部分金融中介职能。在日本央行实行量化宽松政策的2001年3月－2006年3月间，其资产负债表对该国GDP的比率由13%左右升至22%左右。由于美日两国金融体系的根本性差异，美联储资产负债表的规模远小于日本央行——本次危机前，前者的资产负债表对美国GDP的比率仅为6%左右。到10月份，美联储资产负债表已扩大到美国GDP的8%以上，比危机前增长了35%。
要正确地估价联储“印钞”对通胀的影响，关键在于观察广义货币总量的变化，以及“货币乘数”的变化。我们比较了日、美两国M2的变化轨迹。在日本实行量化宽松政策时期，该国的M2货币供应量从GDP的125%左右增至140%以上。就美国而言，9月份至今，M2货币供应量已经由GDP的53%增至57%。不过，虽然近期美国的M2增势迅猛，但当前的M2/GDP比率并未大大超过2003年的水平，当时美联储将联邦基金利率降至1.00%的低点。事实上，近期M2/ GDP比率大幅上升的主要原因在于美国4季度名义GDP的急剧下滑。总而言之，8月份至今，虽然基础货币猛增了34%，但M2仅增长了2.7%。
基础货币和广义货币供应量走势之所以出现严重背离，原因自然在于“货币乘数”的急剧下降。而货币乘数暴跌又反映出，自危机发生以来，美国银行及其他金融机构承担资本中介功能的能力和意愿都大为萎缩。当前美国货币乘数下降的严重程度大大超过了当年的日本：日本方面，货币乘数由资产负债表的10倍左右降至最低点的6.5倍左右，降幅为35%；而在美国，仅仅过去2个月之中，货币乘数就从9以上降至7左右，降幅已达22%。
量化宽松对美元影响有限
那么，美联储的量化宽松政策是否会打击美元？简单说，我们的回答是否定的。但这个回答能否成立，还要取决于很多假设。首先，通缩压力是不可忽视的。虽然分析师对通胀与通缩之争可以有不同的观点，但令人担忧的是，迄今为止，全球决策者依然未能遏制金融危机。认为美国或其他地区不会发生通缩的观点是难以令人信服的。出现持续的通缩的可能性虽然不高，但也远没有低到令人高枕无忧的程度。虽然与当年的日本央行相比，美联储近期的行动要迅速得多，但可以说，美国的问题也比当年的日本更严重，而且本次金融危机已经重创了货币乘数，从而迫使美联储代替私人部门承担了融资功能。简言之，当今的通缩风险比未来的通胀风险更令人担忧，因为我们不敢肯定，我们面临的仅仅是通缩“恐慌”。
我们认为，在通胀和通缩两项风险中，美联储和其他发达国家央行更加担忧后者，其放松银根的举措就反映了这一点。比较而言，新兴市场央行则更担心通胀风险。
投资者已经认识到这种情况，他们下注于七国集团央行能够控制通缩风险。这种情况下，七国集团过度放松的风险就超出了其无法启动经济的风险。换言之，央行越担心通缩，就会采取越有力的措施，因而危机过后出现通胀的风险也就越大。
美元是否会贬值，取决于投资者是否认为中期之内美国的通胀走势是否会失控。我们认为，几乎可以肯定，美联储有足够的时间回收通过量化宽松投放的货币，从而保持对通胀的控制。假设其他条件相同，只要不出现通胀，美元就不会受到打击。虽然流动性陷阱难以应对，但如果宏观经济形势恢复正常，那么对美联储而言，消除货币刺激措施的影响并非难事。当然，上述分析也表明，关键在于，在量化宽松阶段过后，美联储是否有能力消除其影响。
此外，虽然目前很多人仅仅关注着美联储的量化宽松举措和美元，但要不了多久，欧洲的央行很可能也会被迫采取同样的举措。正像美国银行也一样，欧洲大陆和英国的银行体系也不愿承担资本中介功能。因此，正如美联储一样，欧洲央行和英格兰银行很可能也会被迫代替私人金融部门承担一部分金融中介职能。这样，在美联储、欧洲央行和英格兰都采取量化宽松举措的情况下，欧元－美元汇率或英镑－美元汇率存在极大不确定性，而不像一些仅仅关注美联储的投资者所认为的那样，欧元和英镑必定会对美元升值。
值得指出的是，担忧通缩者与担忧通胀者的争论意味着一种倒置的美元微笑曲线。我们的美元微笑曲线理论认为，在美国经济走势强于或弱于世界经济的其它部分时，美元将升值；在二者之间的中间情形下，美元将贬值。而通缩与通胀之争却意味着，在上述两种极端情形下美元将走弱，而在中间情形下美元将走强。
基本面恶化将打击美元内在价值
美联储的“量化宽松”操作会通过通胀影响到美元的名义价值，而美国金融体系的严重恶化则必然会损害美元的内在价值。我们再次强调，汇率是一个相对概念，因此我们需要关注美国之外发达国家的金融体系受到了多么严重的结构性损害，以及此类损害给这些国家造成的财政负担。就此而言，虽然美国银行体系的问题很严重，但相对而言，美国银行业资产负债表的规模却远小于许多其他发达国家。以银行业负债对GDP的比率为例，瑞士的数字是650%，英国是430%，欧元区是320%，日本是150%，而美国只有85%。投资者应记住这一事实。
为了衡量美国金融动荡及官方应对举措对于美元汇率指数的影响，我们用多边美元指数公允价值框架进行了模拟，预测了相对生产率、相对贸易条件、相对财政状况以及美国海外净资产等4项指标的变化对美元汇率指数的定量影响，重点是后2项指标的影响，因为我们认为，银行业与相对生产率和相对贸易条件两项指标间难以找到明确的关联。
根据近期动向判断，一方面，美国的相对财政状况将恶化2%；另一方面，虽然美国经常项目赤字可望进一步下降，但其绝对规模仍然很大，因此美国海外净资产对GDP的比率将下降3%。这两个因素将使美元指数的公允价值下跌5.3%。
总而言之，要评估美联储的量化宽松举措对美元的影响，需要考虑两个因素。首先，量化宽松举措本身通过相对通胀率的变化，只会影响美元的名义价值。量化宽松导致的通胀压力越大，美元贬值的压力也就越大。第二，迫使美联储实施量化宽松举措的根本原因、即美国金融体系受到的打击将使得美元的内在价值下跌5%－7%.因此，虽然我们仍然预计，在全球经济低迷的背景下，美元有所升值，但未来6个月之中，美元对其他主要货币升值的幅度将小于我们先前的预测。另一方面，未来6个月之中，美元将对新兴市场货币大幅升值。
<br><br>
英文原文：The Fed’s QE Operations and the Dollar
Summary and Conclusions
