Tuesday, June 19, 2012

Does High Debt-to-GDP Ratio Inhibit Growth?

A Twitter friend of mine sent me a paper by Reinhart, Reinhart, and Rogoff titled "Debt Overhangs: Past and Present". It presents evidence that periods in advanced economies where the public debt/GDP ratio is above 90% for five or more years are marked by lower growth than periods where the public debt/GDP ratio is lower. The implication, if not the conclusion, is that a large debt overhang can cause GDP growth to be lower than it otherwise should be.

The introduction of the paper reveals the question that its trying to answer: if interest rates for a high debt/GDP economy are low, should the government take it as a sign that it should not worry about the debt and use borrowed money to try to stimulate the economy? The authors seem to argue that the lower growth rate during high debt/GDP periods means that no, that isn't the right approach to take.

The debt-to-GDP ratio has two factors: debt and GDP. For public debt to go above the 90% threshold, either debt had to rise greatly versus GDP, GDP had to fall greatly versus the debt load, or both. It's interesting to see how that plays out with the paper's core data.

The data it presents covers the past 200 years for today's advanced economies. Page 13 of the paper is where the table of it begins. If you look at each period of at least 90% debt-to-GDP ratio, you'll see that nearly all coincide with at least one of four factors: war, financial crisis, the Great Depression, and international recessions. There are two exceptions.  One is when Spain lost the last of its colonies. The other is Greece in the 1800s, and that country, as best as I understand it, wasn't really an advanced economy during that time.

War causes debt to rise as governments mobilize, and it can negatively affect GDP quite a bit if its on your own soil. A financial crisis will not only hit GDP but also cause a government to start running deficits (or much larger deficits) as tax revenues fall and social safety net spending rises. The Great Depression lowered GDP for everyone, and international recessions do the same thing on a smaller scale.

The paper doesn't discuss the possibility that both the high debt/GDP ratio and the sustained period of low growth might have been caused by some other element (war, financial crisis, etc.), which would mean that the solution to slow growth might not have anything to do with reducing the public debt load. It devotes only two sentences to this kind of question of causality:

Another line of reasoning for dismissing concerns about public debt and growth is the view the causality mostly runs from growth to debt.  The multi-decade long duration of past public debt overhang episodes suggests that at very least, the association is not due to recessions at business cycle frequencies.

It's probably not due to the regular ups and downs associated with the business cycle. However I don't think that war, severe banking panics, equity market collapses, the popping of enormous asset bubbles, or things of that nature are part of the regular business cycle either.

Two of the authors, Carmen Reinhart and Kenneth Rogoff, wrote a Bloomberg editorial in 2011 well before this paper was published but along the same lines. They've been working on this issue for years. They do at least admit there that, "Anyone familiar with doing empirical research understands that vulnerability to crises and anemic growth seldom depends on a single factor such as public debt."

In both that editorial and especially the paper, they talk about how the overhang of private debt can be a big problem too. That's definitely for sure. If you're a believer in the idea that we're in a balance sheet recession, you're definitely on board with private debt being a problem. Individuals deleveraging will consume and invest less, depressing economic growth rates.

But how would high public debt take a toll on the economy? One way is if the government also deleverages by cutting spending and/or raising tax revenue to pay down the debt. That certainly could be problematic, but it hasn't been so far in the US. The federal government has not done anything substantial to address its deficits and debt loads. State and local governments have had to though, and it has resulted in a large decrease in the public sector workforce. That is keeping unemployment high and is depressing the growth rate for sure.

The other way is if the private sector savings rate rises due to fears of higher taxes to pay down that debt in the future. I really doubt that's a big factor. Only 61.7% of voting age people voted in the 2008 election, and that was the highest turnout since 1968. A 2007 study showed that only 35% of Americans nationwide qualify for a "high" level of knowledge of current affairs. Furthermore, 43% of US households live paycheck-to-paycheck. They can't afford to adjust their spending based on decades-out tax expectations even if they wanted to. The likelihood that a significant number of people consider future taxation in their purchasing decisions is low.

Finally, I used FRED data to run a correlation between the federal debt/GDP ratio and the personal savings rate. The range is since 1966, the maximum I could do. It came out to -0.743, meaning that as the debt/GDP ratio has risen, the savings rate has fallen. Running a regression yields a microscopic p-value, meaning that we can reject the premise that the two things are related. If lots of people consider the national debt load when deciding their savings rate, they certainly aren't acting on it or at least assuming that they need to tighten up in the face of future taxes for debt payments.

What I was really looking for from the paper was some kind of call to action or policy recommendation. It doesn't contain one other than "don't impose austerity, but don't leave the long term debt question unanswered, and do try to get it below 90% of GDP as soon as possible". That's not all that useful, although even getting the first part correct is somehow difficult for world governments right now.

Ultimately I don't think this paper leads to a real policy recommendation because it doesn't look at all about how countries left periods of 90% debt/GDP or higher. Did higher growth occur before or after the end of the period? Did the countries leave the periods more due to growth or focus on paying down the debt? They don't say. And anyway, the most important question is not what caused the nations' debt loads to fall but what ignited growth. Again, they don't say.

The closest they get to answering the exit strategy for a country (other than noting defaults) is mentioning that Belgium's 1920-26 period of high debt was associated with a rebuilding boom after WWI. It's implied that the country grew out of its high debt problem. So does that then mean that countries should embark on similar build-to-grow campaigns (such as massive infrastructure investments, perhaps) to solve the problem? They don't say. The UK was able to grow quite well from 1830-68 with far higher debt than the US has now thanks to being the largest and most powerful country in the world. Might the US's similar status allow it to do the same, or are the situations too different to be comparable? They don't say.

The correlation between high public debt/GDP ratios and slower growth than normal is compelling, but as always, correlation doesn't imply causation. The Great Depression caused 90% debt/GDP or more in some countries, but it didn't in the US and a few other nations. They languished for well over a decade with low growth without high debt/GDP causing it. That as much as anything proves that a nation can have an extended period of low growth without debt/GDP over over 90% as the cause. Without that causal link, the idea that this paper's central thesis offers any universally applicable practical advice disappears.

It's certainly possible that low interest rates on US debt are not a green light to borrow more to try to stimulate the economy, but I didn't really get that out of this paper.

Friday, June 15, 2012

Two Charts That Illustrate Why Unemployment Is So High Still

The answer to why the employment part of the recovery has been so slow is a very simple one. I'll use two charts to show why: private sector employment and public sector employment. The data is seasonally adjusted and comes from FRED, and the public sector figures have temporary census workers removed (because they're just that: temporary) thanks to data published by Veronique de Rugy of George Mason University.

First up, the private sector:


Click the image to make it bigger. The X-axis is months after the official end of the recession.

Compared to the last two recessions, the rate of job growth from the official end of the recession (June 2009) is actually doing OK. The public sector is by no means "fine", as it lost nearly 8.9 million jobs from its peak of employment (January '08) to its trough (February '10) and it's still about 4.5 million jobs below that peak. That deficit in jobs doesn't even account for the number of jobs needed to keep up with population growth either.

However, its growth is similar to that after the 1990-91 recession, and it's doing better than after the 2001 recession. It would be great if it was growing jobs at a higher rate, but its current rate is not out of the ordinary for a post-recession economy.

Now, the public sector:


There's your problem. Overall public employment has done just about nothing but fall since the end of this recession. The terrible recent jobs numbers can mostly be blamed on the decline of the number of government workers. It's less a federal problem than a state and local problem, but that's your explanation for why unemployment isn't lower.