Traders, today when Signal flipped to the back month and CNBC stayed on the front month, I finally flipped, and decided to implement a program that I'd been meaning to write for a while: namely a little program that given the current SPX spot value, prints the SPX future's fair value, S&P future's premium, along with buy and sell limits. The values printed are for both the front _and_ back month. The 'prem' program is a few hundred lines of C, and compiles on my Unix workstation. Maybe some tweaking will be required for a PC platform. If one of y'all decides to compile the program on PC or Mac, perhaps you'd be willing to share it with others that request the binary .exe file. The program runs along the lines of the algorithm described in the MTA discussion below. Empirically, I'd say the commission plus slippage estimates are a little on the high side, but perhaps with experience you'll be able to tune those values. The program's method for handling dates internally is somewhat opaque (it uses a date conversion method described in Algorithm 199 of the Communications of the ACM, circa 1972). You can probably rewrite this entire program using 5 lines in a spreadsheet, or make the C version simpler by taking advantage of your C library's date conversion support routines. The discussion below was contributed (courtesy of the BBS) by a trader (George Kirikos, gkirikos@epas.utoronto.ca) who subscribes to the Random Walk BBS. You can find this file and several others via anonymous ftp at dg-rtp.dg.com:/pub/misc.invest. Here's an example (today's date is 03/08/95 and SPX closed at 483.14). The program determines the current date, and the number of days remaining until futures contract expiration for the front (nearest) month, and back (next) month. T-bill rate and S&P 500 rate are explicitly in the program. % prem 617.48 Front Month (Dec) ----------------- fair value: 617.81 prem: 0.33 buy limit: 1.48 sell limit: -0.32 Back Month (Mar) ---------------- fair value: 622.91 prem: 5.43 buy limit: 6.58 sell limit: 4.78 spread: 5.10 If you have any questions, suggestions or improvments to the 'prem' program, please send them my way: Gary Funck (gary@intrepid.com) -- -- BBS: The Market Technicians Association BBS 914-478-7311 Date: 11-27-91 (00:40) Number: 225 To: ALL Refer#: NONE (Msg #38 of 46) From: JACK FRICKEY Read: NO (Thread #1 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG I am a newcommer to the MTA BBS. It looks like a great place to learn and share ideas. I'm looking forward to it. I have a question about program trading and the S&P premium: FNN quotes the S&P futures premium and goes on to say, "Fair value is X; buy programs kick in at Y; and sell programs kick in at Z." How are these values (X,Y,Z) calculated? and how do they relate to one another? Shouldn't I (at least in principle) be able to calculate them before the market opens each day? BBS: The Market Technicians Association BBS 914-478-7311 Date: 11-28-91 (22:20) Number: 233 To: JACK FRICKEY Refer#: 225 (Msg #39 of 46) From: JOHN CARDER Read: YES (Thread #2 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG JF>How are these values (X,Y,Z) calculated? and how do they relate to one >another? Shouldn't I (at least in principle) be able to calculate them >before the market opens each day? The easiest thing to do is to simply adjust the previous day's figures. They don't move much from day to day. Keep in mind that X and Z, the levels for buy and sell programs are just estimates. The idea is to guess the point at which entering the arbitrage (futures vs. underlying stocks) is profitable. Why aren't these points precise? Because costs vary. Arbitrage programs come in many sizes. Some participants are comfortable with as few as 160 of the 500 stocks. They look for the ones that they think will track the S&P500 best. Keep in mind that the S&P is cap weighted, so that's not as big a risk as it seems at first. Another example: An index fund already owns the underlying stock. If the futures trade at a discount to cash, they can buy the futures and sell an equal amount of their stock and lock in a profit. They don't have to worry about the uptick rule since they are selling stock that they already own rather than borrowing it to sell (selling short). Another example: An arb entered a sell spread (sold the futures and bought the stocks) at a good price (relatively big premium). If, before expiration, the futures sell at a discount to cash, he can unwind his position early and make at least as much in less time. If he's greedy, he may decide to wait and see if the discount will deepen before unwinding. Then he'll make even more. Once he enters the position, he knows the least he can make (unwind at expiration), so it can be tempting to try to make a little more, since there isn't a downside. I guess I didn't answer your question, but I hope this helped. John Date: 12-04-91 (18:23) Number: 236 To: JACK FRICKEY Refer#: 225 (Msg #40 of 46) From: PARRY HETTARACHI Read: YES (Thread #3 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG JF>How are these values (X,Y,Z) calculated? and how do they relate to on JF>another? Shouldn't I (at least in principle) be able to calculate th JF>before the market opens each day? FAIRVALUE (FV) CALCULATION: 1. FV=Spot Index(S) + Carrying Cost(CC) CC= Differential between interest income(I) and forgone dividends(D) over the holding period. EXAMPLE: Assume a S&P500 futures contract with 60 days to expiration. S=(spot index) 165.00 pts. I=(interest)* +2.70 D=(dividends) -1.35 ------- Net 166.35 pts. I*=I(dtm/365)S=.10(60/365)165 D =D(dtm/365)S=.05(60/365)165 2. Arbs seek to take advantage of mispricings; however, to breakeven, the arbs must cover his round-trip costs