Just two years removed from a $27 billion budget gap, Texas lawmakers find themselves with an $8.8 billion surplus for the next budget cycle.

And the Lone Star State is not alone. After rounds of painful cuts, states are in their best fiscal positions in years. At least eleven states besides Texas expect significant surpluses, including California, which faced a $60 billion deficit only three years ago.

Having plummeted by $78 billion between 2008 and 2010, state general fund revenues are expected in 2013 to surpass pre-Recession levels for the first time, according to the National Association of State Budget Officers. As a whole, state budgets are in better shape than they have been in quite some time. State spending is growing at a nominal rate of 2.2 percent, the third consecutive year of increase, and 26 states are now spending more than they did in 2008.

Figure 1State Budgets are Growing Again

At first glance, this seems like a good thing. But the fiscal strength of a state budget is not simply a question of the size of the plus or minus sign that goes with it. At present, there are at least three reasons to hold our applause on this state resurgence: (1) recent revenue boons are likely transitory, (2) major imbalances loom, and (3) predicting the future is always tricky.

Assessing the state of state budgets is challenging. The landscape is a patchwork. Some are strong, and some are weak. Some are large, and others small. Some are expanding, and others shrinking. While some states have grown their economies strong through agriculture (Iowa and Nebraska), others have credit ratings that resemble a hockey box score (Illinois). Some states have become industrially resurgent (Oregon and Tennessee), yet others still appear to be a bad bet (Nevada). And while some states enjoy oil-induced economic booms (North Dakota and Texas), others are still stuck in the mud (Mississippi).

For states currently in the black, the unexpected cash is not an unambiguously good thing. While it may be true that surpluses are preferable to deficits, both can cause economic harm—especially when the economy is in the early stages of a fragile recovery. Politically, surpluses complicate budget battles and create false senses of security—encouraging the same imprudent choices that got us here in the first place. And they should be a blow to our egos: surpluses are just more proof that we are bad at making forecasts.

Getting state finances right is critical for America’s economic health. State and local governments account for 12 percent of GDP. They employ nearly one out of every seven American workers. While the attention is on Congress, it is lawmakers in statehouses around the country who are making key decisions about our fiscal future. It is essential to understand which states are doing it right, which states are doing it wrong, and what separates the strong from the weak.

Fiscal Strength Defined

Before we can judge whether a state has a good or bad budget, we need to understand what is meant by fiscal strength. This step is not as straightforward as you would think. In particular, analyzing state fiscal strength encounters four challenges: (1) measurement is ideologically subjective, (2) budget rules limit the range of possible outcomes, (3) wide diversity in situations means that it is difficult to compare states, and (4) federal policymaking cannot help but influence state performance

The ideological nature of measurement

First, it is important to clarify that fiscal strength is not the same thing as economic strength. The two are closely related—all else equal, a strong economy makes having a strong budget considerably easier—but distinct.

Economic strength refers to economic conditions within a state. Is unemployment high or low? Are incomes rising or falling? Is the housing market booming or busting? Is the workforce productive, happy, and mobile?

Fiscal strength has to do with the health of state government finances. Are revenues rising or falling? What about spending? Are the trends sustainable? Are long-term liabilities balanced by long-term assets? And perhaps most importantly, is the government providing the quantity and types of services taxpayers expect and desire?

By way of example, no amount of oil wealth (economic strength) can make up for irresponsible governance or playing politics with taxpayers’ pocketbooks (lack of fiscal strength).

Unlike economic strength, where “good” and “bad” are typically uncontroversial (low unemployment is better than high unemployment; high GDP is better than low GDP), what constitutes fiscal strength is somewhat a matter of political perspective and policy preferences.

Increases in spending from one year to the next may be seen as a sign of fiscal health by some observers, while others will see them as a dangerous encroachment of government upon the private sector. Similarly, some will see higher tax receipts or borrowing as a signal of a strengthening economy; others will see them as needlessly crowding out economic growth.

