Why I Don’t Use Stop Orders
Thursday, April 25, 2013
Charles Rotblut, CFA
AAII Journal Editor

AAII Resources

Your Order Please
There are several types of buy and sell orders.

How Your Orders Get Filled
An explanation of how orders actually get filled.

AAII Discussion Boards
Do you use stop orders or price alerts?

Most Popular AAII Articles

  1. “The Liquidity Style: Finding Bargains by Seeking Less Popular Stocks”
  2. “Why We Don’t Rebalance”
  3. “A Pseudo-Life Annuity: Is There a Downside?”


Sentiment Survey

This week’s AAII Sentiment Survey results:
  Bullish: 28.3%, up 1.4 points
  Neutral: 32.9%, up 8.0 points
  Bearish: 38.8%, down 9.4 points

Long-term averages:
  Bullish: 39.0%
  Neutral: 30.5%
  Bearish: 30.5%

Take the AAII Sentiment Survey »




I don’t use stop orders. (Most commonly, these are standing orders to sell a stock if it falls below a specified price.) My avoidance of them has nothing to do with my investing style, but rather because of the influence that software code has on intraday moves of the market. Tuesday’s events give an example of why I think human intervention should remain a critical part of any buy or sell decision.

As you may have heard, hackers hijacked The Associated Press’ Twitter account. The hackers then sent out a tweet at 1:07 ET Tuesday afternoon reporting an explosion at the White House and an injury to President Barack Obama. The reaction in the market was swift, with the S&P 500 falling by nearly 1% over the next three minutes following the tweet.

Though the drop was temporary, it showed the impact that computers have on intraday volatility. Trading algorithms designed to act upon designated phrases kicked in once word of the tweet circulated. Likely, a combination of words similar to “Obama” and “injury” were encoded to trigger sell programs. The logic is that a trader or a fund will want to sell their stocks and potentially short the market if something bad were to happen to the president of the United States. Had the tweet not quickly proven to be false, it is entirely possible that other sell programs would have started in reaction to the market’s decline.

On the surface, there is merit for such programs. A trader (or a fund) following a short-term strategy will want to close out his positions should bad news hit the newswires. A computer program designed to identify specific events and sell in reaction to them can respond much quicker than a human ever could. The problem is that a software program does not know anything beyond what it is programmed to look for and react to. An algorithm’s entire world is a set of programmed formulas with instructions to do “X” if it identifies “Y.” If the goal of an algorithm is to sell quickly when bad news breaks, it seems unlikely that a trader would want precious time to be wasted on confirming a headline.

It is important to acknowledge the role these programs have, given how interconnected the world has become due to technology. A tweet from a well-known source will quickly be circulated on other sources. For instance, Bloomberg—a news service used by many trading firms—gives a constant stream of updated headlines. As anyone who has used a Bloomberg terminal can attest too, the headlines stream the moment news breaks. Though there are advantages to getting updated news as fast as possible, risks form when algorithms can act on the headlines without any human intervention.

Therein lies the problem. A human could have looked at the false AP tweet, realized how extraordinary it was and then sought confirmation from other sources. A human could also have taken into account the probability of the AP’s Twitter feed being hijacked. The New York Times said six prominent Twitter accounts have been hacked in recent months, including three-CBS affiliated twitter feeds last Saturday. Algorithms do not consider the broader picture the way humans do.

Though Tuesday’s decline in prices in reaction to the tweet was modest—quick, but modest—it could have caused stop orders to be triggered. Once a security falls below the specified stop (e.g., $25), a sell order is initiated. Because of the nature of these orders, a stock can be sold even if it drops one cent below the stop price and then quickly rebounds higher.

I should also point out that Tuesday’s intraday volatility occurred when the programs were functioning properly. As we all know, glitches can and do occur. When these glitches have occurred in the past, intraday volatility has increased.

I don’t have a problem with the premise of stop orders. They are a tool that can limit losses. In practice, however, quick, sharp market moves caused by algorithms can cause stop orders to be triggered. This in turn results in transactions an investor may have not otherwise wanted to occur. This is why I think it is better to be alerted if a certain price is reached than to use stop orders. When a price alert is issued, take a look at what is happening and then make the decision as to whether to react or not.

None of this means you should fear the trading algorithms. Rather, you should be cognizant of their role and be prepared to work around their shortcomings. Knowing these programs are prone to being reactive can help you better manage the decision to sell or not to sell. Computers are helpful, but rational human decisions are still important for achieving investment success.


More on AAII.com

The Week Ahead

More than 120 S&P 500 companies will report earnings next week. Included in this group are Dow components Pfizer (PFE) and Merck & Co. (MRK). The two will report on Tuesday and Wednesday, respectively.

The Federal Open Market Committee will hold a two-day meeting starting on Tuesday. The meeting statement will be published on Wednesday afternoon.

The week’s first economic reports of note will be March personal spending and income and the March pending home sales index, both of which will be released on Monday. Tuesday will feature the February Case-Shiller home price index, the April Chicago PMI and the Conference Board’s April consumer confidence index. The April ISM manufacturing index, the April ADP Employment Report and March construction spending will be released on Wednesday. Thursday will feature first-quarter productivity and March international trade data. April jobs data—including the unemployment rate and the change in the nonfarm payrolls, the April ISM non-manufacturing index and March factory orders—will be released on Friday.

Thursday marks the start of the worst six months for stocks. Since 1945, the S&P has averaged a 1.2% return between May and October, according to Sam Stovall at S&P Capital IQ. Sam further says stocks rise just 63% of the time during this period. (Only April through September has a lower frequency of six month gains, at 62%.) I don’t advocate selling in May since the market does tend to rise over the six-month period more often than not, but I do use the semiannual anniversary as a calendar trigger to see if my portfolio needs to be rebalanced.

AAII Sentiment Survey

Though bearish sentiment pulled back to within its normal range, individual investors remain pessimistic overall about the short-term direction of stock prices, according to the latest AAII Sentiment Survey.

Bullish sentiment, expectations that stock prices will rise over the next six months, rose 1.4 percentage points to 28.3%. Even with the increase, optimism remained below its historical average of 39% for the sixth consecutive week. This is the longest such streak since August 30, 2012, through November 22, 2012.

Neutral sentiment, expectations that stock prices will stay essentially unchanged over the next six months, jumped 8.0 percentage points to 32.9%. The historical average is 30.5%.

Bearish sentiment, expectations that stock prices will fall over the next six months, fell 9.4 percentage points to 38.8%. Since spiking to 54.5% on April 11, 2013, pessimism has pulled back by a cumulative 15.7 percentage points. Nonetheless, bearish sentiment remains above its historical average of 30.5%.

The improvement in bullish sentiment puts it right at one standard deviation below its historical average, or the tip of where we would consider it to be unusually low. The decline in bearish sentiment puts it back within the typical range of readings we have seen throughout the survey's history.

Though bearish sentiment declined for the second consecutive week, more individual investors remain pessimistic than optimistic about the short-term direction of stock prices. Concerns that stock prices have moved too far, too fast, the recent increase in market volatility, slow earnings growth, mixed views about the pace of economic growth and ongoing frustration with Washington are all playing a role.

This week’s special question asked AAII members how attractive U.S. stocks are relative to other asset classes, such as bonds, foreign stocks and precious metals. Slightly more than half of all respondents (52%) said U.S. stocks were the most attractive asset class. About 10% said foreign stocks are more attractive. Approximately one-quarter of respondents (26%) described U.S. stocks as overvalued, overbought or not attractive, however.

» Take the sentiment survey