Reinhart and Rogoff: Lesson in statistical terms

Advocates of government spending are enjoying the recent news that errors were discovered in an often quoted 2010 analysis by economists Reinhart and Rogoff that showed countries with debt at levels (resulting from high spending) greater than 90% of GDP had an average GDP growth rate of -0.2%, which was statistically lower than countries with lower debt levels.

Recent corrections show that these countries actually had averaged 2.2% growth, not -0.2%, which is not statistically different from countries with lower debt levels.

Critics have accused R&R’s analysis of spurring irresponsible austerity in government spending and may have prevented more beneficial government stimulus spending around the world.

But, wait. The corrected analysis shows that countries with lower debt had higher GDP growth rates ranging from 3.1% to 4.2%. Yep. In this data set, apparently 2.2% is not statistically different from 4.2%.

That doesn’t mean that the data shows that government spending helps (or hurts) GDP. It also doesn’t mean that ‘austerity’ hurts or helps.

To be clear…it means nothing. Government spending advocates are not wise to use the the corrected stats bolster their case.

Not statistically different means that from the size and sample of this set of data, it cannot be concluded with high confidence that the differences in GDP growth rates are caused by the differences in debt levels.  But, it doesn’t rule it out either.

If anything, the analysis still provides directional support that debt may hurt, rather than help.

Personally, I’m not a fan of GDP. I explain why herehere and here. Basically, it’s because GDP treats an expense like an income.

The multiplier is not prosperity

“I’m doing my part to help the economy!”

I’ve heard many folks make this joke after a big purchase. We snicker. We know they really bought it for the personal benefits they expect to gain. As we’ve been discussing in the comments, they bought it because they valued it more than what they gave up.

The joke implies the multiplier effect — the idea that your purchase stimulates economic activity. You buy a car, which means income for the car maker and workers, they spend that income on suits and shoes, and so on. And, by the time it’s all said and done every dollar of your purchase ‘stimulated’ more than a dollars worth of economic activity, which is measured as GDP.

For some reason, we don’t snicker when economists and politicians make this same claim. We should.

David Henderson, who doesn’t make this claim, does a great job explaining why we should snicker in his aptly titled essay, GDP Fetishism, which I discovered after reading a recent post of his about the ‘multiplier’ of foreign aid.

Also recommended, his latest post about subjective value, which is a topic we’ve touched on here recently in the comments.

Advice for those who want to raise taxes

Here are a few thoughts for folks who want to raise taxes to help pay for the irresponsible spending of politicians.

First, tell me how much more we should pay.  That way when we get there, we will know and you won’t be able to say — we need to pay more — forever and always.

Second, I encourage you to lead by example.  Here are instructions on how you can pay more than you are currently required to pay.  Certainly, your contribution won’t amount to a drop in the bucket, but your act will gain you more credibility in asking others to pay more.

Next, persuade others to do the same, voluntarily.

Consider getting to know more about how your elected representatives feel about spending and vote for those who want to be more responsible.  If you keep voting for the ones who spend more money than the government takes in, then you will want to raise taxes again in the near future once they outspend their limits.  Voting for folks who want to manage government finances responsibly will  also give you more credibility when asking others to pay more

Finally, take a look a close look at this page, which details government spending growth by various categories. If you ask others to pay more, you should at least know why you are asking them to do so.

I especially like the 5th table on the page which details spending changes by category since 2000.  It shows that overall spending has grown 5% annually and some categories have grown much faster, for a total increase of 62%.  At the same time, the economy as measured by GDP, has grown by 45%.  The private sector of the economy has only grown by 31% (if my quick and dirty math is accurate).

So, growth in government has doubled the growth in the private sector.  Since it is the private sector that pays for government, it doesn’t seem like that is a sustainable pattern.

Why C+I is a better measure for the health of the economy

In this post, I claimed that Gross Productive Domestic Product = Consumer Spending + Investment Spending – 2 times Government Spending, or GPDP = C + I – 2G.

Tyler Cowen explains why during this EconTalk podcast:

Considering our economy right now: about 17% of it is health care; about 6% in terms of GDP is education; and with some overlap, 15-20% is what we call government consumption–government activity, not just transfers. At all levels of government, including state and local. Add those all up, take out the overlap, and it’s a pretty big chunk of the economy, like 20-30%. Those are all sectors where there are massive subsidies, massive distortions of incentives, a lot of bad policy; and it’s hard to measure value.

So, when we talk about biases in measuring output and living standards, the bias I worry about the most is we’re spending a lot of money and simply writing it down as value added when it might not be.

GDP measures how much money was spent in the economy, not how much value is created.

What do you do when you spend money that destroys value?  You try not to do it again.

If you buy something and you don’t think the value you received was worth the price, you stop buying it and you might warn others to save their money or spend it elsewhere.  As I pointed out in the post, What is a job?, when we spend our own money we typically think in terms of cost vs. benefit relative to the cost vs. benefit of alternative choices for the spending.   This is the engine of value creation.

But, what happens when third parties, not just governments, spend money and it destroys value?  Often, money continues to be spent because it is being spent for reasons other than value creation.  Sometimes more is spent to try and fix perceived problems.

Third party spending doesn’t have to hold to the same rules as first party spending.  As I also pointed out in What is a job?, there are many reasons for third parties to spend money that have nothing to do with value creation.  This is the engine of value destruction.

This is why I don’t believe C + I + G is a good representation of value created in the economy.

This is also why we should carefully consider the change in incentives when we make changes that moves payments from first parties to third parties, like in the areas Cowen points out — health and education.

Given the incentives around G, I’d rather keep as much of the economy as possible in C and I transactions and use C + I as the measure of the productive, or value creation, estimate for the economy.