Impact of Stock Bubble Burst: An Update

(摩根斯坦利：股市大跌影响的新评估)

from http://www.cei.gov.cn/

 

　提要：近几周，市场再度担忧中国股市泡沫破裂的严重影响。股市暴跌对实体经济的影响有三方面，即对居民消费的负面“财富效应”，企业投资和银行资产负债状况的影响。本文预测，如果股市下跌30%，将使中国的居民消费和GDP分别减少1.2%和0.4%；如果股市下跌60%，将使消费和GDP分别减少2.4%和0.8%。企业投资方面，股市大跌不会对固定资产投资产生很大影响；而银行财务状况受到的影响将是可控的，但需要密切关注贷款增长和实际经济活动之间是否出现脱节。

　　（外脑精华·北京）我们在5月底曾发布《中国股市大跌，影响究竟有几何》一文，讨论股市泡沫破裂可能造成的影响（参见《压力测试：中国股市大跌，影响究竟有几何》一文，2007年6月5日《外脑精华》第40期）。目前，上证指数已经比“530”之后的低点上涨了60%，9月份的静态市盈率已达64倍的历史高点。这种情况下，股市泡沫破裂将会造成的严重后果又一次引起了人们的担忧。投资者尤其担心实体经济有可能受到三方面的严重影响，即负面“财富效应”对居民消费的影响、企业固定资产投资的融资渠道受到的影响以及银行财务状况受到的影响。本文分析两种情形，即较短时期内股市暴跌30%和60%的影响。由于数据不足，无法进行精确的模拟，所以我们只能运用官方数据，并采取较为严格的假设，以便粗略地评估可能出现的最坏情况。

　　“财富效应”日益增强

　　按我们的估算，2007年9月底，中国的居民资产总额为64万亿元（相当于GDP的264%左右），其中股票资产约为8.6万亿元（约占总资产的13%），房地产约为35万亿元（占总资产55%左右）。9月底，中国股市总市值约为26万亿元，但流通股只占三分之一左右，其余股份由少数大股东持有（主要是国家）。换言之，中国居民的股票投资规模远小于中国股市的总市值。

　　9月底，居民持有的8.6万亿股票比2006年底增加了240%。我估计，其中有60个百分点来自从银行转入股市的资金，180个百分点来自股价的攀升。按我的估算，居民金融财富的增长额相当于GDP的12%左右。从理论上说，这会带来可观的财富效应，从而刺激消费。

　　由于缺少足够的数据，因此我们无法对“财富效应”的准确规模做出可靠的定量估算，但一些零散的证据却显示了财富效应的增强。例如，过去，居民收入的增速总是大大落后于实际工资的增长；今年以来，收入增速却大大超过了实际收入的增长。而且，奢侈品和高端消费品的销售增长不仅势头强劲，而且有所加快。之所以如此，股市高涨产生的财富作用可能发挥了一定作用。

　　负面“财富效应”明显但可控

　　股市大幅调整可能对居民消费产生负面“财富效应”，这是市场担心调整的一个重要原因。如果股市下跌，直接持股或通过基金间接持股的居民都会感觉自己变穷了，这可能会导致消费的下降。

　　这里讨论两种情况，即股市在较短时间内下跌30%和60%。在第一种情况下，假设其他条件不变，按我的估算，居民持有的股票价值将缩水到6万亿元。根据其他资产价值不变的假设，居民金融财富总额将减少2.6万亿元，相当于GDP的11%左右。在第二种情况下，居民的股票价值将缩水到3.4万亿元。假设其他资产价值不变，那么居民的金融财富将减少5.1万亿元，相当于GDP的21%左右。

　　要判断上述两种情况可能对消费造成的影响，还需要一个重要指标，即“财富的边际消费倾向”（MPCW）。由于数据问题，中国的MPCW极难估算，我们以美国相应指标作为替代。根据研究，美国的MPCW约为4%。假设中国的数字与此相当，这就意味着，股市财富每减少100元，消费就会减少4元。应该指出，4%代表了中国的MPCW可能达到的上限，这一是因为美国居民的总体消费倾向要高于中国，二是因为美国的金融市场极为发达，因此资产升值转换为现金收入的难度要比中国小得多。

　　假设中国的MPCW为4%，按照我们的估算，在其他条件不变的情况下，如果股市下跌30%，将使中国的居民消费和GDP分别减少1.2%和0.4%，如果股市下跌60%，将使消费和GDP分别减少2.4%和0.8%。就股市下跌60%的情况而言，如果股市泡沫破裂对房地产市场产生传染效应，就可能导致房价大跌。而假设房价下降10%，就会使居民财富的损失增加3.5万亿元。在这种情况下，假设其他条件不变，总体上的直接财富效应将使居民消费和GDP分别减少4%和1.4%。

