Tuesday, January 12, 2010

"Measuring" the effect of Stimulus

The AP recently wrote what can only be called a hit piece targeting the Obama administration that alleged that the stimulus has not created any jobs. This is the result of a study conducted by the AP and reviewed by "independent economists at five universities". Unfortunately, the study itself, which the AP quotes in its own article, does not seem to be available for review by the general public. This is troubling, because the study is certainly methodologically flawed. The claim that the stimulus has not been effective, which is what many readers will take from this article, is not one that an economist can make in good faith.

The claim is suspect for three reasons.

First, the study seems to make a serious error in counting of jobs. A key hint comes from a quote by Emory University Economist Thomas Smith, who reviewed the study:
"As a policy tool for creating jobs, this doesn't seem to have much bite. In terms of creating jobs, it doesn't seem like it's created very many. It may well be employing lots of people but those two things are very different."
Most likely, this study is counting "jobs" as the secondary stimulus effects, or ripples in the private sector created by the public spending. This type of counting is wrong. In terms of stability and growth, there is no functional difference between a job in the private and public sector. The difference is at the level of the firm, meaning that expanding the public sector undermines efficiency at the firm level, assuming that firms function more efficiently than government. Workers who are employed by the government do not spend less or create less demand for production materials than private employees.

Suppose that a law were passed that had the effect of taking $40,000 annually from the single richest person and using it to hire a single laborer - then so long as there is not a shortage of laborers, the number of jobs in the economy is increased by 1. One might here ask whether that wealthy individual will now be deprived of capital that he could use to hire an employee. But two conditions must be met before the wealthy individual will fund such a job - he must have the capital and there must be sufficient demand for the product. In our economic situation, as in most economic situations, the second effect is certainly dominating over the first. In keeping with this analysis, a $20 billion dollar stimulus could be expected to create 500,000 jobs if it was raised totally from taxes. However, in the case of this stimulus plan, the stimulus comes from government issued debts, which are held as safe assets by banks, meaning that they are sold on the market only to wealthy individuals who do not have a better investment option (such as going into business). This makes the stimulus much less likely to trigger capital shortages by potential investors. Certainly, though, there are mitigations, such as disbursement costs that make the initial effect somewhat less. So, this number could be trimmed further, maybe even cut in half, but it is still not zero. If it is not a large enough number, then it is evidence, as many economists have argued, not for abandoning the stimulus concept but for making a much larger stimulus - on the order of $800 billion or so, which is about what the total stimulus adds up to. That is the opposite spin from what many people got from the article.

Secondly, there are persistent questions that need to be asked about the measurement of economic variables. Since many things in the Economy are in constant flux, and since the Economy itself is unpredictable, it is not possible to separate the variables involved without depending on one's economic models. In other words, economics can only ever compare reality to the theoretical. It cannot create or observe any experimental controls. The very term ceteris paribus is a transition to the theoretical.

Many economists sadly use a neoclassical model that posits the economy in a constant trend toward an equilibrium labor employment that is set by a unique equilibrium given by the labor market. Even the Oregon State's official economists use such a model, and as soon as the economy started to go downhill, this model predicted a recovery. In fact, I have seen one rather hilarious graph which shows a series of three predictions, each made two months apart, that doggedly showed the same upward trend in the coming months, each being wrong.

We have known that such a model is wrong since Keynes wrote the General Theory. Though the assumptions, first formulated by Ricardo, about the way labor behaves were almost certainly wrong, Keynes also pointed out that the classical model was based on circular logic. But just to add perspective to this issue, Keynes' primary academic opponent in the General Theory was Pigou, who pioneered many aspects of welfare economics and advocated for subsidies to correct market failures. I highly doubt that were Pigou alive today, he would advocate against government intervention in our economy.

In other words, without knowing if things were going to get better or get worse without the stimulus, it is impossible to tell with any accuracy how many jobs the stimulus created. If one uses a classical model, one is likely to predict a recovery too soon and therefore underestimate the effect. Even finding the stimulus ineffective does not endorse inaction.

Third, the very small nature of the stimulus studied in this article makes it harder to measure. For reasons not explained, this study looked only at a small part of Obama's total stimulus package. In an economy with 300 million people and GDP in the trillions, the aggregate employment effect of a $20 billion disbursement at the national level is going to be rather diluted.

Furthermore, this measure was offset directly by State Governments cutting their budgets. Here in Oregon, we were smart and knew that gutting the State budget would only make our economic situation worse. Even so, we may be forced into such a situation, and we could even lose our super-majorities in the legislature. The tax increases that Oregon passed are really a good option for our economy, but you would never learn this by reading what newspapers like the Oregonian have published. Referring to the disastrous way that budget cuts by Herbert Hoover exacerbated the great depression, Paul Krugman warned of the state legislatures becoming "Fifty Herbert Hoovers".

Sadly, the media remains a pawn of business interests. This manipulative relationship is formed in part by the inability of journalists to see through the economic deceptions created by conservative economists. The best way to describe such economists is ironically through their own language of regulatory capture. I would say that these people have been captured by corporate interests. With their help, corporations can expand their grasp, to capture the media. We must work diligently to build our own information networks for political and economic information, because the evidence is mounting that advertising-based media cannot be trusted.

No comments: