All Terrain Thinking

A Compendium of things I think are Important

Economics: It's not just whats' in your wallet

The first question that must be addressed is the appropriate time period for any analysis. To see the impact of this, let's return to our example of Slippery Slope University where we are asked to evaluate the President on his ability to raise the school's revenue since he took office in 1991. The revenue data for school during his tenure are provided below. Would you vote for an extension of the president's contract based on his success in achieving the revenue goals?

As we found out earlier, this is a surprisingly difficult question given the little bit of data which we have. On the surface, it would appear that he is vulnerable to the critics who point out that in the six years since he took office, revenues have increased a paltry 1% (101 from 100).

The President points out, however, that he inherited the problems of the previous president and that he has been able to turn things around since 1992, his first full year. To support his case he has provided the revenue data for the five years before he took office. It is very clear from these revenue data that the University's revenue was in decline for this 5 year period and that the decline in 1992 could be viewed as a continuation of this trend.

How much does the timing issue matter? If we use the 1991-1996 period we have a revenue increase of approximately 1%, but if we accept the President's view and use only the 1992-1996 period we see that revenue has actually increased 12 percent (101 from 90). As you can see, one's interpretation of the president's performance is dependent upon the time-frame one uses.

For some additional examples of the importance of timing we will turn our attention to some time-series graphs, graphs that measure time on the horizontal axis and the variable being observed on the vertical axis. These are very popular graphs in economics and finance and to extract the information embedded in any time-series graph, you should begin by conceptually decomposing the graph into four separate components: long-term trends, short-term cyclical movements, seasonal patterns, and unexplained fluctuations called noise. For example, let's look at forecasting retail sales in Rhode Island for Wal-Mart. Under the heading of long-term factors would be population which tends to change rather slowly. We would expect that population growth in the State would tend to push sales up.

In any forecast we would also need to account for changes in the financial condition of the people. We would expect sales to be affected by any substantial change in the number of people unemployed. If the economy fell into a recession the rise in the number of unemployed would probably translate into lower sales as these unemployed cut back on their purchases-an example of a cyclical effect. And the there is the seasonal effect which anyone in retail knows. You could not forecast sales for any month without knowing the month. In Newport, a popular summer tourist, your sales forecast is likely to be significantly higher in July then in February, while at the Warwick Mall, you could expect sales to be highest in December.

When decomposing changes in sales of widgets, economists' favorite hypothetical good, it is easiest if you begin with the trend. This would usually be reflected in the 'average slope' which can be determined by the unsophisticated eyeball approach or the sophisticated regression approach [we will talk about this later]. Does the curve slope up, does it slope down, is there a definite change in the direction or the magnitude of the slope over time? These are the questions that you must attempt to answer at the outset of your analysis. Comments such as wages have grown on average approximately 4 percent per year for the past two decades or that labor productivity rose at a rate of approximately 2 percent per year during the post W.W.II period refer to trends. In the graph below, the underlying trend appears to be positive throughout the entire time under review.

It is clear from this graph, however, that there are also significant variations about this trend, variations which are referred to as cycles, business cycles to be more precise. These business cycles are a pervasive feature of capitalist economies. The reoccurring pattern of recessions (R) followed by expansions (E) characterizes nearly all measures of economic performance in nearly all economies. If you have any doubts, just look at the graphs of inflation, unemployment, interest rates, budget deficits, trade deficits, and exchange rates which you will find in your macroeconomics texts. You will see that the peaks and troughs of the various time-series graphs often tend to coincide with each other

What this means to you, is that care must be taken when comparing the performance of any variable at two points in time. For example, you have certainly heard many times about how wages increased 4 percent per year between 1980 and 1989, the unemployment rate in the US decreased 1.6 percentage points between 1980 and 1990, or that interest rates on government securities averaged 8.8 percent in the 1980s. The problem is that the changes are very sensitive to the selection of the beginning and ending time period which are controlled by the person providing the information.

To see what I am concerned about, consider the graph below in which I have constructed a hypothetical time-series graph. In the graph I have chosen two time periods to analyze the growth rate in Y. The first period begins at T0, a peak in the business cycle, and ends at T1, a trough in the business cycle. The second time period begins at T0', a trough in the business cycle, and ends at T1', a peak in the business cycle. It is clear that the rate of growth, as measured by the slope of the line linking the beginning and ending points, is extremely sensitive to the time period chosen. In this specific example, we have an underestimate of the growth rate in the first case, while in the second case we have an inflated view of the growth rate.

One real world example of the importance of this would be the poverty statistics. As you would expect, poverty tends to increase when the economy falls into a recession. As George Bush was running for reelection in 1992, the poverty statistics during his time in office showed a marked increase which was interpreted by many as evidence of increasing poverty resulting from Republican's lack of concern for the poor. In fact, much of the increase could be accounted for simply by the fact that the economy had moved from its expansion phase to its recession phase and many newly unemployed joined the ranks of the poor. This was a case where it would be important, although difficult, to separate out the cyclical from the trend components of change.

The selection of the time horizon can also be valuable when you are attempting to forecast. Consider the accompanying graphs on short-term interest rates on government securities. When looking at the first graph of interest rates in the 1980's, one is left with the impression that interest rates in 1989 are low and that we might expect them to rise as they return to more 'normal' levels.

In the graph with the longer time horizon, however, one gets a very different picture of the situation in 1989. Given the longer time horizon, one gets a sense that the current rates are high by historic standards and that they have considerable room to fall further.

What are you to do in this situation? Which is the correct perspective? Unfortunately. there is no 'correct' answer. All that can be said with confidence is that when you are presented with a table or graph be sure to ask about what you did not see. Often times it is what is not seen that is the most important.

Before we go, let us return to our opening question and look at the graph constructed to describe the effectiveness of Ansylium. What is the difference between the two graphs below? In which one do you find better confirmation of the effectiveness of Ansylium? The graphs differ only in the length of the time series shown. In diagram A, we see only the time period since the introduction of Ansylium which certainly suggests that there has been a pronounced reduction in the blood level of Chemical X. This interpretation is not, however, supported by Diagram B where the time period in Diagram A appears as the shaded area. The recent decline in Chemical X appears to be more a continuation of a long term decline than a reaction to the introduction of Ansylium.

Add to Your Social Bookmarks: Del.icio.us Digg Furl Reddit Simpy Spurl Y! MyWeb - StumbleUpon