　　显然，与5个月之前相比，股市调整可能对于居民消费造成的负面影响已经明显增大了。但考虑到中国的经济增长率高达两位数，总体来说这种影响还是可控的。还应指出，以上分析的只是直接影响，现实中还会有第二轮的乘数效应，因此最终影响会更大。此外，上述分析假设不存在政策反应，而决策者很可能会采取多方面的对策，从而有效地缓解股市大跌对消费的影响。

　　企业投资受到的直接影响不明显

　　股市泡沫破裂会通过两种渠道对企业投资产生负面影响：第一，如果负面财富效应导致消费等最终需求陷入疲软，那么企业就会因业务前景恶化而削减投资；第二，泡沫的破裂将影响股市的筹资功能。第一种渠道关系到负面财富效应对消费的影响，上文已讨论了这个问题。

　　关于第二种渠道的影响，经济学中也有和“财富效应”对应的理论，即“托宾q理论”。根据该理论，如果股市对一项新增资本（即投资）的估值高于其实际重置成本，企业就会进行投资。托宾q值，是资本的市场价值与其实际重置成本的比率。如果q值大于1，投资就是值得的。举例来说，如果某公司的某项投资计划耗资1000万元，而项目完成后能使公司总市值增加1200万元，这项投资就是合理的，因为它能够额外创造200万元市值。因此，股市高涨会刺激企业投资，而股市大跌倾向于导致投资收缩。

　　然而，在当前的价位上，即使中国股市下跌30%、甚至是60%，也不大可能对固定资产投资产生严重影响。首先，无论股市走势如何，中国的托宾q值总是远大于1，原因主要在于中国的资本重置成本很低。投资项目的融资成本即银行贷款的利率，而中国的银行贷款利率大大低于市场出清水平。这种情况下，中国企业投资决策面临的约束条件通常不是融资成本，而是能否获得银行存款。由于利率很低，中国的投资需求总是非常旺盛，从而形成了过度投资的倾向。虽然股市高涨会令中国的托宾q值远远超过1，但股市在当前的价位上大跌30%乃至60%也决不会使q值降至1以下。换言之，托宾q值并非中国企业投资决策的约束因素。这也在一定程度上解释了，为什么中国的IPO审批会受到严格控制。

　　第二，在中国固定资产投资的融资渠道之中，股市直接筹资所占的比重很小。2006年，在固定资产投资的融资总额之中，A股筹资额所占的比重只是略高于2%。今年1－8月，这个比率大为提高，但仍然很低。这意味着，即使股市筹资因市场调整而彻底中断，只要其他融资渠道仍然开放，企业融资条件就不会出现严重的恶化。

　　事实上，如果股市泡沫破裂，其影响可能恰恰相反，可以用于实物投资的资金或许会因此而增加。近年来，由于股市投资的预期回报率一直远高于传统投资活动的回报率，因而中国企业可能把一部分原本会用于固定资产投资的资金投入了股市。这种情况下，如果股市泡沫破裂，就可能促使企业重新把资金投向生产性的实际投资活动。

　　关注贷款增长与实际经济活动的关系

　　股市大跌对银行资产负债状况的影响将是可控的。我尤其关注的是，银行存款增长与实际经济活动之间是否出现了脱节，因为这种情况意味着，银行发放的部分资金正在以某种方式进入股市，即便没有明确证据能够说明银行贷款直接进入了股市。这种情况之所以值得关注，是因为资金具有“替代性”：如果企业把自有资金投入股市，同时增加银行贷款，用于“合法”用途，银行承受的间接风险也会增大。迄今为止，银行信贷增速和实际经济活动之间还没有脱节。但国际经验表明，股市泡沫持续的时间越长，银行承受的风险就可能越大。

　　本文的分析非常粗略，而且由于数据不足，我们的估算存在很大不确定性。但文中的分析可以说明，如果股市继续迅速上涨，那么股市泡沫破裂的潜在影响就会明显增加。

　　英文原文：Impact of Stock Bubble Burst: An Update

Renewed fears of stock bubble bursting

We published a note on May 28 asking the question, “What is the risk to China’s economy if the A-share market bubble bursts?” The Shanghai A-share composite index has surged by 60% from the lows reached in the aftermath of the stamp duty hike on May 30 and now hovers around the 6,000 level. The average trailing P/E reached a historical high of 64 times in September. Against this backdrop, fears of serious consequences from a potential bursting of the stock market bubble appear to have re-intensified in recent weeks. In this context and with acceleration of property prices, similar concerns have been expressed by clients about the impact of a major correction in property prices. Investors are especially concerned about the potentially serious real economic impact of a negative ‘wealth effect’ on household consumption, the financing channel for corporate investments and banks’ balance sheets. I examine two scenarios under which the stock market plunges by 30% and 60%, respectively, in a relatively short period. Based on investor feedback, there are orders of correction that could trigger different degrees of concern about the implications for the wider Chinese and even global economies. Since the paucity of data prevents us from conducting a thorough or precise simulation of such a scenario, we rely on officially published data and make several rather strong assumptions with a view to gauging how bad things could get.

