In this New York Times article she makes a good argument that the structural unemployment rate has not increased. If the unemployment rate stays high for a long time then the story changes - some people could lose their skills and become long-term, structurally unemployed. Some economists have argued that there already has been an increase in the number of long-term, structurally unemployed due to problems in the housing and financial sectors.
Romer points out that there were only 1.3 million people working in those sectors before the crisis. Even if none of them ever work again (but keep looking so they still count as unemployed), there are too few of them to have a big impact on the national unemployment rate. This ignores the possibility that other people who made stuff for new houses (like refrigerator manufacturers or landscapers) might also be permanently out of work.
For me, her most convincing point is that people who lost jobs in the housing and finance sectors are able to find new jobs at about the same rate as people who lost jobs in other sectors. In other words, people who worked in housing and finance are able to switch sectors - they are not much more likely to stay unemployed than anybody else. Once the recovery is fully underway they, and the rest of cyclically unemployed, should be able to get new jobs.
For more on this, and the effect of underwater homes see:
NYT (4-10-2011): Jobless Rate is Not the New Normal by Christina Romer
Jan Hatzius, Chief Economist for Goldman Sachs, agrees the natural rate (= structural + frictional) has not increased much and is probably now about 5.5%. (from 9-1-2010 interview on CNBC, 5 minutes)
Crash course on structural, cyclical and frictional unemployment (2 minutes)