Institutions matter — something that economists, who might like to dismiss them as inconsequential sand in the gears of clever theories, often seem to forget. As Brink Lindsey of the Niskanen Center reminds us in a recent essay, even the economists who lead the free-market intellectual movement often lose sight of the fact that, in order to achieve what they are capable of, markets need to be embedded in supportive legal and political institutions. Adopting an institutional approach, however, does not have to mean forsaking the rigorous measurement tools that are the hallmark of modern economics. Crucial institutional concepts like quality of government can be quantified and compared across countries. The numbers powerfully support the argument that quality of government is a key to democracy, freedom and material prosperity.

What Is Quality of Government?

As used here, the term “QoG” focuses on the impartiality, integrity and technocratic competence of the institutions of government — a perspective outlined by Bo Rothstein of the University of Gothenberg.

Quality of government in this sense is a narrow, pragmatic concept that is neutral with respect to ethics. Others, notably Marcus Agnafors (another Swedish academic), have objected that this approach could generate a high-quality rating for, say, a government that wrote clear laws to guide a campaign of ethnic cleansing of its territory, enforced them impartially, and offered formal due process to those marked for forced resettlement. But although such a combination of high-quality means and reprehensible ends is theoretically possible, it happily turns out that, by and large, high-quality governments by the Rothstein definition tend to respect freedom and human rights.

Our pragmatic concept, quality of government, is also somewhat narrower than the related concept of “state capacity,” a term used by writers such as Noel D. Johnson and Mark Koyama of George Mason University to mean the “ability of a state to collect taxes, enforce law and order, and provide public goods.” As I see it, high quality of government is a necessary condition for high state capacity, but the latter concept encompasses substantively good policy outcomes as well as competence and impartiality. Instead of rolling good outcomes into our definition, here I will define quality of government narrowly in order to be able to pose the question of whether better quality leads to better outcomes.

In my analysis, I make use of data issued annually by the Legatum Institute, a Londonbased, libertarian-leaning think tank. The Legatum data include a dozen or more elements that plausibly appear to bear on quality of government, including measures of rule of law, government integrity (that is, absence of corruption), protection of property rights, contract enforcement, protection of investor rights, executive constraints, government effectiveness, regulatory quality and government accountability. From these, I extract a summary measure of quality of government that I will abbreviate as QoG.

QoG and Democracy

Many who live in democratic countries tend to think of “good government” and “democracy” as synonymous. However, the relationship between quality of government (as I’ve defined it) and democracy (a set of procedures that give citizens a voice in selecting public officials and holding them accountable) is far from simple.

On the one hand, if a country is to be considered more than superficially democratic, procedural rules for the selection of leaders must be embodied in a system of law and elections and must be carried out impartially. In that sense, some minimal level of QoG is a necessary condition for democracy. On the other hand, there are plainly countries in which voters enthusiastically elect and reelect leaders who are known to be corrupt, or partial to their own tribe, or both. In that sense, the external trappings of procedural democracy are far from a sufficient guarantee of government quality.

To make a quantitative assessment of the relationship of quality of government to democracy, we need a measure of the latter as well as the former. Freedom House, a nonpartisan group, is a good source of data on democracy. Using three indicators from that source of the freedom and fairness of the electoral process and four indicators of freedom of political participation, I construct a measure of procedural democracy that I call DEM. (DEM, which does not include measures of human rights, draws on only a subset of the data that Freedom House uses to rank countries as “Free,” “Partially Free” or “Not Free.”)

The simple statistical correlation of DEM with QoG is 0.62 (where 1.00 is lockstep correlation and 0 is no correlation) for a sample of 162 countries for which both variables are available. The figure to the right shows a scatter plot of QoG versus DEM. Scores are normalized so that the axes show how far each country is above or below the average value for each data point, using the standard deviation of each variable as the unit of measurement.

Fear not; this all becomes more intuitive. The dotted trend line shows that there is a tendency for countries that are more democratic to have relatively high QoG ratings. Countries above the trend line have higher QoG than would be expected based on their level of democracy alone, while the opposite is true for those below the trend line.

The diagram is divided into quadrants, with notable outliers labeled. Looking first at the “northeast” quadrant, we find all countries that have better-than-average ratings on both procedural democracy and government quality. No shock: Norway wins the QoG blue ribbon.