The Fed has commenced QE （quantitative easing）. In this note, we review the concept of QE and analyse the likely impact of this extraordinary operation on the dollar. The upshot is that, since monetary policy, including QE, is a ‘nominal’ operation, the operation itself should not have significant implications for the real value of the dollar. The nominal dollar value should, thus, only be affected if QE alters the outlook of inflation in the US over the medium term. Also, whether QE by the Fed should erode the value of the dollar should be assessed relative to what other central banks do. To the extent that the ECB and the BoE also conduct QE - which is the case - the impact of QE on the dollar is not necessarily negative.
Having said this, though QE per se should affect the dollar through relative inflation as well as inflation expectations, the underlying structural problems that forced the Fed to conduct QE in the first place should alter the fundamental value of the dollar, relative to those of other currencies. The parlous state of the US financial system should, in theory, be reflected in a lower value of the dollar, had it not been for its hegemonic reserve currency status propping the dollar up during this deleveraging phase. The bloated fiscal deficits （which we assume will exceed those of the G7 countries） will further weigh on the intrinsic value of the dollar.
In sum, whether QE by the Fed is negative for the dollar depends on the inflation outlook of the US and the resulting inflation expectations. But at a fundamental level, the dollar's intrinsic value has indeed deteriorated with its severely weakened financial sector. We maintain our core view that the dollar should continue to appreciate as the world slows - which we assume will last until next summer - but could give back some of the gains when deleveraging stops and the recovery phase for the US economy proves to be more protracted and treacherous than for other economies. The size and vigour of the dollar rally against the majors in the next six months or so are also likely to be more tempered than we have had in mind, in light of the deteriorating fundamentals in the US. Our call on EM currencies remains unchanged.
Background Discussion on QE
As inflation falls, and the unemployment rate rises, the Fed is likely to embark on QE to sustain monetary stimulus even when the FFR approaches zero. Broadly speaking, central banks can ease by either altering the price of money （i.e., interest rates） or the quantity of money. While policy orthodoxy these days is focused on the former lever, when short-term nominal interest rates approach zero as inflation falls, central banks could in principle use quantitative channels through which to impart monetary stimulus. Since it is the real interest rate that affects economic activities, the zero bound on nominal interest rates exposes an economy with deflation to persistently positive real interest rates. This was the situation faced by Japan in 2000, that interest rates were cut to zero but there was still not enough demand for money for monetary policy to work. That was an old-fashioned case of a 'liquidity trap'.
