Economic statistics is composed of two interrelated fields, those related to data collection and those to data analysis. In fact, economic statistics is differentiated from other fields of applied statistics due to its unique data collection methods, and because of the scope/scale of analysis.
Most economic data is collected by governmental or large-scale pseudo governmental agencies. These include the United Nations, the World Bank, and International Monetary Fund (IMF), and the various regional development banks. These are often compilations of data provided by the various countries’ central banks. By its nature, macro level data is nearly impossible for individual researchers to collect. However, the governmental provision of economic data is increasingly extended to microeconomic data. In the United States, for example, the most comprehensive individual, or micro level, data is compiled by the Census Bureau and the Bureau of Labor Statistics.
Economic statistics have traditionally been centered upon directly measurable concepts, or to concepts that are potentially well defined. For example, there is less ambiguity in the definition or proper measurement of “income” than there is in defining the concept of “happiness.” For this reason, the statistical problems implied by large measurement errors require less attention in economics than in fields such as marketing or psychology.
Many economic phenomena can be measured in different ways; in fact, they are often defined in different ways as well. Some countries, for example, compute the inflation rate by adjusting each good’s price by a quality improvement factor; others do not. For this reason, several agencies and companies have specialized in producing datasets that are internationally comparable. These include the Penn World Tables, the International Financial Statistics database compiled by the IMF, much of the data reported by the Organisation for Economic Co-operation and Development, and several databases compiled by the United Nations.
Econometrics is the application of statistical techniques to the analysis of economic data and their interrelationships. Physical scientists, and some social scientists, can often rely upon carefully crafted, controlled experiments with which to collect data and test competing theories. Because large-scale controlled experiments are not feasible on a national level, econometrics required a unique set of tools.
In earlier statistical research using the regression methodology, the role of regression was to estimate the correlation between an exogenous variable X on an endogenous variable Y, while holding the other exogenous variables constant. This is accomplished in a statistical sense, since controlled experiments are rare in economics. This is done simultaneously for many endogenous variables within a single equation: Y is a function of Xs.
The concept of General Equilibrium, however, required that the economic variables of interest are jointly determined. That is, X causes Y, but Y also causes X. This phenomenon is often termed the “endogeneity problem.” Since all economic data in the United States, for example, are determined within the same national economy, the analysis is significantly complicated. Such endogeneity is, in fact, the cornerstone of economics, as embodied in the Supply and Demand graphs, a system of two, not one, equations. Market prices and quantities are determined by the interaction (indeed, intersection) of supply and demand. Thus, one cannot hold price constant in order to isolate the effects of X on quantity.
Econometrics, as a unique field of study, arguably began with the formation in the early 1930s of the Cowles Commission, the Econometric Society and their journal Econometrica, and in the 1940s with the Department of Applied Economics at Cambridge. As fitting the world’s preoccupation with the global macroeconomic problems of the time, economic theory and economic statistics became understandably macro-oriented. Especially at Cowles, the aim was to study systems of equations much larger than the simple two-equation supply-and-demand system. Rather, dozens of such systems were incorporated into large-scale models of economies, with each sub-market influencing and being influenced by all other markets. During this time, the mainstream economic school of thought was Keynesianism, according to which there is a large role for government in controlling the economy. Thus, measurement and analysis were prerequisites to control. The Keynesian macro econometricians sought to estimate the parameters of their many economic equations. These parameters were thought to be constants just as there are physical constants in the hard sciences. Once all of the economies’ parameters were estimated, fine-tuned economic prediction and control could be exercised.
Governmental institutions became engaged in developing truly massive systems of hundreds of equations with which to model their home economies; this, in an attempt to predict the likely outcome of proposed economic policies. This method of analysis remained the dominant technique until the 1970s, when very simple time-series models were found to outperform their large-scale brethren. These simple models, developed largely by George Box and Gwilym Jenkins, were usually univariate time-series models which leveraged the inertia in economies by putting lagged dependent variables as the key terms used for prediction.