Emerging ‘wealth effect’ on personal consumption

We estimate that Chinese households had about Rmb64 trillion in assets （i.e., about 264% of GDP） by end-September 2007, of which about 13% （i.e., Rmb8.6 trillion, or 35% of GDP） are in the form of stock holdings and 55% （i.e., Rmb35 trillion, or 144% of GDP） are in the form of property. China’s headline stock market capitalization stood at about Rmb26 trillion at end-September, but only about one-third of this is actually ‘marketable’, i.e., available for investment by retail investors. The rest remains held by a small number of large shareholders （mostly the state）. In other words, household exposure to the stock market is much smaller than that suggested by headline stock market capitalization.

The value of households’ stock holdings as of end-September reached Rmb8.6 trillion, an increase of 240% from end-2006, of which, I estimate, 60 percentage points were contributed by the rebalancing of households’ financial asset portfolios from bank deposits to stock holdings and 180 percentage points were contributed by market appreciation. Such a massive increase in household financial wealth, which amounts to about 12% of GDP by my estimate, should in theory generate a substantial wealth effect that helps boost consumption.

Although we do not have sufficient statistics to make a reliable quantitative estimate of the exact magnitude of the wealth effect, we do find some anecdotal evidence suggesting an emerging wealth effect. Specifically, in the past, household income growth has consistently and significantly lagged real wage growth; however, income growth started to outpace real wage growth by a large margin this year. Real consumption growth has followed a trend similar to that of real income growth. Moreover, sales of luxury and high-end consumer discretionary goods have been strong and have accelerated this year, suggesting that the wealth effect underpinned by the extraordinary stock market performance may have played a role.

Potentially significant impact on consumption

A typical concern about the fallout from a major stock or property market correction is the impact of the negative ‘wealth effect’ on household consumption. Households that hold stocks directly or indirectly through mutual funds feel poorer when the stock market falls. This can result in a decline in current consumption.

We look at two scenarios under which the stock market plunges by 30% and 60%, respectively, in a relatively short period. I estimate, under the first scenario, that were the stock market to plunge by 30% from current levels, the value of households’ stocks would, ceteris paribus, shrink to Rmb6 trillion. With the value of other assets unchanged, total household financial wealth would decline by Rmb2.6 trillion, or about 11% of GDP. Under the second scenario, were the stock market to plunge by 60% from current levels to the level reached in the aftermath of the stamp duty hike in late May, the value of households’ stocks would, ceteris paribus, shrink to Rmb3.4 trillion. With the value of other assets unchanged, total household financial wealth would decline by Rmb5.1 trillion, or about 21% of GDP.

To determine how such a decline in household financial wealth might affect consumption, we need one important parameter – the ‘marginal propensity to consume out of wealth’ （MPCW）. As we did in the previous study, we use estimated MPCW for the US – which is around 4% based on academic literature – as a proxy, given the extreme difficulty of estimating this figure for China, due to poor data quality and availability. In other words, we assume that every Rmb100 decline in stock wealth will lead to a Rmb4 reduction in consumption. I would caution that we view the MPCW of 0.4% as a high-end figure for China, given US households’ greater propensity to consume in general and the highly developed financial markets in the US – which make it much easier to translate financial gains into cash income.

Based on an MPCW of 4%, we estimate that the direct negative wealth effect could, ceteris paribus, cause personal consumption and GDP to decline by 1.2% and 0.4%, respectively, if the market were to drop by 30% from the current level, and by 2.4% and 0.8%, respectively, if the market were to drop by 60%. Furthermore, under the second scenario （i.e., a drop of 60%）, if the bursting of the stock market bubble were to have a contagion effect on the property market, causing average property prices to fall by, say, 10%, it would add Rmb3.5 trillion to the loss of household wealth. In that event, we estimate that the overall direct negative wealth effect could, ceteris paribus, cause personal consumption and GDP to decline by 4% and 1.4%, respectively.

Clearly, the negative impact of a potential market correction at the current juncture has become substantially larger than five months ago, although it still appears broadly manageable by China’s double-digit growth standards, in my view. Moreover, I should point out that we estimate only the direct impact, and in practice there would be second-round multiplier effects that would amplify the ultimate impact. Nor do we assume any policy reaction in our estimate, while, in practice, policy makers may well react with various measures that would effectively mitigate the impact.