Other indicators may be equally problematic. A fully funded pension system can be seen as evidence of fiscal responsibility—but in a sluggish economy, long-term discipline may compromise necessary short-term stimulus. A well-endowed rainy day fund can provide critical cushion against future economic shocks—but try to rebuild your reserves during a recovery and you may find yourself back in recession.

The narrow rules of state budget finance

States are more challenging to analyze than the federal government because state constitutions confine fiscal policymaking within fairly narrow parameters. Unlike the federal government, which can operate with budget deficits, states (excepting Vermont) are required to balance their budgets and thus have limited fiscal flexibility.

Whereas almost any discussion of national finances begin with surpluses and deficits, for state purposes, notions of black and red are largely irrelevant. The result is a classic “begging-the-question” predicament: because states are legally obligated to achieve budget balance for operating expenses, balance will be achieved. Any departure from perfect balance is a result of inaccurate projections or poor planning. Bottom lines measure forecast accuracy more than they measure financial performance.

Of course, states can (and do) engage in all sorts of budget gimmickry to paper over deficits or borrow from their future selves. Shifting a payroll by a matter of days can save a state millions. Similarly, surpluses can be hidden from public view by prepaying debts or accelerating other multiyear obligations.

But the potential variation in state budgets is smaller than that in federal budgets—and this small range of outcomes makes identifying differences between “good” and “bad” state budgets more difficult. In theory, one could compensate by tracking the magnitude and impact of budget cuts, but distinguishing between reductions, restorations, additions, economic conditions, accounting gimmicks, tax changes, and broader structural forces—over multiple budget cycles—becomes increasingly difficult, especially when budget changes can be as political in origin as they are financial (and are typically memorialized in impenetrable appropriations bills). Tracking a dollar that was not spent is more difficult than accounting for one that was.

The problem of cross-state comparability

One of the hallmarks of American democracy is the independence guaranteed to individual states. As a result, states have strikingly divergent cultures, values, economies, and political preferences. These varied attitudes and ways of life cannot help but play out in state budgets. And these attitudes change over time. Comparing apples and oranges can yield misleading conclusions.

Fortunately, there are two pretty simple ways to make state budgets comparable. The first is to look at percent changes. In this approach, the benchmark for a state’s performance is its own past. It does not matter if a state spent $1 last year, or $1 million; what matters is the difference between this year’s spending and last year’s, divided by the level of last year’s spending. Assuming state spending preferences are relatively constant over short periods of time, we can interpret a percentage increase in expenditures or revenues as fiscally positive and a decrease as fiscally negative. One state may have a preference for higher taxes and spending than another, but measured in percentage terms, what matters is increases or decreases, scaled by their reference period. Using this method, the size of a state does not matter: percentages are unitless measures.

Percentages can also be useful as static measures, rather than as measures of changes over time. For example, the absolute size of a rainy day fund is relatively meaningless. But, put in the context of state spending—rainy day fund size as a percentage of general fund expenditures—it becomes a useful measure of state reserves.

A second way to put states on equal footing—to relativize their budgets—is scaling. Here, measurements are expressed as ratios, which are like percentages, but relate two distinct quantities. Per capita scaling expresses fiscal quantities in terms of population—for example, spending per capita. It takes only a moments’ inspection to recognize such a scaled measure is much more relevant for determining standards of living than spending totals.

For other indicators, scaling by aggregate economic activity—the most common measure of which is gross domestic product, or GDP—makes more sense. Debt burdens, for instance, are better expressed as debt-to-GDP ratios; $1 billion of debt implies very different things in a trillion-dollar economy than it does in a billion-dollar one.

The impact of the federal government

In our federalist system, the relationship between the federal government and state spending is a close one—a third of state spending consists of transfers from the feds. Because the federal government is free to borrow, and therefore can spend as much as it likes, it can counteract state downturns with stimulus, as it did during the Great Recession. These boosts, while absolutely crucial for state well-being, can make state budgets look artificially good. In a similar way, federal austerity can make overall state finances look artificially bad. The problem in both cases is that state budget-making is conflated with federal budget-making (or lack thereof), and state fiscal health is bound up in federal fiscal health. If evaluating state fiscal strength in and of itself is our goal, one way to minimize federal influence is by focusing on state “general funds,” which in most states in the name given to the set of financial accounts comprising government activities financed by state taxes.

Introducing a Fiscal Quality Index

Making sense of this fiscal patchwork requires first taming the complex nature of fiscal assessment. To manage these complexities, I have created a “fiscal quality index,” which includes nine measures of fiscal strength, weighted equally. (See Table 1 for an overview.)

To make states comparable—and to control for unobserved state characteristics, like fiscal conservatism or past practices—the index emphasizes percent changes for indicators where baseline levels are important. And since at this particular moment in time we are primary concerned with the recovery from the Great Recession, I have focused on how states have changed since 2008, the last year before many states started feeling the deep effects of the economic contraction. Along similar lines, for indicators where scaling plays a role—that is, where size matters—magnitudes are measured in per capita terms.

The index also strives for ideological balance. Four indicators are likely to correspond more with the Democratic notion of fiscal strength: percent change in general fund revenues between 2008 and 2013, percent change in general fund expenditures between 2008 and 2013, per capita general fund spending in 2013, and percent change in per capita general fund spending between 2008 and 2013. Four indicators are likely to appeal more to Republicans: rainy day fund quality, pension fund quality, credit rating, and debt-to-GDP ratio.

The index includes a final indicator—whether or not a state is expecting a 2013 surplus—because surpluses and deficits are perhaps the most common way of discussing fiscal health. As I’ve discussed above, state surpluses are strength in a very narrow sense: the government takes in more revenue than it plans to spend, giving the state flexibility in the near-term.

Besides striving for ideological impartiality, the index is also designed to be balanced temporally. That is, five indicators reflect current fiscal conditions (revenue percent change, expenditure percent change, surplus, per capita spending, and per capita spending percent change), while four are oriented to the long-term (rainy day fund quality, pension fund quality, credit rating, and debt-to-GDP ratio). Just as a tanning salon aficionado may look good today while imperiling her future, states that look good today may be doing so only through compromising their futures. When assessing fiscal health, we must care about both now and later (even if our political institutions are ill-incentivized to do so).

Table 1. Fiscal Quality Index

For those indicators that measure binary yes/no assessments, such as whether the benchmark for pension fund quality is met, a state received a score of 1 if the condition was present and 0 if not. For continuous indicators, such as expenditure percent change, a state received a 1 if its variable was above the median for all states, and 0 if it was below. States receiving a score of 6 or higher across the 9 indicators were considered “strong states”; 16 states met this threshold. The remaining 34 states were classified as “weak states.”

Strong States versus Weak States

Using this system, I was able to create a picture of what strong states look like, in terms of geography, population, demographics, economics, and politics. Before I discuss the results, two words of caution.

First, the sample size is small, consisting of just one observation for each of the fifty states, using the most recently available data for each characteristic, or where appropriate, its change since the beginning of the Great Recession, in 2008. As a result, the statistical power of any comparison is low, meaning it is likely some true variations between strong and weak states are masked by large margins of error. At the same time, small samples are easily corrupted by large outliers.

Second, the relationships between state characteristics and fiscal strength are not necessary causal. Just because a fiscally strong state has a larger than average number of senior citizens does not mean senior citizens make states fiscally stronger. Nor do such relationships necessarily offer policy prescriptions. States with high GDP may be fiscally strong, but that correlation alone does not tell you how to increase GDP. Without a theoretical model designed to explain cause and effect, it is best to interpret the patterns in the data as associations. My findings are descriptive rather than explanatory—places for beginning future research instead of launching points for wild claims.

Nonetheless, strong states are different from weak states in several important ways, as detailed in the following table. On a whole, strong states are less poor and better educated. Their residents tend to be older and whiter. Health insurance coverage is more widespread. Geographically, they tend to cluster in the Midwest; in addition, Northeastern states tend to be weak. Perhaps most important, strong states have strong economies, with substantially lower unemployment and considerably higher GDP on a per-person basis. Strong states also tend to rely less on government as a source of employment (public sector jobs were among the hardest hit by the recession). There is also some evidence their economies rely less on low-wage jobs such as in the food service industry.

