Economics is a very interesting subject. The scope of economic domain is vast. Economics deals with market structure, consumer behavior, investment, growth, fiscal policy, monetary policy, the roles of the bank, etc. The list can go on for quite some time. It also predicts how economic agents behave in response to changes in economic and noneconomic factors such as price, income, political party, stability, and so on. The economic theory, however, is not specific. For example, the theory proves that when the price of a good increases the quantity supplied increases, provided all the other pertinent factors remain constant, which is also known as ceteris paribus. What the theory does not and cannot state is how much the quantity increases for a given increase in price. The answer to this question seems to be more interesting to most people than the fact that the quantity will increase as a result of an increase in price. The truth is that the theory that explains the above relationship is important for economists. For the rest of the population, the knowledge of that relationship is worthless if the magnitude is unknown. Assume for 10% increase in price the quantity increases by 1%. This has many different consequences than if the quantity increases by 10%, and totally different consequences if the quantity increases by 20%. The knowledge of the magnitude of change is as important, if not more important, than the knowledge of the direction of change. In other words, predictions are valuable when they are specific.
Statistics is the science that can answer specific issues raised above. The science of statistics provides the necessary theories that can provide the foundation for answering such specific questions. Statistics theory indicates the necessary conditions to set up the study and collect data. It provides the means to analyze and clarify the meaning of the findings. It also provides the foundation to explain the meaning of the finding using statistical inference.
In order to be able to make an economic decision, it is necessary to know the economic conditions. This is true for all economic agents, from the smallest to the largest. The smallest economic agent might be an individual with little earning and disposable income, while the largest can be a multinational corporation with thousands of employees, not to mention governments. Briefly, we will discuss some of the main needs and uses of statistics in economics and then present some uses of regression analysis in economics as well.
The first step in making any economic decision is to gain knowledge of the state of economy. Economic condition is always in a state of flux. Sometimes it seems that we are not very concerned with mundane economic basics. For example, we may not try to forecast what the price of a loaf of bread is or a pound of meat. We know the average prices for these items; we consume them on a regular basis and will continue doing so as long as nothing drastic happens. However, if you were to buy a new car you would most likely call around and check some showrooms to learn about available features and prices because we tend not to have up-to-date information on big-ticket items or goods and services that we do not purchase regularly. The process described above is a kind of sampling, and the information that you obtain is called sample statistics, which you use to make an informed decision about the average price of an automobile. When the process is performed according to restrict and formal statistical methods, it is called statistical inference. The specific sample statistics is called sample mean. Mean is one of numerous statistical measures at the disposal of modern economists. Another useful measure is the median. The median is a value that divides observations into two equal halves, one with values less than the median and the other with values more than median. Statistics explains when each measure should be used and what determines which one is the appropriate measure. Median is the appropriate measure when dealing with home prices or income. Applications of statistical analysis in economics are vast, and sometimes they reach to other disciplines that need economics for assistance. For example, when we need to build a bridge to meet economic, social, and even cultural needs of a community, it is important to find a reliable estimate of the necessary capacity of the bridge. Statistics indicates the appropriate measure to be used by teaching us whether we should use the median or the mode. It also provides insight on the role that variance plays in this problem. In addition to identifying the appropriate tools for the task on hand, statistics also provides the methods of obtaining suitable data and procedure for performing analysis to deliver the necessary inference.
One cannot imagine an economic problem that does not depend on statistical analysis. Every year, the Government Printing Office compiles the Economic Report of the President. Although the majority of the statistics in the report are fact-based information about different aspects of economics, many of the statistics are based on some statistical analysis, albeit descriptive statistics. Descriptive statistics provides simple yet powerful insight to economic agents and enable them to make more informed decisions.
Another component of statistical analysis is inferential statistics. Inferential statistics allows the economist and political leaders to test hypotheses about economic condition. For example, in the presence of inflation, the Federal Reserve Board of Governors may choose to reduce money supply to cool down the economy and slow down the pace of inflation. The knowledge of how much to reduce the supply of money is not only based on economic theory, but also depends on proper estimation of the final outcome.
