Malamud and Pop-Eleches chose the Euro 200 program because it solved the apples-to-oranges problem. While Euro 200 didn't exactly hand out computers at random, it came pretty close. The program provided vouchers worth 200 euros (about $240 at the time, nearly $315 today) for computer purchases by poor families with kids. (The income cutoff was $50 per month per household member.) But there weren't nearly enough vouchers to go around. In 2005, for example, nearly 52,000 qualified families applied, but the government had funding for only 27,555 of them. As a result, vouchers were given only to families with incomes below $17 per household member. This means that some of the families that got vouchers—those with, say, incomes between $16 and $17—were basically identical to some of those that didn't (families with $17-$18 incomes). These families all have similar computing aspirations (they all applied to the program) and differ only in which side of the $17 cutoff they happened to sit on. (Economists call this a "regression discontinuity.")
So what happens when good fortune delivers vouchers (and hence computers) into the homes of Romanian youths? Obviously a lot more time logged on to a computer—about seven hours more per week for vouchered versus unvouchered kids. Much of this computer time came at the expense of television-watching: Children in families that received a voucher spent 3.5 fewer hours in front of the tube per week. But computer use also crowded out homework (2.3 hours less per week), reading, and sleep. Less schoolwork translated into lower grades at school—vouchered kids' GPAs were 0.36 grade points lower than their nonvouchered counterparts—and also lower aspirations for higher education. Vouchered kids were 13 percentage points less likely to report an intention to attend college. And, interestingly, vouchered students who were college-bound were not more likely to express interest in majoring in computer science.
When my friends and I figured out how to transform my PET from a learning tool to a proto-video-game console, my parents stepped in to make sure Space Invaders didn't crowd out homework. Where were Romania's parents? The voucher program was specifically designed to help poor households, and their dire financial circumstances meant that these families were probably less able to afford after-school care or otherwise see to it that the computers were used for learning and not just recreation. Indeed, the authors found that when they looked specifically at families with stay-at-home moms who may be more present and able to police computer use, the negative effects of vouchers were greatly reduced.
Perhaps not surprisingly, the lesson from Romania's voucher experiment is not that computers aren't useful learning tools, but that their usefulness relies on parents being around to assure they don't simply become a very tempting distraction from the unpleasantness of trigonometry homework. But this is a crucial insight for those tasked with designing policies to bridge the digital divide. The express intent of Euro 200 was to give a boost to poor kids' educations. Through programs such as One Laptop per Child, governments around the world have similarly committed to purchasing millions of computers to improve computer access for children. But Malamud and Pop-Eleches' results suggest that merely providing access may be more of a curse than a blessing. If we really want to help poor kids, whether in Romania, sub-Saharan Africa, or America's housing projects, we may want to focus on approaches that provide structured, supervised access through after-school programs or subsidies that bring technology into low-income schools. But just giving kids computers? Might as well just ship them PlayStations.