Table 2. Strong States vs. Weak States

Just as interesting is what is absent from the table. Strong states and weak states are not significantly different in terms of population size or density, share of foreign-born residents, safety net program utilization, housing market conditions, economic inequality, or perhaps most surprisingly, political party dominance and polarization. And although there are some small systematic differences in industry composition, by and large there are few persistent patterns in the economic makeup of strong and weak states.

Then again, it is important to recall that my index was created in a particular way and was designed to control for differences in baseline state characteristics. As a result, we should perhaps not be surprised that strong states and weak states are not systematically different politically or industrially, since for the most part, we are measuring a state’s performance, in terms of spending, revenues, rainy day funds, and the like, relative to its prior history or relative size. It is equally important to remember there is more than one way to skin a cat: no one party or collection of industries holds a monopoly when it comes to strong fiscal performance.

Even more to the point, as I cautioned above, the findings presented in this table are correlations: they are patterns, made meaningful only in an appropriate theoretical context. Causality can run both ways—a well-run state can promote economic growth, just as economic growth can strengthen state budgets. And correlations provide no protection against an econometrician’s nemesis: omitted variable bias.

Omitted variable bias occurs whenever a relevant factor is excluded from a model; its exclusion means the relationship detected in the data is a mismeasured one. For example, the finding that physically large states are fiscally strong may simply be an artifact of a few large states happening to be located on top of fortuitous deposits of a valuable natural resource (say, oil). The presence of oil, not the state’s geographical expanse, is the source of its strength.

All Else Equal

One way to address the threats posed by spurious correlations is multivariable regression, which can examine relationships among variables of interest while holding constant other observable factors. With potentially confounding factors held aside, we can be more confident in the validity of our findings. However—and this is an important point—although multivariable regression is superior to simple correlations, it still does not ensure causality, because (without additional theoretical and methodological procedures) it cannot control for unobserved factors that may contaminate our results.

(Warning: The next two paragraphs are fairly technical. The mathematically disinclined reader may safely skip ahead to the results.)

My regression analysis included both linear probability and logit models. Linear regression using ordinary least squares (OLS) is the most commonly employed econometric method. OLS regression examines the strength of linear relationship between an outcome of interest (here, whether or not a state is fiscally strong) and one or more potentially explanatory variables (for example, age, race, or educational attainment). The math is relatively simple and the assumptions are intuitive. By finding the line that most closely fits the data points in our sample, we can make predictions about how varying the explanatory variables will affect the outcome of interest. In effect, we are minimizing the error between our “best-fit line” predictions and the data we actually observe.

However, linear regression has a serious shortcoming when the outcome of interest is binary: the best linear fit of the data can predict outcomes that fall outside the mathematically viable range. That is because when the outcome variable is binary—meaning it can take on a value of 0 or 1 only—unit changes in the outcome as an explanatory variable changes are changes in the probability of the outcome occurring (which by definition can also vary only between 0 and 1). But linear probability models—because they seek only to identify the most accurate association between the data points—can yield predictions outside the 0–1 range, which are nonsensical (it is impossible to have a probability greater than 100 percent or less than 0 percent). The solution is to impose a constraint on the model to force the outcome to be limited between 0 and 1. Functions particularly well-suited to this are cumulative probability distribution functions, of which the logistic distribution utilized by the logit model is one. Rather than minimizing the errors in the data, logit regression uses a mathematical approach to identify a model that maximizes the likelihood of finding the relationship we observe in the data.

These technical considerations aside, the results of my regression analysis confirmed many of the correlations discussed above—and unearthed a couple of additional relationships. The key results are in the table below. (As you will notice, the findings are fairly consistent across the models, which should increase our confidence that something useful is being measured.)

