we learn some interesting facts about his trading strategy, and some of the behavioral biases to which he was subject.1 In the interview, Citron indicated that he had $2.5 billion under management, which he was investing in U.S. Treasuries (bills and two-to-five year notes), U.S. agencies, certificates of deposit, commercial paper, banker’s acceptance and medium-term notes, as well as time deposits with banks and S&Ls.
In light of the subsequent bankruptcy, the interview is telling about both Citron’s actions and his thinking. Two excerpts follow. In the first Citron discusses the riskiness and performance of his investment strategy.
Q: Do you invest in the stock market?
A: County and city treasurers by state law are not permitted to invest in the equity markets or corporate bonds.
Q: So your money managing decisions revolve around the kinds of notes or bonds you’re willing to buy?
A: Yes. I have around $900 million in U.S. government securities, from two to five years in maturity.
Q: Do you actively trade, based on fluctuations in interest rates?
A: I’m not a trader. … I buy and hold, and I do matched reverse repurchase agreements. … It’s not trading, because the spread is always locked in.
Q: How much money has this technique earned you?
A: Through the first seven months of the fiscal year, we have earned $4.5 million on reverse repos alone.
A: From July 1, 1986, to June 30, 1987, we earned 8% on all the funds. If we hadn’t been doing reverse repos, our return would have been only about 7.2%.
Q: How does that 8% return compare to other counties?
A: The state of California has a local agency investment fund in which any local agency, county, city or special district can invest money. … (The fund) had an average yield of 7.4% for the fiscal year.
In the second excerpt, Citron provides his predictions for interest rates and the overall economy. As you read this portion of the interview, think back to the discussion in chapter 5 about the strategists’ market predictions. See whether you can spot any similarities.
Q: What is your outlook for the economy?
A: The current business cycle has already extended far beyond the average length of three to four years, to over five years. And there’s little doubt that a correction is rapidly approaching in the form of a recession. …
Q: When can we expect to see a recession?
A: Events in the summer months of 1988 will indicate the beginning of a recessionary period. …
Q: What’s your outlook for interest rates?
A: We are in a volatile market, so we’ll have strong swings. But on average, I think rates will go up.
Q: Why?
A: Because of inflationary pressures. The Fed will react by doing more tightening than loosening.
Q: So we are going to see a recession with rising interest rates?
A: Yes.
Q: Is the Fed going to succeed in dampening inflation with this restrictive monetary policy?
A: Yes, in the summertime.
Overconfidence and Gambler’s Fallacy
Were you able to spot examples of heuristic-driven bias in the second excerpt? Citron’s 1988 forecast, of a coming recession with rising interest rates that will dampen inflation in the summer, appears to have been made with a great deal of confidence—some might say overconfidence. And why was Citron forecasting a recession in the summer of 1988? Because the economic expansion at that time was a year or so longer than average. Is there a classic bias underlying this forecast? How about gambler’s fallacy—the law of small numbers? After tossing five heads in a row using a fair coin, is a tail due? Was a recession due? And did a recession with rising interest rates come about? Yes to rising interest rates, but no to the recession, at least not for two years more.
Well, one out of two may not seem bad. But the overconfident interest rate forecast, in combination with other issues identified in this 1988 interview, combined to produce a disaster. The other issues revolve around Citron’s investment strategy. He mentions his use of two- and five-year Treasury securities. Figure 14-2 illustrates the yields on the two and five-year securities between November 1985 and December 1993. The top curve represents the five-year Treasury yields.
Figure 14-2 Yields on 2-Year and 5-Year Treasury Bonds, November 1985–December 1993
Robert Citron borrowed at the two-year rate and invested at the five-year rate. The chart shows why, between 1990 and 1993, Orange County’s portfolio profited from the widening gap between two-year and five-year yields.
