The NBER is a massive consumer of econometrics, so it needs at least a group or two devoted to producing econometrics. Hence I'm thrilled that the "Forecasting and Empirical Methods in Macroeconomics and Finance" group, now led by Allan Timmermann and Jonathan Wright, continues to thrive. Timmermann-Wright is strongly and appropriately time-series in flavor, focusing on developing econometric methods for macroeconomics, financial economics, and other areas that feature time series prominently.
In my view, there's a strong and obvious argument favoring creation of a second NBER working group in econometrics, focusing on micro-econometrics. Quite simply, econometiric methods are central to the NBER's mission, which has both macro/finance and micro components. Timmermann-Wright addresses the former, but there's still no explicit group addressing the latter. (The Bureau's wonderful and recently-instituted Econometrics Methods Lectures include micro, but the Methods Lectures are surveys/tutorials and hence fill a very different void.) An ongoing working group led by Chernozhukov-Imbens-Wooldridge (for example -- I'm just making this up) would be a fine addition and would nicely complement Timmermann-Wright.