Another widely used application of statistical analysis is in policy decision. We hear a lot about the erosion of the middle class or that the middle class pays a larger percentage of its income in taxes than the lower and upper classes. However, how do we know who is the middle class. A set dollar amount of income would be inadequate because of inflation, although, we must admit even a single dollar amount must also be obtained using statistics. However, statistical analysis has a much more meaningful and more elegant solution. The concept of interquartile range identifies the middle 50% of the population or income. Although interquartile range was not designed to identify the middle 50% and is not explained in these terms, the combination of economics and statistics is used to identify the middle 50% for economics and policy decision purposes.
The knowledge of statistics can also help to identify and comprehend daily news and events. Recently, a report indicated that the chance of accident for teenage drivers increases by 40% when there are passengers in the car that are under 21 years of age. This is a meaningless report. Few teenagers drive alone or have passengers over 21 years of age. Total miles driven by teenagers when there passengers under 21 years of age far exceeds any other types of teenage driving. Other things equal, the more you drive, the higher the probability of an accident. This example indicates that the knowledge of statistics is helpful in understanding everyday events and in making sound analysis.
When an economic phenomenon is changed to produce a desirable income, we need more powerful tools than simple statistics. Regression analysis is one of the most widely used statistical tools at the disposal of economists.
In regression analysis, the effect of one or more factor is measured to determine another factor. The first group is also known as explanatory variables, while the latter is known as endogenous variables. In economics it makes sense to refer to explanatory variables as policy instruments. Policy instruments are variables that economists and policy makers can change or control. The supply of money is a policy instrument controlled by the Federal Reserve. The Fed has to collect data first, which is done on a periodic basis. These statistics inform the Fed that there is a problem in the economy, such as inflation. The Fed decides to reduce the supply of money. It will wait for the economy to respond to the change in supply of money. Then economic indicators are measured again and tested against the target set by the policy. If the policy objectives are not met, the action is repeated until the desirable outcome is obtained.
When working with a regression model, one might wonder if it was designed to serve economists. Even some of the commonly used terminologies are the same in both fields. For example, both subjects use explanatory variables to measure the response variable. Typical regression models do not consist of one explanatory variable and one response variable. Instead, in addition to explanatory variables, the model has additional variables known as control variables. Control variables are actually the same thing as economics shifters. Shifters in economics refer to variables that are assumed to remain constant for the sake of identifying the impact of the explanatory variables on the response variables. In fact, every economic theory seems to have the famous ceteris paribus, which means other things being equal. When other things are not equal and change, they do not distort the relationship between explanatory and response variables. They simply shift the magnitude up or down, depending on the direction of the impact. Estimation of demand provides a good example. Economic theory states that an increase in price reduces the quantity demanded, ceteris paribus. The regression model for this economic theory can be written as
Qd = β0 + β1P + ε(I.1)
where ε is the error term, which will be explained later. To complete the process, we need to test the hypothesis that the coefficient of price, which is also the slope of the demand curve, is negative. So we use statistics to test the following hypothesis:
H0: β1 = 0 H1: β1 < 0
The model, however, is not complete, because it is not subject to ceteris paribus as it does not control anything. Simple control variables consist of price of a complementary good, a substitute good, and income, to name just a few important ones. The theory predicts that the effect of a change in the price of a complementary good is inverse, the effect of a change in the price of a substitute good is direct, and the effect of change in income is direct. Thus, model (I.1) should be modified as below.
Qd = β0 + β1P + β2 Pc + β3 Ps + β4 Y + + ε,(I.2)
The theoretical claims are written as
H0 : β1 = 0 H1 : β1 < 0
H0 : β2 = 0 H1 : β2 > 0
H0 : β3 = 0 H1 : β3 < 0,
where the subscripts use the first letters of complementary and substitute, and Y represents income. The regression model clearly and perfectly matches the economic theory from expected effects of each variable to the concept of ceteris paribus.