Dark Clouds

In my view, there was nothing earthshaking in the November December forecast that MMB released a couple of weeks ago (12/4/2024), unless you’re a Pollyana who expects ever rising surpluses. Admittedly, recent forecasts may have conditioned one to expect that, along with the consistent collections above forecast.

The forecasted surplus for the 2026-27 biennium was down, despite the economy continuing to chug along. But the change was modest with higher forecasted spending (0.9%) and lower revenues (-1.5%) compared to the end-of-session numbers. The net change was a reduction of $1.6 billion or -2.4% relative to the previous forecast of either revenue or spending. It’s significant but well within the swings one expects over an 8-month period. Unfortunately, both revenues and spending went in the wrong direction, so they did not offset each other. That’s part of a bigger story (more below).

The national economic outlook is about the same or slightly better than last February (see graph on page 13 of hard copy), although the Minnesota situation looks a little worse (see graphs on pages 33 and 38 on labor force participation and wage growth). As is often the case, a more important driver of revenues is the detail provided by the new sample of income tax returns. The real story tends not to be in net taxes paid (i.e., monthly collections compared to forecast) or small changes in economic growth projections, but rather how the mix of income (wages, interest, dividends, capital gain, etc.) reported on returns changed and how that change interacts with the forecast of economic variables in MMB’s models. That requires having the income detail from the sample of returns. In this case, those new data pulled down the forecast, even though economic growth estimates were about the same. A bright side is that a rising forecast of corporate profits partially offset the individual income tax decline (caveat: corporate profits are volatile, and forecasts of taxes on them are subject to more uncertainty than sales or individual income tax revenues).

So, there remains a surplus forecast for the coming biennium, it’s just smaller.  With split control (thanks to a tied House, at a minimum pending election contests) that should augur for a status quo 2026-27 budget with the usual fighting about how to cut up the pie.

Dour outlook

However, if that were the main takeaway from the forecast, it would be short sighted. The key issue is the unbalanced budget structure that is most starkly revealed by the planning estimates for the out biennium. The planning estimates show a $5.1 billion budget shortfall for FY 2028-29. Even worse, the difference between revenues and spending is a negative $5.7 billion because the forecast assumes about $600 million carries over from the previous biennium. The $5.7 billion is the best measure of structural imbalance. You can only spend down prior surpluses for so long; revenues need to equal spending. Even if you ignore discretionary inflation in spending (i.e., assume policy makers will hold nominal spending constant, cutting overall spending in real terms), the imbalance only drops to $3.5 billion. See the table on page 11.

The planning estimates, it should be noted, are subject to a much wider confidence interval than the forecast for the coming biennium. The greater uncertainty results both from the higher degree of difficulty in forecasting farther into the future and from the various shortcuts or rules of thumb MMB uses for planning estimates (after all, these numbers are not being used directly to enact a budget, so less effort is appropriate). But they are our best guide.

How did we get here?

I wasn’t working for or hanging around the legislature when the crucial decisions were made in 2023. So, the following is speculation (both as to the intention behind decisions and to their actual effects). Take them in that vein. In any case, I think the big structural imbalance results from basic economics – state revenues tend, over the long run, to grow more slowly than auto-pilot spending does. The big surplus in 2022-23 could have covered that gap (assuming I’m right) for a very long time. But politics (left, right, and center) would never allow that: they require policymakers to spend it or give it back through tax cuts. Of course, that is what their priors tell them to do anyway.

The interesting financial or economic question is to what extent the 2023 decisions exacerbated the problem. That would require careful and detailed analysis of the growth rates (elasticities) of the 2023 permanent changes in revenue and spending. That’s beyond my capacity or desire to do. Most of the issues are on the spending side, which I never worked on and in which I have no expertise. My instinct on the revenue side is that the changes slightly increased elasticity and volatility, but it’s just intuition, and the main story is on the spending side.

In any case, what follows are some of my observations about the background situation and decisions made in the 2023 legislative session. Again, they’re largely just my speculations as an outsider but someone who observed legislators and governor deal with analogous but less dramatic surpluses (1979 and 1999-2000 come to mind).

