Do any of you gentle readers out there know and understand demand based forecasting? It is a simple process in its basic concept. A manufacturer produces a good or variety of goods to sell to wholesalers who in turn sell to retailers. Now sometimes the wholesaler is the retailer, not that it makes too much difference for the purposes of our discussion. It a product is new, one that has not been in the public eye until now such as instant orange juice, then forecasting becomes rather difficult. Typically, forecasting based on demand is really trying to figure out how much of an item or service is needed in the future based on past sales. If I bake and sell loaves of bread then I have a history of sales that I can consult and determine if there is an increasing trend of individuals and families buying more bread for their personal consumption. Of course we know that bread tends to be less generic and more brand based. The reason for this is due in part to a perceived difference in quality and taste and in part to the marketing effort to convince buyers that Brand X is superior to Brand Y or Brand Z. Thus customer loyalty must be a factor in forecasting. Some items are more commodity like. Nails are manufactures by several different companies and usually only one company’s brand is stocked in the various hardware stores. And most customers rarely care what the brand name is as long and the nail can be driven through wood without bending, assuming proper skill is applied. On the other hand when it comes to the higher cost items such a washing machines or automobiles, demand based planning can become a bit trickier. Usually individuals or families do not buy several washing machines every year. If the average life of a washing machine is ten years than that is the average replacement rate for an individual or family. The perception that one needs to trade up after a few years to a higher priced washing machine is difficult to create in the minds of many customers.
As you might start to perceive, this is a problem of economics, one of supply and demand. The supply is based not only on price but perceived value. In the case of bread, that perceptions can change very rapidly. If the bread you like is a speciality bread, something that everyone doesn’t buy such as a whole wheat rye, then the price you will pay is going to be higher than the popular plain enriched white bread. You pay the higher price for the rye because you perceive that it has greater value in terms of quality and taste. Now if you are the baker who bakes this bread how much of a price break would you need to give to increase sales? Assuming that an increase in sales beyond a certain point would provide a lower price in ingredients you buy from the flower mill owners, then you would need to work out the cost of baking more rye and buying more flour. Given that flour is a commodity, per se, then an increase in your ordering could be done slowly. That drop in price might not generate the sales needed to immediately obtain the flour mill discount and you may need to rely of reduced gross income until such discounts do occur. One way you might determine the increase demand is through a promotion special. Two loaves for the price of one or this week only, twenty percent off. So our calculation of future demand is still based on past sales. Not exactly a fool proof method of forecasting future sales for future demand.
Now oddly enough this line of forecasting, demand based planning, also applies to investing. That is, based on the past history of a company’s earning as applied to share prices, and its cash flow, its debt load, and so forth, as well as the company stock price history, one should be able to forecast the future price as well as future earnings which may include dividends for the current stockholders. Demand based planning or forecasting depends on all other things being equal. There is no recession or depression or financial panic or general world war on the horizon, so as the population of consumers increases so should demand. Statistical analysis is one of the main tools people use in forecasting. One analyses the sales data, the stock price over the course of a year, the earnings history, and a host of other histories. Notice that I said histories for a reason. Statistical analysis can only use historical data, events that have occurred. You can try and compute the probability of future events but that relies on historical trends. And probable outcomes are not real events. If one tosses a fair coin in the air what is the probability that it will land head side up? What is the probability that six tosses of a fair coin will conclude with the head side up each times it lands? The difficulty is in the actual tossing of the fair coin and having it land head side up six times in a row. Do you see the difference?
So regression to the mean, a wonderful tool, merely points out the tendency of some process to revert to the mean average. A stock’s price can fluctuate day by day and month by month, but the extremes in its highest prices and its lowest prices will tend to revert back to the mean average price. But only when the stock market is normal. Unfortunately, we have seem the various markets, stock, bond, commodity, and so forth upset from their normal workings through an increase in credit used, meaning an increase in debt load in the world, and bad monetary policy such as Quantitative Easing which pumps money into speculation. The money which the Federal Reserve has spent on buying an enormous quantity of US Treasuries, Mortgage Backed Securities, Stocks and Corporate bonds combined with reducing interest rates to near zero rates has pushed investment into seeking high risk investment vehicles. It has created opportunities for the borrowing of immense quantities of funds for the purpose of speculation. And it has produced inflations in asset prices and some consumer necessities such a as food. Right now, all the statistical calculations one might do are useless because speculation is not a logical behavior. It is gambling and the intermittent reward based structure does not allow sufficient history of a valid nature to do anything but compute probabilities of future behavior.
The free market’s ability to contribute to demand based planning has been seriously compromised. It is rather pointless to listen to the financial gurus tell us how the recession is over and we are entering a period of growth in GDP. First, there always seem to be a good many revisions to the data. Second, there is a good deal of fudging the numbers. The fact that unemployment is primarily computed from unemployment benefit data does not mean that others who do not receive unemployment payments nor have jobs are not unemployed. Third, when inflation and speculation are rampant in various ares of the economy but not all, then one can hardly make true forecasts into the future. The data has become so distorted as to be almost useless. Beware political pundits and economic experts bearing false predictions and forecasts.