Ibis, Redibis
One may think that astrology and the dark arts belong to a realm far removed from financial markets. Yet forecasting markets is, in the eyes of its detractors, little more than divination; economists prefer to call it educated guessing based on probabilities. One of the oldest recorded bets on a forecast is a case in point. According to Aristotle, Thales of Miletus, having concluded from his knowledge of astronomy that the coming olive harvest would be abundant, paid small deposits to secure all the olive presses of Miletus and Chios, then rented them out at a handsome profit when the harvest came. Whether his forecast was sound matters less than how he bet: had it failed, he would have lost only his deposits. Most investors today consult economists and models rather than the stars. Not all of them. More than two and a half millennia after Thales, a public treasurer hailed as a wizard was still looking to the sky, with none of Thales's science and far more than a deposit at stake.
The Wizard of Orange County
In early December 1994, the senior officials of Orange County, California, met behind closed doors. The county was days away from what was then the largest municipal bankruptcy in American history. At that meeting, according to grand jury testimony later obtained by the Los Angeles Times, the finance director learned that the treasurer, Robert Citron, took interest-rate predictions from a mail-order astrologer and also consulted a psychic. The psychic had reportedly told him that December would be a bad month, but that his money worries would be over after that.
Citron was no fringe figure. First elected to county office in 1970 and re-elected seven times, his record seemed to prove his gift: between 1991 and 1993 his investment pool earned more than 8.5% a year, against about 7% for bond funds. Cities and school districts rushed to hand him their money; some even borrowed to do so.
His method rested on a single bet: that rates would stay low, or that any rise would be brief. At the end of 1993, short-term rates were below 3%, while five-year notes paid around 5.2%. Citron borrowed short term against his bonds and used the cash to buy more, turning some USD 7.5 billion of public money into a USD 20.5 billion portfolio. When the Fed began raising rates in February 1994, he reassured an anxious investor that his strategy allowed for the "inevitable but unsustainable rise in short-term interest rates." The Fed raised rates six times that year, from 3% to 5.5%. His lenders demanded ever more collateral, until Credit Suisse First Boston refused to renew USD 1.25 billion of loans. On 6 December the county filed for bankruptcy; its losses would reach USD 1.64 billion.
Not in Our Stars
The psychic, it turned out, had been right. But prophecies, even those dressed up as economic forecasts, are treacherous in nature. They are seldom simply false; more often they come true in a way no one expected. The Latin formula Ibis redibis non morieris in bello, traditionally attributed to an ancient oracle, promises a soldier that he will return from war, or that he will perish in it, depending on where one places a comma.
That December was indeed a terrible month, and the storm eventually did pass: soon after the county sold its portfolio, rates fell by about 250 basis points, costing it an estimated further opportunity loss of some USD 1.4 billion. By then Citron had resigned, and the county could only watch.
What, then, is the real lesson of 1994? It is not simply "avoid bonds when central banks tighten." Most of Citron's bonds matured within five years; on their own, they carried only modest risk. It was disproportionate and reckless borrowing that swelled the portfolio to 2.7 times the money entrusted to it and multiplied its sensitivity to rates by the same factor. Worse, that borrowing was renewed from one day to the next, with no assurance that lenders would keep renewing it. The fault, as Shakespeare's Cassius would have put it, lay not in his stars but in himself: in how much his strategy needed him to be right, and how little time it gave him if he was wrong.
Thirty-two years later, the sky has again offered its omens. In May 1994, in the middle of the bond market rout, an annular eclipse crossed the United States; on 12 August this year, the first total eclipse over mainland Europe since 1999 swept from Iceland to Spain, and the moon's shadow dimmed skies across the continent. Within weeks, central banks were raising rates again, and investors were once more trying to divine what comes next. On 10 September the ECB hiked rates for the second time this year, to 2.5%, to stop the energy shock from spreading to wages and prices. The next day, drone attacks led Saudi Arabia to shut its East-West pipeline, its main export route while the Strait of Hormuz remains effectively closed. Brent rose to around USD 108 a barrel. On 16 September the Fed followed with its first hike since 2023, and most of its members expects another before year-end.
For monetary policy, this is in our reading a recalibration of rates rather than a regime shift. The path beyond it is anything but clear: for 2027, eight Fed officials project another hike, six a hold and four cuts.
Planets in Evil Mixture
If central banks are moving in small steps, bond markets are not. Sovereign bonds have been caught in a storm, a sharp selloff driven by more than monetary policy. As our house view has argued, central banks can no longer systematically look through supply shocks when inflation stays above target. Then comes an inauspicious planetary alignment for bondholders: public debt, persistent deficits, geopolitical uncertainty and a resilient US economy, all pushing long-term yields higher.
Following the Fed decision, our Investment Committee concluded that the selloff had created an opportunity and decided to reduce our underweight in sovereign bonds through the purchase of short-maturity government bonds in all reference currencies except the Swiss franc. With the SNB policy rate at 0% and the two-year Confederation yield close to zero, the Swiss short end still pays too little to justify the move.
The arithmetic is simple. On the day of the decision, the two-year Treasury yielded 4.74%, only 28 basis points less than the ten-year, with about a quarter of its sensitivity to interest rates. In other words, almost the same income for a fraction of the risk. Roughly speaking, yields would have to rise by about 2.5 percentage points within a year before losses on a two-year note wiped out its income; for the ten-year, a rise of about 0.65 points would be enough.
Three Tomorrows, One Lesson
We do not claim to read the stars, nor do we rely on prophecies, even the self-fulfilling kind that markets sometimes produce. Our own educated guesses for the months ahead take the form of three scenarios, each with a probability attached, and 1994 helps read each of them, if only by contrast.
The most likely of these futures, a soft reflation to which we give a 50% probability, looks much like 1994: solid growth, inflation pressures and more rate hikes to come. It would test the short end, and a first test has already come: after our decision, markets briefly priced a 75% chance of another Fed hike in October, before that probability halved to below 40% at the end of September. Long yields did not follow; the ten-year Treasury yield has since risen above 5.3%, its highest since 2007. Short maturities, as 1994 showed, are resilient, not immune. Two-year yields had risen some 200 basis points by early May, more than long-term yields; yet by mid-November, one- to three-year Treasuries had lost less than 5% of their price, against 20.5% for twenty-year bonds. And unlike 1994, when few expected so much tightening, several hikes are already priced today: the two-year yield sits 90 to 120 basis points above the Fed's policy rate. We are buying after the storm broke, not before it.
The omens could also be darker, in a way 1994, with its strong growth, never knew. In a stagflation, which we put at 30%, inflation would stay high while growth slows, and central banks would hold rates rather than raise them. Short-dated bonds would simply earn their yield; long-dated bonds would suffer from persistent inflation and fiscal strain. This is why we remain prudent, and why we have not bought long maturities.
The opposite turn would be a mild recession, to which we give 20%. Here too 1994 offers no exact parallel, since 1995 brought a slowdown rather than a recession, but what followed it shows what happens once tightening ends: rates fell by about 250 basis points within a year. This is the scenario in which our choice would look too cautious. Long bonds would gain most, and a portfolio still underweight in sovereign bonds would lag. Short maturities would gain too, however, and without leverage they could be held to maturity, as Orange County's could not. It is a cost we accept, given the low odds we attach to that outcome.
A Margin for Error
The difference between a prophecy and a probability is that the first names one future, while the second prepares for several. In two of our three scenarios, short-dated government bonds earn at least their yield; in the third, our central one, they have a cushion of about 2.5 percentage points before a year's income is lost. Long bonds, by contrast, only clearly win in the least likely scenario. This is why we have increased our exposure to government bonds where it asks for the least prophecy, and nowhere else.
And because our central scenario still carries upside risk to rates, while stagflation would punish duration, we have halved our underweight rather than closed it. A cushion of that size is one no astrologer has ever offered, but by no means a dispensation from foresight.
Thales risked his deposits; Citron risked the county. Neither lacked a forecast; only one had a margin for error. Today's oracle of Delphi sits in Washington, consulted by all and, now more than ever, sibylline.
Amid misleading stars and misread oracles, a short-dated government bond held to maturity is the rare promise whose comma is already in place: whatever comes to pass, the capital goes, and it comes back, with interest.
Ibis, redibis.