In 1976, an influential paper by Nobel Prize–winning economist Robert Lucas introduced what is now known as the “Lucas Critique.” This critique, in effect, pulled the theoretical rug out from under large-scale econometric modeling. Lucas argued that even the estimated parameters were the results of the economic process; the parameters were not unchanging and structural, they were also endogenous. From that point onward the systems approach has been largely abandoned in favor of a return to single-equation models, though these are considerably more complex than the univariate time-series models of Box and Jenkins. (Interestingly, this occurred at largely the same time that other social sciences turned from single-equation models to multiple “structural equations” models.)
Econometrics is at the intersection of economic theory and economic data, where the priority of one over the other remains in dispute. For some economists, the primary role of econometrics is to test the validity of economic theories. In the physical sciences, where theory is well established, the functional forms of the equations to be estimated are well defined. These well-defined forms have not been found in the social sciences. For many econometricians, proper practice requires developing a formal model with micro foundations (utility functions, production functions, etc.) as a necessary step prior to estimation. If one theory, for example, maintains that there is a positive relationship between X and Y, but it is estimated that the relationship is negative, then it can be argued that the theory has been falsified. On the other hand, a completely different conclusion can be drawn.
There is often little testing that can be done regarding whether the equation that is estimated is properly specified in the first place. Thus, many researchers advocate using economic theory as a guide to model selection. In this vein, if an equation is estimated and it is found that there is a negative relationship between X and Y, this has not falsified the theory, but rather, it has called into question the equation that was said to represent the theory. Thus, a competing use of econometrics is the illustration, not testing, of economic theory. These researchers adopt a more intuitive approach to model selection. Finally, adherents to Chris Sims’ theory-free approach eschew theory altogether, preferring the “data to speak for themselves.” Sims’ approach recognizes the endogeneity of all economic variables, and estimates all of their interrelationships, without placing restrictions on what these relationships would be (regardless of what economic theory may imply).
While much of this entry has been devoted to macro econometrics, this is not to say that micro level econometrics was not practiced throughout this time. However, most of this data was at the industry level, and so the data were necessarily aggregated to some extent. Increasingly, truly micro level data—that is, data collected at the individual level—are being examined. In the United States, for example, popular micro level datasets are collected by the Bureau of Labor Statistics and the Census Bureau. Moreover, under the guidance of Vernon Smith, experimental economics has established controlled experiments as a valid means of collecting microeconomic data. Increasingly, economists have become freed of the governmental macro level databases, and have begun generating their own micro level data, tailored to their own research needs.
From the 1930s to the present, econometrics has been shedding its macroeconomic roots, and is largely indistinguishable from the other branches of applied statistics that use the regression approach. The differences lie in the questions that are asked, not in the techniques they use to answer these questions.
- John Abowd and Lars Vilhuber, “The Sensitivity of Economic Statistics to Coding Errors in Personal Identifiers,” Journal of Business and Economic Statistics (v.23/2, 2005);
- Bernard Baumohl, The Secrets of Economic Indicators: Hidden Clues to Future Economic Trends and Investment Opportunities (Wharton School, 2008);
- George E. P. Box and G. M. Jenkins, Time Series Analysis: Forecasting and Control (Holden-Day, 1970);
- Adrian Darnell and J. Lynne Evans, The Limits of Econometrics (Edward Elgar, 1990);
- Norman Frumkin, Guide to Economic Indicators (M.E. Sharpe, 2006);
- David F. Hendry, Econometrics: Alchemy or Science?: Essays in Econometric Methodology (Blackwell, 1993);
- Robert Lucas, “Econometric Policy Evaluation: A Critique.” Carnegie-Rochester Conference Series on Public Policy (n.1, 1976);
- Edward F. McKelvey, Understanding US Economic Statistics (Goldman Sachs Economic Research Group, 2008);
- Chris A. Sims, “Macroeconomics and Reality” Econometrica (v.48, 1980).
This example Economic Statistics Essay is published for educational and informational purposes only. If you need a custom essay or research paper on this topic please use our writing services. EssayEmpire.com offers reliable custom essay writing services that can help you to receive high grades and impress your professors with the quality of each essay or research paper you hand in.