A new analysis by the Government Accountability Office finds that in 2010 large U.S. corporations paid an average effective tax rate on their worldwide income of 22.7 percent and U.S. federal tax of only about 16.6 percent.  The federal rate was less than half of the 35 percent statutory rate. Large firms that made a profit that year paid an even lower effective rate—an average of 16.9 percent in worldwide taxes and only 12.6 percent in U.S. federal tax.   It isn’t news that these firms pay low effective tax rates. Journalists and academic researchers have been reporting this for years. But for the most part they have been hamstrung by limited access to good data. Publically available financial statements include information about taxes, but there is often a huge gap between what firms report on their public books and what they actually pay. And when asked about their true tax liability, firms mostly respond with a curt, “none of your business.” But GAO has access to new IRS data from a form called the M-3. That schedule requires firms with assets of $10 million or more to reconcile their book and tax reporting in a way that gives a much clearer picture of their actual tax liability. These data are still not ideal. For one thing, GAO only had access to aggregate data, not firm-level information. And there are some big differences hidden behind those averages. A company’s effective tax rate can vary dramatically, depending among other things on what investments it makes in plant and equipment and how it finances those purchases. GAO was not only unable to look at individual firms, it could not even compare different types of firms, such as U.S. controlled corporations or foreign controlled corporations. The second problem, which will go away with time, is that M-3s are only available for 2009 and 2010, two years when corporate profits were extremely low (or non-existent).  It will be interesting to see whether the pattern of tax liability changed as companies returned to profitability. Finally, as my Tax Policy Center colleague Donald Marron points out, the GAO uses a worldwide rate that includes state and local taxes. Thus, a fair comparison should not be against the 35 percent federal rate alone but against the combined federal/state rate, which averages closer to about 39 percent.    The new study doesn’t tell us what the “right” corporate rate ought to be.  Neither does it tell us anything about marginal tax rates, which are what matter to corporate decision-makers. But it does give us important new information. And it is further evidence that some large U.S. companies have figured out how to make the crazy U.S. corporate tax system--and the even-crazier international system—work for them. They’ve learned to manipulate income and expenses in ways that minimize their taxes, not only in the U.S. but everywhere they do business. Lawmakers talk about corporate tax reform that would eliminate many business preferences in return for reducing the top rate from the current 35 percent to somewhere in the high 20s. But why should firms paying rates that average 17 percent give up preferences in exchange for a 28 percent rate? For many, it would still be a tax increase.   Sure, the current system makes them do strange things—such as Apple borrowing $17 billion to help pay a dividend instead of repatriating (and paying tax on) its own foreign earnings. But, for many large firms, it works. Why would they change it?