• In enacting the 2024-25 budget, Governor Walz and 2023 legislature had a $17.5 billion surplus (table on page 11). That resulted mainly from a combination of (1) direct federal pandemic aid to state and local governments that offset spending normally financed with own-source revenues and (2) federal pandemic stimulus paid to businesses and individuals (e.g., PPP loans/grants, rebates, etc. etc.) that indirectly stimulated economic activity and state revenue growth. Most forms of the aid, although clearly income to the recipients, were not taxable, so the effect on revenues was indirect but substantial. They stimulated growth in wages, capital gains, purchases subject to sales tax, etc., while federal aid helped pay for state spending. (Note: decisions made by both Republican and Democratic administrations and Congresses in response to the pandemic were instrumental in effecting a national economic recovery. The rest of the developed worlds provided less aid and generally had more anemic recoveries and similar or higher inflation.)
• State policymakers in 2023 recognized that most of this revenue was one-time and they used much of it for one-time purposes. But estimating the permanency of revenues is difficult, highly inexact. And policymakers prefer to make permanent policy changes, not give out rebates or enact one-time spending. That uncertainty and preference encourages, even if subconsciously, shading toward or outright tilting toward permanent policy changes. So, human psychology says errors will tend to be made in that direction, even if one assumes the best of intentions and, of course, rationalizations that further one’s goals are usually easily at hand.  All of that encourages tilting toward too much permanency whether in tax cuts or spending.
• The MMB forecasts cover a 4-year period. That means in the 2023 session, the last year in the budget window was FY 2027). In a textbook case of trying to enact sustainable permanent changes, one would aim to have revenues and spending in FY 2027 be about equal. A high-level look at the end-of-session estimates suggests that $300 million or so of it was spenddown of prior budget balances. So, it wasn’t wildly structurally unbalanced. The end-of-session spreadsheets don’t formally report structural balance, of course, and I’m off, for sure, maybe by a lot. By the November 2023 forecast, the structural imbalance (in FY 2026-27) was $1.4 billion (i.e., more than doubling on a biennial basis). One year later, it essentially doubled again to over $3 billion for FY 2026-27. November 2024 forecast (table p. 11). The $5.7 billion number is for the FY 2028-29 biennium.
• This is pure speculation on my part, but I suspect that in making a judgment about the magnitude of permanent budget changes that could be made during the 2023 session, policymakers thought that revenues would grow at least as fast as spending and probably a little faster. That made them comfortable with enacting more permanent changes. They likely thought revenue growth would outpace spending growth and with modest budgetary adjustments in the 2025-2027 sessions, everything would be okay. (Other possible, perhaps plausible explanations: They ignored it; they decided to strike while the iron was hot and put their long-desired policies in place while they had a trifecta, imbalance be damned; they expected they would be able to make modest adjustments in the programs and/or increase taxes to cover shortfalls in 2025 and later; etc. My experience is that legislative leaders and governors – the ones who make these decisions – are sober minded and generally prudent about fiscal balance, even though the need to resolve legislative negotiations pull them in the opposite direction. Many backbenchers are not, of course.)
• The opposite is forecast to happen. Spending is projected to grow faster than tax revenues and not by a small amount. For example, the table on page 54 of the forecast document shows biennial general fund spending (FY 2028-28 over FY 2026-27) growing by 4.7% annually compared with 2.9% for revenues.  That’s a whooping 1.8 percentage point differential per year or a $2.6 billion gap for the biennium (some spending in the prior biennium was covered by carryover surplus!). If accurate that is not sustainable and portents difficulty choices ahead.

Random Observations

Some of my general reactions are:

