The Fed Model: Fix It Before You Use It
Thanks to two months of slumping share prices, stocks are now a much better value than bonds. But long-run stock returns could still be thoroughly mediocre.
This good news-bad news story is based on the Fed model, one of the most widely used and abused stock-market valuation models.
Fancy yourself as an amateur investment strategist? Here's what the Fed model is all about -- and how you can try it at home.
Model Behavior
In the 1980s and 1990s, market commentators argued that falling interest rates justified higher and higher share prices. This notion became formalized in the Fed model, so called because it is purportedly favored by Federal Reserve Chairman Alan Greenspan.
For investors, there's a lot to like about the Fed model, which involves comparing the yield on the benchmark 10-year Treasury note with the earnings of the Standard & Poor's 500-stock index. Not only did the model work well in the 1980s and 1990s, but also it's so simple that almost anybody can use it.
To do so, you first have to find out the S&P 500's forecasted earnings. That figure is available at
www.spglobal.com, the Web site for Standard & Poor's, a unit of McGraw-Hill. Click on the link for "S&P 500 Index Earnings" and then call up the spreadsheet labeled "S&P 500 Earnings and Estimate Report."
According to the site, 2005's reported earnings should come in at around 67.40. This earnings estimate is calculated so that it's comparable with the index value for the S&P 500, which lately has been trading around 1150.
Usually, investors would divide that 1150 by the forecasted earnings of 67.40. That would give you the market's price-to-forecasted earnings multiple, which is currently 17. But for the Fed model, you reverse the calculation, dividing the 67.40 earnings by the 1150 index value, thus getting an "earnings yield" of 5.9%.
You then compare this 5.9% to the 10-year Treasury's 4.2% yield. When the earnings yield is above the Treasury yield, as it is today, that suggests stocks are cheaper than bonds.
To be sure, bond investors get their interest in cash, while shareholders receive only a sliver of earnings as dividends. The rest of a company's profits are plowed back into the business, with a view to clocking additional growth. Still, presumably management could stop pursuing growth and instead pay out pretty much all of a company's earnings as dividends.
Failing Grade
Seem reasonable? Unfortunately, there are a bunch of problems with the Fed model.
For starters, the model is often biased toward stocks. Instead of reported earnings, fans of the Fed model often use operating earnings, which are higher because they ignore one-time accounting charges. To make matters worse, analysts are typically too optimistic in their earnings forecasts, further boosting the stock market's apparent earnings yield.
Until 2001, the S&P 500's earnings yield was often compared with 30-year Treasury bonds. But when the government announced in late 2001 that it would stop selling 30-year Treasurys, Fed model users gravitated to the 10-year note instead. Because the 10-year note had a lower yield, that immediately made stocks more appealing -- or so said the Fed model.
But the model's biggest problem is that bond yields and earnings yields really aren't comparable. After all, Treasury-bond interest is fixed for the life of the bond and you can count on receiving it every year.
Meanwhile, corporate earnings are iffier, but they should rise over time. Indeed, even if a company paid out all of its earnings as dividends, the company's profits would still tend to climb along with inflation.
Measuring Up
There is a way to fix these problems, contends Jeremy Siegel , a finance professor at the University of Pennsylvania's Wharton School and author of a new book, "The Future for Investors."
Instead of using conventional 10-year Treasurys in the Fed model, he suggests substituting 10-year inflation-indexed Treasury notes. With these inflation bonds, you earn the inflation rate plus a small additional yield, currently 1.6%. Because inflation bonds and the interest they pay grow along with inflation, this 1.6% is truly comparable to the S&P 500's 5.9%.
Today, as you can see, the earnings yield is much higher than the inflation-bond yield. This is no great surprise. Stocks are riskier, so the expected return ought to be higher. The key question: How much extra should stocks yield?
Over the past eight decades, the S&P 500 has outpaced government bonds by five percentage points a year, according to Chicago's Ibbotson Associates. But Prof. Siegel reckons the margin of victory will be somewhat smaller in the years ahead.
In fact, he figures stocks would be fairly valued if the earnings yield were two to three percentage points above the yield on inflation-indexed Treasurys. The implication: With stocks today yielding four percentage points more than inflation bonds, either stocks are cheap -- or, as Prof. Siegel suspects, bonds are expensive.
Cliff Asness, managing principal of hedge-fund manager AQR Capital Management in Greenwich, Conn., has been a vocal critic of the Fed model. He says using inflation bonds, rather than conventional Treasury notes, is a big improvement. "Jeremy's changed it from mathematical garbage to a legitimate model," Mr. Asness says.
He also says that it would be reasonable for stocks to be priced to deliver two to three percentage points a year more than bonds. "There's a good argument that says people have done too well in stocks historically and they ought to accept less," he says. "The question is, will they?"
As Mr. Asness sees it, investors use the Fed model to justify their bullishness on stocks. But he isn't sure they are willing to accept the implied lower returns.
Suppose annual inflation comes in at 2.5%, inflation bonds give you 1.6 percentage points more than that and then stocks outpace inflation bonds by four percentage points a year. Add it up, and you are looking at an annual stock-market return of just over 8%, and possibly less if today's earnings forecasts are too optimistic.
