[That is evident in the chart below, which displays the fact that virtually none of the orthodox forecasters predicted the scale or timing of the biggest economic event of the half century.] Fully five months after the Great Recession began, more than half of the professional economists in one poll had still not called it.
Forecasting skill, notably, does not follow the money. One of the first acts of Ben Bernanke's tenure, after moving from the George W. Bush White House to the Fed, was to institute a formal display of projections by the FOMC members. At the time it was taken as the beginning of inflation targeting, but it turned out to be the official record of an abysmal performance in economic prognostication. This comes in spite of very large staffs and enormous statistical resources.
[Following are two series of charts, one displaying the Fed governors projections at the advent of the Great Recession vs. actual outcomes, and the second displaying projections and outcomes two years later.] Since that time the economy has settled into stagnation, but it has helped the Fed's board only slightly in their calculations of the future. Vigorous disagreements are routinely reported in the media between those who think rampant inflation is the threat and those whose greatest fear is a depressionary deflation. If, as Karl Popper said and George Soros is wont to repeat, "Predictions and explanations are symmetrical and reversible," the architects of monetary policy do not have a good explanation of how the economy is put together.
That this is not more fatal to their credibility is a function of time. Forecasting is about the future. When the future gets here, excuses have been made and attention is again on the future. In the meantime, as actual results have come in, longer term forecasts have been ratcheted down or up, so that when the data arrives, the current prediction is not so far off. This is most evident in the absurd exercise of forecasting the latest releases of the governmental statistical offices up until the day of that release. But these data are for the previous months or quarters, so the exercise is about forecasting the past. (The participants, however, gain credibility for their precision, even here.) Those data points from the Bureau of Economic Analysis or Bureau of Labor Statistics are often moving targets, as well, as advanced and preliminary and final numbers often dance around after their initial presentation.
The relevant consideration is one again made by Keynes, "It is better to be roughly right than precisely wrong." Unexpected events can jolt one month's or one quarter's numbers. It is the trend and the scale that are important for public policy and private decision-making.
What does it mean?
The failure to come close in the GDP, inflation, unemployment and other metrics has serious implications. Decisions based on these projections were bad decisions, obviously. More troubling is that these errors reveal the basic misunderstanding and ignorance of the forecasters. And even more troubling than that is the absence of accountability for the error. Accuracy is not valuable at all if the accurate forecasts are ignored, or the explanations used for them are not incorporated into the general understanding. And that is precisely what has happened over the past half dozen years. Those who predicted the crisis and described its causes and cures are no more influential now than they were before. Those who minimized the dangers and eventually steered the economy onto the rocks are still in charge. And the explanations survive without amendment!
The Error
The dominant explanation of the economy is wrong. Very wrong. The models used by forecasters are often DSGE, (dynamic stochastic general equilibrium) models that do not include money or a financial sector and that constrain outcomes to an equilibrium that we have not seen in the real world. That is, the crises we have been experiencing over the past decade are BY ASSUMPTION not possible. Plainly these models are wrong. YET THEY ARE STILL USED.
The exercise in this framework then becomes calculating what sorts of economic "shocks" are likely. A "shock" is an event external to the economy, which affects activity, and it us used to characterize such things as the financial meltdown of 2008, a credit shock, and the policy debate of 2013, the "fiscal cliff," a demand shock. These events and their results are assumed to be outside the economy, when they are simply outside the explanations of the economy used by the forecasters. The nature and severity of these "shocks" varies depending on the data they need to describe. They are enlisted in an ad hoc ex ante manner, as needed after the fact, to excuse bad forecasting.
It will soon be revealed that the positive "shock" of cheap oil and gas from fracking turns into a negative "shock" as the planet has to retool and reform and rebuild to survive climate change. Population growth and aging, demographics that are clearly visible decades in advance, are another "shock" to some. It is not that these events are hard to see or anticipate, nor that their implications are not completely discernible with only a moment's reflection, that makes them "shocks." It is that they are not included in the assumptions.
Integral to this problem and the limitation by assumptions is that the models – in spite of their mind-numbing complexity – omit enormous swaths of economic activity and ignore economic values such as resources and infrastructure in a systematic way. The models (not just the DSGE models, but others, as well) are concerned only with monetized economic activity.
