Each morning, CNBC commentators note whether futures weigh in above or below fair value. What is fair value as it applies to the indices, anyway, and why does it appear to be such a confusing topic? According to H.L. Camp and Company, more callers to CNBC's Squawk Box ask about fair value than any other topic. The term doesn't hold the same meaning for all, it turns out, and some would argue that the CNBC-type accounting of fair value proves to be a sloppy and useless measure. An article written by Mark Hanes and contributed first to CNBC.com and then reprinted on another source argues against that conclusion.
Most CNBC watchers want to know each morning whether S&P 500, Nasdaq and Dow futures are trading above or below fair value. They want an idea of how sentiment lies as the open approaches. If futures trade above fair value, they feel that early trading might have an upward bias. If futures trade below fair value, they prepare for early trading with a negative bias. Wise traders don't expect that bias to always appear or to carry much beyond the opening moments, however, with other factors then kicking in to contribute to the behavior of the markets.
The CNBC version of fair value differs from the strictest interpretation of that term, according to some sources. According to Investopedia, the CNBC-type meaning references the difference between the actual cash value of an index and the futures contract on that index. More specifically, Herb Greenberg, Senior Columnist for TheStreet.com defined the CNBC-type term as the difference between the closing cost of the component stocks of an index the day before and the closing cost of the futures contract.
That definition of fair value will garner complaints from purists, however, as it did for Herb Greenberg back in 1998 when he briefly answered a reader's question about fair value. One correspondent wanted to redefine that "CNBC-type" fair value, as many sources tagged it, as the "the closing spread," a different value entirely from the more respected "fair value" meaning. Greenberg's next column was comprised of a number of excerpts from readers who wanted to expound upon or argue against Greenberg's brief answer, but none contested Greenberg's characterization of CNBC's use of fair value. Hanes' definition, however, comports with the more traditional definition of fair value, as it turns out.
Investopedia, brief as its definitions tend to be, offers a more analytical computation for this equilibrium price or fair value for a futures contract, different from its "CNBC type" definition. This site notes that fair value is calculated by figuring in spot price, compounded interest paid and dividends lost until the expiration of a futures contract. Futures owners do not collect dividends, as do owners of the actual stocks.
Index arbitrage traders may be more interested in that traditional computation than in a closing spread. Here's how it works. If futures are priced at fair value, then they're theoretically at the same value that cash indices are in the absence of transaction costs. To calculate fair value, financing charges, a function of interest rates, would be added to the spot prices, and dividends would be subtracted to account for those transactions costs. The Chicago Mercantile Exchange's Rule 813.D sets forth the calculation and rules for fair value. HL Camp & Company uses the following formula:
FV = S[1 + (I - D)], where "FV" is fair value, "I" is the interest a trader would pay to borrow enough money to buy all the SPX stocks, and "D" is the dividend that would have been collected if the stocks had been owned instead of futures.
There's a time factor applied to the calculation of the interest, as the interest must be calculated from the date of the FV calculation until the expiration of March, June, September or December expiration of a futures contract. Carrying costs will drop as a futures contract draws nearer to its settlement day, so the fair value will more closely approach the actual spot price of the index in question. Those interested in following an exact computation of fair value, an exercise that might be useful for all to follow at least once, can find such a sample computation at www.cme.com/trading/prd/equity/fairvalue2543.html on the Chicago Mercantile site. Following such a computation provides a greater grasp of how fair value might change as interest rates or dividends change.
Hanes' article makes note of the same computations in the CNBC's version of fair value. At the time his article was written, CNBC was receiving fair values from Prudential Securities, and Hanes noted that the formula used was as follows:
F = S[1 + (I-D)t/365], where F is the fair value for futures; S, the spot index price; I, the interest rate, expressed as a money market yield; D, the dividend rate, expressed as a money market yield; and t, the number of days from the spot value date until the settlement of the futures contract. That formula argues against detractors who claim that the "CNBC type" fair value is a less useful measure of closing spread.
Useful or less useful to index arbitragers, that is, although many everyday traders might be interested in the times when futures and index values get so out of synch that program buying or selling programs are likely to be triggered. For that reason, the discussion of fair value probably wouldn't be complete without a discussion of premium. One of Greenberg's correspondents described premium as a threshold amount above or below fair value, at which buy or sell programs might be triggered.