Most members of the Organization for Economic Cooperation and Development (OECD; members are shown as red dots) belong in this quadrant. That is not surprising, since the OECD was founded as a club of prosperous democratic countries. Taiwan lies farthest to the northeast among economies outside the OECD. A few countries with good DEM scores, such as India, have higher-thanaverage QoG scores. But they fall below the trend line, indicating that their quality of government is not as high as we would expect given their level of democracy.

The northwest quadrant contains economies with higher-than-average QoG but lower-than-average DEM scores. Singapore and Hong Kong are the most conspicuous outliers, although one suspects that Hong Kong has slipped in both categories since these data were collected. Lower and farther to the left we find countries, including the United Arab Emirates and China, that are less well governed but still have higher-than-average QoG scores, even though they are even less democratic than Singapore or Hong Kong.

Part of the southwest quadrant lies above the trend line. Russia is found there, showing that although its quality of government is below average, it is nonetheless a bit better than we would expect from a country so undemocratic. The same goes for Turkey, which is the least-democratic country in the OECD. The part of the southwest quadrant that falls below the trend line includes countries often classified as failed states — Somalia, Venezuela and Haiti, among others.

Finally, the southeast quadrant includes countries that are more democratic than average but less well governed. Mexico is the only OECD country in this quadrant, with a QoG just below the average. The extreme outliers in this region are mostly very small countries, such as Cape Verde, Suriname and Madagascar. Two large countries, the Philippines and Ukraine, also fall into this quadrant but are closer to the center of the grid.

QoG and the Resource Curse

Next, we turn our attention to the relationship between QoG and GDP per capita. There is a strong tendency for wealthier countries to have better governments. For the 162 countries in our sample, the correlation of government quality with GDP per capita is an impressive 0.80.

The figure to the left shows a scatter plot of QoG quality against GDP per capita. (This time, the vertical axis shows QoG scores on a scale from 0 to 100 instead of the standardized units above or below the mean in the first figure.) It is clear at a glance that the countries fall into two distinct groups. To emphasize that pattern, the figure highlights 20 outliers as red dots.

Often referred to as “petro states,” all 20 are strongly dependent on gas and oil exports. Petro states outside the Middle East include Russia, Kazakhstan, Equatorial Guinea, Angola and Ecuador. (Norway, the United States and Canada, though all large oil producers, have lower exports as a percentage of GDP than do the true petro states.)

QoG is positively related to GDP within both the petro and non-petro states, but the trend line for the petro states lies well below that for the non-petros. In both cases, the fit between QoG and GDP per capita is pretty snug.

The relatively poor quality of government in the petro states illustrates the so-called “curse of riches.” Much has been written about this concept. Institutional weakness is one of the most frequently noted weaknesses of resource-rich countries. A popular explanation is that in such countries, politics is reduced to a struggle to capture the resource gravy (economists call them “rents”), whereas in resource-poor countries, governments must motivate people to create wealth and must provide the institutional framework in which they can do so.

Strictly speaking, there is nothing in the data shown in this figure that would allow us to draw firm causal conclusions regarding the relationship between wealth and government quality. Fans of free markets do not hesitate to express a belief that rule of law, protection of property rights and so on are the cause of superior economic performance.

However, there is probably feedback in the other direction as well: it seems likely that more prosperous and better-educated citizens in countries higher up the curve tend to have less tolerance for corrupt and unresponsive governments.

Bigger is Better?

We all know the hoary saying that “the government is best that governs least” (probably coined by the mid-19th-century magazine columnist John O’Sullivan — not John Locke, Henry David Thoreau or Thomas Jefferson). But the available data show a strong tendency for governments with high QoG ratings to be relatively large, as measured by the share of government revenue or expenditure in GDP. Across all countries in our sample, the correlation coefficient for QoG and government revenue share is a pretty convincing 0.80. If we omit the petro states, which have a dynamic of their own, the correlation rises to 0.87. Incidentally, if government expenditures rather than revenue are used as the measure of the size of government, the correlations are slightly lower but still strongly statistically significant.

However, a question immediately arises regarding the interpretation of these correlations. It seems that both QoG and the share of government revenue in GDP increase as GDP per capita increases. As a result, the simple correlation of quality with size of government could be spurious. Conceivably, even the sign of the correlation could be misleading. It might be the case, for example, that rich countries have high-quality government despite the fact that their governments are large, and that small governments might be inherently better, holding GDP constant.