As can be seen from figure 14-2, the spread was typically positive during this period. Citron used a reverse repurchase agreement strategy in which he purchased five-year Treasurys obtained with money borrowed for two years at the two-year rate. So, for example, in December 1993 Citron might have been borrowing at the two-year rate, 4.24 percent, to purchase five-year securities that were yielding 5.2 percent, thereby capturing a spread that was almost 100 basis points. Such leveraging corresponds to purchasing stock on margin.
Overconfidence and Hindsight Bias
The 1988 summer recession predicted by Robert Citron finally arrived in the autumn of 1990. But by 1993, the slowdown was over and the economy was again growing robustly—so much so, in fact, that the Federal Reserve Board was concerned about a rekindling of inflation. In February 1994, the Fed raised short-term interest rates for the first time in five years. It raised them again in March.
On December 6, 1993, Martin Mauro, fixed income strategist at Merrill Lynch, predicted that the Federal Reserve would tighten monetary policy modestly in 1994.2 Consequently, he expected a flatter yield curve, with higher short-term rates and lower long-term rates. The forecast was way off. Indeed, 1994 was an extraordinary year for the yield curve. In a 1995 survey article that appeared in the Journal of Economic Perspectives, John Campbell described what made it so. As short-term interest rates rose throughout the year, so too did long-term rates. Furthermore, the spread stayed about the same.
The behavior of the spread was of considerable importance to the value of the Orange County Investment Pool portfolio. The rising yields reduced the value of both the two-year Treasurys and the five-year Treasurys. A well-known feature of bond prices is that the longer the duration of the bond, the more sensitive the bond price is to a change in yield. Specifically, the prices of five-year Treasurys are more sensitive to yield changes than the prices of two-year Treasurys. Hence, the rise in the yield curve during February and March lowered the value of the Orange County portfolio.
In July 1993, Citron had confidently predicted that interest rates would not go up. In response to a question from an investment banker about how the value of his portfolio would be affected by a rise in interest rates, he responded that interest rates would not rise. When the investment banker asked Citron how he knew this, he is reported to have replied: “I am one of the largest investors in America. I know these things.”3
Citron described his outlook further in his September 1993 annual report, stating: “We will have level if not lower interest rates through this decade. Certainly, there is nothing on the horizon that would indicate that we will have rising interest rates for a minimum of three years. We believe that our comparative higher interest earning rate yields over the next three fiscal years is insured.”
Interestingly, Citron later claimed to have anticipated the February interest rate jump. On March 9, responding to the administrator of the county’s public employee retirement system, Citron stated: “The recent increase in rates was not a surprise to us; we expected it and were prepared for it.”4 In light of all the public statements Citron made before February 1994, I find it difficult to believe that he was not surprised. But I do believe that, after the fact, he reconstructed his past beliefs. I believe he displayed hindsight bias. Nevertheless he went on the record that March as saying that he thought further Fed tightening to be likely, with any run-up in interest rates short-lived.
Remember Charles
Clough, the Merrill Lynch strategist on whose views Citron was relying. The April 6, 1994, issue of USA Today reported Charles Clough’s outlook on interest rates, an outlook that resurfaced in court proceedings later.
Bonds a buy: After the sharp rise in interest rates, Merrill Lynch strategist Charles Clough is bullish on the bond market again. Tuesday, he told clients to start buying 30-year Treasury bonds when their yields are higher than 7.3%. T-bond yields were 7.40% Monday but fell to 7.24 Tuesday. Clough made one of the great bond-market calls ever in 1988, when T-bond yields were over 9%. He forecast a long period of disinflation and said long-term rates would fall further than most people thought possible. … “This is entirely different, partly because we’re still in an economic expansion so rates don’t have that far to fall,” Clough says of his renewed call to bonds.