Table 3. Traits of Strong States

Controlling for a state’s baseline economic strength (2008 per capita GDP) and population, I found that the Northeast is the weakest fiscal region by far. Midwestern and Southern states have especially outperformed their New England and Mid-Atlantic brethren in recent years. Fiscally strong states are also older, in the sense of having a smaller share of children, and somewhat whiter. They are better educated and have fewer high school dropouts. In addition, they have significantly higher rates of health insurance coverage; interestingly, these differences in health insurance were after controlling for poverty—and they were more pronounced than were differences in safety net program utilization. This result should make opponents of Obamacare think twice. There is also some evidence that strong states have fewer poor people. But Democrats should not get cocky, either: after controlling for these factors, Republican states (as measured by the governor’s party) were fiscally stronger than Democratic ones.

And the Winner Is…

So who is number one? According to my scoring system, the fiscally strongest state in America is a four-way tie between Indiana, Iowa, North Dakota, and Tennessee. Each state scored 8 out of 9. Tied for second, with scores of 7, were Delaware, Nebraska, and Ohio. Nine states scored 6—a favorable rating on two-thirds of the indicators.

The makeup of top performing fiscal states offers a lesson in diversity. Because the index is a qualitative aggregate—all indicators were either binary or transformed into binary criteria—it can mask variation among states with similar scores.

Among the seven top-tier states, North Dakota stands out as a clear front-runner. With its economy bolstered by a huge oil boom, North Dakota avoided the Great Recession and continues to grow. The “roughrider” state boasts the lowest unemployment rate in the nation and the third-highest per capita GDP. With this oil wealth has come strong growth in other sectors, including real estate, construction, and transportation. This economic strength has translated into fiscal strength: it has the fastest growing expenditures of any state and the second fastest growing revenues. Its per capita spending and rainy day fund also rank in the top ten. Its weakness, which is slight, comes with pension funding, where it ranks twenty-fourth.

Delaware and Nebraska are also standouts. Each ranks in the top fifteen on six out of seven rankable dimensions of fiscal quality (surplus and credit rating, as binary variables, offer little variation between states and so are excluded from the rankable list). Each struggles on one dimension—Delaware with debt, and Nebraska with per capita spending. Both are also strong economically: Delaware has the highest per capita GDP and Nebraska has the second lowest unemployment. But how they have gotten there is different: Nebraska’s economy, known for farming, is widely diversified, with strong manufacturing, real estate, finance, retail, transportation, health, and government sectors. Delaware, by contrast, thrives mostly on finance, which accounts for more than a third of its economy—the largest share of any state.

Indiana, Iowa, and Tennessee—the other three states that share first place with North Dakota—offer a different lesson. All are mostly unexceptional. Tennessee has the lowest debt-to-GDP ratio, Iowa has the ninth-best rainy day fund, and Indiana is eighth in revenue growth—but all fail to crack the top ten on any other measure. However, they are consistently above average: all are in the top half of states on six out of seven rankable indicators. In other words, they are solid rather than spectacular.

Figure 2 How Does Your State Stack Up?

Table 4. How Does Your State Stack Up?

Another important finding is that fiscal quality does not necessarily go hand-in-hand with economic quality. For example, by some measures, Texas has one of the fastest growing economies in the country—but in terms of my fiscal quality rankings, Texas is tied for just seventeenth. One of the fiscally strongest states—Indiana—ranks just thirty-first in per capita GDP and thirty-sixth in unemployment. Another, Tennessee, has the sixth-lowest median household income and the twelfth-highest poverty rate. You could argue my index is flawed (most things are)—but that would ignore the larger point that fiscal strength and economic strength are not the same thing. (If you would like to explore an alternate system, the folks at 24/7 Wall St. have devised their own rankings.)

Similarly, it is interesting to note that states that are strong on one fiscal quality indicator are not necessarily strong on others. If you examine the top ten states for each of the fiscal quality indicators listed below, you will notice there is definitely overlap—but as much as you might expect. Massachusetts has the fastest growing revenues, but the worst debt. Alaska has the highest per capita spending and the largest rainy day fund, but has the third-worst revenue growth and the seventh-worst debt. In fact, across all nine components of the fiscal quality index, the average correlation is 0.31, which indicates that although some indicators are closely related (for example, growth in expenditures and per capita expenditures are tightly linked), other indicators have almost no relationship (for example, revenue growth and debt-to-GDP ratio). This suggests that in the absence of rapid economic growth, achieving fiscal objectives requires trade-offs.

