Seeking Alpha
Peter F. Way, Blockdesk (712 clicks)
ETF investing, CFA, portfolio strategy, long/short equity
Profile| Send Message| ()  

What can we learn from market-maker hedging in this market cycle (since June, 2007)?

Let's take the Nasdaq-100 Index [NDX]. Whenever market-makers hedged their at-risk NDX positions the way they are right now - and they did so 132 days out of 1302, 10% of the time - the index was higher 3/4ths of days in the next 3 months than it was the day of the hedging.

The average gain of those up-days was +8%. That's an annual rate of at least +36%. But we didn't know that was the case, before the fact.

The aggregate market-making community, as indicated by their hedging, now is betting that the NDX in the next 3 months will see a possible rise of +8.4%, and a possible further decline during that time of -1.7%.

That's somewhat optimistic. Mildly reassuring. But is it worth a bet?

Suppose we set up a test of what could have happened in the past, in a way that minimizes any after-the-fact knowledge of what has actually taken place. To do that, let's use a procedure that was established more than twelve years ago (before the year 2000), and use it on just the last 5+ years.

The test is to see what market-maker forecasts have been for the NDX daily, since mid-year 2007, shortly before that index's earlier high of 2238 on October 31, 2007. (It's now 2584, and has been as low in the interim as 1044.) We will be impressed by the balance between their upside and downside prospects seen on each day, and select all those days where the up-to-down balance was at least as favorable (to a buyer) as it is today.

The bet proposition is that, for each day selected, in the next 3 months the index will rise to a sell target at least as high as the top of the forecast range of that day. At the first day's end-of-market price where that happens, that position will be closed out and the gain and elapsed time required (in market days) will be recorded. If that has not happened by the end of the 3 months, the gain or loss at that point will be added to the record, along with the 63 market days (21 a month) consumed.

The test produced 109 days from the 132 (84%) with recorded gains, 80 of them by reaching their sell targets, and the remainder by ending the 3 months at a price above the forecast day's. The average gain of those 109 was 10.8%. The record shows that of those separate 132 starting days selected, the average upper price forecasted was +13.1% higher than the market quote of that day.

The losing 23 positions, when averaged into the winners' scores by geometric mean, produced an overall, by position, gain of +7.2%. The average market-day holding periods for each of the 132 were 50.8 days, and their average daily gains produced an annual rate of gain (recognizing the 252-market-day year) of +41%.

Wow! That sounds pretty good. What's the trick? There must be one?

The NDX is now higher than its prior peak by more than 15%. A lot of these selected dates must have been in those market-depressed dates in early 2009. They should have generated some pretty fancy numbers by now from when the index was pushing its low in March. There ought to be lots of doubles in there.

Well, let's check that out. We know all the 132 buy dates and NDX prices then. We know what the Index is now. We can calculate a buy and hold annual return for each. On an average daily return basis we can get an annual rate for all of them to compare with the 41% bogey.

Surprise! Here's a picture of the annual rates of return for each of those 132 buy-and-holds. See how easy it is to forget what time does to rates of return. It's now over 3 ½ years after March, 2009.

(click to enlarge)

Not one of these even approaches the 41% bogey. Their combined average is +18%.

But wait. This analysis is only the return side. How about the risk involvement?

When we look at the selected dates and their test holding periods, the average maximum drawdown from cost was -6%.

Let's do the same for the buy-and holds. It turns out to be -6.5%, so we'll call it even.

So it turns out that in terms of what market-makers see now as prospects for future gains in the Nasdaq-100 Index, this may not be a bad time to make a generic bet, since the trend average growth in the index since the start of the century (2000) has been negative, and the past five years has been only + 3 ¾ % a year.

That's if you are a buy-and-holder and +18% a year looks good to you. The ETFs are (QQQ) for an equal-to-the-index tracker, (QLD) for a 2x leverage, and (TQQQ) for a 3x leverage. But leverage cuts both ways and makes less emotional sense for a buy-and-holder.

If you are an active investor, you know your game already, and +41% may not be competitive with the kinds of prospects indicated in our recent article.

Source: Can Market-Makers Help Us Time The Market? You Better Believe It