• It’s important to reiterate that the planning estimates are very approximate and uncertain. That does not mean they can or should be ignored, but it makes doing so easier to rationalize. I hope that doesn’t occur, but I won’t be surprised if it does (more on the political dynamics below).
• As I noted above, my view is that state tax elasticity is lower than spending elasticity for most periods of time. That means if they thought (as I think they likely did) that revenue growth would be higher, they fell prey to recency bias from the parade of surpluses in the last string of forecasts.
• Future events could make the situation much worse. Two good possibilities: (1) a recession or just a mild slowdown in economic growth (tariffs and a TCJA extension financed with debt are growth inhibiting) and/or (2) Congress cuts Medicaid in a way that reduces state reimbursements. The former are almost certain policies of the incoming Trump administration; it’s just a question of the configuration. The latter seems especially likely to me given the need to find budget offsets for a TCJA extension and other tax cuts and the various paths available for cutting aid primarily to blue states (e.g., cutting the federal percentage of the ACA expansions, which multiple red states have still not adopted, or rewriting reimbursement formulas that make them more sensitive to capacity to pay which is consistently higher in blue states). MCFE has pointed out those risks.  I consider some version of this almost inevitable unless the Republican Congress turns out to be totally dysfunctional (unlikely but a possibility). The $5 billion gap growing to $10 billion is plausible under any of a variety of scenarios, none of which is just a tail risk.
• The 2023 tax changes made the state’s tax revenues more dependent on taxing capital income by raising taxes on corporations and capital gain and other passive income of very high-income individuals (>$1 million). This will typically make state tax revenues subject to bigger swings when a slowdown or actual recession occurs. The desired benefits of progressivity come with costs of less reliability and more volatility. (I’ll write more about this in the future.)
• The prudent response is to begin immediately, even if gradually, to realign spending and revenues. That means trimming spending and enhancing revenues, to use some euphemisms, in the coming biennial budget. Putting off those decisions will just make the day of reckoning in the 2027 session more difficult. That is what a well-managed private organization, whether for-profit or nonprofit, would do.
• State government budgeting, however, is determined by politics, not prudent management principles. And the political considerations push against making prudent, but painful decision, certainly sooner than absolutely necessary. Consider the recent presidential campaigns: The country is facing unprecedented peacetime deficits and debt, yet both campaigns pitched free stuff to the voters, a true plethora of tax cuts from Trump and a somewhat lesser of amount of new spending and targeted tax cuts from Harris. Turning to the state situation, all legislators and the governor are on the ballot in 2026. That’s bad. My experience is that in biennia in which the Senate is not on the ballot, they show more prudence and political courage than the House; in years where they are on the ballot, even less than the House typically does. That’s likely an artifact of the reality when you run less frequently, doing so produces higher anxiety. In any case, political courage for painful decisions is likely to a low ebb in 2025-26.
• In a Panglossian world, the split in control would provide cover for making painful decisions. The two parties would agree that the bleak budget situation requires inflicting pain and if they join forces and agree on that, neither will be held solely responsible. They could fight (campaign over) their philosophical principles about how much and what government should be doing but would recognize that the 2027 legislature should not be faced with a massive budget gap and so compromise in 2025 is best. I see almost zero chance for that. Republicans will be unwilling to raise taxes, because their core principle is tax aversion and doing so in this context would help fund (rightly or wrongly as an objective matter) the Democrats’ overspending in 2023. Democrats, by contrast, after an incredibly close 2024 election (not that much different than 2022 at the legislative level), will be unwilling to do that without some tax increases. Republicans, by contrast, view getting any lever on power means that they must block any tax increase (they’ve explicitly said this, I believe) and that creating a bigger gap in 2028-29 will likely lead to more spending cuts then (assuming they retain some control and hope always springs eternal in politicians’ minds).
• I hope I’m wrong but have a fair degree of confidence that I’m not, especially since the always popular kick-the-can down the road option will likely be available (assuming no more adverse budget impacts).

Postscripts

#1: This is comparing apples and oranges to a certain extent, but the size of the budget swing in this forecast for next biennium (> $1 billion) without any material change in the economic outlook provides some context on the size of the reserve fund. My intuition is that we should not fool ourselves into believing that an unprecedented reserve (>$3 billion) provides much in the way of protection if we have a serious recession. Modest forecast variability wiped out a large portion of that. Of course, we’re not into the biennium yet, and the reserve is designed to protect against swings that occur during the biennium. But still.

#2: Much of the political rhetoric will focus on the 2023 legislature’s (alleged) overspending as the cause of the gap. In my view, this is inevitable and meaningless (maybe not politically meaningless, though). Thought experiments:

• Assume that the GOP had a 2023 trifecta (improbable as that may seem), would they also have overcommitted budget resources by enacting permanent tax cuts (e.g., repealing the income tax per Scott Jensen’s plan or somewhat less audacious but still major cuts)? Probably. Would those changes have reduced the elasticity of tax revenues by as much as the DFL’s spending increased the elasticity spending? Harder to say, but good chance. The income and corporate taxes are the engines of state revenue growth and those are the taxes the GOP prefers to cut.
• Assume split control in 2023, would some mix of permanent budget commitments for DFL programs and GOP tax cuts equivalent to the 2023 actual permanent commitments have been enacted? Probably. (I acknowledge that split control could have led to more one-time stuff, because both parties hoped that their leverage would increase in 2025 to get their permanent agenda items.)

Thus, the real arguments are or should be about the policy content of the agenda, not fiscal management or prudence, at least that is my view.

Another thought experiment along the same lines: Assume split control in 2023 resulted in gridlock and the money was left in the bank or used to pay down debt. That would make future budgets look much better. (Without additional state funding for special ed and LGA, property taxes likely would be higher. Similarly, without wage increases for long-term care workers, that industry might be in crisis. That’s the like narrative DFL leaders would push, anyway.) Would the public have been very unhappy about that state of affairs? Almost surely. The general point is that fiscal prudence has no political constituency. But that doesn’t mean charges of fiscal imprudence will not have political salience. The reality for political leaders is that being cautious and prudent has no intrinsic political reward, but if bad times come, you’ll still be blamed and maybe held accountable for the mess (see Al Quie).