What if investors decide that sort of return is inadequate? The Fed model may say stocks are a bargain. But that doesn't mean shares won't get a lot cheaper
慎用Fed模型
多亏了过去两个月的低迷行情，现在股市的估值要比债市好得多。但股票的长期回报率恐怕仍会让人提不起精神来。
这样好坏参半的结论是基于美国联邦储备委员会(Fed)的模型作出的，它是使用最广的股市估值模型之一，也是最常被滥用的股市估值模型之一。
有没有把自己想像成一位业余投资策略师？以下便是Fed模型的具体内容－－以及你如何在家运用它。
模型行为
在20世纪80年代和90年代，市场评论家争辩称，利率持续下跌意味著股价不断走高。这种观念在Fed模型中被定形，之所以称之为Fed模型，是因为据称它得到Fed主席格林斯潘(Alan Greenspan)的青睐。
对于投资者而言，Fed模型有许多值得一提的地方，例如，它把基准10年期国债的收益率与标准普尔500指数成份股的收回报率进行了比较。这个模型不仅在80年代和90年代运作得很好，而且它十分简单，几乎人人都可以运用自如。
首先，你得找出标准普尔500指数成份股的预期收益。有关数字可以在标准普尔公司(Standard & Poor's)的网站
www.spglobal.com上找到。标准普尔公司是McGraw-Hill的子公司。点击链接“标准普尔500指数成份股收益”('S&P 500 Index Earnings')，然后就可以找到名为“标准普尔500指数成份股收益和预测报告”('S&P 500 Earnings and Estimate Report')的表格。
根据这个网站的数字，2005年的收益应该在67.40左右。该预测已经经过调整，以便可以与标准普尔500指数的价值相比较，该指数最近一直在1150点附近徘徊。
通常，投资者会用1150除以预期收益67.40，得出市场本益比，目前市场本益比为17。但根据Fed模型，你要进行反向计算，即用67.40除以1150，这样便得出“公司收益率”为5.9%。
然后你把这个“公司收益率”5.9%与10年期国债的收益率4.2%相比较。当“公司收益率”高于国债收益率，就像现在一样，这表明股市比债市便宜。
当然，债券投资者对现金感兴趣，而股东仅得到少量股息。公司剩余的利润又被投入业务，以推动进一步的增长。尽管如此，推测起来，管理层也可能停止寻求增长，而将公司收益的相当一大部分作为股息派发给股东。
失效
看起来合情合理？不幸的是，Fed模型存在著大量的问题。
对于新手而言，这个模型经常偏向于股票。Fed模型的追随者通常使用营运收益、而不是净收益，营运收益因为没有包括一次性会计支出项目要高于净收益。更糟糕的是，分析师通常对他们的收益预测都过于乐观，从而进一步提高了股市表面上的收益率。
在2001年之前，标准普尔500指数成份股的收益率通常与30年期国债收益率相比较。但2001年年末，政府宣布将停止发售30年期国债，Fed模型的使用者便转而将10年期国债作为参照。由于10年期国债收益率较低，这立即使得股票更具有吸引力起来－－或者说根据Fed模型来看是这样。
但这个模型最大的问题在于债券收益率和公司收益率实际上是不可比的。毕竟，国债利息在其期限内是固定的，你每年都能指望得到利息。
与此同时，公司收益则更加不确定，但应能随著时间的推移不断增加。确实，即便公司将全部收益都用来派息，其利润仍然趋于与通货膨胀一道攀升。
改进
宾夕法尼亚州大学沃顿商学院(University of Pennsylvania's Wharton School)金融学教授Jeremy Siegel声称，有一个方法可以解决这些问题。Siegel是新书《投资者的未来》('The Future for Investors')的作者。
他建议用10年期通货膨胀保值国债替代传统的10年期国债。运用通货膨胀保值国债作比较的话，你得到的收益率就是通货膨胀率外加一个小幅收益率，目前为1.6%。由于通货膨胀保值国债及其支付的利息随著通货膨胀率的攀升而增加，因此这里的1.6%可以在真正意义上与标准普尔500指数的收益率5.9%相比较。
你可以看到，今天的公司收益率要远远高于通货膨胀保值国债的收益率。这并不足为奇。股票的风险更大，因此预期回报应该更高。关键的问题是：股票的收益率应该高出多少？
根据芝加哥的Ibbotson Associates提供的资料，过去80年中，标准普尔500指数的收益率较国债的收益率每年要高出5个百分点。但Siegel教授认为未来几年两者的收益率之差将有所收窄。
实际上，他认为，如果公司收益率较通货膨胀保值国债收益率高出两至三个百分点，股票的价值将趋于合理。其中的含义就是：由于今天的股票收益率较通货膨胀保值国债收益率要高出4个百分点，因此，要么就是股票价格较低－－要么就是像Siegel教授怀疑的那样，债券价格过高。
AQR Capital Management的对冲基金主管Cliff Asness一直在对Fed模型提出直率的批评。他说，使用通货膨胀保值国债替代传统的国债是一大改进。Asness称，Jeremy将Fed模型从一个数学垃圾变成了一个合理的模型。
他还表示，股票收益率较国债收益率每年高出两到三个百分点是合理的。他说，有一种确凿的论点说，历史上投资者在股票方面做得太出色了，他们应该接受较低的收益率。“问题是，他们会吗？”
正如Asness预见的那样，投资者运用Fed模型来证明他们看涨股市是合理的。但他不确定投资者是否原因接受较低的收益率。
假设年通货膨胀率在2.5%，通货膨胀保值国债的收益率较之高出1.6个百分点，而股票收益率较通货膨胀保值国债的收益率还要高出4个百分点。加起来后，得出股市的年回报率为8%以上，如果今天的收益预期过于乐观，股市的回报率可能会稍微低一些。
如果投资者认为这样的回报率不够怎么办？Fed模型可能会说股票值得逢低买盘。但那并不意味著股价不会进一步大幅走低。