which include commissions and bid/ask slippage. Commissions (r-t) +/- .70 pts. Bid/Ask slippage (r-t) +/-1.00 ---------- Total +/-1.70 pts. Arb-Free Zones = 164.65 to 168.05 I hope this helps Jack. Formula credits go to Greg Kipinis, Donaldson, Lufkin & Jenrette. *-Parry-* Date: 12-08-91 (12:24) Number: 242 To: JOHN CARDER Refer#: 233 (Msg #41 of 46) From: JACK FRICKEY Read: YES (Thread #4 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG Thanks for your response. An interesting discussion on how the big boys operate. I'm trying to determine if I can (and should) factor the movement of the premium into a day trading algorithm for the S&P futures. There have been several occasions recently that my indicators would say that a down trending market was ready to turn up when CNBC/FNN would say sell programs kicked in and the market would move lower. Any thoughts on the subject? Date: 12-08-91 (12:28) Number: 243 To: PARRY HETTARACHI Refer#: 236 (Msg #42 of 46) From: JACK FRICKEY Read: YES (Thread #5 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG Thanks. That's exactly what I was looking for. Now that I see it, it is very logical and I should have been able to intuit it (at least philosophically). I would not have been able to come up with the numbers for r/t costs, though. Are most of the 1.70 pts. in costs associated with buying and selling the "representative basket of stocks?" So, depending on the content of the basket, the costs may vary a bit. (Hence, as John mentioned in msg #233, the numbers are not precise.) Date: 12-08-91 (12:16) Number: 246 To: JACK FRICKEY Refer#: 242 (Msg #43 of 46) From: RICK BULLEN Read: YES (Thread #6 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG JF>There have been several occasions recently that my indicators JF>would say that a down trending market was ready to turn up when CNBC/FNN JF>would say sell programs kicked in and the market would move lower. Any JF>thoughts on the subject? I have noticed this myself. This is an area that is not well understood and hopefully will stay that way. :-) It would be a shame to lose such a good indicator. [just kidding] Are you familiar with the Technician program from Equis? There is an attractive composite sentiment indicator in the program and, would you believe, one of its components is the premium/discount in the 1st deferred NYFE futures (picked, no doubt, for its [relative] stability); this measures the same phenomenon that you have observed, ie, that futures depressed sufficiently to invite program selling reflect a state of excessive bearishness and may suggest the appropriateness of a contrarian viewpoint. Got any ideas on why this works? Date: 12-08-91 (15:52) Number: 248 To: JACK FRICKEY Refer#: 242 (Msg #44 of 46) From: JOHN CARDER Read: YES (Thread #7 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG >There have been several occasions recently that my indicators >would say that a down trending market was ready to turn up when CNBC/FNN >would say sell programs kicked in and the market would move lower. Any >thoughts on the subject? First, I'm not a daytrader. But I'm always happy to dispense an opinion. I don't know how you could day trade the S&P without incorporating the premium. It has such a dramatic effect. Parry uploaded the general formula. Keep in mind that the interest rate is the rate currently paid on T-Bills maturing near the S&P expiration. Good luck. Date: 12-15-91 (02:19) Number: 254 To: RICK BULLEN Refer#: 246 (Msg #45 of 46) From: JACK FRICKEY Read: YES (Thread #8 of 9) Subj: S&P PREMIUM/PROGRAM TRADE Status: PUBLIC MSG Yes, I have the Technician. It uses, at best, daily data. Maybe it is my narrow viewpoint, but it seems to me that all the interesting implications of premium variation is in the intraday action and would be washed out in daily data. Am I missing something? Date: 12-15-91 (08:33) Number: 255 To: JACK FRICKEY Refer#: 254 (Msg #46 of 46) From: RICK BULLEN Read: YES (Thread #9 of 9) JF>Am I missing something? In my experience, yes, but perhaps I am too. I dont use it intra day, because I dont (intentionally) day trade, and I find intraday moving averages and oscillators of the price effective enough in calling turns for timing position entry and exits. Also, like all leading indicators, sell program activity has a way of repeating itself multiple times before turns, and I find this type of "nagging" by an indicator to be distracting. BTW, John Murphy displayed an especially interesting chart on tv a couple of days after your original message (I believe it was 12/11, did you see it?) showing intraday premium/discount in the S&P with buy program and sell program limits shown on the chart as horizontal lines. This was shown below a chart of the intraday s&p. The chart showed several waves of sell programs during the first half of the day with the market going lower with each wave, then a *coincident* turn in the premium and the market in midday, followed by waves of buy programs during the second half of the day with the market getting stronger with each wave. This, it seems to me, is the type of action that people focus on when they say that program selling is bearish and program buying is bullish. Now, if we turn our attention to the daily data, you will notice that the low in the spread was on 11/22/91, when the futures were at a sharp discount to cash (-64 in spz). This was followed by a couple of days of basing and then the market took off. While the absolute value of the premium doesnt, a priori, tell you that there was or was not program trading, the fact that this indicator hit a low on a closing basis a couple of days before the actual market bottom suggests it is worthy of study as a leading market indicator. -- end --