Internet Access Increases Stock Trading

02/01/2001
Summary of working paper 7878
Featured in print Digest

Although the introduction of Web trading seems to increase total trading, it does not appear to increase short-term trading or last-hour trading, suggesting that the Web may not be an important catalyst for speculative trading.

The growing availability of online trading technologies is believed to have had a large effect on stock market volume and volatility. But is the increase in trading coincidental to the spread of the Web or a direct result of it? It may simply be that online traders would have executed their trades on other channels if the Web had not been available. In Does the Internet Increase Trading? Evidence from Investor Behavior in 401(k) Plans (NBER Working Paper No. 7878), James J. Choi, David Laibson, and Andrew Metrick argue that introduction of a Web-based trading channel in a 401(k) plan does cause dramatic increases in the number and dollar volume of total trades.

The authors use trading data from the participants of two large corporate 401(k) plans. Both plans opened a Web trading channel in 1998, adding to pre-existing phone trading. Within18 months of the initiation of Web trading, Web transactions had grown to approximately 60 percent of all transactions. The total trading rate of participants - adding up phone and Web channels -- had quadrupled from its pre-Web level.

Even after the authors control for factors that might affect trading volume, such as stock price volatility or increased overall market volume, the Web effect is very large: daily trading frequency nearly doubles (a 96 percent rise) and daily turnover -- the fraction of balances traded -- increases by 55 percent. The transaction frequency effect is greater than the turnover effect because Web access lowers average transaction size and increases Web trading on low-balance accounts belonging to young traders.

The authors find that young, male traders with high salaries and plan balances are most likely to try the Web, which is consistent with societal stereotypes of Web traders. Retired and job-terminated participants are less likely to trade using the Web. The authors speculate that since ex-employees are out of the workplace "loop" they are less likely to know about plan changes like the new Web-based trading opportunities. Participants who traded frequently by phone prior to the introduction of Web trading are also less likely to try the Web.

Choi, Laibson, and Metrick find that most participants who try the Web stick with it. Of those participants who try the web, 88 percent make their next trade on the web. Conditional on a first and second Web trade, 94 percent make their third trade online, and 96 percent of Web-Web-Web traders make their fourth trade online.

Of those participants who try the web and revert back to the phone for the next trade, about half continue using the phone. Of the Web-Phone traders, 57 percent make their third trade by phone, and 77 percent of Web-Phone-Phone traders make their fourth trade by phone.

The lower effort cost associated with Web transactions leads traders to execute smaller dollar value transactions in their Web trades relative to phone trades. For example, small trades that would not be worth a time-consuming phone execution, would be worth executing quickly on a Web-based trading system. The average phone transaction in this data is 75 percent larger than the average Web transaction --- for example, respective transaction averages of $70,000 (phone) versus $40,000 (Web) for one of the two firms and $105,000 versus $60,000 for the other firm. These differences also are driven by demographic contrasts between the typical phone and Web traders. While high balance plan participants are most likely to try the Web, low balance participants are most likely to trade frequently on the Web. Hence, low balance participants conduct a relatively large share of Web transactions.

Finally, the authors find that despite popular reports about excessive internet-based day trading, the availability of a Web-based trading channel does not increase short-term trading. They define short-term trades as trades that are "reversed" within five days of the original trade.

The authors also analyze "last-hour trades": trading orders that are placed in the hour before the market closes. All 401(k) trades are executed at closing prices, so last-hour trading is the only way for a 401(k) trader to trade with up-to-date information. Naturally, short-term market-timers would be disproportionately likely to be last-hour traders. The authors find that a greater proportion of phone trades are last-hour traders and that the introduction of Web trading does not increase the number of last-hour traders. Although the introduction of Web trading seems to increase total trading, it does not appear to increase short-term trading or last-hour trading, suggesting that the Web may not be an important catalyst for speculative trading.

-- Noshua Watson