Are Macroeconomic Forecasts Informative? Cointegration Evidence from the Asa-nber Surveys

This paper assesses the time series properties of the ASA-NBER forecasts for several US macroeconomic variables. Specifically, we examine whether forecasts exhibit a weak form of rationality which fulfills the following conditions: (i) are the actual and forecast series integrated of the same order; (ii) are they cointegrated, and; (iii) is the cointegrating vector consistent with long run unitary elasticity of expectations with respect to the actual series. We term the condition wherein all three characteristics hold as "consistency". We also examine whether forecasts respond to disequilibria, such as estimated error correction terms. These tests are applied to both the final (revised) and the preliminary (unrevised) versions of the actual data. The findings are as follows. The (final) Treasury bill rate, housing starts, industrial production, inflation and most of their respective forecasts appear to be trend stationary. The corporate bond rate, GNP, the GNP deflator, unemployment and most of their respective forecasts appear to be difference stationary. About half of the unit root pairs are cointegrated. In only one of these cases is the unitary elasticity restriction rejected: the 1-quarter ahead GNP deflator forecast. The forecasts appear to behave in a reasonable manner; they respond to disequilibria (defined by estimated cointegrating vectors). This finding is robust to the use of an imposed (-1 1) cointegrating vector, rather than an estimated one. Similar results are obtained when using the originally-reported data (unrevised). However, while there is more evidence of cointegration, there is also a greater rate of rejection of the unitary elasticity restriction. We also find that as the forecast horizon increases, the importance of the long run information also rises relative to the information contained in the short run dynamics. We conclude, in line with much recent research, that survey data do fulfill some weak conditions for rationality. Nonetheless, very few forecasts pass all three of our criteria.