It is all but inevitable that the 2023 legislature will exempt more social security benefits from taxation. Thankfully, it appears it will not exempt all benefits, limiting the exemption (technically a subtraction) to filers with no more than low-six-figure incomes. It’s probably the best one could hope for under the political circumstances. But that’s not the end of it.
Because the subtraction favors one class of taxpayers (seniors) with one type of income (social security benefits), it leads or will lead to similarly situated taxpayers crying foul and pleading for comparable treatment.
First in that queue are governmental retirees who are covered by pension plans that are exempt from social security. They’re stuck paying taxes on all their retirement income. Unfair. The main categories of those pensions are public safety retirees (federal, state, and local) and legacy CSRS federal retirees (some covered by CSRS are still working, of course). Both tax bills preemptively address those claims by extending the bad policy of exempting SS benefits via a qualified retirement subtraction for those public pensions that very roughly mirrors the expanded social security subtraction. H.F. 1938, art. 1 § 22 (third engrossment); H.F. 1938, art. 1 § 22 (1st unofficial engrossment).
These provisions allow a married joint filer with a qualifying public pension to exempt up to $25,000 of it ($12,500 for a single filer), subject to an income phase-out that parallels that under the social security subtraction. I did not carefully compare the House and Senate language, but the provisions appear identical. It seems inevitable that the provision will be in the final tax bill.
A few observations about this exercise:
The best policy would be to ignore the pleas. My mother used to tell me that two wrongs do not make a right. I get that their claims are hard to ignore, especially since many are cops and firefighters – favored public employees for both political parties. Federal law compels giving equal treatment to federal pensioners, so federal retirees cannot be left behind if Minnesota pensions are exempted.
The complicated and unique nature of the social security program makes it impossible to create an exemption for these recipients that provides identical treatment to the social security subtraction. These pensions are designed to replace the equivalent of social security benefits and a standard pension that is coordinated with social security. Calculating what part of their pension replaces social security is not feasible and, in any case, would be maddingly complicated. That is why (I assume) the designers of this opted for a lower dollar limit ($25,000 annual is much lower than the maximum social security benefit – my benefit is over $50,000, e.g., and is not the maximum). It’s a rough compromise. But it overcompensates those with small pensions.
No effort is made to coordinate this subtraction with the social security exemption. Many/most of these recipients will also collect social security, either because they had covered employment or receive spousal benefits. Married couples may have one spouse with social security covered employment and one with a qualifying pension. These individuals will be able to claim both subtractions. Options to coordinate the two could be done by: (1) requiring filers to opt for one or the other or (2) reducing the social security exemption by any amount excluded under the qualified retirement contribution or vice versa. A coordinating provision would reduce the cost of the provisions obviously.
The higher limit for married filers is independent of who receives the pension. In most cases, one spouse will have earned the pension. Death will cut the survivor’s subtraction potentially in half.
No age restriction applies. Old age social security benefits cannot be collected before age 62. Imposing a similar age restriction (other than for pensions based on disabled status) would seem to provide some equivalence. Of course, a prime reason that cops and firefighters have resisted pensions that are coordinated with social security is that they want to retire early (e.g., well before age 62) with unreduced pensions, when social security benefits are unavailable. But that doesn’t mean we have to give them more beneficial tax treatment too.
Recipients of pensions from non-Minnesota state and local governments are denied the subtraction unless their state provides similar treatment to Minnesota pensions. I’m not sure about the rationale for this and wonder if it violates the Commerce Clause. Retaliation and reciprocity rules in state insurance taxation are constitutional because Congress authorized them under the McCarron Ferguson Act. Without such authority, I doubt a reciprocity requirement is constitutional because appears to discriminate against interstate commerce (pension income earned in another state). But I have never had cause to research the issue.
Complicated mess. After the 2023 tax bill is enacted, Minnesota will have four special income tax preferences for senior taxpayers:
The social security subtraction
The qualified retirement subtraction
The higher standard deduction for seniors
The subtraction for elderly and disabled
A better approach? These provisions overlap and none is coordinated with any of the others. A better approach would be to forgo the subtraction for qualified retirement income and to combine ## 3 – 4 into one comprehensive subtraction for seniors that is reduced by any subtraction allowed for social security. One way to do that would be to:
Increase the additional standard deduction dollar amount (by $5k to $10k, for example);
Subtract from the higher additional standard deduction the social security benefits that are exempt from Minnesota tax (other than 15% rule); and
Further reduce the additional standard deduction by AGI (and probably add tax exempt interest for good measure) over specified thresholds that vary by filing status.
The goal would be to set the dollar amounts so that the total costs equal that for the existing additional standard deduction for seniors, the proposed qualified retirement subtraction, and the existing subtraction for elderly and disabled. That would be simpler and fairer. It would raise taxes on some higher income seniors (like me) who will lose their additional standard deduction. It would treat seniors regardless of whether they derive their income from pensions, earnings, or taxable investment income more equally.
A comparison of two similar senior taxpayers illustrates why I think the tax bill provisions are unfair: Taxpayer #1 is a 70-year-old widow who cannot live on her $12,000 in social security benefits so she supplements that by working as a cashier, earning an additional $25,000 in wages for total income of $37,000. (I often see folks working in retail stores who I imagine are in this situation.) Her social security benefits are entirely exempt, but she must pay over $500 in Minnesota income taxes on her wages. By contrast, Taxpayer #2 has a $32,500 public pension that qualifies for the new qualified retirement subtraction and collects $5,000 in social security benefits based on his covered work or his spouse’s. (These benefits are subject to reduction called windfall elimination program. That is designed to prevent these recipients from falsely looking like they had low lifetime earnings, when they were working in uncovered employment, for purposes of social security’s redistributionist benefit structure. A similar rule, called GPO, reduces spousal benefits.) Taxpayer #2 pays no Minnesota tax despite having the same income as Taxpayer #1. To add insult to injury, Taxpayer #1 pays federal income tax and FICA. So, she has less after-tax income than Taxpayer #2 and must pay non-deductible work-related expenses, such as commuting and clothing.
At this point in the process, it is obviously too late to redesign these provisions (a project for the 2024 session?). But the subtraction for elderly and disabled should be reduced by the amount social security benefits that are nontaxable under the Minnesota subtraction, rather than just the federal exemption as it is now, and by the amount subtracted under the new qualified retirement subtraction. That would save a trivial amount of revenue and would prevent double dipping.
Exempting social security benefits is among the highest profile tax issues in the 2023 legislative session. Fully exempting benefits was a central plank of the 2022 GOP unsuccessful campaign to gain control of the legislature. Although the DFL won full legislative control, several its successful candidates also supported a full exemption. Given the DFL’s narrow control of both houses (only one vote in the Senate), expansion of the existing exemption is inevitable, even though only about half of taxpayers pay tax on any benefits and most of those that do pay little. The only question is whether DFLers conclude they must go Full Monte.
It’s a matter of political calculations
I hope not. As I have discussed previously (here and here), I see no principled policy case for expanding the exemption. Just to review the basics, doing so would be:
Unfair because it is horizontally inequitable (unequal treatment of equals) by singling out recipients of one type of income for special treatment. It also treats seniors unequally – e.g., those who collect little or no social security with more taxable earnings or retirement benefits pay relatively more.
Regressive because it primarily benefits higher income individuals. The existing subtraction could be made more progressive and simpler at minor or no additional revenue loss or the elderly exclusion expanded to benefit low-income seniors generally.
Expensive, reducing revenues significantly and the opportunity to use that money instead for actual improvements in tax policy or direct spending, rather than making the tax less fair.
Inefficient (in the economic sense) because all else equal the narrower tax base requires higher tax rates and increases the tax wedge/negative effects that are a necessary evil of higher tax rates.
Ineffective in improving Minnesota’s tax competitiveness with regard to either residency or investment decisions. It will make things worse on both counts because it will impose implicitly higher tax rates on the desired residents (young workers) and investment, while doing little to keep seniors in Minnesota. I made that case in my first post and MCFE has done an even better job.
It’s mainly, if not all, about political survival for DFL legislators in swing districts. I have no expertise in that regard. My amateur instincts are that the expansion in the tax bill released by the House might work politically, for the reasons suggested by MCFE. It’s unfortunate that it has to come to this.
I think most everyone who has spent time thinking about the policy has come to accept that it is all a matter of politics, not policy. But there are two specious arguments for full exemption repeated, almost mantra-like, by Republican legislators and others that I cannot resist writing about/debunking, especially the second which is what this post is about:
Most states fully exempt social security from their income taxes. Therefore, Minnesota should too. This is the lemming argument. It’s comforting to go with flow, to run with the herd right over the cliff into the sea … until your fur is wet and your lungs are filling with water. Mindlessly aping other states is … mindless.
Social security contributions are taxed, so it is double taxation to also tax the benefits they fund. Because half of social security contributions (i.e., employee-paid FICA) were taxed when made, this argument has superficial appeal and requires a bit of explanation as to why it is hooey. My explanation, which is typically long and plodding, follows. The short answer is that the 15% exclusion of benefits in federal law is as good or better than how a comparable traditional pension which is one-half funded with after-tax contributions would be taxed.
It’s not double taxation
Inexact concept. Double taxation is an oft-made, sloppy argument opposing tax features that resonates with many. But tax systems are replete with examples of double taxation. For example, I earn income and use it to buy something, paying both income and sales tax in the process. Cigarettes are subject to both excise and sales taxation. Alcohol is subject to the excise taxation, special sales tax, and general sales tax (triple taxation?). Most are untroubled by any of that. So, as a matter of context, it’s hard for me to get too worked up about any of this. My examples involve different taxes and here we’re talking only about one, the income tax.
Recovery of capital. A well-designed income tax taxes income earned to a taxing unit (however the tax defines that, such as an individual, married couple, or corporation) only once. Multi-year provisions, such as income from depreciating capital investments or retirement arrangements allowing for deferred taxation, create complexity requiring a method of recovering a capital that was already subject to taxation. For example, if an annuity is purchased with after-tax income, a portion of its periodic payments are allowed as tax exempt recovery of that investment or purchase price.
Thus, the obvious question is how these recovery of capital rules should be applied to social security benefits, if at all. The right answer is not simple or obvious. Exempting all benefits because some of contributions were taxed is clearly wrong. That would be like fully exempting a pension or annuity because some part of it was funded with after-tax contributions. The task is to determine what social security benefits are most like and then, determine if the current federal rules provide at least as beneficial treatment. Spoiler alert: they are more generous.
Caveats: Social Security sui generis
The task of finding a comparable is challenging because social security is government program and unlike any private financial arrangement that I am aware of.
Atypical benefits provided. Social Security (specifically the Old Age, Survivor, and Disability Insurance or OSADI element of it) is a social insurance program. It is widely thought of as a retirement program, but it is more. For example, it also provides:
Life insurance – die and your minor dependents collect monthly benefits and your surviving spouse collects a small lump sum and potentially spousal benefits.