If futures contracts are trading at fair value, big funds or institutions don't care whether they own stocks or futures. Arbitrageurs aren't going to find an opportunity to exploit the difference. If futures are trading at a threshold amount above fair value--if it would cost a threshold amount more to buy and hold futures until their settlement than to buy the equities and pay the transactions costs until that settlement day--then futures contracts might be sold and equities bought. If equities are trading at a threshold amount above futures, arbitrageurs might take the opposite action. HL Camp and Company claims that index arbitrage actions currently account for only about 10 percent of all program trading activity, however.
For those interested in watching for program trading activity, premium value should also be understood. Premium values are spreads that measure the arithmetic difference between futures and spot index values. So there's that plain old spread showing up again, rather than the more complicated formulas for fair value. Hanes says that CNBC defines "premium" just as it's defined here. Premium value can be monitored on many charting or quote services. Traders should check the symbol help section on their services to find the appropriate levels.
So how is that spread or premium used on CNBC or elsewhere? HL Camp & Company provides program selling levels each day for those traders interested in setting alerts for potential program selling or buying signals. On our site, Jeff Bailey usually provides those numbers for the benefit of readers who want to watch or set alerts. For example, on Friday, July 15, the company's fair value for the S&P 500 was set for +$3.57, with their computers set for program buying at +$4.71. The same day, another service listed fair value at +$3.52 and suggested that traders buy the SPY if the premium rose above +$4.85, so obviously the various companies issuing these program trading signals might offer slightly different results.
While these sources include premium, as defined as that spread, in their decisions as to where program buying or selling might occur, they do not use premium alone. Rather, they include proprietary studies on where program selling is most likely to occur with respect to the fair value and the premium of the futures to the cash index. They're making proprietary guesses, if computer-assisted and sophisticated ones, as to the probability that a buy or sell program would be triggered at a certain premium level. Hanes' article indicates one reason that some buy or sell programs might not be triggered when expected and why those proprietary studies might be needed. Because arbitrageurs use borrowed money and since each might have a slightly different interest rate, not all arbitrageurs might act at the same premium level.
HL Camp and Company describes premium as the spread between the most active of the available S&P 500 futures contract, usually the front-month contract, and the actual cash index value, with the cash index value subtracted from the futures contract. Most times, the spread varies between +/-$5.00, the site notes. This company's numbers are projections for the amounts at which they believe program trading will kick in and index stocks might be bought or sold as a result. The company corroborates that premium values cannot be used as a pure indicator of when a buy or sell program might kick in, as the company's extensive research has shown some occasions when PREM signaled that a buy or sell program should occur but didn't.
Some traders find it helpful to monitor premium values and match that to a preferred source's projections of when buy or sell programs might be hit, to gauge their own buy or sell decisions on these best guesses. However, as Hanes himself cautioned, the action of buying stocks when futures are a certain threshold above fair value helps to narrow that spread. That action brings futures closer to fair value. The narrowing can happen quickly, as it can when futures might have been at a discount to the indices, and equity selling occurs. Those who have seen pre-cash-market futures values far above or below fair value are often surprised to see action at the open. That upward or downward bias sometimes doesn't last long, sometimes not even until the cash open. Use those comparisons as one tool only.
If Hanes' characterization is correct, it turns out that those who disparage "CNBC type" fair value might need to be a little fairer in their criticisms, but let's look at actual numbers. On the morning of July 21, CNBC noted that fair value for the SPX was +$2.05. The SPX close the day before had been 1235.03 while futures saw a 4:00 closing value of 1236.75, a 3:15 value of 1237.50 and a 5:15 closing value of 1236.25, none of which translated into a closing spread of +2.05. However, HL Camp & Company noted more traditional calculation of fair value at +2.08 for the day, roughly analogous to the CBNC-announced version. It appears Hanes was right. Early in the morning, futures had been positive, but below fair value, indicating that expectations were for a flat or slightly lower open, but the Chinese announcement about the unpegging of yuan and the news of new bombing attempts in London changed the outlook, with equity futures shooting all directions. Fair value comparisons probably proved less useful than other factors.
This article was intended to provide an overview of these often-confused terms rather than to promote a trading style based on potential program selling or buying points. Some find such tactics useful; others, less so. Traders should monitor premium and expected buy or sell programs for a time to determine applicability to their trading styles before basing decisions on those signals.