Moderate direct impact on investment

Investment would be negatively affected by a bursting of the stock market bubble through two channels: 1） If the underlying final demand （e.g., consumption） weakens due to a negative wealth effect, investment would be cut back due to a less favorable outlook; and 2） a bursting of the stock market bubble would halt the market’s function of capital raising. The first has to do with the impact on final consumption from a negative wealth effect, which we addressed in the previous section. Regarding the second, I do not believe that a correction of the stock market by 30% or even 60% from current levels would change the desire of Chinese companies to invest now for the purpose of a public listing in the future.

Since my view on this particular point has remained unchanged, the paragraphs below are based on the relevant section in our previous note with updated data （see China Economics: What Is the Risk to China’s Economy if the A-share Market Bubble Bursts, May 28）.

The theoretical counterpart to the ‘wealth effect’ on consumption is the ‘Tobin’s q theory’. According to this theory, a firm will invest if the stock market’s valuation of new capital addition （or investment） to the firm is higher than the actual replacement cost of the capital. Tobin’s q is the ratio of market valuation of the capital to its actual replacement cost. When q is greater than one, investment is worthwhile and encouraged. For example, if a firm has an investment project that costs Rmb10 million and the completion of the project will increase the overall market value of the firm by Rmb12 million, it makes sense for the firm to carry out this investment, as it creates an additional Rmb2 million of market value. As such, a booming stock market encourages investment; conversely, a plunge in stock market prices tends to lead to a contraction in investment.

However, we believe that a 30% or even 60% correction in the stock market from current levels would be unlikely to have much of an impact on fixed-asset investment in China. First, we argue that the value of Tobin’s q in China is always substantially greater than one regardless of the performance of the stock market, mainly reflecting the low replacement cost of capital goods in China. The financing costs of investment projects （e.g., the interest rate on the bank loans） are kept substantially below the market-clearing level. The binding constraint on investment decisions is usually not the cost of capital but the availability of bank loans. Reflecting low interest rates, there is always strong demand for investment, generating the tendency for over-investment. While a booming stock market should drive Tobin’s q far above one, a 30% or even 60% decline in the stock market from current levels would by no means take it below one, in our view. Put another way, we do not believe that Tobin’s q is a binding consideration when investment decisions are made in China. This explains in part why approvals of IPOs in China are still tightly controlled and even rationed.

Second, as a source for investment financing, the amount of capital raised directly from the stock market remains very small. In 2006, only slightly over 2% of total financing for fixed-asset investment came from the A-share market. Although this ratio has increased substantially in the first eight months of this year, it is still very small. This suggests that even if stock market financing is shut off completely in the wake of a correction, corporate investment financing conditions would not deteriorate significantly, as long as other channels of financing remain open.

In fact, at the current juncture, there could be a rather perverse outcome: A bursting of the stock market bubble may make more funds available for investment in physical assets. With expected returns from investing in the stock market substantially and persistently higher than those from traditional investment activity, firms may have channeled the funds that would otherwise be used for fixed-asset investment into the stock market. A bursting of the stock market bubble should help firms to refocus on and make funds available for real and productive investment activity.

Other impact

My view on the other impact of a bursting of the stock market bubble （e.g., banking sector, social and global） remains broadly unchanged （see China Economics: What Is the Risk to China’s Economy if the A-share Market Bubble Bursts, May 28, and China Economics: An ‘Untimely’ Question: What Could Go Wrong with the Economy? July 23）.

In particular, I continue to believe that the impact on bank balance sheets will likely remain manageable for the reasons I discussed in two previous notes. I would become more seriously worried if bank loan growth started to get out of line with the underlying real activity, as this would indicate that some funds from the banks are somehow being channeled to the stock market, even if there is no clear evidence that funds borrowed from banks are invested directly in the stock market. This is because money is fungible: Banks could still be indirectly exposed to the stock market if the corporate sector invests its own funds in the market and simultaneously increases borrowing from the banks for other ‘legitimate’ purposes. In this regard, I have not detected a disconnect between the growth rate of bank credit and that of real underlying activity. At the same time, I shall remain vigilant on this front, as cross-country experience suggests that the longer the bubbly stock market lasts, the more exposed the banking sector will likely become.

Risks

This type of exercise is rather crude, and our estimate is subject to considerable uncertainties due to paucity of data. Nonetheless, it illustrates that the potential impact of the bursting of the stock market bubble will increase substantially if the market continues rising rapidly. We shall keep a close eye on stock market developments and their impact on the real economy and convey to clients our regular updates on the potential impact of the stock bubble bursting.

来源：摩根斯坦利，2007.10.24，作者：Qing Wang

作者：Qing Wang