The connection between closed-end fund discounts and small stocks extends to IPOs, a subject I discuss in chapter 18. Lee, Shleifer, and Thaler point out that if the number of new IPOs goes up when individual investors become more optimistic, then a narrowing of discounts should coincide with an increase in IPOs. They do find that this is the case.
Certainly, the popular financial press is aware of the connection between the returns to small stocks and the frequency of IPOs. Below is an illustration from the Wall Street Journal:
Underwriters say the pace of IPO registrations appears to favor an even stronger fourth quarter, provided the demand for small caps continues.
So far this month, 41 IPOs have gone public, generating first-day gains averaging 17.6%, the strongest one-day average since the 25.2% gains of May 1996, says Securities Data Co. More than 77% of the deals in the past three months have been priced within or above their originally targeted price ranges, compared with 64% at the beginning of the year and 88% during last spring’s IPO frenzy.8
The article by Lee, Shleifer, and Thaler (1991) generated a heated academic debate. The June 1993 issue of the Journal of Finance contained a critical comment by Nai-Fu Chen, Raymond Kan, and Merton Miller. They suggested that the relationship between fund discounts and small-firm returns is neither robust over time nor impacted by the degree of institutional ownership. The same issue contained a reply by the original authors, who were joined by Navin Chopra. In their reply, the authors demonstrated that for 90 percent of small-firm stocks, lower institutional ownership stocks do better than higher institutional ownership stocks when discounts narrow.
Bhaskaran Swaminathan (1996) has extended the Lee, Shleifer, and Thaler analysis and investigated the issues they raise in greater depth. He replicates their findings for a broader data set, obtaining additional findings that are consistent with closed-end fund discounts being an indicator of sentiment. For instance, he finds that the discounts serve as better predictors of future returns to the closed-end funds than returns to the underlying assets (NAV). He notes that the discounts have a much stronger relationship with returns to small stocks than returns to large stocks. And he finds that the information embedded within discounts is not subsumed by other variables such as the default premium on bonds, the term premium on Treasury, book-to-market ratios, and dividend yields.
But Swaminathan also finds that discounts are informative about the future value of corporate earnings and macroeconomic fundamentals. This leads him to question whether the discount on closed-end funds is a pure sentiment index. He suggests that discounts may be driven both by sentiment and by fundamentals.
Country Funds
Country funds offer some vivid examples of how sentiment is perceived in closed-end fund discounts. Consider the following excerpt from a Barron’s article entitled “Closed-End Mania Is Ba-a-a-ack.”
One convenient way to buy a package of foreign stocks is via closed-end funds that are denominated in dollars and listed on the New York Stock Exchange. Unlike mutual funds, these single-country and regional funds boast a fixed number of shares that trade at a premium or discount to their underlying net asset value.
Last week, discounts shrank and premiums rose. In the latest mania for country funds, 30 of them hit 1993 highs and many posted huge percentage gains. According to Michael Porter, who covers country funds for Smith Barney Shearson, their average premiums to NAV now are flashing a warning signal.9
A dramatic example involves the Germany Fund. Figure 13-3 below illustrates how the price of this fund behaved in the period July 1986 through April 1998. Take a close look at this figure. Can you tell from the chart when the Berlin Wall fell?
Most people guess that the spike in figure 13-3 marks the day, November 9, 1989, when the Wall came down. That would not be a bad guess. But actually, the peak came several months later.
In retrospect, what proportion of that spike appears to have been sentiment as opposed to fundamentals? You decide. In jumping from a price of $13.75 to $24.50, the premium went from 17 percent to 100 percent!
A taste of the exuberance associated with this price movement can be gleaned from a January 1980 Wall Street Journal article titled “Capital Floods into Germany at Frantic Pace.” Consider the following excerpt, showing that not only individual investors got caught up with the excitement of the moment.
Figure 13-3 Share Price of Germany Fund
Price path for the Germany Fund, a closed-end country fund. Looking at this figure, can you pick out about when the Berlin Wall came down? Is there a lesson from the rapid rise and subsequent decline in the chart?
Germany Fund, a publicly traded U.S. fund that invests in German stocks, has soared to where it trades at a sky-high 37% premium to the value of its holdings.
Michael Tintelnot, a managing director at Deutsche Bank Capital Management USA Corp., the New York–based institutional money-management arm of Deutsche Bank AG, is scheduled today to make a presentation on bond portfolios to a U.S. public pension fund. But on Friday, Mr. Tintelnot says, the chairman of the fund’s investment committee called to ask if he would also talk about Germany Fund. Mr. Tintelnot told him Germany Fund was primarily designed for individual investors. “He replied: ‘That’s fine, but can you do something like that for us?’” Mr. Tintelnot recalls.10
Lee, Shleifer, and Thaler (1991) emphasize that sentiment is like background noise, affecting whole classes of securities—closed-end funds, small cap stocks, and IPOs. The end of the cold war, with the fall of the Berlin Wall as its symbolic event, had implications far beyond the price of the Germany Fund. If we looked at the price histories for the closed-end country funds in Italy, Spain, Thailand, and the Republic of China, we would observe similar patterns. That is, they all exhibited spiked patterns at the same time.