However, for the US, Europe and the UK, the current situation is somewhat different. Interest rates are still above zero. The problem monetary authorities face is not quite deflation and inadequate demand for money （i.e., a deficient aggregate demand problem） – though this problem could emerge if demand slows further. Rather, monetary policies lack traction mainly because there is no longer a smooth transmission mechanism from the short-term policy interest rates to broader monetary aggregates, credit and aggregate demand. As a result, Japan had QE with ZIRP, but in the US, Euroland and the UK, there is QE without ZIRP.
<br><br>
There are essentially three broad channels through which a central bank can conduct unconventional monetary easing. First, a central bank could ‘do things’ （i.e., through communications or QE） to foster the expectation that short-term interest rates will stay low for an extended period of time. Indeed, this was the primary aim of the BoJ during its QE episode between March 2001 and March 2006. The FOMC statement in August 2003 which stated for the first time that "policy accommodation can be maintained for a considerable period” is another example of this type of commitment to monetary easing. Second, a central bank could increase the size of its balance sheet to foster an expectation on the future path of inflation. （Currently, the Fed’s balance sheet is around US$2 trillion, up from US$900 billion in August.） The intuition of this ‘money printing’ method of QE is obvious. Third, a central bank could alter the composition of its balance sheet. Assuming that investors treat different assets as not perfect substitutes, central banks' purchase operations of selected assets could materially alter their prices. The best example is long-term US Treasuries. In theory, the Fed could purchase large amounts of Treasuries to cap the yields, just as the BoJ did through its rinban operations during the QE period. These three different methods of QE are conceptually distinct but operationally fungible. As the US economy slows further, we expect that the Fed will put all three methods of unconventional easing into practice.
The Fed began expanding its balance sheet in September. It may be useful to consider two motivations for QE by the Fed. The first is to take on market securities in an attempt to 'jump-start' the banking system; the second is to take on some of the intermediation duties that the private sector has refused to conduct. We looked at how the expansion in the Fed's balance sheet compares with the experience of the BoJ during the QE period that spanned from March 2001 to March 2006, during which time the BoJ’s balance sheet expanded from around 13% of GDP to about 22% （we measure the balance sheet by the monetary base for both countries）. Due to fundamental differences in the financial systems of the two countries, the Fed’s balance sheet has historically been substantially smaller than that of the BoJ - it averaged around 6% of GDP prior to this crisis. By October, however, the Fed's balance sheet had been expanded to more than 8% of GDP - a 35% increase.
To properly assess the impact of 'money printing' by the Fed on inflation, it is important to track the evolution of broader monetary aggregates, and observe how the 'money multiplier' - the ratio between broad money and the Fed's balance sheet - changes over time. We looked at the trajectories for M2 of Japan and the US. During Japan's QE period, its M2 expanded from around 125% of GDP to more than 140%. In the US, M2 has expanded from 53% to 57% of GDP since September. While the latter is a sharp surge, M2/GDP is not substantially higher than it was during 2003, when the Fed drove the FFR towards 1.00%. Indeed, most of the surge in the figure comes from the anticipated sharp drop in 4Q nominal GDP. To summarise, while base money has increased dramatically （by 34% since August）, M2 has grown by only 2.7%.
The reason for this, of course, is that the ‘money multiplier’ （MM） has collapsed, reflecting a severe breakdown in the ability and the willingness of the bank and non-bank entities in the US to intermediate capital to the extent they had done prior to the crisis. The collapse in the US MM is significantly more severe than in the case of Japan, during which time Japan’s MM fell from around 10 times the size of the BoJ’s balance sheet to around 6.5 times （a fall of 35%）. In the US in the last two months, we have seen the US MM falling from more than 9 to around 7 （down 22%）.
Will QE by the Fed Weaken the Dollar?
Our short answer is ‘no’. But whether this answer is right depends on a number of assumptions. （1） Deflationary pressures cannot be dismissed so easily in this cycle. Analysts can have their views on the inflation/deflation debate; however, to us, the fact that, with half a dozen ‘nuclear options’, the world’s policymakers have not yet succeeded at halting or containing the crisis worries us. It is, to us, very difficult to argue convincingly that ‘it’ won’t happen in the US or elsewhere in the world （see Vice Chairman Don Kohn’s speech last week）. While the probability of this risk of sustained deflation is not high, it may be too high for comfort. Though the Fed did act much sooner than the BoJ, the underlying problem is arguably more serious and the financial crisis has severely shocked the 'money multiplier', forcing the Fed to intermediate on behalf of the private sector. In short, we fear deflation now more than we fear inflation tomorrow, because we simply have no confidence that this will merely be a 'deflation scare'.
