@article{Saving-Investment:4238,
      recid = {4238},
      author = {Cuddington, John},
      title = {The Saving-Investment Approach to the Current Account},
      publisher = {Bank of Canada},
      address = {1982},
      pages = {1 online resource (vi, 25 pages)},
      abstract = {This report summarizes the results of an investigation of  the potential usefulness of focusing on the gap between  national saving and investment when analyzing the current  account. From a theoretical viewpoint, the conclusion is  that such an approach for short-run analysis is likely to  be extremely misleading because it does not give sufficient  emphasis to the income determination process over the  course of the business cycle. For intermediate and long-run  analyses, economic theory suggests the importance of  interest rate effects on saving and investment and the  stock-flow interactions between the capital stock and  capital investment and between wealth and saving.  Notwithstanding these theoretical considerations, the  attempt to isolate statistically significant interest rate  and stock accumulation effects in aggregate saving and  investment equations was quite unrewarding. It appears that  the saving-investment approach can provide little  assistance in improving forecasts of the current account  over the short run. In part this result may be due to the  quality of data series available on financial wealth and  capital stock variables. In part it may reflect a need to  apply more sophisticated theories of consumption and  investment behaviour based on intertemporal optimization.  For example, it may be necessary to distinguish between  permanent and temporary income and to allow explicitly for  interest rate shocks in order to relate current account  movements to saving and investment in a satisfactory  manner. Of course, one may still want to impose theoretical  properties suggested by the approach in intermediate-run  models. It is my view, however, that model builders will  have to be willing to impose a large amount of a priori  theoretical knowledge to obtain satisfactory results for  estimated coefficients.},
      url = {http://www.oar-rao.bank-banque-canada.ca/record/4238},
      doi = {https://doi.org/10.34989/tr-32},
}