One way to sort out this three-way relationship is to use “multiple regression” analysis. (I said it was getting easier, but we’re more or less still heading in the right direction.) Here, I try to find the statistically best fit comparing QoG with both GDP and the share of government revenue in GDP. It turns out that for the sample of non-petro countries, the analysis reveals a strong statistical relationship between the size and quality of government even when isolating GDP as a variable. In short, there is no evidence here for badboy anti-tax lobbyist Grover Norquist’s famous maxim that we can make government better by keeping it “small enough to drown in a bathtub.”

To be sure, the fit between QoG and size of government is not airtight. There are outliers, as shown in the figure below. The graph covers all non-petro countries, with selected outliers highlighted in red. Here, again, the axes show standardized values, with scales indicating standard deviations above or below the means for each variable.

Strictly speaking, there is nothing in the data shown in this figure that would allow us to draw firm causal conclusions regarding the relationship between wealth and government quality. Fans of free markets do not hesitate to express a belief that rule of law, protection of property rights and so on are the cause of superior economic performance.

However, there is probably feedback in the other direction as well: it seems likely that more prosperous and better-educated citizens in countries higher up the curve tend to have less tolerance for corrupt and unresponsive governments.

Bigger is Better?

We all know the hoary saying that “the government is best that governs least” (probably coined by the mid-19th-century magazine columnist John O’Sullivan — not John Locke, Henry David Thoreau or Thomas Jefferson). But the available data show a strong tendency for governments with high QoG ratings to be relatively large, as measured by the share of government revenue or expenditure in GDP. Across all countries in our sample, the correlation coefficient for QoG and government revenue share is a pretty convincing 0.80. If we omit the petro states, which have a dynamic of their own, the correlation rises to 0.87. Incidentally, if government expenditures rather than revenue are used as the measure of the size of government, the correlations are slightly lower but still strongly statistically significant.

However, a question immediately arises regarding the interpretation of these correlations. It seems that both QoG and the share of government revenue in GDP increase as GDP per capita increases. As a result, the simple correlation of quality with size of government could be spurious. Conceivably, even the sign of the correlation could be misleading. It might be the case, for example, that rich countries have high-quality government despite the fact that their governments are large, and that small governments might be inherently better, holding GDP constant.

One way to sort out this three-way relationship is to use “multiple regression” analysis. (I said it was getting easier, but we’re more or less still heading in the right direction.) Here, I try to find the statistically best fit comparing QoG with both GDP and the share of government revenue in GDP. It turns out that for the sample of non-petro countries, the analysis reveals a strong statistical relationship between the size and quality of government even when isolating GDP as a variable. In short, there is no evidence here for badboy anti-tax lobbyist Grover Norquist’s famous maxim that we can make government better by keeping it “small enough to drown in a bathtub.”

To be sure, the fit between QoG and size of government is not airtight. There are outliers, as shown in the figure below. The graph covers all non-petro countries, with selected outliers highlighted in red. Here, again, the axes show standardized values, with scales indicating standard deviations above or below the means for each variable.

Starting in the northeast quadrant, we are not surprised to find that four Nordic countries have the most prominently large and high-quality governments. New Zealand, Switzerland (abbr. CHE) and the United States have governments that are also very high quality, but only a little larger than average as measured by revenue. Italy and Greece also fall in the northeast quadrant, but below the trend line. Their governments, although large, are not as good as we would expect them to be based on their size alone.

The less heavily populated northwest quadrant consists of governments of higherthan- average quality and smaller-than-average size. These economies are given high marks on the economic freedom rankings by libertarian organizations, including the Fraser Institute, the Cato Institute and the Heritage Foundation. There is no doubt they are prosperous and well governed, especially the five that are highlighted as red dots. But the comparatively small size of their governments, as ranked according to the share of revenue in GDP, should not necessarily be accepted at face value.

Ireland is a special case, because more than a quarter of its GDP consists of the profits of international companies like Apple and Google, which use Ireland as a tax haven. If we were to adjust for this “phantom” GDP, Ireland’s ratio of tax revenue to income would be high enough to move it over into the northeast quadrant.

