Earlier this month I met Pedro Chan at his small apartment above an evangelical church in Brooklyn's Sunset Park neighborhood. Chan, who shares the place with three others, is short and muscular. He has a quiet voice and a patient demeanor that seems to have served him well on his journey to New York. In 2002, he left his Guatemalan village for a long trip through Mexico and, with the help of a smuggler, across the Texas border. In 2004, he made it to Brooklyn, where his uncle helped him find work on small construction crews.

These days, Chan helps skilled (and fully documented) carpenters, electricians and stucco installers do their jobs by carrying heavy things and cleaning the work site. For this, he earns up to $25,000 a year, which is considerably less than the average entry wage for New York City's 100,000 or so documented construction workers. Chan's boss, who spoke on the condition of anonymity, said that unless he learned a specialized skill, Chan would never be able to move up the income ladder. As long as there are thousands of undocumented workers competing for low-end jobs, salaries are more likely to fall than to rise.

As Congress debates the contours of immigration reform, many arguments have been made on economic grounds. Undocumented workers, some suggest, undercut wages and take jobs that would otherwise go to Americans. Worse, the argument goes, many use social programs, like hospitals and schools, that cost taxpayers and add to our $16 trillion national debt. Would deporting Pedro Chan and the other 11 million or so undocumented workers mean more jobs, lower taxes and a stronger economy?

Illegal immigration does have some undeniably negative economic effects. Similarly skilled native-born workers are faced with a choice of either accepting lower pay or not working in the field at all. Labor economists have concluded that undocumented workers have lowered the wages of U.S. adults without a high-school diploma — 25 million of them — by anywhere between 0.4 to 7.4 percent.

The impact on everyone else, though, is surprisingly positive. Giovanni Peri, an economist at the University of California, Davis, has written a series of influential papers comparing the labor markets in states with high immigration levels to those with low ones. He concluded that undocumented workers do not compete with skilled laborers — instead, they complement them. Economies, as Adam Smith argued in "Wealth of Nations," work best when workers become specialized and divide up tasks among themselves. Pedro Chan's ability to take care of routine tasks on a work site allows carpenters and electricians to focus on what they do best. In states with more undocumented immigrants, Peri said, skilled workers made more money and worked more hours; the economy's productivity grew. From 1990 to 2007, undocumented workers increased legal workers' pay in complementary jobs by up to 10 percent.

I saw this in action when Chan took me to his current work site, a two-story office building on Coney Island Avenue. The skilled workers had already installed wood flooring in a lawyer's office and were off to the next job site. That left Chan to clean up the debris and to install a new toilet. As I looked around, I could see how we were on one end of an economic chain reaction. Chan's boss no longer had to pay a highly skilled worker to perform basic tasks. That lowered the overall cost of construction, increasing the number of jobs the company could book, which meant more customers and more money. It reminded me of how so many restaurants operate. Without undocumented labor performing routine tasks, meals, which factor labor costs into the price, would be more expensive. There would also be fewer jobs for waiters and chefs.

Earlier that day, I was reminded of another seldom-discussed fact about immigrant life in the United States. Immigrants spend most of the money they make. Chan had broken down his monthly expenses: $400 a month in rent, another $30 or so for gas, electric and Internet. He sends some money home and tries to save a few thousand a year in his Citibank account, but he ends up spending more than $10,000 annually. That includes the $1,400 or so he pays the I.R.S. so that he can have a taxpayer I.D. number, which allows him to have a credit score so that he can rent an apartment or lease a car.