Disability insurance – if you become totally disabled, you can collect benefits.
Marriage retirement bonuses – marry a spouse with much lower or no covered lifetime earnings (e.g., a stay-at-home spouse) and “earn” spousal benefits to augment your own. At no extra cost!
Divorce insurance – divorce a spouse to whom you were married for 10 years or more and he/she is entitled to spousal benefits on your account with no effect on your own benefits. You can do this multiple times. Two or three sets of spousal benefits can be “earned” with one contribution.
These features undercut the validity of simply considering social security to be equivalent to a standard pension. No pension system that I know of provides similar benefits. Yes, some pension programs provide disability and survivor benefits, so those elements are not totally unique. But not with the scale and generosity of social security. Moreover, benefits for spouses typically must be opted into and their cost offset by reducing the worker’s benefits. Some portions of social security contributions should be considered to fund those benefits that typically provide exempt income or income to others (e.g., former spouses and dependents), not the person who paid FICA or SECA (the “after-tax contributions” that are purportedly being double taxed). Thus, they should not be considered retirement contributions that need to be recovered free of income tax to avoid “double taxation.”
Redistributive social insurance. A more fundamental point: as a social insurance program, social security is redistributive. It uses income from the better off workers to fund benefits for those with lower lifetime earnings. Yes, it calculates its benefits based on a worker’s earnings (lifetime contributions); higher contributions = higher benefits. So, it is true that you “pay in” to get your social security retirement benefits. (Making social insurance look like an earned retirement program was the politically brilliant stroke of FDR that allowed the program to survive and flourish, despite implacable opposition from the Liberty League and conservative Republicans in the 1930s and 1940s. Much like the opposition to the ACA.) But the formula that determines benefits is very progressive; as lifetime earnings rise, the percentage “earned” in benefits declines. The lowest covered wages earn benefits at 90%, while the highest at 15%. Thus, although the contribution/tax is regressive because it does not apply to wages over a threshold (about $160K now), the overall program is progressive taking the tax/contributions and benefits together.
Put more simply, part of the contributions of higher income workers are not really “contributions” that earn them benefits but rather a tax that helps pay for benefits for the less fortunate. If that were done through the federal income tax, the state would not allow it to be deducted or recovered as a previously taxed retirement contribution to avoid “double taxation.” It was just a redistributive federal tax.
Disclosure: I feel compelled to make philosophical points like this to satisfy my theoretical urges, even though they are irrelevant politically. My doing so used to irritate some colleagues and legislators. With good reason. I was wasting their time. In any case, neither is determinative of anything but they provide useful context for considering claims that taxing social security benefits is double taxation.
Best fit: taxation of defined benefits pensions
Putting my caveats aside, the obvious analogy is the income tax treatment of retirement benefits – more specifically, when contributions are made, at least partially, with wages that were subjected to income tax. There are two flavors of tax rules to consider, those applicable to –
Defined contribution or DC programs like 401(k)s and IRAs, either traditional or Roth plans; or
Defined benefit or DB programs like traditional pensions that pay out monthly benefits under a formula (or an annuity purchased with after tax income).
The two sets of rules differ somewhat and overlap a bit. The obvious closest analogue to social security benefits is a DB pension plan.
With a DC plan, the account holder fully owns the account and bears investment risk but does not bear mortality risk. That is, the contribution and account owner or his/her heirs will get all the contributions and investment return, even if he/she dies before collecting a dime. Thus, the account holder or participant will fully recover his/her contribution (capital) unless it was eaten up by investment losses. This is nothing like the social security program. It would have been for benefits paid under a privatized social security system as foggily envisioned by George W. Bush. That idea died before a real plan could be formulated. The DC rules are obviously not an appropriate fit.
Under social security, the amount of benefits is dependent, in part, on longevity and investment returns are irrelevant (unlike in the DC case). The longer you live the more you get. There is no investment return, just a program parameter that provides a benefit based on actuarial estimates and political decisions. In this sense, social security is more like an old-fashioned DB pension, which pays a recipient a periodic benefit under a formula. A DB plan, like social security, bears mortality and investment risk, the recipient neither. DB plans use actuaries to estimate that risk and scale contributions accordingly. social security rules are set by law with Congress (supposedly) adjusting them if its finances become out of whack. If Congress doesn’t and tax collections lag, benefits will automatically be scaled back by law – unlikely politically, I’d think. Congress will step in and stop it rather than bear the political consequences of shorted benefit checks.
In any case, the question is how the DB tax pension rules that allow recovery of pretax contributions tax free compare to the taxation of social security benefits, which is the subjects of the next two sections.
Tax rules for DB pensions with after-tax contributions
There are two rules – the general rule and the simplified method – governing how after-tax contributions to qualified retirement plans are allowed to be recovered tax free. The provisions are described in lay terms in IRS Publication 575. Both rules allow a portion of the monthly or other periodic pension payments to be excluded from tax.
Why not exempt pension payments until all after-tax contributions are recovered? This simple approach is likely what would come to the mind a nonfinance person who wants to be taxpayer friendly. So, it is useful to consider why the tax system no longer allows it. (It was allowed most recently for federal Civil Service Retirement System covered employees. The Tax Reform Act of 1986 repealed it for CSRS members retiring after 7/1/1986.)
That approach would ensure that recovery of capital has first priority. As such, it is inconsistent with the reality that each periodic payment is funded partially by (1) pretax contributions, (2) after-tax contributions, and (3) investment return. A neutral approach would apportion those contributors to each payment. Moreover, it is inconsistent with the reality that the pension plan assumes longevity risk. That is so because it front loads the tax benefits of capital recovery, essentially giving the retiree the best of both worlds. If he/she outlives the actuarial tables, the plan picks up the tab. If he/she dies early, more tax benefits are recovered. In any case, retirees collecting under qualified retirement DB plans cannot use this option, so it would not have made sense for Congress to use for social security benefits in 1993 with it expanded income taxation.
Simplified method. The simplified method is the default rule that applies to most recipients who made after-tax contributions to qualified DB plans. This quote from Publication 575 captures the essence of how it works:
Under the Simplified Method, you figure the tax-free part of each annuity payment by dividing your cost by the total number of anticipated monthly payments. For an annuity that is payable for the lives of the annuitants, this number is based on the annuitants’ ages on the annuity starting date and is determined from a [life expectancy] table. For any other annuity, this number is the number of monthly annuity payments under the contract.
In other words, a dollar amount is determined by dividing total after-tax contributions by the expected number of payments. If the term is fixed that number is clear. If payments are made for life, as with SS, then the number is based on the average life expectancy based on the age benefits began being paid.
General rule. Despite its name, the general rule applies to a narrow group of recipients. Those who are older than 75 when they begin collecting and who are guaranteed an amount or those receiving payments from nonqualified plans are the two big categories. Here is how Publication 575 describes the rule:
Under the General Rule, you determine the tax-free part of each annuity payment based on the ratio of the cost of the contract to the total expected return. Expected return is the total amount you and other eligible annuitants can expect to receive under the contract. To figure it, you must use life expectancy (actuarial) tables prescribed by the IRS.
Conceptually it is identical to the Simplified Method, but the general rule is not the best comparison to OADSI benefits because no one waits until age 75 to claim social security (if they do, they’re forfeiting benefits). Thus, the appropriate comparison is the Simplified Method.
SS Actuary’s resolution – the 15% solution
So, the obvious approach would be to have the Social Security Administration (SSA) calculate the amount under the Simplified Method and report it to each beneficiary (e.g., on a Form 1099SS). That would treat social security benefits just like benefits from a qualified DB type plan. But the nature of social security benefits makes that impossible, not just impractical.
One person’s social security contributions may provide benefits for multiple individuals as a result of spousal and dependent benefits, as noted above. The extent that this will occur cannot be known even when initial benefits are paid. For example, a divorced spouse may later claim based on the ex’s account, or a single recipient may marry after claiming benefits entitling the new spouse to benefits. Recipients can switch from their own benefits to spousal benefits after a death as another example. Thus, SSA, even if it had the best computer database and software (which it doesn’t), could not accurately calculate and report the amount to recipients. It’s unknowable.
This was initially irrelevant because OASDI benefits were fully exempt from federal and state taxation for decades. The Bureau of Internal Revenue or BIR, IRS’s predecessor agency, exempted OASDI benefits from federal income tax on unclear premises in 1941. Brunson and Johnson, Good Intentions: Administrative Fiat and the General Welfare Exclusion, pp. 6- 9 (2/2/2022). Text of BIR rulings. So, there was no need to address the issue.
But in the early 1980s, Congress determined it should begin taxing some benefits to shore up the program’s financing. SSA, Research Note #12: Taxation of Social Security Benefits. In 1983, it did so by subjecting up to one-half of benefits to taxation. That did not raise the issue. Recovery of capital would always be less than one-half. In 1993, Congress determined to more fully tax benefits for higher income recipients. Doing so required developing a practical way of ensuring that the financial benefits of after-tax contributions could be recovered tax free by applying the DB pension tax rules. That task fell to the SSA actuary.
The result can only be described as making sure that virtually all after tax contributions are recovered free of tax. The actuary estimated average percentages of benefits that represented recovery of previously taxed contributions under the DB pension rules. Stephen C. Goss, Unpublished Memo (Social Security Office of the Actuary, 4/7/1993).
The actuary’s memo does not appear to be available online. But here is how a Congressional Research Service (CRS) publication describes its analysis and the method:
In 1993, the SSA’s Office of the Actuary estimated that, if pension tax rules were applied to Social Security, the ratio of total employee Social Security payroll taxes to expected benefits for current recipients (in 1993) would be approximately 4% or 5%. For workers just entering the workforce, the actuaries estimated that the ratio would be, on average, about 7%. Because Social Security benefits replaced a higher proportion of earnings for workers who were lower paid and had dependents, and because women had longer life expectancies, the workers with the highest ratio of taxes to benefits would be single, highly paid males. The estimated ratio for these workers (highly paid males) entering the workforce in 1993 was 15%.
In other words, the Actuary took the worst-case scenario, high-income unmarried males, and estimated the percentage necessary to recover their contributions tax free and applied it as a general rule for all recipients. Congress codified that 15% estimate in the tax law, which carries over to Minnesota tax via its use of federal adjusted gross income. Note that these calculations appear to assume that all employee FICA are the equivalent of retirement contributions. That is, my niggling points about benefits going to third parties (like divorced spouse) or the redistributive tax built-in to social security are irrelevant. Much more important, because it set the rule for the group whose members have the lowest returns on their contribution, everyone else pays lower taxes than they would if it were a DB pension. The population average is 5% or only one-third of what the general rule (15%) is. That ignores the income phase-in and Minnesota’s subtraction that exempts even more benefits for most.
What that means is that so long as 15% of benefits remain tax exempt, recipients are getting as good or better treatment than if their employee social security taxes (i.e., FICA or SECA for a self-employed person) were after-tax contributions to a standard pension plan. There is no double taxation.