Table 5. The Top 10s

What Did We Learn?

The variation between high-performing fiscal states—geographically, culturally, politically, and economically—underscores what is perhaps the central finding of this analytical enterprise: there is no one “right” way to achieve fiscal health. Rather, states achieve fiscal strength by following some basic budgeting best practices within the context of local conditions, traditions, and preferences. Fiscal success is more gymnastics than it is golf: positive outcomes can be achieved with diverse styles. Some of it is luck, some of it is skill, but a lot of it comes down to pursuing creative strategies with determination, discipline, and a bunch of practice.

Budgeting is also a lot like poker: you are forced to play the hand you are dealt, but it is your strategy, intelligence, respect for the rules of the game, and knowledge of the mathematical fundamentals that determine victory and defeat. There is a place for taking risks and making bold bets, but there is also a place for ever-tempered restraint. Your opponents’ hands matter, too—whether these adversaries are economic conditions or political winds—but clever tactics and prudent planning make long-run success more likely. There is no single way to win, but there are better and worse ways of playing.

Of course, debate over what constitutes the “right” way to play is one of the reasons why budget battles are so common. Should we expand health insurance or hire more teachers? Reduce college tuition or increase public assistance benefits? Would we rather pave roads or keep parks open later? Build bridges or collect garbage twice a week? Do we need more affordable houses or fire houses?

Equally controversial, as we have seen, is defining what “success” looks like. Should we raise taxes or cut wasteful spending? If we raise taxes, do we apply it to income, property, or the things people buy? If we cut spending, do we close prisons or reduce benefits to the unemployed? If we find ourselves with surpluses, do we save it for pensions or spend it now to stimulate the development of new industries? Is is the state’s job to create more jobs—or get out of the way and let small businesses do their thing? Unlike poker, or even gymnastics, the objective is not objective. Budget balance does not mean budgets are balanced. In this regard, judging fiscal policy is more like being a movie critic than an umpire.

But what is beyond debate is that state fiscal strength matters. With so much of our political discourse today dominated by ideological battles over the public good, it’s easy to lose sight of the fact that, without strong state governments, we would have far fewer public goods—the types of non-rival, non-excludable assets and services society would underproduce if not for government prodding. The reason is positive externalities: because more benefits accrue than are compensated by the market, too little of the goods are produced. Among the best examples are police officers and fire fighters. Once a first responder force exists, everyone is safer, regardless of their level of taxation; nor does one person’s safety reduce anyone else’s. The same goes for clean air and good water: no one can be excluded from enjoying the benefits, once they exist, and one person’s consumption does not preclude another’s.

Not all case are so clear cut, but most of what state governments do involve some degree of positive externalities. The benefits of a highly educated population extend far beyond the students themselves. Highways and public transportation are lubricants of commerce and pathways to higher standards of living. State parks and greenspaces offer all sorts of intangible advantages. And one need not be an infectious disease specialist to appreciate the group rewards of health insurance and easy access to medical care. Indeed, some economists go so far as to say public goods are the justification for government existence.

With federal budget debates increasingly intractable, and austerity seemingly inevitable, the importance of state fiscal policy is becoming even greater. If states were once conceived as laboratories for democratic experiment, today they are the point of purchase. As Congress cuts back, it is states that will be forced to step forward. If fiscal policy is a poker game, it’s one where our chips are constantly at the center of the table. But when we focus on winning the pot, we miss the larger objective: making sure the pot keeps growing. Whether we like it or not, the states are all in.

Fiscal fights are substantive, not semantic. Budgets can be government at its most contentious, and with good reason: they are at the heart of what government does, the link between policy theory and practice. As technically impenetrable as they may be, appropriations are societal values in dollar form—the expression of collective preferences through the allocation of resources. Politics is about who gets what; budgets are how we get there. The practice of fiscal policy is a much a process as it is an outcome: the one thing all participants in the state fiscal game can agree on is that they want the competition to continue. The journey is the destination. We live to fight another day.