The spike is the most salient feature in figure 13-3. Indeed, the fall of the Berlin Wall was a highly salient event. Are premiums and discounts affected by the saliency of events? Are the investors who trade closed-end country funds subject to availability bias?
In a fascinating study, Peter Klibanoff, Owen Lamont, and Thierry Wizman (1998) suggest that discounts on closed-end country funds, which are traded on U.S. exchanges, are in fact affected by the saliency of news events in the U.S. press. What these authors did was measure saliency by looking at the column width of stories that appeared on the front page of the New York Times.
Here is an example from their study, involving the First Israel Fund. On December 18, 1992, a three-column Times story appeared with the following headline: “Israel Expels 400 from Occupied Lands; Lebanese Deploy to Bar Entry of Palestinians.” In the week that followed the appearance of this article, the discount on the fund widened from 14.87 percent to 15.55 percent.
Three columns are less salient than six columns. The September 14, 1993, story announcing the peace accord between the government of Israel and the Palestine Liberation Organization (PLO) was six columns wide. Its headline was “Mideast Accord: The Overview; Rabin and Arafat Seal Their Accord as Clinton Applauds ‘Brave Gamble.’”
During the week this story appeared, the premium on the New Israel Fund jumped from 1.7 percent to 7.7 percent!
Klibanoff, Lamont, and Wizman argue that in the absence of salient news, country fund prices are sticky, meaning that investors underreact to news that is not salient—news that does not appear on the front page of the New York Times but that is germane to underlying NAV. For example, consider the period just prior to the appearance of the December 19, 1992 story on Israel’s expulsion of the 400 Palestinians. There were no front page stories about Israel between December 11 and December 18. During this seven-day period, the NAV of the New Israel Fund declined from $14 to $13.95, yet the price of the fund was unchanged at $11.88.
Figure 13-4 displays the price path of the New Israel Fund. The salient news events are not difficult to identify. Moreover, like the Germany Fund, the contribution of sentiment relative to fundamentals is clear, at least in hindsi
ght.
Events that occur in foreign countries usually receive more press coverage in the country in which they occur than in the United States. Hence the salience of those events is typically different in the foreign country. For the most part, the NAV of a fund is determined on the
Figure 13-4 Price of New Israel Fund
Price path for the New Israel Fund, a closed-end country fund. Looking at this figure, can you pick out about when Itzhak Rabin and Yasser Arafat shook hands on the grounds of the White House? Is there a lesson from the rapid rise and subsequent decline in the chart? home country market, and these market features are less vulnerable to the sentiment of American investors than they are to home country investors. But closed country funds traded in the United States are more vulnerable to the sentiment of American investors than they are to home country investors. In fact, work by James Bodurtha, Dong-Soon Kim, and Charles Lee (1993) demonstrates that the premiums on U.S. closed-end funds tend to move in tandem with the return on U.S. markets; however, changes in the net asset values of these funds do not.
Jeffrey Pontiff (1997) compares the volatility of closed-end fund returns in the Lee, Shleifer, and Thaler (1991) study with the volatility of the returns to the fund’s underlying assets. Given figures 13-3 and 13-4, which do you think is more volatile—the return to holding the fund or the return to holding its underlying assets? Pontiff finds that the return to holding the fund is 64 percent more volatile than the return to the underlying assets.
More Puzzles
There are yet more puzzles associated with closed-end funds.
Dividends: In Bary’s article for Barron’s, discussed earlier, the following intriguing statement appears: “The closed-end market drives efficient-market theorists crazy because similar funds can trade at appreciably different prices.”11 Table 13-2 illustrates the point using the three Nuveen Premium Income Funds discussed earlier.
Notice that the yields, taxable equivalent yields, and fifty-two-week market returns of the three funds are quite similar to one another, yet the discounts are markedly different. Moreover, although the annual total return on NAV also differs across the three funds, this does not explain the discounts. NPI has both the lowest annual return on NAV and the lowest discount. Only the difference in dividend seems to explain the discount pattern.
Because people think about money in segregated terms rather than in integrated terms, dividends are important. Remember the comments made by closed-end fund manager Martin Zweig, whom I quoted in chapter 3 As was discussed in that chapter and in chapter 11, dividends get evaluated in their own mental accounts. Investors find high, stable dividend payouts attractive; all else being equal, the fund with the superior dividend profile will feature a higher price.
In his Barron’s article, Bary raises two other interesting points, both pertaining to frame dependence. The two points involve leverage and rights offerings.