A few caveats could be made:
This will no longer hold if Congress cuts benefits to shore up social security finances (e.g., raising the retirement age) or raises the FICA tax rate or earnings threshold. In that case, the return on contributions will be lower and Congress should consider whether to revise the 15% rule.
Demographic changes could make this worse (e.g., reduced life expectancy) or better (increased life expectancy). Until COVID, increases in life expectancy since 1993 likely made the 15% rule too high even for the worst case, high earning single males – just a guess. Despite COVID’s effects, US life expectancy is still higher than in 1993, so 15% is probably higher than the data the estimate was based on. See KFF Health System Tracker, How does U.S. life expectancy compare to other countries? (Although COVID caused a 2.7-year drop, at 76.1 it is still higher than 75.5 in 1991, the likely data used by the SSA actuary in 1993.) The silver lining in the US’s COVID vaccine hesitancy is a small improvement in social security’s finances. Not what you want obviously.
Yes, for some outlier groups – the very highest earning never-married males who are smokers, NFL players, sky divers, or similar with lower average life expectancies – the 15% assumption is too low. Using averages will necessarily disadvantage some cohorts (not a big deal if that is based on choice rather than immutable characteristics) and, as I noted above, the program is intentionally designed to be redistributive in other ways.
Bottom Line
There is no double taxation. The 15% exclusion built into federal and state tax law ensures that social security benefits are taxed no more heavily than a DB pension. Taxable social security benefits are income, not return of previously taxed contributions. As such, they presumptively should be taxed just like wages, interest, and pension benefits are. Beneficiaries, on average, get multiples of their after-tax FICA payments.
Sadly, in the current political environment none of this matters. Key Republican legislators (thinking of former chairs of legislative tax committees, e.g.) surely know that their oft-repeated claims of double taxation are false, but they still make them and are sure to continue doing so. It gives a patina of policy respectability to the crass political appeals. The complexity of why they are false and the strong self-interest of recipients in believing them to be true guarantee many will believe them. In a world in which many do not believe in the reliability of something as simple as counting ballots in a carefully monitored presidential election, accepting a government actuary’s estimates of a complex program has a snowballs chance. That’s particularly true when confirmation bias and self-interest tell you not to do so.
Postscript – 3 snarky asides
1. Implicit tax rates become more explicit
While working, I often made the point that reducing the tax base – e.g., by exempting social security benefits from income taxation – results in implicit higher tax rates. Many, including most legislators, considered that to be a throwaway wonky or academic point without practical consequences. The House’s tax bill, which includes both base narrowing and rate increases, illustrates why it is more valid than they cared to admit. The table shows three estimates for the bill’s FY 2025 change in revenues (when the annual costs are fully reflected in the estimates).
The numbers show that the bill’s revenue from the new top tax rate almost exactly offsets the loss from expanding social security exemption and the related new public pension exclusion. The net difference for the provisions is only $10 million or 4% of the gross. It looks like the bill is a case of classic income redistribution – raising taxes on millionaires (new top rate applies to married couples with a million or more in taxable income) to give to tax breaks to middle-income retirees. (Okay, there are a lot of other provisions in the bill, so you can’t trace revenues but you get the point.) A bad idea in my book for a number of reasons I don’t need to go into. Maybe the long-term economic risks of a double-digit top rate are justified if delivery of important services are at stake but not to pay for poorly targeted tax breaks.
2. Stumbling into bad policy
A 3-decade partisan chess match explains how we got to this point. Consider this series of moves (oversimplified to keep the narrative short) for my snarky take on what happened:
Move 1. GOP adopts the tax pledge orthodoxy in the 1990s saying something like: “We promise to never raise and always keep cutting your taxes.” The economic boom generated a surfeit of revenues throughout the 1990s and gave superficial plausibility to that slogan. The 2002 recession and collapse of the 35W bridge caused some minor backsliding – 50 cent/pack cigarette “fee” and allowing a DFL sponsored gas tax increase with a little GOP support to override Governor Pawlenty’s veto. But the consensus of skittish DFLers was that much of the public, especially key swing voters, opposed virtually any tax increase. GOP opposition to taxes goes so far as to result in previously unthinkable borrowing to fund operations in the 2011 session, stretching the constitution to the limit.
Move 2. DFL responds by promising to tax the guy behind the tree (the “rich,” corporations, and smokers specifically), none of whom most voters are. Yes, this is partially consistent with their fixation on progressivity. When they gain trifecta control after the 2012 election, they implement this strategy in the 2013 tax bill, raising significant revenues from a new 4th tier income tax rate, as well as corporate income and cigarette tax increases. The GOP finds voters are not very troubled by this turn events. So, they have to come up with a counter move.
Move 3. Following the 2014 election, GOP regains full legislative control. Instead of attempting to reverse the 2013 DFL tax increases, it turns to various special tax favors, including exempting social security. This was a sort of break-the-glass moment that abandoned conservative principles (across-the-board rate cuts, e.g.) in favor of handing out tax cut candy to a targeted group of high-turnout voters, seniors. (Note: prior to this, GOP majorities had not attempted to cut the tax on social security and, in fact, had approved the decisions to tax benefits – House majority in 1985 and Governor Carlson in 1994. The House GOP majority in 2011, in the face of a multi-billion budget deficit had proposed an income tax rate cut, rather than exempting social security, for example.) Touting the synthesized unfairness (double taxation) as justification for plump tax cuts for a core constituency was very popular, a reality recognized by both parties. That caused the GOP to make it a signature issues in the 2018 election and onward.
Move 4. The DFL reading the political tea leaves, not surprisingly, accedes to a water-down version of the full exemption (or worse) but funds it with its proven approach of taxing the guy behind the tree. Voila: the 2023 House Tax bill.
On one level, it is all too predictable. On another, it seems almost inescapable with no exit ultimately ending in a full exemption. Oh, for the days when the political debate was about the need for and quality of the government services and whether and how to best pay for them. I’ve become a crotchy old man longing for the good-old-days of my early career when the GOP was conservative and the DFL was not afraid to advocate for broad-based tax increases to fund or preserve services. Hey kids, get off my lawn.
3. Convenient hypocrisy
While eating lunch a few days ago, I heard an interview on MPR with a member of the GOP House leadership who decried the House’s human services bill for its failure to adequately increase state funding for nursing homes and, in turn, stated implacable opposition to the House tax bill because it did not fully exempt social security. As if annually reducing revenues by more than one-half billion would not make it harder to find more funding for nursing homes. Being in the minority permits the luxury of advocating for politically appealing but fiscally inconsistent policies.
Whether ’tis nobler in the mind to suffer The slings and arrows of a special session, Or to take arms against a sea of campaign troubles
We’ll have to wait and see how Governor Walz and, of course, the legislators he is negotiating with resolve that question. I’m sure he will not foolishly call a special session without an ironclad agreement on what will be passed. If he has contrary thoughts, he should chekc iwth Tim Pawlenty as to how that works.
Based on my experience hanging around legislative politicos, I’m sure the political tacticians on both sides are in hyperdrive trying to figure out the advantages and disadvantages of a deal or no deal. When they say we need some time off to decompress, they really mean “we need to do some formal and informal polling on how this may play in November.”
I have mixed feelings. On the one hand, going through the 2023 filing season without federal conformity is more than an inconvenience for taxpayers, preparers, and DOR. On the other hand, I still think the revenue uncertainty is palpable (see below for more). Although I haven’t spent the effort to analyze it, my instinct is that much of the $4 billion committed to reserves is onetime money (i.e., most of the spending and tax cuts are permanent commitments leaving little structural balance). And there is the supreme stupidity of the social security exclusion. Delaying that by even one year would be a blessing on the slight chance some more semblance of reason may prevail (unlikely, I know).
What will happen is anyone’s guess. I certainly have no special insight.
Lotterman Column on SS taxation
I like very much Ed Lotterman’s columns in the PiPress. He typically applies economic analysis to policy and other issues in ways that the average newspaper reader can easily grasp and that (to me anyway) make intuitive sense. I am glad that his column has endured for so long, given the cutbacks and general carnage that has been visited on the paper by its tightfisted hedge fund owners (old WaPo story). Columnists are cheaper than reporters, I’d guess.
His most recent column is on one of my favorite topics, the foolishness of the state exempting social security benefits from taxation, Parsing Social Security taxation. It’s worth reading but underwhelmed me.
I think it is fair to distill his column, drawn from MFCE and Minnesota Budget Project publications, down to two basic points. A full exclusion of benefits would:
Be regressive.
Have little effect on concerns about seniors moving to low-tax states.
Both points are valid, of course. For an economist, the classic policy tradeoff is between efficiency (essentially maximizing well being for the overall population) and equity (fairness). Typically increasing equity comes at the cost of reducing efficiency (e.g., Okun’s leaky bucket metaphor). Here, a full exclusion is simultaneously inequitable (measured vertically) and inefficient (doesn’t work as an incentive to counter tax-based migration?). No need to worry about tradeoffs. It flunks both tests, a no-brainer to reject. As a good economist, Lotterman also points out the opportunity cost of the resources used for an exclusion; better uses are available whether your priors bend left or right. I whole heartily agree. Great points.
So, why was I underwhelmed? The best case against the exclusion is stronger in my view:
From an equity perspective, the big objection is horizontal. By favoring one type of income (social security) and one class of taxpayers (seniors), it treats similarly situated people unequally. It favors the old over the young and the retiree over the worker, regardless of age. To my lights, a bigger deal than its regressivity, which could be offset by restructing rates to neutralize the regressivity (not the tax cut) IF the exclusion made any policy sense. It doesn’t. MCFE (not the publication Lotterman references) has made the horizontal equity point, just to be fair (no pun intended).
Trying to structure the state’s tax system to attract people to live here is a fool’s errand. It would require attracting people who make net contributions to the state’s economy and budget, a sort of laser-targeted migration mercantalism policy that is impossible in practice. But if you’re going down that path, focusing on seniors has to be the wrong target. The state by most accounts faces a looming workforce shortage. So, we’re focusing our incentives on retirees? It’s not just that it doesn’t work; it’s aimed at the wrong target. And it means, all else equal, that workers will pay higher tax rates for the same level of state and local services.
Putting this in opportunity cost terms, it’s not just that we lit a bunch of dollar bills on fire, preventing their use for something more productive. It’s worse – to extend the metaphor to the breaking point – we burned them in the fireplace when we were already running the a/c to keep the house cool. Sigh
CBO Budget Outlook
The Congressional Budget Office (CBO) is out with its Budget and Economic Outlook: 2022 – 2032 (May 2022). As anyone paying attention knows, the feds have had strong revenue flows similar to what Minnesota has been experiencing. Striking evidence for that is this statement:
CBO expects individual income tax receipts to rise by 28 percent in 2022, to $2.6 trillion. At 10.6 percent of GDP, that total is expected to be the highest amount of individual income tax receipts recorded since 1913, when ratification of the Sixteenth Amendment authorized the federal government to begin collecting income taxes.
