您在這裡

Variables, data and index numbers

16 十二月, 2015 - 15:49

Economic theories and models are concerned with economic variables. Variables are measures that can take on different sizes. The interest rate on a student loan, for example, is a variable with a certain value at a point in time but perhaps a different value at an earlier or later date. Economic theories and models explain the causal relationships between variables.

\mid  Variables: measures that can take on different values.

Data are the recorded values of variables. Sets of data provide specific values for the variables we want to study and analyze. Knowing that the Don Valley Parkway is congested does not tell us how slow our trip to downtown Toronto will be. To choose the best route downtown we need to ascertain the degree of congestion—the data on traffic density and flow on alternative routes. A model is useful because it defines the variables that are most important and to the analysis of travel time and the data that are required for that analysis.

\mid  Data: recorded values of variables.

Sets of data also help us to test our models or theories, but first we need to pay attention to the economic logic involved in observations and modelling. For example, if sunspots or baggy pants were found to be correlated with economic expansion, would we consider these events a coincidence or a key to understanding economic growth? The observation is based on facts or data but it need not have any economic content. The economist’s task is to distinguish between coincidence and economic causation.

While the more frequent wearing of loose clothing in the past may have been associated with economic growth because they both occurred at the same time (correlation), one could not argue on a logical basis that this behaviour causes good economic times. Therefore, the past association of these variables should be considered as no more than a coincidence.

Once specified on the basis of economic logic, a model must be tested to determine its usefulness in explaining observed economic events. The earlier example of a model of house prices and mortgage rates was based on the economics of the effect of financing cost on expenditure and prices. But we did not test that model by confronting it with the data. It may be that effects of mortgages rates are insignificant compared to other influences on house prices.