What does the future hold?  This is a question we all would like to know the answer to, which may help explain why business is good for Tarot card and palm readers.  Our interest in the future has also been good for meteorologists, who get some prime-time news to tell us what will happen tomorrow, and for political scientists and pollsters who appear every two years to forecast election results.  The widespread interest in seeing into the future was a theme picked up by John Casti, in his 1990 book Searching for Certainty: What Scientists Can Know About the Future.  Casti looked at a wide array of forecasts by researchers / scientists in a number of fields and attempted to evaluate their forecasts.  At the head of the class were Quantum mechanics and Celestial mechanics where the forecasters received an A.  Those specializing in forecasting the stock market found themselves in the middle of the pack with a grade of C+,  slightly better than the C- received by those in the business of forecasting war. 

All of these forecasters, however, received better grades than the economists who got a D for their efforts, which put them slightly ahead of the D- given to Evolutionary biologists.  Maybe it is not surprising that economic forecasters fared so poorly.  If you look at the economic press in the mid 1990s you will find stories predicting the Asian century and trumpeting the Asian secrets to growth, and very little advance warning for the economic crashes of the late 1990s.  And then there was Professor Franklin Fisher's opinion that stock prices had reached "what looks like a permanently high plateau," which he offered just prior to the stock market crash of 1929 that ushered in the Great Depression. 

To understand the grade, it is essential that we understand what it is that economists are getting graded on.  In this unit we will be looking at economic forecasts.  More specifically we will be looking at the three interrelated questions: When in the future are they forecasting? What are economists forecasting? and How do they develop their forecasts?

Let's start with the, When in the future question.  If you return to the work of the earliest economists you will find that any forecasts they made dealt with the long-term - what will happen in twenty, fifty, or one hundred years, not what will happen in the next month or quarter. Adam Smith, taking the very long view, saw the market system as providing the world with something new - economic growth.  This growth was not something that would happen in the near future, and it was not something that would be forecast with much degree of accuracy, but it was certainly a forecast. And it was an optimistic one at that.  Following up on this long-term theme, Thomas Malthus came away with a decidedly more pessimistic view when he peered into the future. What Malthus saw was a world where the geometric growth of population would quickly outstrip the arithmetic growth of food.  The result was the world's people would always be condemned to live at a subsistence level.  The pessimism implicit in this forecast was not lost on others and it was this forecast that got Economics dubbed "The Dismal Science." 

There is still interest in the long view.  One example would be Peter Schwartz, who in the Art of the Long View, pushes scenario planning where a forecaster sets out to identify the drivers, the agents of change that "influence the outcome of events." Schwartz compares developing a forecast to solving a murder mystery.  Once you know the motives, you are well on your way to solving a murder mystery, and once you have identified the drivers, you are well on your way to developing a forecast. One of the examples he mentions is the ancient temple priests' forecasts of the Nile River. 

"Every spring, the temple priests gathered at the river's edge to check the color of the water.  If it were clear, the White Nile, which flowed from Lake Victoria through the Sudanese swamps, would dominate the flow.   The flooding would be mild, and late; farmers would produce a minimum crop.   If the stream appeared dark, the stronger waters of the Blue Nile, which joined the White Nile at Khartoum, would prevail.  The flood would rise enough to saturate the fields and provide a bountiful harvest.  Finally, if the stream showed dominance by the green-brown waters of the Aterba, which rushed down from the Ethiopian highlands, then the floods would be early and catastrophically high.  The crops might drown; indeed, the Pharaoh might have to use his grain stores as reserves.

The driver behind the flooding of the Nile was the relative strength of the three rivers, and with information on this the priests had their flooding forecast.  A second example offered by Schwartz is the work of Pierre Wack for Royal/Dutch Shell in the late 1960s and early 1970s.  Wack observed the increasing dependence of the US on imported oil, the growth in oil demand, and the resentment toward the West of OPEC countries due to the Arab-Israeli war of 1967.  Wack concluded these forces could very well lead to rather dramatic oil price increases and Shell should plan accordingly.  By 1973 the scenario was playing itself out as the OPEC countries cut off the supply of oil to the west and the price rose dramatically.

Where does one look for the 'drivers'?  Schwartz identifies five factors: Society, Technology, Economics, Politics, and Environment.  Hamish McRae, in his book The World in 2020, identifies two drivers - demographics and technology.  One example which highlights the importance of demographics, and one you should be familiar with, would be the forecasts of the future of social security.  We have all heard of the dire forecasts of difficulties encountered when the baby boomers begin to retire.   It is fairly easy to see that many factors will have an impact on the retirement decision - health, income, wealth, occupation, and marital status at any time - but that the real driver is age, and the number of retirees will be dependent upon the number of people over age 65.   In the graph below it is very obvious that in the decade ending in 2020, the real 'crisis' will emerge as the boomers begin to hit age 65.   In that decade the number of people aged 65 or older will increase nearly 80 percent (solid blue columns) while the working age population will actually decline (maroon-lined columns).  Similar analysis is behind the explanations for the decline in crimes experienced in American cities in the late 1990s.  Crime is down simply because there are fewer teenagers who are in the high crime rate age bracket.  

Technology is also a prime driver and history is filled with examples of where technology altered the pattern of growth.  The adaptation of the steam engine to textile mills moved the center of the textile industry from Lowell, MA to Fall River, MA; the automobile was behind the suburbanization of the US population; and air conditioning contributed to the explosive growth of the Southwest.  In 1998 one of the favorite topics is the impact the internet will have on various industries ranging from banking to retail shopping to education.

By the 1960s we were seeing the early stages of a movement toward short-term forecasting.  This movement could be attributed to four factors - the theory, the numbers, the machine, and the market.  The theoretical basis for concern with the short-run was provided by John Maynard Keynes.  In the 1930s, when Keynes published his General Theory, the world was suffering through the Great Depression and there was a sense of immediacy to addressing the problems of falling output and employment.  As we saw in our discussion of the 1930s, Keynes worked hard to direct the profession's attention toward the short-run and away from the long-run.  More importantly, Keynes' theory offered the basis for government intervention as his "multiplier" replaced "crowding-out" in discussions of macroeconomic policies. 

And then there were the data the US government had begun to collect and publish after W.W.II - the numbers needed for the short-term forecasting.  During the Great Depression there were very few statistics available to monitor the economy's performance - none of the wide array of statistics that you hear in the news today. To remedy this situation, Simon Kuznets was given the task of leading the effort to construct the national income accounts for the US.  The direct result was the GNP statistics published by the Commerce Department, but certainly there was a spin-off as other government agencies began to collect and publish various measures of the macro economy and its four primary markets (output, labor, capital, and foreign exchange).  By the mid 1960s, adequate macroeconomic data existed that provided economists with the basis for constructing the statistical models that would bring realism to the theoretical work of Keynes. 

For example,  Keynes talked about the marginal propensity to consume (MPC) and highlighted the importance of this link between consumption and income in explaining the government multiplier (the change in income (DY) generated by a change in government spending (DG). The formula in the simplest Keynesian macroeconomic model for the government multiplier was:  DY/DG = 1/(1-MPC).  For this model to provide a useful forecast of the direction of change in income, or the effect of government spending on income, we would need to know the value of the multiplier.  This, in turn, meant we would need to know the value of the MPC.  One way to establish the value of the MPC would be to look at the data on income and consumption and 'test' for the nature of the relationship, and by the mid 1960s the data sets existed. 

The primary focus of economists' forecasting efforts have been on the performance of the aggregate markets - the quantity and price measures of the output, labor, capital, and foreign exchange market.  It is highly unlikely you will get through a week of reading the financial press without running across some economic forecasts of macro aggregates such as interest rates, inflation, GDP, unemployment, and the exchange rate.

Businesses also care about the performance of the macro economy since their "bottom line" can be very much influenced by the economy's ups and downs.  In late 1998 Alan Greenspan, chair of the Federal Reserve (Fed), lowered interest rates in the capital market with the hopes the lower rates would spur spending in the output market.  This is the reverse of what we saw in the early 1980s when Paul Volcker, chair of the Fed drove up interest rates, bringing home mortgage rates to near 20 percent range and effectively putting the brakes on the housing industry.  And in 1998, the concern hanging over the industrialized world was the decline in the Asian economies would translate into lower export sales by US companies.  The forecasts of lower corporate earnings based on these economic forecasts were responsible, at least in part, for the drop in the stock market. 

This was the situation confronting Sam who had his market located in western RI.  The accompanying graph describes the pattern of sales at Sam's market during much of the 1970's.  As you can see, in 1973 Sam's suffered a substantial drop in sales and never regained the loss. Sam, the owner of this intermediate size super market was looking for an explanation for the loss.  Some of the obvious candidates would be new super markets moving into the area, changes in advertising strategies, or major construction in the area that reduced customer access to the store. 

Another place to look would be at the 'macro' environment.  Anyone familiar with macroeconomics would tell you that 1973 marked the beginning of the OPEC induced recession that hit the New England region hard.   The result was many businesses suffered losses. In the diagram below, indexes of sales of Sam are contrasted with the overall sales in the region and it is clear that Sam's troubles simply reflected the troubles in the broader, macro economy. [for a discussion of indexes see Benchmarking].   The prevalence of this type of causality from the macro economy to the individual firms and industries has created a very real demand for macroeconomic forecasts.

Another example would be forecasts of prices. Just as it would be difficult to forecast sales without some sense of the underlying strength of the economy, it would be impossible to forecast or explain prices without an understanding of the underlying inflation rate.  One of the best examples would be wages - the cost of labor.  If you look at the following graph to answer the question, in which decade did wages grow fastest, the obvious answer would be the 1970s.  This is what students have said for years when confronted with these data.

What was driving this was not the condition in the labor market, but rather the underlying rate of inflation.  Wages were rising because prices were rising.  To understand the above average wage increases in the 1970s, it is essential to understand that the 1970s was also a period of above average inflation as you can see from the graph below. 

The situation would be the same if you were dealing with interest rates.  If you wanted to forecast interest rates, you would need information on inflation rates. This relationship between the inflation rate and interest rate (on short-term government securities) is apparent in the diagram below.  The 18 percent mortgage rates that we saw in the early 1980s could not be explained without being placed in the context of the double-digit inflation rates.

At this point we have looked at two of the questions - how far into the future, and  what will we forecast.   Now we will turn our attention to the how question.

To highlight the different forecasting techniques, let's look at a very simple problem - the situation at Itzibitzi motorcycles.  The questions facing the company is: What will sales be over the next 5-6 years? Will these sales be adequate to justify a new production facility?  In this section we will look at three techniques - time-series analysis (the Ruler), econometrics (the Relationship), and barometric forecasting (the Experts).