Hong Kong is another special case. For one thing, it is not an independent country, so it does not bear costs such as national defense and maintaining a full worldwide diplomatic presence. Also, as suggested previously, it seems likely that its high QoG rating is under threat as Beijing moves to snuff out its democratic institutions.

The other three East Asian economies highlighted in the northwest quadrant have a trait in common that tends to exaggerate the difference between the size of their governments and those of their European peers. Both the East Asian and the European countries highlighted in the figure have high-quality, comprehensive health care systems. But they are financed differently. European health systems are supported directly by taxes, so the costs of maintaining them are reflected in the share of government revenue in GDP. In contrast, their East Asian counterparts are based on compulsory private insurance.

Replacing a tax with a mandatory premium paid to a regulated private insurance company means that the stream of payments bypasses the government budget. But is there any real difference between a compulsory premium and a tax? Some might say there is not. (Indeed, the question is a great point of contention in the Republican obsession with ridding the country of Obamacare.)

Turning now to the southeast quadrant, we find countries that have large but lowquality governments. There are relatively few of these. In some, like Serbia and Ukraine, citizens pay higher-than-average taxes without getting much in return by way of quality government. Others, like South Sudan and Eritrea, display even worse governance without even the compensation of low taxes. China also falls into this quadrant, but it is very close to the center.

Finally, the southwest quadrant contains a large number of countries where taxes are low but the quality of government is low, too. Yemen, Haiti and the Democratic Republic of Congo offer little for fans of small government to admire. Nigeria and a few others at least manage to make it above the trend line, showing slightly better governance than we would expect, given their poorly funded public sectors.

QoG, Human Needs and Freedom

So far, this statistical world tour has focused on factors that tend to strengthen the institutional foundations of governance. However, quality of government is not an end in itself. Political theorists and activists throughout history have promised us that institutions like rule of law and strong property rights are good, not just for their own sake, but because they make life better. This section tests the proposition that basic human needs are better satisfied and personal freedom is greater where QoG is high.

For a measure of the degree to which human needs are satisfied, I again draw on the Legatum data. The result is a variable for human needs that I call NEEDS, which is derived from data on health care systems, longevity, nutrition, shelter, protection from crime and other indicators.

A statistical analysis relating NEEDS to QoG, controlling for GDP per capita and the size of government, generates a very tight fit. It is possible to tease out the separate impacts on NEEDS of QoG and the size of government by using statistical methods that show what would happen if just one of the independent variables were changed while the others were held constant. NEEDS, it turns out, is also statistically linked to government revenue.

In short, we find that:

• The satisfaction of basic needs increases with higher-quality and larger size of government.
• QoG has a significant independent effect on the satisfaction of needs, even controlling for the effects of higher GDP and higher government revenue.
• QoG has a somewhat stronger effect on need satisfaction than does size of governent.

Finally, we turn to the relationship between quality of government and personal freedom. To measure personal freedom, or PFREE, I again drew on the Legatum data, this time using indicators related to freedom of assembly and association, freedom of speech and access to information, absence of legal discrimination, and a group of indicators Legatum calls “agency” that measure individual rights, freedom of movement, women’s rights and forced labor.

Here the statistical analysis suggests that QoG explains a lot of the country-to-country variation in PFREE, even when controlling for differences in GDP and the size of government. The partial correlation of government revenue with PFREE — controlling for GDP and QoG — is modest, but doesn’t seem to be a statistical accident. Personal freedom tends to increase as the size of government increases. Overall, though, the quality of government has a much stronger effect on personal freedom than does the size of government.

Now Let Us Praise QoG

This statistical portrait gives us a better idea not only of what quality of government is, but why it is worth striving for. On the whole, countries that score highly on measures of integrity, impartiality, competence and rule of law tend to be more democratic. To no one’s surprise, we have found that rich countries tend to be better governed than poor countries, although countries that are heavily reliant on natural resource wealth are distinct outliers in that regard.

Finally, and more controversially, better governments tend to be bigger, as measured by the size of their budgets relative to GDP. The data hardly support the idea that slashing government budgets is a reliable way to improve government quality.

We have also found that life is better in countries that have high-quality governments, and even more so when those governments are both higher quality and larger. That is true both when a “better life” is defined in terms of the satisfaction of basic human needs and when it is defined in terms of human freedom.

To be cute about it, government is best that governs best.