Budget and Economic Outlook, pp. 86 – 87.
Federal individual income tax revenues typically have been about 8% of GDP since the early 1960s. CBO is forecasting them to be just under 10% for the entire period. (Small detail: CBO assumes – but contrary to the GOP game plan – that TCJA’s individual income tax provisions will expire in 2026 as provided by current law. That is 0.8% of GDP’s worth of revenue. If the experience with EGTRRA’s expiring tax cuts is any guide, I would not bet on that. At best, some portion that affects high income filers seems more likely and that is premised on the Dems controlling the presidency or one house of congress, no sure thing.)
Like most of us, CBO is not sure what has caused this revenue growth. It estimates that capital gain realizations (“significantly above that historical average in the past two years”) is much of the explanation. But some mystery is afoot:
[R]eceipts from individual income taxes in the past few years have been larger than expected given currently available data on economic activity and the past relationship between tax revenues and the state of the economy. Those larger-than anticipated receipts may have resulted from several possible developments. They may reflect higher wage or nonwage income than the Bureau of Economic Analysis has reported for the past two years. Realizations of capital gains in those years may have been larger than CBO has estimated, or a larger share of those realizations may have come from sales of assets held for less than a year, which are subject to higher tax rates. Also, the temporary tax provisions enacted in response to the pandemic may not have been as widely used as anticipated. And the distribution of taxable income may have been skewed more toward higher-income taxpayers (who, on average, are subject to higher tax rates) than CBO estimates.
Budget and Economic Outlook, p. 89.
In other words, we don’t have much of a clue. CBO is estimating over time that these effects go away, as one would expect for a phenomenon the cause of which is unknown. It’s safer to bet on “regression to the mean” than “this time is different.” CBO also, in a piece of bad news, estimates wages, salaries, and pensions will decline as percentages of GDP. Lets hope they’re wrong, but I would not bet on that.
With regard to corporate collections, they are also rising nicely – up 61.5% for 2022 over the pre-pandemic (FY 2019) level. But CBO is similarly uncertain why:
Corporate tax collections were larger in 2021 and early 2022 than can be fully explained by currently available data on business activity for those years. The factors that contributed to the unexplained strength in receipts will not become fully apparent until information from tax returns becomes available over the next two years. Depending on the factors that caused those larger receipts, their effects might be expected to continue indefinitely, end suddenly, or even change direction.
Budget and Economic Outlook, p. 91.
Because Minnesota has forgone taxing GILTI or conforming on bonus depreciation, state corporate collections will not track federal as closely as they have traditionally, for whatever that is worth.
Bottom line: more than the typical level of uncertainty seems to be the order of the day. CBO reports its typical (mean) error would be 1% of GDP for 2023 and 1.8% for 2027. I personally would not bet on individual income tax revenues continuing at 25% higher than their historical average without seeing a longer trend than two years. My instinct is that aggressive federal fiscal policy in response to the pandemic is a big part of the explanation. And it’s going away now.
Tax expenditure tidbit from the report: “The exclusion of pension plan contributions and earnings is the single largest tax expenditure in the tax code; including effects on payroll taxes, the tax expenditure resulting from that exclusion is estimated to equal 1.8 percent of GDP in 2022.” (p. 97). Stepped up basis on death equals 0.2% of GDP.
TCJA as the cause?
As I noted, CBO is uncertain as to why federal corporate income tax revenues are up so much. TCJA’s acolytes are not similarly sanguine and think it’s responsible. See Tyler Goodspeed and Kevin Hassett, The 2017 Tax Reform Delivered as Promised (WSJ May 8, 2022) for an “I told you so” account. Brookings economists have doubts. William G. Gale, Kyle Pomerleau, and Steven M. Rosenthal, The booming economy, not the 2017 tax act, is fueling corporate tax receipts (June 3, 2022) (“TCJA is not a plausible explanation for the large recent increase in corporate tax receipts”). Count me a skeptic about TCJA being the cause, although there is some evidence it increased in-bound international investment, a primary goal of its international changes.But not enough to offset the rate cuts and certainly not a 62% bump for 2022 over the last pre-pandemic year, 2019. Nice try – good enough for the WSJ editorial page, I guess.
With just a week until adjournment (no bills can be passed on 5/23), Governor Walz and the legislature face key budget decisions regarding the remaining $7 billion or so surplus (after refilling UI fund and adding more hero pay). It’s useful to think about how complex and fraught these issues are for legislators and the governor.
Specifically, how much of the surplus should be:
Spent or left for the 2023 legislative session?
Used for one-time versus ongoing or permanent items – e.g., used for rebates or one-time capital projects versus ongoing tax cuts or permanent spending?
Allocated for tax cuts (e.g., rate cuts) versus spending (e.g., tax expenditures or direct spending)?
Divided among differing priorities – e.g., education, housing, human services, general tax cuts, etc.?
Permanence and reliability of revenues
For issues #1 and #2, a key issue is how much of the surplus is structural or permanent and will support ongoing commitments. MMB has judged that a good portion is likely permanent. But there is uncertainty around assertions like that. Probably more than usual in this case.
Structural change? The tax system has changed little in the last 20 years or so; the 2013 addition of a higher top income tax rate provided a modest elasticity dividend and conformity to TCJA shuffled the deck with unclear effects. Pre-pandemic revenue growth was less than recently experienced. State tax revenue growth in 2015 – 2019 was essentially flat; the February forecast projects annual growth for 2020 – 2025 at 4.5%. Did the pandemic or the response to it permanently change the economy (1) to put it is on an enduring higher growth path or (2) change its structure to yield more state tax revenue (w/o changing the tax rules)? Making more than modest permanent commitments to either tax cuts or spending increases seem premised on some combination of those assumptions. That may prove true (I can think of reasons why it might be), but it is highly uncertain.
Recession coming? Aside from issues of whether some economic change is yielding materially higher state revenues, a related issue is whether a recession is looming. The stock market seems to think so or maybe not (Market Watch – additional 500-point S&P 500 drop needed to signal recession). Of course, as Paul Samuelson famously quipped, the stock market has predicted nine out of the last five recessions. In any case, the Fed has little experience reigning in inflation without causing a recession (maybe immediately after WWII). Although Minnesota’s reserve accounts are full, that is cold comfort. It certainly would not have been enough in three out of the last four recessions to prevent severe cuts, the only exception being the short pandemic recession.
Inflation. Inflation has ramped up a bit more since February. That will surely fuel some more cost of current services than no one seems focused on. The siren allure of new stuff makes it easy to ignore the risk that inflation poses to maintaining existing services.
Under the circumstances, prudence and caution would be in order. But neither is rewarded in the political realm. Avoiding unknown bad things happening does not get politicians attaboys, much less reelected. For what is worth, the April revenue report continues to show booming revenues ($1 b above forecast). So, that will not give anyone reason to hit pause.
Priorities dictated by partisan priors
One’s positions on #3 and #4 are heavily influenced by whether you wear a red or blue partisan jersey, and the two parties are miles apart on policy priorities.
Republicans favor tax cuts to downsize state government; they think it is too big and is doing much of the wrong stuff. When they address policy priorities, they spend through the tax code pushing tax credits for things like childcare or parental leave. Note: that’s not a considered judgment about which instrument is more effective or better suited to the task (see here for my take on that, written for House Research), but largely an ideological matter and a desire to make intervening in the private market look like a tax cut.
Democrats, by contrast, favor bigger state government and direct spending programs (e.g., for public education or similar) generally. But they will use tax expenditures as a second-best alternative if necessary to get a deal or appease voters’ perceptions of a preference for tax cuts.
Resolving these policy issues will not be easy, assuming the first order issue can be resolved – how many dollars to commit and how much of that should be temporary versus permanent.
Political tactics – roll the November dice?
To further complicate matters, tactical considerations affect the negotiations in a big way with everyone up for election in November. Both parties have a mixed set of policy and political goals that must be traded off within and between the partisan caucuses and the governor. Should I go for half a loaf now and give half a loaf to my opponents? How will that affect my and my opponents’ election prospects? Could I get a better deal, say three-quarters of a loaf or the whole loaf, in 2023?
Some potentially relevant tactical considerations or questions:
Both parties want to advance their policy priorities (duh).
They can do that best by winning control of all three power centers (Governor, House, and Senate) so they are unfettered by the other party’s wildly different views (duh again).
Failing to win all three, they absolutely do not want the other party to win all three, since that would allow them to dictate the agenda in 2023. Democrats think something like Scott Walker’s Wisconsin would takeover Minnesota government; GOPers likely assume some unholy combination of socialism and lawlessness would appear on the Minnesota prairie.
What are each side’s probabilities of winning total or maintaining partial control in November?
How is that affected by what the legislature does or doesn’t do in 2022? The participants (perhaps, deluding themselves) think legislative actions have strong ramifications for the November election.
Who will voters hold responsible if there is no deal? How much do they even care?
How will it affect their base voters and the other party’s base? Will it energize or demoralize them?
How will it affect swing voters (potentially quite differently than base voters)? Each party wants to energize it base and to attract swing voters. Frequently the two are at odds with each other but strong partisans often don’t believe that.
How will whatever they do or don’t do get spun by the regular media or on social media?
Is it all meaningless because national trends (inflation, the SCOTUS abortion decision, Ukraine, Biden’s unpopularity, etc.) are determinative?
These are just some of the factors and questions decision makers will be mulling over. The complexity and unknowability are obvious. Resolving these issues is difficult even with one party control, because legislative caucuses are not monoliths that agree on policies and tactics. Far from that, they disagree and vigorously debate them internally. There are three separate decisional entities, five for a bonding bill which requires supermajority approval.
Given all this, it is impressive that anything gets done. In a budget year, continued functioning of government compels action. In an off year like 2022, it does not. Of course, doing nothing (deadlock) is always easier than compromising.
No predictions
I have no special insights or expertise to predict what will happen. When I was working and daily talking with legislators and partisan staff, my predictions were close to as good as chance. Now I am sufficiently out of touch that a monkey throwing darts (unfettered by thinking he knows something) would likely be more accurate.
Other states
Many states, regardless of partisan control, are cutting taxes. Damn the torpedoes, full speed ahead.
Alan Rappeport, States Turn to Tax Cuts as Inflation Stays Hot (NYT, May 10, 2022) details cuts enacted and proposed in other states with Republican, Democratic, and split control. The article quotes Jason Furman, Obama administration economist, as opining the cuts are evidence federal pandemic aid to state was too large and that the cuts will contribute to inflation (probably true on both scores). Effects on inflation are irrelevant to state politicians fighting for their political survival and policy agendas, of course.
See the Tax Foundation’s useful tracker of 2022 state tax changes, misleadingly labeled “tax reform.” Two common trends appear to be flattening rates (w/o expanding bases) and exempting more pension and retirement income. Neither qualifies as reform, in my book. Throwing money at retirees is happening in many states.