Still, he believes the recent jolt in rates will slow the economy enough to force rates back down late in the year.5
Frame Dependence: Reference Points and Layered Pyramids
Later that April, the Federal Reserve increased interest rates again. This lowered the value of Citron’s portfolio even more. Now, the Orange County treasurer’s position is an elected one, and 1994 was an election year. Interestingly, for the first time since 1970, Robert Citron found himself with an opponent. Given that the environment of the time was characterized by rising interest rates, I am not surprised that a spirited debate ensued about the merits of using leverage.
Citron’s opponent was John Moorlach, later to become his successor. During the campaign, Moorlach wrote a letter to the Orange County Board of Supervisors, criticizing Citron’s investment strategy. In his letter, Moorlach stated: “Every prudent investor chooses safety of principal as the top priority,” he wrote. “Next comes the need for liquidity. The last priority is achieving yields. … Mr. Citron has these priorities inversed.”6 Is this approach reminiscent of a layered pyramid, with a hierarchy of securities matched to a hierarchy of goals? The notion of a layered pyramid forms the behavioral basis for constructing portfolios (see chapter 10).
An article that appeared in the April 30, 1994, issue of the Los Angeles Times offers additional insights. The article quotes R. A. Scott, the head of the county’s General Services Agency, who has known Citron for more than twenty years and paints the following revealing portrait. “He’s competitive, and if he returns a greater rate on short term money than most people, he considers that winning. … It’s pride. It’s being above average. When he’s trading, he’s all business.”7
Pride, being above average, and winning: is there a reference point effect here? How about overconfidence about his relative skill? Did Citron’s reference point correspond to earning a higher return on short-term money than others earn? We know from chapter 9 that when the probability of ending up below a given reference point is large enough, people tend to seek risk. Indeed, the Los Angeles Times article points out that Citron concedes he takes risks but insists they are prudent ones.”
Reference points and loss aversion operate in peculiar ways. Despite the risky investment approach he followed in managing the Orange County Investment Pool, Robert Citron never owned a single share of stock in his personal portfolio. Paradoxical? The difference reflects different mental accounts and different reference points.
Forecasts and the Illusion of Validity
What about Citron’s interest rate forecast at this point, April 1994? Was it similar to Clough’s April perspective? The article states the following:
In recent weeks, Citron has received “collateral calls” that forced him to pony up an additional $215 million in collateral, because the value of the securities he used to borrow money has skidded as interest rates have climbed. His assistant, Matthew R. Raabe, said he expects yet more collateral calls, but he is not worried because the county’s $7.5-billion investment pool has about $1 billion in liquid assets.
Citron and Raabe … believe interest rates will inch up for a short time and level off. If they are wrong, they say they have a contingency or “exit” plan, although they are unwilling to share the particulars. Raabe said the only way the county could get into trouble is if “short-term interest rates are going to continue to go up and … we don’t react. But that’s not going to happen.”8
At this stage, Robert Citron was plagued by a lot of questions: Would he win reelection against John Moorlach? Which investors would withdraw funds from the county pool? In what direction would interest rates move? To whom did Citron turn for guidance? To psychics and a mail-order astrologer. One psychic adviser predicted that Citron would encounter financial difficulties in the month of November, but that these would be over in December.9 Some of these predictions turned out to be correct. Might there have been some behavioral phenomenon at work here? Indeed, yes: Try the illusion of validity.
Conservatism and Loss Aversion
Still anchored to the downward trend that had brought his investment strategy so much success, Citron appears to have been convinced that short-term rates would level off, so much so that he continued to hold inverse floaters. Inverse floaters are structured to pay lower interest as interest rates rise. The investor makes money when interest rates decline, but loses money when they rise. In fact, it appears that he actually used the inverse floaters as collateral in order to purchase medium-term securities.