The most commonly cited statistic is GDP, gross domestic product. The change in GDP is "growth." GDP does not capture services in households (child care, food preparation, health care, etc.), but it does capture those same services when money changes hands. GDP describes the depletion of natural resources, even oil and gas reserves, only in a very limited way. This metric overlooks virtually all effects of climate change, from the acidification of oceans to more frequent extreme weather, until those effects somebody to spend money.
Related to this problem is the fact that GDP treats "bads" as the equivalent of "goods," so it cannot be considered a measure of well-being or prosperity. That is, alcohol is the equivalent of nutrition, jails the equivalent of schools, wars the equivalent of public utilities, treatments for obesity the equivalent of cultural events, and so on.
This last point is true across the economic metrics that drive policy. GDP, employment, inflation, all of these are not related to well-being except passively. Growth does not mean improvement in well-being. Employment does not necessarily mean a better life. Inflation can be a remedy for ills and can accompany an improving economy, or it can be a scourge to the consumer and accompany a stagnating economy. It is analogous to body temperature, which can just as often accompany healthy physical activity as it does serious illness. Yet the indicator is considered on its own account, whether growth, employment, inflation, or any of the rest.
This dissociation of economic metrics from objective experience of well-being is also manifested in the typical treatment of income distribution. Orthodox economists see no difference in the vitality or stability of an economy of one rich person/ninety-nine peons vs. one hundred middle class citizens. Again, it is not necessary to enlist the reflexive moralistic reaction. In strictly economic terms, it has been shown that (1) greater income equality leads to far better social outcomes in terms of more goods than bads being produced (see Wilkinson and Pickett (2011)), (2) historically high rates of income inequality coincide with economic and financial crises, and (3) the vitality of the economy as measured by response to policy or investment is much greater (the multiplier is higher, in economic-speak) in more equal societies. These are demonstrable phenomena, yet they are ignored. They offend a particular political point of view, and they frustrate the sterilization by mathematics.
In addition to these general defects are specific defects in measurement. National accounts and hence the economic models built on them treat output from the public and private sectors differently, and even from one aspect of the private sector to another. An egregious example of the latter is health care. Output from the health care delivery system is assumed to be equivalent to the costs of inputs. In GDP terms, the US has twice the output of health care as other advanced nations. This contradicts, of course, the many surveys and measurements which show actual health outcomes in the US as being below virtually all other advanced nations. Instead of the robust health and long lives implicit in GDP numbers, you find bad teeth, bad backs, obesity, child mortality and shorter life spans. Were the obvious correction made, US GDP would drop by seven percent overnight.
Similarly, the output of the public sector – everything from roads to education to courts to police and fire – is assumed to be the cost of the inputs. This is again in spite of cost/benefit analyses and common sense. The value of a road is not the cost of the road, but the number of vehicles it carries and the increase in economic activity it provides. The value of an education is calculable in terms of its returns to the person, her family, her employer, and the society, not the number of dollars that are paid to the teacher.
So, primarily because they do not fit neatly into the mathematical models, orthodox economics omits a vast swath of economic activity, ignores well-being and economic health in favor of easily quantifiable metrics, and mischaracterizes much of what it does include.
As if that were not enough, the Demand Side forecast strays further from the neat mathematization of economic activity. The economy is not a machine which operates according to immutable laws, but a construct of institutions which is often turned to the purposes of the powerful. Money itself flows to power, not to value, as John Kenneth Galbraith has noted (and contrary to the contention of one sect of the orthodoxy). To explain the workings of the society as if it were a closed, thermodynamic system made of molecules of economic variables is an exercise in absurdity. The economy is the way the society organizes competing interests, motivations, human behaviors. Nobody has ever seen a unit of output, or a unit of inflation, or a unit of productivity. These are aggregated measures reduced to money equivalents.
Rather than its steering being controlled by the invisible hand, the automatic pilot of an intelligent system, the economy is controlled by institutions. Decisions and policy frameworks are distorted and incomes and wealth diverted based on the temporary arrangements of a political nature – the power of corporations dominating in the current context – not by any fundamental efficiency or true value. Thus, Demand Side forecasting is Institutionalist in the tradition of John Kenneth Galbraith and James K. Galbraith.