According to TPC, states are reporting weaker revenue growth for FY 2022. It characterizes FY 2022 and 2023 state forecasts as “alarmingly weak.” For fiscal year 2023, it says “personal income tax revenues are expected to increase by only 0.4 percent, while corporate income tax revenues are expected to decrease by 8.6 percent. and sales tax revenue growth is only expected to be 1.9 percent. As a result, total tax revenues are expected to increase by just 0.1 percent.” It notes this growth is in nominal terms a big concern with the current inflation rate. As noted above, the anemic growth pattern has not shown up here, where both the forecast and collections keep growing (albeit subject to the confusion resulting from the new PTE tax option). Minnesota’s forecast tax revenue growth for FY2023 over FY2023 is over 5%, but the drops to 2.5% and 1.9% for the two succeeding years.
Connecticut, a deep blue state, enacted a $600 million tax cut (claimed to be the biggest ever in the state). But also claims (per governor’s statement) to be contributing $3.3 billion to reduced unfunded pension liabilities. Count me skeptical without seeing the details, because it is characterized as an “anticipated contribution[.]” If accurate, it would be a truly unusual act of fiscal responsibility and might unexpectedly validate Mitch McConnel’s expressed concern that federal pandemic aid would turn into a blue state public pension bailout (?!).
Horner whiffs
Tom Horner, Republican (I think) and former independent candidate for governor, has an op-ed in the Strib, For an aging state in an aging country, future could be bleak (May 10, 2022). He points out that a big challenge facing the state and the legislature is the declining population of college-aged young adults. He chastises the legislature for ignoring the problem. In his words:
It’s not just that legislators have ignored the reality of losing young adults, it’s that they are oblivious to its implications. Look at the proposals coming from St. Paul. Public safety? Hire more cops. Education? Hire more teachers. Child care? Hire more caregivers.
What world are they living in? Certainly not the Minnesota of 2022.
Tom Horner, For an aging state in an aging country, future could be bleak, Star Tribune (5/10/2022)
He proposes using the budget surplus to redesign and better deliver public services, all good general ideas but lacking in specificity. So, what’s my issue with the piece? Why do I consider it a policy strike-out?
He failed to call out Republicans for a core and hugely expensive piece of their tax cut – full exemption of social security. That is the first thing he should have done emphatically – especially since he is a Republican (former if not current) and I assume he has somewhat more influence in that quarter. It’s like a Hypocritic Policy Oath – first do no harm. As noted in the quote above, he expressed a general concern about the GOP’s permanent tax cuts. But if any of the current legislative proposals is antithetical to his premise, exempting social security from taxation is it. And both parties are proposing it, just in different amounts and configurations.
If we want to attract/retain young people, why would we raise their relative cost of public services? Exempting social security income from taxation (whether totally or just somewhat more) means Minnesota will tax young, working people more heavily for existing public services. That is so because the lost revenue could have paid for an across-the-board tax cut that would have benefited workers as well as retirees. Instead, we’d surcharge them to benefit more affluent seniors (lower income recipients are already exempt). Never mind that seniors already consume a disproportionate share of public services (think nursing homes and home health care). Preferentially taxing seniors who can easily pay is policy malpractice, even if it is politically shrewd.
If there is a tax bill, some version of an expanded social security exemption will be in it. Given Horner’s basic premise, strongly pointing out the wrong headedness of that should have been the first thing he did. Instead, he promoted a bucket of amorphous services – not helpful when bills are already written and have been passed by one or both houses.
The expected happened; the state has a big surplus. What was unexpected was how BIG it was. I haven’t done the research to verify this, but it must be the biggest in both absolute (unquestionably) and relative terms. What makes it even more impressive is that the headline number, $7.7 billion, reflects the discount for refilling the budget reserve and reversing shifts. The real surplus is closer to $9 billion. (Aside: I never thought that I would see reversal of June accelerated sales tax shift. It happened.) Good news, right? Well, it is better than a big deficit but the track record for a divided legislature doing good things with big surpluses does not inspire confidence. Big surpluses often result in improvident budget commitments. It took more than a decade for the budget to recover from the orgy of spending and tax cutting in 1999-2001 sessions spurred by big surpluses (“boatloads of revenue” in the memorable terms of a former MMB commissioner).
This post collects some of my random thoughts on the topic, sorted into five categories:
The context
What the GOP is likely to propose
What the DFL is likely to propose
What should be done I would do
What will happen (spoiler alert: nobody knows, especially not me)
The Context
A year and half ago, no one (certainly not I) thought MMB would forecast own-source state tax revenues for the 2022-23 biennium rising 11% over the current biennium and 22% over the 2018-19 biennium, the last one unaffected by the pandemic. Because those are increases in own-source revenues, they ignore the federal aid gusher that has poured into the state’s coffers. So, despite the ups and downs caused by the pandemic, the state’s own-source tax revenues continued on an upward trajectory. They’re growing as fast or faster than in the pre-pandemic period, raising two related questions:
How could this happen (minor chord, how did the estimators get this so wrong)?
Is this sustainable?
Unfortunately, I’m not competent to intelligently answer either question. That won’t stop me from speculating/babbling. But modesty about the lack of (most anyone) knowing the answers should inform how the legislature decides to use the surplus.
How could this happen?
One obvious reason is that the federal government’s pandemic response that poured aid into businesses. households, and state and local governments was very effective. Unlike during the Great Recession, Congress ignored concerns about deficits and fiscal restraint and the economic effects showed it. (As an aside and as I have opined before, Congress overdid it. There was no need to give money to Social Security recipients since their benefits were unaffected by the pandemic shutdown. Too much money was given to businesses via PPP loans (at least the tax exemption) etc. Congress overdid aid to state and local governments once the Democrats took over. See this WaPo article for details. Of course, overdoing pandemic relief was preferable to underdoing it, as with the Great Recession.)
My second observation is that the pandemic played into the old economy bias of state and local tax structures by redirecting economic activity away from lightly taxed services (more risk of getting infected by going out to eat, getting a haircut, or travelling) to more heavily taxed goods under states sales taxes. That helped state revenues hold up in the face of a pandemic slow down and has been widely recognized.
Finally, the pandemic’s worst effects have been on the lower income strata. Minnesota with a modestly less regressive tax structure probably benefited slightly, compared to states with more regressive structures and vulnerable economies. This effect is very small, which is reflected in the fact that fiscal conditions of most states are in good shape, regardless of the progressivity of their tax structures.
Why didn’t the estimators see this coming?
Macroeconomic forecasting in normal times is devilishly difficult. A once in a century pandemic raises the degree of difficulty by several orders of magnitude. See, e.g., this econofact podcast. Translating macroeconomic forecasts into estimates of tax revenues is itself very difficult and introduces yet more possibilities for errors. (That I am familiar with since I helped MMB economists when I was working. I know how difficult their jobs are.) That estimators missed by as much as they did should be expected. It would have been easy to miss by even more.
Can we expect it to continue?
The more important question is whether these revenues and the resulting surpluses are sustainable. Is it prudent to rely on them to make permanent policy changes that reduce ongoing revenues or fund ongoing programmatic commitments? (Note: a first order question is how far off the underlying macro forecast is. For example, are the extraordinary growth rates in wages built into the forecast really going to happen? Will the effects of the omicron variant reduce GDP growth? For example, how much will expiration of federal pandemic spending, such as the child credit, slow down growth? Those questions are more in the nature of short run concerns for which there will be more information by the February forecast and when the final budget decisions are made in May. As a result, I’m less concerned about them. If things turn down significantly in the short run, it will relieve the natural impetus to make permanent commitments.)
The numbers are so large and unexpected that my cautious nature warns me to be skeptical. But I do not have any special insights qualifying me to make a competent judgment on the issue; it’s simply my gut reaction.
One frame of reference to test that anxiety would be to calculate the pre-pandemic growth rate of revenues from income, sales, and corporate taxes, adjusted for changes in law (i.e., using the 2021 law to recalculate revenues for all years), and then see if the FY 2023-25 revenue projections fall roughly along the same path. If they are materially higher that likely reflects an assumption that something fundamental in the relationship between state taxes and the economy changed or that the macro forecast somehow is estimating growth exceeding the pre-pandemic rate. If either is the case, I would be very cautious about assuming the relationship is permanent. If not, I would feel somewhat more comfortable with the forecast.
Unfortunately, that is not a simple task because of the challenge of re-estimating how much the tax cuts in 2017 and 2019 sessions would have reduced pre-FY2018 revenues. Nevertheless, I made some crude assumptions about how much the 2017 tax cuts reduced income tax revenues and compared the pre-pandemic growth rates in actual revenues with those predicted by the forecast. (I used the 6 fiscal years ending in FY 2019, the last year unaffected by the pandemic, and compared it to the subsequent 6 fiscal years under the forecast. Those time frames conveniently start after the effects of the big 2013 tax increase are built into the income tax revenues.) The forecast growth rates for the individual income and corporate taxes are about the same as those for the pre-pandemic period. So, essentially the recovery following the pandemic recession has gotten revenues back up to their old path but by enough to average out the recession’s (FY 2020’s) dip. A real V-shaped recovery!
More remarkably, the growth rate for sales tax revenues during and after the pandemic are higher by a fair amount (~2 percentage points). That may (emphasis on “may” given the crudeness of the test) suggest that the forecast is predicting something more fundamental is going on with sales tax revenues. For example, the shift to relatively more purchases of taxable goods and services is partially a permanent phenomenon? Inflation, of course, gives a helping hand to nominal sales tax revenues too and we have more inflation now. Fortunately, sales tax revenue growth is a much smaller contributor to the surplus (and sales tax revenues to the budget generally) than the income tax.
Both factors – a very robust recovery and some sort of secular shift in consumption – are certainly plausible and I’m not competent to question that. Moreover, I am making inferences from a very crude test. But it does not cure my instinct to be cautious. Of course, growth after the Great Recession was anemic, so that may cut against the caution.
Budget reserve thin protection
The forecast completely refilled the budget reserve, which is at an all-time high in nominal terms – $2.4 billion (an additional $350 million is in the cash flow account and over $100 million in the stadium reserve, both of which support general fund spending). That is the level that MMB’s methodology says is fiscally prudent. However, I have two observations about that:
Shortfalls in revenue following the 2002 recession and the Great Recession were both much larger than $2.4 billion for a budget that was smaller. Even a garden variety recession (pandemic recession is an outlier) will likely reduce revenues by as much or more than the reserves, necessitating spending cuts or revenue increases. Increasing spending or cutting the revenue base (that is what the surplus is likely to cause the legislature to do) increase that risk and arguably support putting more in reserve. The legislature won’t (I say that with confidence).