Table 13-2 Comparing the Characteristics of Three Nuveen Municipal Bond Closed-End Funds
Leverage: First, municipal bond fund companies often use leverage in their closed-end funds. They achieve leverage by issuing both common stock and preferred stock, split, say, two shares of common for each share of preferred. In usual circumstances, the yield curve will be positively sloped, and the rate on municipal bonds will be higher than the short-term interest rate. The leveraged fund will set the preferred dividend equal to the short-term interest rate. Since the holders of preferred stock receive a lower rate than the holders of common stock, this structure enables the fund to augment the dividend it pays to common stockholders. And as I discussed elsewhere, some investors care a great deal about dividends. But leverage also leads to volatility. A 100-basis-point decline in the long-term rate leads the dividend to fall by 150 basis points.12
Bary asks: “Given the risks, why do funds bother to use leverage? Because without leverage, they’d have a hard time staying competitive with the larger, open-ended muni-fund market.”13 Do fund shareholders understand the leveraged character of their holdings? Well given the way the leverage is structured, do you think that investors are presented with a transparent frame or an opaque frame?
Rights Offerings: Second, closed-end funds tend to use rights offerings in order to add assets. Now, rights offerings are framed to appear attractive, offering each fund shareholder an opportunity to pay less than the current price to acquire more shares. But other than the fund’s shareholders, who makes up the difference? Nobody. Therefore, the share price of the fund should fall immediately after the rights offering. Although it seems that no shareholder is being coerced into purchasing more shares, the fact is that shareholders who do not purchase additional shares find that the value of their fund holdings will decline as a result of the purchases by other shareholders.
Summary
The fact that closed-end funds trade at prices far from net asset value has long been a puzzle for proponents of market efficiency. The puzzle has four parts. The underlying determinants of the puzzle are heuristic-driven bias and frame dependence. In particular, heuristic-driven bias manifests itself in the form of investor sentiment, and volatile investor sentiment is the main driver behind discounts. This is particularly evident in the case of country funds, where salience is an issue. However, there are other puzzles. Fund shareholders have been apathetic about voting to eliminate discounts through open ending. There is also wide variation in discounts across similarly structured funds.
Chapter 14 Fixed Income Securities: The Full Measure of Behavioral Phenomena
When it comes to interest rates, investors are especially slow learners.
In 1938, Frederick Macaulay published a book that described how interest rate movements were both puzzling and notoriously difficult to predict. Little has changed since 1938, at least in this respect. But some have been slow to learn what Macaulay pointed out more than fifty years ago.
This chapter discusses the following:
• how overconfidence, gambler’s fallacy, and betting on trends set the stage for disaster in the case of the Orange County Investment Pool
• how conservatism, hindsight bias, loss aversion, and regret came into play during and after the crisis that led Orange County to declare bankruptcy
• the theoretical issues that underlie the expectations hypothesis of the term structure of interest rates
• the evidence suggesting that the expectations hypothesis fails
• why underreaction to changes in inflation, stemming from anchoring-and-adjustment, interfere with the expectations hypothesis
This chapter covers some of the successes and failures associated with the management of fixed income securities, first by presenting a case and then by discussing some general issues associated with yield curves. The case, which focuses on the first behavioral themes of heuristic-driven bias and frame dependence, details the experiences of the Orange County Investment Pool. It is exceedingly rich in behavioral phenomena.
The discussion of general issues focuses on the third theme, market inefficiency. Many scholars believe that in an efficient market, yield curves should satisfy a property known as the expectations hypothesis. But the evidence indicates that in practice, yield curves fail to satisfy this property, possibly because the most important behavioral element is conservatism. Specifically, anchoring-and-adjustment gives rise to underreaction, particularly in connection with expectations about future rates of inflation. It may be that underreaction interferes with the forces that would otherwise induce yield curves to satisfy the expectations hypothesis.
Case Study: The Orange County Bankruptcy: Setting the Stage
In December 1994, Orange County, California, filed for bankruptcy. The largest municipal bankruptcy in U.S. history occurred as a result of the investment strategy followed by its treasurer, Robert Citron. The case provides an excellent vehicle for a discussion of interest rate forecasts.
One month after the bankruptcy declaration, Robert Citron appeared before California legislators. At tha
t time, he testified that he relied almost exclusively on the advice of Merrill Lynch officials, including the interest rate outlook from its chief investment strategist, Charles Clough. Citron stated that in 1993 Clough had forecast flat or falling interest rates for three to five years.
Charles Clough was known for having made one very important interest rate prediction. Consider figure 14-1, which displays the behavior of the yield on the ten-year Treasury bond between 1985 and 1993. Focus on 1988. In that year, when Treasury bond yields were over 9 percent, Clough forecast a long period of disinflation and said long-term rates would fall further than most people thought possible. As figure 14-1 shows, his forecast turned out to be very accurate. Merrill Lynch heavily promoted the forecast, basing an advertising campaign on it.
In 1988, Robert Citron had already been county treasurer for seventeen years. In a revealing interview with the Los Angeles Times that year,
Figure 14-1 Yield on 10-Year Treasury Bond, November 1985–December 1993
In 1988, Merrill Lynch’s Charles Clough predicted a long period of disinflation and said that long-term interest rates would fall further than most people thought at the time. This chart shows him to have been right. Robert Citron, treasurer of Orange County, California, made several leveraged bets based on his belief that interest rates would decline. The chart also shows why his bets paid off handsomely through 1993.
Beyond Greed and Fear Page 25