Extensive, but necessary adjustments have to be made to extricate the economic forecast from these limitations:
First, a capitalist market economy is constrained by demand. Demand Side operates its analysis and explanation from the demand side. Government and the investment goods sector of the private economy are the means by which demand is managed.The most general difference between Demand Side and orthodox forecasting will be seen in the attention to public policy. It is not true that the current stagnation is here by necessity nor mandated by a decree of immutable economic forces. It is here because policy response is what it is, constrained today by political actors and by economic ignorance. It is not true that we must bequeath to our children a dead planet. It is our present choice (again constrained by an ignorance, perhaps fomented by powerful entrenched interests). Public policy is in the open, albeit colored by political paint.
Second, credit, money and banking generate a primary component of demand, and are included in the Demand Side models. This follows from Hyman Minsky and Steve Keen. The Demand Side model emphasizes, rather than ignores, the real effects of this sector.
Third, also following from Keynes, Minsky, Keen and others, Demand Side forecasting rejects the assumption of equilibrium and therefore dispenses with the need for an ever-expanding menagerie (?) of "shocks." As Hyman Minsky taught, a theory which does not allow the possibility of financial crisis or economic depression is not a theory relevant to a market capitalist economy. Yet that is precisely what equilibrium forecasts do.
Fourth, the real events of the world – climate change, the depletion of resources, the wars and poverty that are diseases of society – are not so much difficult to see as ignoring them is essential to the conceptual frame of the supply side orthodoxy. Thus, it is relatively simple to incorporate them into long-term forecasts. There is a similar bias in favor of entrenched corporate interests, not so difficult to see, particularly in a relatively open democracy, but not amenable to change. Thus we adjust the forecast for the relevant time horizon for the easily anticipated collapse of the environment, the probably chaotic attempts to deal with that phenomenon after the fact, the vigorous obstruction by the power elite and the absence of meaningful policy response. The value of a healthy livable planet is taken as a given, and so the terms of the forecast will contrast with the orthodox forecasts, where the monetized economic activity of attaining climatic stability or destroying the habitable environment may be more or less the same.
Fifth, the issues of income equality, correct valuing of resources and investment and activity, impacts the Demand Side calculations in ways that are absent from other analyses. Households who provide their own services when they cannot afford to buy them, for example, are adapting to their realities. Businesses which depend on income from these households will contract, incomes will diminish, and the tendency to more non-market activity will continue. (Meanwhile the financial economy will amplify the movement, which feeds back to our first point – the economy is constrained by demand, i.e., by incomes and credit.)
Sixth, Demand Side uses the multiplier extensively to estimate the impacts of investment, public spending, income equity, tax policy, debt levels, and so on.
The Demand Side view, the model, the explanation, is proven and empirically validated and historically consistent. The orthodox view is, sadly, that of a morose Pollyanna. "This must be the best, because it is what we have." In fact, we have the product of a bad economics and a repressive corporate state.
So we have set ourselves a difficult task. In order to make the Demand Side forecast relevant, we need to describe it in its own terms, that is, portray the condition of the actual economy and how it is likely to progress. (We would like to use the term "real economy" here, but real often means "inflation adjusted," which is quite distant from "actual.") And we would like to contrast the results with those of the orthodoxy to demonstrate which view is actually superior, which explanation fits best. Unfortunately this will mean translating the Demand Side terms into the orthodox terms, talking in terms of GDP and inflation and unemployment rates. It will mean comparing our "approximately right" to the "precisely wrong," and trying to talk in shorter terms than are actually useful. (Short-term metrics, the object of precision for most macro forecasters, are not only bumpy because of events, but the statistics are often revised long after the fact and long after the discussion has moved on.
To do this, we have developed some customized metrics, such as "Net Real GDP," and we will bring to the fore the forecasts of others in our medium- and long-term frames.
Engaging in this manner, unfortunately, is like engaging in a debate on the correct positioning of the deck chairs on the Titanic. The appropriate discussion was not had at the appropriate time in the captain's cabin, and now that the orthodox charts have put the ship on the rocks, the seating arrangement is likely in other vessels.
We will do our best. Simplifying and making parallel the salient discussion. Keeping the long view. Providing comparative statistics.
Posts