MMB’s reserve methodology is based on textbook economics – that is, estimates of the variability of tax revenues over the budget period. That methodology ignores the political economy of state government budgeting. When shortfalls hit, politicians are quick to tap reserves and to temporize (use “shifts” or deferrals of spending and acceleration of revenues); they are reluctant to make the hard decisions (cutting spending or raising taxes) if that can be avoided and it usually can for a while. The reserve methodology implicitly assumes that hard choices will be made, and the fiscal ship will be righted for the next budget period – permanent spending/revenue adjustments and refilling the reserve. That does not happen. Deferrals and shifts drag the budget problems into the next biennium, and the legislature does not refill the budget reserve, except if revenues recover above later forecasts. Thus, the legislature typically starts the first biennium after a recession underwater with depleted reserves. Moreover, the recession’s effects (lower revenues and higher demands for spending) may linger. In short, the budget reserve does not provide protection against multi-biennium budget effects, which is the standard case. Politically, it is impossible to do that, of course. The public and, therefore, politicians (right or left) won’t tolerate the state carrying such large reserves.
Resurgent inflation. A big economic story of 2021 was the reappearance of significant inflation. Regardless of the cause or how permanent it is (both subjects of vigorous debates), it will affect the spending side of the budget, which is only partially reflected in the forecast of expenditures. This modestly deflates the size of the surplus because the state will need to pay more for labor and other inputs. Some wonks argue for not including inflation in the forecast because omitting it is a sort of “zero based budgeting lite” and it allows the legislature some slack in making spending decisions. (I personally think that is modestly dishonest intellectually but recognize the reality of human decision making and the power of inertia, anchoring, and similar that biases decisions toward the status quo, which is not good as an abstract policy making principle.)
Political context
In November the governor and all legislators will be on the ballot. That will make them very sensitive to the political ramifications of their decisions. Framed another way, it creates a strong incentive to enact policies that pander to their bases or/and attract swing voters or ideally, both. However, with divided government DFLers and Republicans must compromise to do so. A big surplus creates the risky prospect of doing both-and rather than either-or or nothing-at-all.
To summarize, the parties have deeply polarized views – Democrats believe that government should act to improve society’s lot through more spending on education, social services, childcare, etc. By contrast, the GOP strongly believes government is mostly the problem and that it already spends too much and that advocating tax cuts is the ultimate way to appeal to both their base and most swing voters. Although in their heart-of-hearts, DFLers disagree, many of them recognize the visceral appeal of tax cuts, especially with the state projecting a massive surplus. Tactical/strategic political considerations enter into the parties deciding whether to compromise or to advocate for a position that is unacceptable to the other side, hoping the voters will give them complete control in November. This segues into my next two sections.
What GOP likely will propose
In predicting what I think each party will propose, it is useful to consider three factors: (1) the party’s base supporters, (2) its political philosophy of taxation, and (3) what they think appeals to swing voters. That typically dictates what they propose.
GOP base
Old
White
Rural
Business owners including more prosperous farmers
Tax policy credo
State government is too big (a lot of wasteful spending, particularly on urban problems and to promote social welfare rising to outright redistribution) and needs to be limited.
Minnesota’s high state and local taxes are a big drag on economic activity, causing people to move out of state, restraining business investment, and crimping job growth.
Income and estate taxes are among the worst taxes because they penalize (hold back) society’s most successful and hard working members.
Applying these basic principles means constantly advocating for tax cuts of any and all flavors, but especially income and business tax cuts.
What they think appeals to swing voters?
Tax cuts – given their governing philosophy and their perception that most people want to keep as much of their money as possible and perceive only marginal benefits from existing or expanding government services, this follows as day follows night.
Likely 2022 tax proposals
Given this background, my best guess is the GOP legislative caucuses will propose a host of tax cuts:
Income tax exemption for social security benefits (skewed toward older white people; rural areas are both much whiter and older than the metro). My reaction: as I have opined this is bad policy because it is unfair, looks backward rather than forward, and creates a large and growing hole in the tax base as the baby boom ages.
Pay off the deficit in the Unemployment Insurance (UI) fund (business). My reaction: this isn’t handled by the tax committees, but businesses and the public think of as a tax issue. It’s a one-time cost and would be a good way to provide a tax cut without an ongoing budget obligation. The governor now seems willing to accept this.
Repeal the state general tax, the property tax paid by businesses and cabin owners (business). My reaction: This is a bad idea, because it is the best business tax the state imposes. It applies to all businesses that own or lease property and is correlated with benefits they received in government services. Cutting the corporate tax would cut a narrow and volatile tax that discriminates against businesses based on the type of organization (C corps only) and whether they are profitable. The state general tax, by contrast, adds some stability to state revenues, which are increasing subject to economic ups and downs and applies to any business of consequence. But DFLers oppose corporate cuts because they perceive it would benefit big businesses that have done okay during the pandemic and the tax is one of the few that polls well.
Income tax rate cuts (base and swing voters because of its simplicity and PR appeal). This is a staple of GOP policy.
Large cuts in or full repeal of the estate tax (business owners are prime beneficiaries). Me: Although the estate tax applies to very few people (less than 2% of the population), it seems to be the tax GOP legislators on the tax committees hate the most. That is based, I think, on the misperception that it is a “double tax.” This thinking is based on thinking it applies mainly to saved income (wealth) that has been taxed under income tax. But the majority of taxable estates are comprised of unrealized capital gains that will never be taxed without an estate tax.
A potpourri of tax expenditures supported by interest groups that fund their campaigns or are part of their base. In sessions since 2010, GOP legislators have moved away from cutting taxes for everyone (e.g., across the board rate cuts) in favor of special deductions, exemptions, and credits advocated by interest groups. With the big surplus, it will be difficult for them to say no to more of these. Since I’m now several years removed from working in the process, I don’t know what is likely to be proposed specifically. But there are always a host of these proposals. My reaction: They reflect the reality that despite their anti-government rhetoric, many GOP legislators run because they want government to do stuff; they just don’t want it to look like government spending. Hence, they turn to tax expenditures, which are government intervention in the economy but do not show up on the books as spending. Most of these are bad ideas, whether done through the tax code or as direct spending, but they need to be evaluated individually and whether it is better to do them as direct spending should also be evaluated but rarely is. Their big appeal is that they are tax cuts and typically help a favored group/constituency.
What DFL will propose
DFL base
Urban
Minorities (well, less white than GOP)
Younger (compared to GOP)
More educated typically
Tax policy credo
Robust government services are needed across the gamut (education, health care, social services, housing etc.), which requires imposing higher taxes.
Progressivity – taxes should be designed so that high income people pay more as a percentage of their incomes – this is the prime tax principle, which dictates which taxes they look to cut (regressive taxes like property and sales) and increase (income taxes particularly on the “rich”).
Higher taxes have little effects on behavior, such as migration or business investment; people are attracted to a state more by its quality of life (translation: good government services and amenities) than by low taxes and services.
What they think appeals to swing voters?
In line with their views of the value of government services, DFLers are much less convinced that swing voters respond to tax cut promises/proposals. But the Reagan era legacy makes many of them (particularly the less lefty types and those representing swingy suburban and Greater Minnesota districts) unsure about that and more likely to tout tax cuts, if only rhetorically or as a defensive counter to the tax cut tattoo that Republican drums consistently beat out. Similarly, they’re not afraid of proposing to tax the rich and corporations, two ideas that don’t automatically get negative responses in their polls and make funding government programs an easier sell. (Side observation: both parties’ core messages involve a healthy dose of free lunch economics.)
Likely 2022 tax proposals
It’s harder to predict what DFLers will propose on tax issues, because their core agenda is programmatic spending with taxes consisting mainly of means to those ends. But some guesses:
Expand low-income tax credits. Making the state’s earned income credit, the Working Family Credit, more generous is a consistent DFL proposal. It is essentially a wage subsidy for low-income workers (it’s refundable so it does more than offset income tax liability – a case can be made that it offsets the sales and property taxes those families pay directly or indirectly).
Since a key national Democratic proposal is the expanded federal child credit (a refundable credit for families with young children), I would not be surprised with such a large surplus, if DFLers propose that Minnesota adopt some version of this credit. Unlike the Working Family Credit, it is not tied to wages or self-employment income.
In a defensive move against inevitable GOP proposals across-the-board income tax rate cuts, DFLers may propose more modest rate cuts on the lower brackets, partially offset by higher taxes on the upper end.
Tax expenditures targeted at their base (young and low income) – e.g., childcare credit expansion, more help for student loan debtors, first time housing breaks, etc.
What I would do
Now that I’m no longer advising legislators, what I think or would recommend is totally irrelevant. (When I was working it largely irrelevant.) But after 40+ years, I can’t stop myself from going through the exercise.
Focus on one-time tax cuts and other items. My general mode in situations like this was to urge caution. With big surpluses the tendency to overexuberance, to get out over your skis, is irresistible for legislators or at least the general process pushes them in that direction. In a year when they are all on the ballot, that tendency will be even greater. As an operational matter, caution means doing one-time items that do not commit a future legislature to continued spending or tax cuts. It is even better if those one-time items reduce future liabilities, essentially pay off implicit debts, since that it will create fiscal slack for future legislatures when inevitable downturns occur or new opportunities arise. The pandemic is not over; there is a good deal of uncertainty as to how much of the revenue increase is permanent; and the budget reserve, as I suggested above, provides less than robust protection against downturns. One-time tax items can be put into three categories – (1) paying off deferred obligations, (2) shift reversals, and (3) explicit one-time tax cuts.
Clean up deferred tax obligations. All these options are business tax cuts and will increase revenues in later tax periods, a twofer. Also, because these provide tax benefits to businesses that have invested in new equipment, they channel money to business that are a bit more likely to be expanding or investing.
Immediately allow previously disallowed section 179 deductions from tax years before 2020 – these amounts are due to be claimed in five annual installments after the year the federal deduction was claimed. These amounts could be allowed in full in tax year 2022. This will have the effect of accelerating when the deductions are claimed and increasing revenues in later years (2023 and later).
Conform to bonus depreciation. Minnesota has not conformed to federal bonus depreciation because the resulting revenue reduction (tax cut) was too big. Now is the time to do that with the very large surplus. Because bonus depreciation is a temporary provision (set to start phasing down in tax year 2022), this is not a permanent tax cut. (Of course, if Congress extends it, the legislature will need to decide whether it can afford to do so as well.) Thus, it qualifies a one-time revenue reduction.
Immediately allow previously disallowed bonus depreciation deductions. This parallels the suggested treatment of disallowed section 179 amounts. These deductions are also claimed in five annual installments after the federal allowance. This could be accelerated into tax year 2022 or in two installments in 2022 and 2023. This makes most sense if the state conforms to bonus depreciation.
Reverse more shifts. The November forecast did not reverse all the shifts that were adopted in the early 1980s. At least two more remain that could be candidates for one-time tax cuts or spending (not sure how one should characterize them):
Pay some LGA and CPA in the first half of the state’s fiscal year. In the old days (late 1970s), the state made LGA and homestead credit reimbursement payments (now replaced by CPA) in March. To defer that obligation outside of the state’s biennium (i.e., to “shift” it), the legislature changed the payment schedule, so all aid is paid in July through December. That yielded one-time budget savings equal to the former March payments. Because cities, towns, and counties must wait until over halfway through their fiscal years (they operate on calendar year bases) for their state aid, it requires them to carry large reserves to cash flow their spending. Reestablishing March payments would be a onetime cost (a June payment could also be made) and would give the local governments more flexibility, enabling them to carry lower reserves.