Our guess is that the Fed and other developed country central banks have a similar asymmetry in fear of the two tail risks: this risk-management approach to monetary easing indeed reflects the asymmetry. （EM central banks have lingering concerns about inflation.） In a way, investors realise this and are betting on the G7 central bankers to overwhelm and neutralise the left tail risk, leaving the risk of the G7 ‘overdoing’ （i.e., over-easing） it more prominent than the G7 failing to re-ignite their economies. In other words, the greater the fears of deflation, the more will be done by central banks to avoid that scenario and, as a result, the higher the risk of inflation in the out-years, or so it is thought. （Ironically, it was this stance the Fed had in 2001-03 that arguably sowed the seed for the ensuing credit bubble.） Having said this, this thought process is logical but very subjective, with no clear right answer, in our view. But tactically, we believe it is wise to bet that the deflation scare will be more powerful a market force as the global economy falters and as central banks fight the 'icing problem'. Whether there will be inflation is a debate for another day, probably three months from now, and it will probably not be reflected in market pricing, we suspect.
Whether the dollar depreciates depends on whether investors believe that there will be runaway inflation in the US over the medium term. Our view is that the Fed will most likely have time to retract on QE in time to keep a lid on inflation. Without inflation, the dollar cannot be weakened by nominal operations, ceteris paribus. Our guess is that, while dealing with a liquidity trap is difficult, removing stimulus when macro conditions normalise should not be a major problem for the Fed. But this discussion also highlights the importance of the Fed having a credible 'exit strategy' for its QE operations. This is essentially the view of Minneapolis Fed President Gary Stern.
Further, while many are focused on this question of QE by the Fed and the dollar, we argue that the ECB will likely be forced down the same path soon. The European and the UK banking systems are just as unwilling to intermediate capital as the American banking system. The ECB and the BoE, therefore, will likely be forced to do some of the intermediation on behalf of the private sector, just as the Fed has been forced to do. So if the Fed, the ECB and the BoE adopt QE, what is the net result on EUR/USD or cable? We think it is very unclear that EUR or GBP will necessarily rally, as is presumed by some investors when they think about the Fed.
<br><br>
Ironically, the tug of war between deflation and inflation fears indicates a reverse Dollar Smile mechanism of sorts. Recall that our Dollar Smile framework suggests that the dollar rallies when the US economy is stronger or weaker than the rest of the world. In the intermediate state, which we call the gutter, the dollar is weak. In the deflation versus inflation debate, on the other hand, both these extremes imply a weak dollar, while the intermediate state would foster dollar strength.
The USD's Fundamental Value Undermined?
While the Fed's QE operations affect the nominal value of the dollar through inflation, the severe deterioration in the US financial system must have had an impact on the intrinsic value of the dollar. We stress again that exchange rates are a relative concept, so we need to be sensitive to how much structural damage the financial systems of other developed countries may have sustained. This also includes the fiscal burden that these countries have to take on because of this damage. Notwithstanding all the problems in the US banking system, the size of banks' balance sheet is much smaller in the US than in many other countries. For example, total bank liabilities are around 650% of GDP for Switzerland, 430% for the UK, 320% for the Euroland, 150% for Japan and 85% for the US. Investors should keep this fact in mind.
In any case, as a rough guesstimate of how the USD major index may be affected by the breakages in the US financial system and the remedial policy actions, we simulated, using our multilateral USD index fair value framework, how the major USD index might be affected by prospective changes in relative productivity, relative terms of trade, relative fiscal positions and the US NFA （net foreign asset） position. Our simulations are centred on the last two variables, as we believe that it is difficult to draw a clear link between the banking sector and relative productivity and the terms of trade.
It looks likely, based on recent developments, that the relative US fiscal position may deteriorate by 2% of GDP, and its NFA position may also deteriorate by 3% of GDP （the US C/A deficit is expected to shrink but will still be large）. This scenario maps to a 5.3% depreciation in the fair value of the major USD index, according to our valuation model （the impact of NFA is small due to the small magnitude of the estimated coefficient in our model）.
Bottom Line
There are two considerations in thinking about the impact of the Fed's QE operations on the dollar. First, QE per se can only affect the nominal value of the USD through relative inflation. The greater the inflationary pressures generated by QE, the more the dollar could weaken. Second, the underlying reasons why the Fed has been forced to undertake QE operations may undermine the intrinsic value of the dollar by 5-7%. Thus, while we still expect the dollar to appreciate as the global economy falters, the size of the USD rally against the majors in the next six months will likely be more modest than we had thought. Against the EM currencies, we continue to expect significant USD strength in the next six months.
来源：摩根斯坦利，2008.11.28，作者：Stephen Jen ，Spyros Andreopoulos |
狗仔卡
发表于 2009-3-19 01:44 AM
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