Unfortunately, short rates did not level off. The Fed increased rates again in August. The losses to the Orange County portfolio were mounting. Just after Labor Day, a concerned president of one of the local water districts in the county learned that Robert Citron had not reduced the degree of leverage. In fact, quite the opposite. An article appearing in USA Today reports that Citron actually increased his borrowing “in a desperate bid to recoup his losses.”10 A few days earlier a piece by Ronald Picur had appeared in the Chicago Tribune describing this behavior in the following terms: “True to a gambler’s mind-set, Citron increased the size of his ‘wagers’ by using leveraged funds.”11 According to county records, in March 1993 the principal fund run by Mr. Citron was leveraged 2.4:1. But by August 1994, its leverage had grown to nearly 3:1. By then, Mr. Citron had borrowed nearly $13 billion, using most of the proceeds to purchase notes from Merrill Lynch. (If we needed further convincing about there being a reference point effect at work here, this should do the job.)12
A Wall Street Journal article on Citron was aptly titled “Hubris and Ambition in Orange County: Robert Citron’s Story.” The following excerpt describes the effect of loss aversion on Citron’s behavior.
However, the losses were still only on paper, and Mr. Citron was apparently convinced that he could still weather the typhoon. For years, he had boasted about how he almost invariably held securities until maturity, when he could cash them in at face value. That way, he explained, he was able to avoid the losses that come from selling a security that has been adversely affected by a rise in interest rates.
He clung fiercely to that philosophy in his last annual report, on Sept. 26. Mr. Citron noted that there was concern over “paper losses” due to rising interest rates. But he said that the county didn’t plan to record any such losses and didn’t plan to sell its securities.13
In chapter 9 I described why, tax issues aside, the distinction between “paper losses” and “realized losses” is more psychological than real. This is a framing issue. Holding securities to maturity so as not to realize a loss does not imply that wealth has not declined. But the foregone wealth is framed as an opportunity loss, rather than an out-of-pocket loss. The “hold until maturity frame” simply obscures the opportunity loss.14
In any event, despite Citron’s stated intention not to sell any securities, the reverse occurred. The Fed raised interest rates again in November. Investors who held Orange County Investment Pool securities as collateral anxiously saw those securities decline further in value. In reaction, they began to sell that collateral, thereby forcing Orange County to realize losses. On December 1, Robert Citron and assistant treasurer
Matthew Raabe told reporters at a county press conference that the fund faced a $1.5 billion paper loss.
On December 4, Robert Citron resigned as county treasurer. Two days later the county filed for bankruptcy under Chapter 9 of the bankruptcy code, in an attempt to prevent its creditors from liquidating the collateralized securities. This was the largest municipal bankruptcy in U.S. history. At that time, the loss to the Orange County Investment Pool was estimated at $2 billion. On April 28, 1995, Citron pled guilty to six felony counts. He was sentenced to a year in county jail and ordered to pay a $100,000 fine. Matthew Raabe was convicted of the same charges and became the only public official sent to state prison for his role in the bankruptcy. He was sentenced to three years in prison and ordered to pay a $10,000 fine.
Regret, Hindsight Bias, and the Confirmation Bias
Citron’s appearance before state legislators in January 1995, offers an example of regret and responsibility shifting. The Los Angeles Times reported the issue as follows: “Here was Citron—proclaimed a financial expert for years both by himself and his fawning Board of Supervisors—now reduced to saying how, ‘in retrospect,’ he would have done a lot of things differently.”15
Regret is the pain felt from recognizing that one could have done things differently. The intensity of the regret depends on the degree to which a person feels responsible for the decision that was taken. One way people attempt to shift responsibility is by playing the “blame game.” This game, whereby a client picks someone regarded as an expert and relies on him or her for advice, is usually set up in advance. If things go well, the client takes the credit, attributing the success to his or her own skill. But if things go badly, then the client can attribute the blame to the expert, thereby reducing regret by shifting responsibility for the negative outcome.16 Of course, for this to work, the person to whom responsibility gets shifted must be seen to have expertise. Otherwise, the client will feel just as much regret for having relied on a novice for advice. In this regard, Citron said during his testimony: “I understood Clough to be the preeminent expert in the field of investment strategy.”17
Beyond Greed and Fear Page 26