Allow renters PTR to be claimed on the income tax. Before 1981, the renters’ property tax refund was allowed as a refundable credit against the income tax. (That was also true for homeowner refunds paid to senior citizens – because of property tax changes that delayed when statements are sent out, it is not practical to return to doing that.) Another early 1980s shift moved those refunds to August. It might be possible to go back to allowing renters to claim their refunds on the income tax. I’m not sure if that is administratively feasible, given DOR’s information systems. But that is what most states do and the fact that Minnesota PTR program counts as spending rather than an income tax reduction pushes up Minnesota’s income tax ranking in national comparisons using census data. Even organizations as sophisticated as the Tax Policy Center do not appear to recognize this difference as I have blogged about.
Provide one-time tax cuts. The simplest way to do this is accede to the business community’s request that the state use the surplus to pay off the UI fund deficit, shortcutting the pending increase in the UI tax. As I noted above, that’s not within the tax committees’ jurisdictions but it is a tax cut (preventing a tax increase that otherwise would occur) and Governor Walz appears to be open to the idea. Another option would be to pay tax rebates as was done under Governor Carlson (1997 and 1998 property rebates) and Governor Ventura (1999, 2000, and 2001 sales tax rebates). All of those were popular and helped the legislature to use the mega surpluses of the late 1990s without permanently committing to either spending increases or tax cuts that it could not keep when revenues returned to normal. There does not appear to be any political interest in rebates now, though.
Making longer term tax cuts that will help the state’s economy – i.e., “investments” (to use Democratic jargon for public policies) that arguably will help spur economic growth. The traditional way to do that is to provide tax breaks for capital investment (e.g., allowing sales tax exemptions on equipment and construction materials, investment credits, or similar). However, the economic evidence for the efficacy of such provisions is thin, at best. So, the case for them as true public investments is weak. Sales tax exemptions for businesses purchases are the correct tax policy, so there is that. One big advantage is that they are relatively easier to repeal than many special tax breaks.
What now appears to be holding back business, as much as anything, is access to high quality labor, not inadequate or too expensive financial capital. It is crucial to attract and retain young workers, especially STEM professionals and similar. To help with that problem, it might make sense to work on tax policies that encourage workers to come to or stay in Minnesota. Expanding the tax credit for student loan payments might be one area to consider. Simple changes making Minnesota’s credit more generous are unlikely to be effective, if objectively measured. At least, I would be surprised if many young workers evaluate whether to move to or stay in a state based on a state tax credit that modestly helps them pay off their student loans.
Out-of-the-box changes to the credit, perhaps on a pilot basis, could help. For example, the state could provide a much more generous student loan credit and guarantee it for a period of years (at least 3), but only if the employer agrees to participate by matching the credit. This could have three good effects – (1) shifting some of the cost to the employer via the match, (2) getting employer buy-in so that they market the credit to employees and potential employees as part of their recruitment and retention efforts (marketing is crucial, in my opinion – the employees need to see it in their compensation offers or their salary increases, which does not happen for regular tax credits), and (3) a multi-year guarantee seems crucial to induce individuals to rely on it in making important decisions about where to live or work. The state cannot legally guarantee to maintain tax provisions (that violates article X, section 1 of the Minnesota Constitution), but a de facto guarantee can be achieved by setting aside money in a dedicated account for participating businesses. An out-of-the-box idea.
Repeal the corporate AMT. I would be remiss if I did not point out the practical need to repeal this orphan feature of the Minnesota corporate tax. The TCJA eliminated the federal tax that the Minnesota tax is based on. The legislature chose not to follow that lead when it conformed to the TCJA because of the revenue cost. The surplus now makes that possible. It is not practical for Minnesota to continue requiring the complex calculations necessary for this tax. There is, of course, little political constituency for doing so (beyond employees of DOR, corporate tax departments, and accounting firms) – the corporate tax is relatively invisible and is paid by a small subset of businesses (profitable C corporations). The AMT affects an even smaller group, typically profitable businesses that heavily invest in depreciable equipment. Repeal is something that should be done but probably won’t, unless one of the key players (governor would be the most likely) insists on it. (Congress may solve the problem if BBB is enacted with a new federal AMT that Minnesota can conform to. The federal proposal is a suboptimal idea but it would be administrable.)
Along these same lines, adopting legislation that conforms to the federal changes in the tax base made since the legislature passed conformity legislation (in 2021) should also be done to make life tolerable for both taxpayers and tax collectors. That’s often a difficult matter when there is a tight budget. It should not be with a massive surplus.
Ditch the stadium reserve and use the money to shore up pensions. I have previously made clear my thoughts about the stadium reserve. The reserve serves no legal or financial purpose as security for payment of the bonds. It was an accidental side effect of the design and politics of the stadium financing. The November forecast now predicts the reserve will exceed the outstanding stadium bonds in FY 2024 and by over $100 million more in FY 2025. Whatever you think about the burden of the stadium debt, it is fixed and declining. There is no need for a reserve; no other general fund appropriation bond issue has one. By contrast, state and local pensions impose implicit debt obligations (future promises to pay benefits) that are highly uncertain and potentially growing burdens on state, school, city, and county governments. A better use of the money committed to the stadium reserve would be to reduce the risk public pensions pose to Minnesota’s fiscal well being.
Recent news on pension investments has been good (see here for 6/30/21 actuary report showing MSRS contribution rates are adequate to fund the plan). But those valuations are based on an aggressive return assumption of 7.5%. After a dozen years of very healthy investment returns, regression to the mean is to be expected. Although it goes against human nature (behavioral economists refer to this as recency bias – the tendency to think the recent past will continue), now is the time lower expectations of future returns and reduce the investment return assumption. That is precisely what the MSRS actuaries recommended in their report:
In our professional judgment, the statutory investment return assumption of 7.5% used in the report deviates materially from the guidance set forth in Actuarial Standards of Practice No. 27 (ASOP No. 27). In a 2021 analysis of long-term rate of investment return and inflation assumptions, GRS suggested that an investment return assumption in the range of 5.71% to 7.00% would be reasonable for this valuation.
Reducing the investment return assumption will increase unfunded pension obligations. Participants in the pension plans are sure to oppose that because it will prevent reducing contributions (for active members and employers) or increasing benefits (for retirees). Redirecting moneys currently expected to go into the stadium reserve, however, could reduce that opposition, while de-risking Minnesota governments’ pension obligations. It would also invest the money in higher return assets (pension investments versus the almost zero that the state’s idle cash reserves earn).
It would be silly to continue to make the Byzantine stadium reserve calculations. The simple approach would be to repeal the reserve and to make annual statutory appropriations, divided among the state’s pension funds in some appropriate manner, equal to the current estimates for growth in the stadium reserve. Those appropriations would end when the deposits to the funds, as estimated by MMB or the actuaries, are sufficient to offset the effects of reducing the assumed investment return.
Obviously, none my suggestions have any political sex appeal and, thus, are presumptive nonstarters.
What will happen
As I have said before, predicting what the legislature will do is a fool’s errand. Divided government increases the degree of difficulty. This is so because you have five parties (governor and legislative caucuses in two houses) playing a complicated game of poker where advancing one’s policy and political agenda depends upon getting agreement from opponents who are trying to do exactly the opposite. Intraparty dynamics within legislative caucuses and lack of knowledge about the other side’s thinking makes this poker even more complex. So, the more reliable and accurate predictions are, the more banal and meaningless they become. I may write a more detailed post about the factors involved and the complexity before the beginning of session (or not).
In any case, here are some banal predictions, which may or may not come to pass:
The surplus is so big (of course, the February forecast may trim it some or a lot) that it seems inconceivable that some sort of tax bill will not pass that cuts taxes. Yes, the parties have wildly different views on the advisability of and how to cut taxes, but it will be hard to let those differences result in a stalemate. DFLers have deep reservations about tax cuts, especially larger permanent ones; they know that it is much easier for future legislatures to cut spending than to enact tax increases (most anyway) and their agenda is mainly about spending. But they will fear the popular perception that the surplus is so large that some of it must go back in tax cuts. And the GOP Senate is unlikely to agree to much of anything unless they get some meaningful tax cuts that check their political boxes.
Funds will be transferred to the UI fund to forestall imposing taxes to pay off its deficit. Governor Walz’s signaling his support should make this inevitable.
Some sort of expansion of the income tax exemption for social security also seems inevitable. This bad policy (in my view) will be dictated by the political power of older voters, the fact that the GOP has made exemption their mantra ad nauseum, and that some/many DFLers think that exempting social security is either okay or politically inevitable. (One meaningless data point supporting the last point is that I have observed multiple lefty commenters on MinnPost supporting the exemption – despite the fact that it objectively goes against their typical tax fairness principles. I assume these commenters are DFL base activist types of the elderly variety.)
Further reduction or repeal of the state general tax also seems likely because the GOP will insist on it. They have many owners of small businesses or commercial real estate in their base (reliable contributors to their legislative campaigns) and the Chamber will continue to goad them to get it done. Since this is a fixed cost (non-growing budget hole), DFLers are more likely to accede.
Some sort of modest cut in lower bracket income tax rates and myriad of minor sales tax exemptions (many project-specific) would be another good guess but somewhat less likely.
Of course, one should never underestimate the ability of events, personalities, distrust, and other factors to get in the way of compromise, yielding gridlock. That might be the most fiscally responsible thing to happen, so it likely won’t.
WaPo is reporting that last week JP Morgan sent a note to its clients saying that there was a 90% chance of a recession, according to market indicators. Events on Sunday (i.e., Saudi Arabia’s oil moves) and movements in the bond market today (the entire Treasury yield curve was below 1% when I wrote this!) provides more evidence for that. Of course, the stock market is imploding, but I regard the bond market moves as more telling. None of this means that it is a sure thing that the economy is heading into a recession, but the probability is much higher than it was a week or month ago. It’s unlikely to be 90%, but that’s probably closer to reality than the 29% or 37% reported in February forecast.
The implications for state budget policy seem obvious. With regard to the projected budget surplus, only two things should be done:
Appropriate money to help prepare for and mitigate the effects of the coronavirus. According to media reports, the legislature is doing that. It should speed things up.
Move the rest of the surplus to the budget reserve or leave it on the bottom line.
If a new budget forecast were done now (well, after a new GII estimate is available), it seems obvious the surplus would be gone and there would a shortfall of some size. It is imprudent, as a result, to spend general fund money or cut taxes – anything that diminishes general fund resources – unless you would be willing to do that without a surplus. Cutting unnecessary stuff would be wiser, but likely politically impossible. Giving the governor authority to unallot without meeting the stringent requirements now in law (i.e., exhausting the budget reserve) would be a good move, but probably would be opposed by DFLers and so will not be seriously considered.
