17 August 2012

GDP Doesn't Measure Economic Well Being

Gross national product aka GNP (or the slight variant on it, gross domestic product aka GDP) are measures of aggregate national production in the monetary economy.  Two calendar quarters of GDP decline is the standard definition of a recession.  GDP growth is the standard measure of a healthy economy that has the virtue of being readily comparable internationally and between units within federal states. 

GDP per capita is a reasonable useful first order approximation of economic standard of living, and of value creation, but is flawed in both respects, and one does not need to resort to something far more subjective and hard to compare like Bhutan's National Happiness Index to address these flaws.

Intra-Household Production

One of the big flaws of GDP is that it fails to adequately measure intra-household production.  Paying for child care or a maid service or eating out increases GDP.  Caring for your own children, or cleaning your own home, or adding value by cooking, serving and cleaning up after your own meals as part of a family, even if it provides an identicial or superior value to paying for these services, does not increase GDP.  Selling agricultural products in a marketplace increases GDP, while growing food for personal consumption as a gardener or subsistance farmer does not.

As long as the proportionate share of intra-household production over time, or between compared regions stays more or less constant, this isn't all that serious a problem.  The trendlines and comparisons remain valid, even though the absolute numbers are wrong.  But, it is problematic if there is a material shift from market production to intra-household production (or visa versa), or if two compared regions have very different levels of intra-household production.

This is serious problem in development economics, because GDP is a very poor measure of economic production in less developed countries where a very large share of economic activity, for example, involves subsistance farming.

Notably, GDP already includes one important category of intra-household value creation: imputed rental value of homes occupied by homeowners.  Extending GDP measures to impute value to intra-household personal services and other forms of intra-household production would not be a revolutionary reform of the measure.

Leisure and Other Non-Monetary Goods

Another big issue, which isn't so much a flaw in the metric as it is a flaw in how to use the metric, is that it doesn't capture tradeoffs between work and leisure.  It is possible to look at GDP per hour worked, or per average number of working days per year, which measure productivity and which also implicitly place an economic value on leisure.  People in a nation where the average worker voluntarily goes on vacation for one month a year are both more productive and more affluent than people in another nation where the per capita GDP is the same, but the average worker voluntarily goes on vacation for just two weeks a year.

A similar problem arises with positive and negative externalities that the monetary economy fails to capture.  GDP does not take into account air quality, for example, or if it does, may actually treat air pollution as something that increases GDP because it increases health care spending.  A cost-benefit analysis valid air regulation that reduces health care spending by more than it costs to implement the pollution control measures could very easily reduce GDP even though it improves quality of life and makes people more economically prosperous.  But, because the benefit from the non-market good (clean air) doesn't get traded monetarily, it doesn't go into the GDP computation.

Quality

A third big issue, which was a huge problem in non-market Soviet economies that were trying to boost GDP, and is also a big problem in making intertemporal comparisons, is measuring changes in quality over time.  If you can buy a better TV than could have ten years ago for the same price, a nation is better off even if the amount of money spent producing TVs in the same now as it was ten years ago, but GDP does not capture this reality.

Similarly, if two countries spend the same amount per capita on health care, but one country uses modern medical methods and gets good outcomes from that spending, while the other uses leeches and snake oil and gets bad outcomes from that spending, nothing in the GDP measure distinguishes between the two cases.

In general, quality issues tend to overvalue past economic production relative to current economic production, and also tend to be particularly problematic for comparing GDP between regions when a good or service is not easily traded. 

In  theory, market mechanisms in international trade will tend to resolve cases where the same dishwasher sells for $100 in one country and $1,000 in another, for the same good.  But, prescription drugs and medical equipment, for example, present just that conundrum.  An MRI machine sold in the United States contributed much more to the U.S. economy than the same MRI machine sold in Japan where identical medical equipment is much cheaper. 

Many services like health care and education, are not easily traded or commoditized, so it is particular hard to compare them, even though at some level it may be clear that better outcomes are being provided for similar expenditures in different places.

Zero Value Or Inefficient Production

Perhaps the most difficult correction to GDP, conceptually anyway, is the notion that some activity in the monetary economy may have zero or negative value.  More generally, GDP doesn't measure the efficiency with which inputs and converted into outputs.  This raises many of the same issues as the quality issue.

Health care is one place where this can be well illustrated.  Suppose that one country spends $2,000 per capita to receive a particular set of healthcare outcomes, and another country spends $5,000 per capita to receive the same healthcare outcomes.  By this measure, the country with the more efficient health care system looks more prosperous by GDP measures than the one with the less efficient health care system, all other things being equal, when in fact, their prosperity would be identical.  The marginal $3,000 per capita that the second country is spending to receive the same product is adding no value to its economy.

In the Soviet economies, again, this was a huge issue, because the Soviet economies often produced too much of goods that no one needed, and too little of goods that people did need.  But, GDP isn't designed to measure the use to which good are produced are put.  If the economy churns out too many televisions and people start to use them as dinner plates, GDP doesn't care, even though the actual value the excess televisions are adding to the economy has plummeted.

In a well functioning market economy, the private sector is supposed to produce goods at efficient prices and in efficient amounts, so neither of these distortions are material.  But, all economies are mixed economies to some extent, and the private sector may have systemic distortions.

Another place this comes into play is in the guns v. butter issue.  GDP, in general, assumes that production purchased by government is worth what government adds value to the economy proportionate to what government pays for it.

Many economists aren't too worried about the zero value production issue in the private sector because they assume that goods produced for consumption in the private sector are appropriately valued for aggregate economic purposes based on actual production at market prices because the mechanism of a reasonable functional market economy place natural checks on the production of useless goods or overpriced goods.  In a reasonably functional market economy, this isn't a grossly unreasonable assumption and has immense practical administrative value in calculating GDP.

But, few economists are equally comfortable in saying that a decision of the elected officials who make spending decisions for government is as an effective a means of preventing the production of useless or overpriced goods as a private market mechanism.  A bridge to nowhere that costs $100 million to build adds the same amount to GDP as a bridge that is used constantly.  A $1 billion dollar warship adds $1 billion to GDP whether it greatly enhances the nation's military might or is actually worthless.  Likewise, if the government spends $1 billion on a military cargo plane when an identical one could be purchased in the private sector for $100 million, because the Defense Department is bad at getting good prices on defense contracts, GDP goes up by $1 billion rather than $100 million.  Certainly, military spending, in general, does not directly impact consumer stadnard of living.  Prison spending on a wrongfully convicted inmate adds the same amount to GDP as prison spending on someone who committed a serious crime and would reoffend if not incarcerated.

Valuing all spending at actual transaction prices makes GDP computation much less subjective.  But, if ignores the capacity of the economy to add value by increasing efficiency or quality rather than merely increasing production in the narrow sense.  The problems related to zero marginal value production in GDP computation are essentially a more subtle variation on the problems with using Marx's labor theory of value, which economists have long derided.

No comments: