Going beyond good and bad governance: still using debt as an example

I think the past few blog posts on debt may be evidence of me getting too engaged in one area. Oh well. I think it is super-interesting and important to think of how a government's debt management reflects on its governance. I've had some online and offline responses to my posts that have kept me thinking in this area, and I thought I'd share the thoughts. The risk, as always, is that the fresh ideas make less sense than I'd like. Not a bad risk, however, for those who believe in PDIA: We have a problem (thinking constructively about governance and debt) and my goal is to think about it iteratively to solve it, with (hopefully) lots of feedback and learning.

So, back to debt.

The always-smart and inspirational Natalia Adler reminded me to focus on outcomes when talking about governance, which I think is right (see long-past posts). I think debt levels are an outcome or output indicator. Debt data reflect the way government uses its authority to raise external funds on behalf of citizens. The process of incurring debt is certainly a process, but the final level of debt is a stock that has major implications for many things: How much a country can do now (because of the funds raised), how much the country must pay now, what must be paid in future.

Philipp Krause reminded me that debt levels are not a simple measure of governance. It is hard to say that high levels of debt are good (or great, as Philipp suggests) or bad. It is contextually relative.

Nadim Mata noted that saying 'governance is relative' is a dangerous thing. My word last week was 'uncomfortable'.

So: where do we go with the idea of governance…focusing on outputs or outcomes that people care about yet in a way that accounts for contextual factors…yet is not saying 'all is relative'?

Here is my idea. Love to know your thoughts.

First, I ground my reasoning in theory: Governance is about the exercise of authority by government on behalf of citizens. Drawing on Dani Kaufmann's work, it is about the way governments work within the context of established authorizing mechanisms and rules of the game. In respect of debt, I argue that governments have authority to raise money to further the interests of citizens. They are given this authority because it is clear that there are situations where debt is needed. The authority is not limitless, however, as there is a downside to too much debt (its cost). Different contexts are likely to have different thresholds of what is acceptable in terms of levels of debt a country can incur. This is determined by informal cultural variables, norms, practical constraints, formal rules that might restrain debt, etc. When we look at debt governance, therefore, we look at the way governments exercise the authority to raise debt within the rules of the game in a given context.

Second, I looked for an indicator: I wanted the indicator to be easy to understand, for starters, and contextualized as well. The standardized content comes in the form of the level of debt (as % of GDP). This is the number most people talk about when reflecting on debt issues, which makes it an easy starting point. The conventional approach is to show that high debt = bad and low debt = good (at least this is how I see the conventional position…). But we know it is not so simple, because some countries (like Japan) have had really high debt to GDP ratios for a long time and seem to continue amassing debt. The Japanese experience shows that some groups of citizens authorize their governments to run up high levels of debt. They do so because of various other characteristics (including the fact that Japan is also one of the world's largest creditors…). So, the amount of debt a government can run up before becoming 'bad' is relative to its contextual characteristics and how much and what type of authority it is given.

Third, I asked how we could capture the contextual authority governments have. I started working out some political and social characteristics of different countries and working out the patterns associated with these characteristics. Countries with natural resources like oil seem to have lower debt levels, for instance, and countries that introduced fiscal rules because of major fiscal meltdowns also have lower debt (Sweden is an example). These stories make sense, and for some countries I found myself developing useful categories that helped locate the kind of debt governance one might expect and how this would equate with debt levels (oil independent like Azerbaijan = limited debt; fiscally restrained like Sweden = limited debt; constrained debt like many developing countries who cannot raise debt = limited debt; net creditor like Japan = higher debt). But one only has simple stories with some countries when looking at these kinds of variables, and I found a number of countries really hard to make sense of. Partly, I believe, because the contextual characteristics influencing governance arrangements are imperfectly apparent…like an iceberg we only see some factors.

So, I asked what evidence there may be about how rules of the game manifest in different places. And I thought, simply, to look at what seemed to be 'acceptable' levels of debt in different nations, given actual data from the prior ten years. A country like Japan has averaged over 100% of debt to GDP in this period, for instance, with relatively little standard deviation from this mean. Switzerland has averaged about 22% and there is very little variation between years. What I take from this is that Japanese authorities are authorized to incur high debt levels but Swiss authorities are not. Ireland had an average of about 45% but the variation between years was really high, indicating a major decrease in debt/GDP levels from 61% in 1998 to the high twenties in 2007. What I take from this is that Irish authorities were authorized to decrease debt to GDP ratios in the 2000s, after experiencing high levels in the 1990s. A return to high levels would not be well received, given that this was beyond what was now being authorized.

I developed a variable out of this analysis that shows how much the 2010 debt levels deviate from the level that seems 'accepted' in the past decade. I used a similar basic formula to that employed in a 2010 Oxford Development Studies article, which is essentially the z score (2010 debt = average debt)/std. deviation. I know that I am assuming a range of things about distributions etc. that create problems but I'm forcing myself to simplify so for the time being let's forget this. Japan scores 1.51, suggesting that the level of debt is 1.5 standard deviations above the 98-08 average. Ireland is 2.35, indicating that its 2010 debt levels were nearly 2.5 standard deviations above the 98-08 average. Note that Japan's debt levels were about twice those of Ireland at the time (174% of GDP as compared with 8% of GDP). Theoretically, however, it would appear that Ireland's authorities were way above their authorized range…even at a lower level of debt…and one would have expected some kind of change.

Consider the following figure that shows 2010 debt to GDP levels and deviations from 'acceptable' levels in about 70 countries.

I like this way of thinking about governance in respect of debt management. It allows us to look at an output or outcome of government activity (the debt it has incurred) but in a contextually specific and relative manner (examining if the government is behaving within the scope of authority it enjoys, as reflected in revealed performance over the past decade):

It might help explain why the Ireland has responded to its smaller debt burden quicker than Greece has responded to its large debt burden (the Greek debt is more accepted, given the history of high debt and high variation).
It could also explain why debt levels are a key issue in countries like the USA and UK even though many economists say both countries have 'technically' got space to absorb more debt. Given the recent history, citizens in both countries have become used to having lower levels of debt than they had in, for instance, the 1980s. This means that governments in these contexts are operating in contexts where higher debt to GDP levels are not going to be easily authorized and even a debt to GDP level of 88% draws attention in the USA. At this level government is operating about 2.5 standard deviations above its average debt level in the past decade, stretching its authority to raise debt.
The analysis also allows us to look at countries that use debt sparingly, and we can better understand the stories here. A country like Sweden has debt to GDP levels of about 40% but this was about 1.5 standard deviations below the 98-08 average. This reflects the fact that Sweden has become a more restrained government in this period, where authority to spend and incur debt has been systematically reduced. Cyprus, which I believe is about to default, is also in this area below the line because of higher debt ratios in prior years.

There are many questions I have about the details in the figure above, but what I like is that it allows us to tell a story about how governments are performing in light of the revealed authority in each context. One might use this to say that governance is more risky as it moves further away from the 0 point on the vertical axis; governments aware of this risk will be expected to either reduce their debt or begin serious political communication about the need for more authority to incur debt. This is, I think, what you see now in the USA. It's not that governance is good or bad in this context: Just that authority is being exercised at an extreme…