With regard to bonding, by contrast, it is fine to proceed with a robust bill. The lag between passing a bill and ramping up projects means that doing so can’t be justified under a rationale that it will allow taking advantage of the big dip in interest rates. But these extraordinarily low rates may persist for some time (lets hope not, because that means the economy is in bad shape) and although it is counter to one instincts (based on personal/household budgeting rubrics), borrowing and making public capital investments during a recession is a good thing. So we might as well get the projects and bonding authorizations on the books to allow that to happen if the recession is a long one or if low rates persist for other reasons.
On a separate issue, Bernard P. Friel wrote a letter to the editor responding to John James’ and my Strib op-ed on taxing social security. Incongruously (to me, anyway), the Strib made it the spotlight letter on Sunday. His essential point is that state taxation of social security is inconsistent with both the original purpose of social security itself and the 1983 law that subjected benefits to federal taxation. The former was intended to promote the general welfare by providing old age benefits and the latter to shore up the financing of those benefits. He goes on to write “There is nothing in that preamble or in the [social security] act suggesting that part of the benefits provided should be used to reform a state’s tax system.”
This response to our op-ed is a total non sequitur, of course. The purpose or policy behind a federal law does not dictate state tax treatment. Congress is perfectly capable of limiting state taxation and has done so repeatedly (e.g., prohibiting states from imposing individual income taxes on interest paid by Treasury securities, taxing railroad retirement benefits, limiting taxation of railroads, air carriers, and so forth). It has not done so for social security benefits. Consider a counter example: Congress’s purpose in paying civil servants and their pensions in no way relates to state taxation and state taxes on those payments could be viewed as making it more difficult for the federal government to achieve its underlying objectives. But that wouldn’t lead one to conclude that states should not tax federal employee pay or pensions under their income taxes would it? Well, for some years Minnesota did not tax federal salaries (until the late 1930s, I believe) and did not tax federal pensions until the 1970s. This was due to an expansive, old view of intergovernmental tax immunity doctrine. I wonder what Mr. Friel’s views were of those decisions as policy matter?
John and my views are not intended to “reform” the Minnesota tax system. Rather, we seek to preserve the status quo, which has been in place for 35+ tax years. An entirely different matter. We’re trying to prevent “deforming” of the state tax system, something much less ambitious than reform.
I mention all of this only because I used to know Bernie. At least, I assume that the Bernard P. Friel who wrote to the Strib is the same Bernie Friel that I knew as a long-time partner in the bond department at Briggs with several blue chip clients when I started working at the House. I can remember him calling up to lecture me on how TIF law had gone astray when it was being used to subsidize low-income housing and another time on how the Dorsey bond lawyers we’re pushing the envelop in some deal that they had signed off on. I’m glad to see that he is still kicking around and as opinionated and feisty as I remember him. He must have retired over 20 years ago and be in his 80’s.
Note: this post has been updated. The changes corrected various minor errors and did not make any substantive changes. If you previously read the post, there is no reason to re-read it.
The Strib published John James’ and my op-ed on legislative proposals to exempt more social security benefits from taxation. After writing the op-ed and my earlier blog post, I had a couple more thoughts about the topic and this post records them, focusing on two items: revenue volatility and “declining returns” on social security for younger workers. Space limitations, as well as a lack of general interest by newspaper readers, did not make either of these topics appropriate to address in the op-ed.
Revenue volatility. A point that I failed to make in my blog post was that exempting social security benefits from taxation would make Minnesota’s tax revenues more volatile, as well as less elastic (the latter point was made in both pieces).
A little discussed fact is that the last three major states tax changes –the 2013 increases by the DFL, the 2017 GOP tax cuts, and 2019 compromise tax bill – all made Minnesota’s tax revenues more volatile. The combined effects of those changes was to increase volatile revenue sources, while reducing more stable revenue sources.
To my knowledge, no one has tried to measure the resulting effects on volatility. When we’re in a period of steady economic growth (the situation since 2010), interest in volatility or how much revenues rise and fall with the business cycle declines. One way to calculate volatility is with standard deviations (or any of a variety of related statistical measures of variance) or with the cyclical swing index, a measure which a couple of coauthors and I calculated about 30 years ago for some of Minnesota tax revenues [House Research Department publication not available on the Internet].
Here’s why the 2013 – 2019 tax changes increased volatility:
2013 tax increase. The 2013 increase had three main components: (1) adding a fourth rate bracket at the top the individual income tax rate schedule, (2) a cigarette excise tax rate increase, along with indexing the rate for inflation in cigarette prices, and (3) a corporate tax increase implemented by expanding the base mainly to more income related to foreign operations. Two of these changes yield relatively volatile revenues –
The income tax rate increase applies to filers with the highest incomes (> $250K). These filers have a very large share of the most volatile sources of income – capital gains and business income. The volatility of revenues from taxing capital gains is quickly revealed by skimming a few MMB forecast write-ups – both how difficult they are to predict and how much they fluctuate. It’s like tying your revenues to the stock market.
The corporate tax is the most volatile of the state’s major sources of tax revenue. Twice over the last four decades it has experienced 50% drops in its revenues – in the double dip recession of the early 1980s and during the Great Recession. In any recession, it typically declines in nominal terms.
The third change, a large increase in the cigarette excise tax rate, by contrast, yields stable revenues since it is a per cigarette amount and smokers are addicted, dampening down their ability to avoid the increase by stopping or reducing their smoking. (The tax is a declining revenue source, however, because smoking is declining despite its addictive nature.) The 2013 bill did eliminate the fixed dollar feature of the tax by indexing its rate for price increases in cigarettes, making the revenue less inelastic. Most of the money in the 2013 tax bill was in the income tax rate increase.
2017 tax reduction. When the GOP controlled both houses of the legislature in 2017-18, they maneuvered Governor Dayton into signing their 2017 tax cut bill. This bill focused much of its reduction on the most stable source of state tax revenues by reducing the state general tax levy and eliminating its indexing. The state general tax is as a fixed dollar levy that is paid almost without regard to underlying economic conditions. By repealing the indexing of both cigarette taxes and the state general tax, the bill also reduced the elasticity of tax revenues.
2019 tax bill. This was a compromise bill whose principal features were to conform to TCJA’s changes (the 2017 federal tax act) and to enact a modest individual income tax rate cut. The rate cut is pretty much neutral on volatility, although income tax revenues are more volatile than sales tax revenues (so a sales tax rate cut would have increased volatility more than an income tax rate cut). The bigger effect resulted from TCJA conformity. TCJA changes included a substantial number of business base expansions, which were used to help offset its revenue losing provisions (mainly rate cuts and faster cost recovery). Minnesota conformed to these base expansions – in other words, corporate and business income tax increases – and used the revenue to help offset the individual income tax changes. The net effect makes Minnesota’s tax revenues more volatile because of the volatility of taxes based on corporate profits and business income.
Thus, the legislature has been increasing volatile revenue sources and cutting stable ones, while taking some of the elasticity out of the system by repealing indexing provisions that preserve the tax base from eroding with inflation. With regard to the latter, no one seems to be concerned about or attempting to repeal the indexing provisions that protect taxpayers in the individual income tax – suggesting the GOP aversion to indexing is just really a feature of their general tax aversion, not some sort of philosophical principle (e.g., that the legislature should be compelled to vote on any and all inflation adjustments because otherwise tax policy would be on autopilot or something like that).
By contrast, the tax on social security benefits is a very reliable and stable revenue source. Since these federal government payments are required by law, they are almost as reliable as collecting treasury bond interest. Not quite – they depend upon the recipient’s other income sources and if the social security trust fund is ever allowed to run dry, the law provides for an automatic haircut in benefits. That is now estimated to occur in about 15 years, but Congress seems unlikely to let it occur because of the political blowback.
Bottom line: exempting social security benefits from taxation does not score well on the reliability element of the revenue adequacy principle. By eliminating a very stable revenue source, exempting more social security benefits from taxation will make Minnesota tax revenues more volatile. No one seems to ever talk or care much about these issues until a deep recession hits and, then, concerns rise. Changes over the last 7 years have tended to make Minnesota tax revenues more volatile. Given recent events related to COVID-19, I would not be surprised if this doesn’t become a topic for legislative discussion sooner, rather than later.
Relevance of “declining returns” on social security for younger workers.
A few people responded to John and my op-ed with disbelief that the 15% allowance actually allows a fair and full recovery of their employee contributions to social security, negating the double taxation argument that is often made. This skepticism is fueled by media coverage about how bad a “deal” social security now is for current workers with “negative returns” for some, particularly those with high lifetime earnings. I can easily see how the typical media stories could lead to the expressed skepticism about John’s and my assertion. So, I decided it would be useful to gather and state my thoughts on it.
It is a case of comparing apples and oranges. The Social Security Actuary’s estimate of the 15% allowance and the estimates of how “good a deal” social security is for today’s workers are measuring or calculating two related but crucially different things. That is apparent by analyzing what each is doing and why.
Estimates or calculation of the “return” on social security. I tend to see estimates of these hypothetical “returns” presented in three different contexts:
The Social Security Actuary has long prepared estimates of “Money’s Worth Ratios” for various hypothetical Social Security participants. Here’s a link to the most recent version that I could find on SSA’s website (published January 2019). You can see in its tables (##1 – 6) how younger workers are paying in more and getting out less than older generations.
Economists evaluating Social Security make similar calculations. Social security is, of course, not a pension plan, but rather a complex social insurance program that by design redistributes income from some groups to others. In addition to retirement income, it provides de facto life insurance and disability protection and is a pay-as-you-go program, not a pre-funded insurance or pension plan. Economists, of course, are interested in measuring exactly what that program does and in evaluating its effects. See here for an analysis along those lines by economists at the Urban Institute. These analyses, like the SS Actuary’s, show declining returns if getting a “return” for your contributions (as well as your employer’s contributions on your behalf) is how you want to characterize social insurance protections. I tend to think of it as society making sure that all of its members have a basic safety net protecting them – whether because of matters beyond their control or simply because they failed to appropriately prepare. The latter, of course, irritates those with more libertarian sensibilities than mine.
When privatization of social security was a hot topic (e.g., during and shortly before the second Bush administration), advocates of it often made a big deal about how workers would be financially better off if they could just keep their money and invest it in private accounts like a 401(k). There are many heroic assumptions in all of that, of course. These sorts analyses were popular publications by right-leaning or conservative organizations like Heritage (see the “A Bad Investment” section).
What all these estimates or calculations have in common is estimating or measuring the sum of three things: (1) employee contributions (e.g., the after-tax employee FICA payments), (2) employer contributions (e.g., employer FICA or the deductible portion of SECA), and (3) some sort of assumed return or present value calculation. Different approaches are used for (3). Advocates for private accounts base their calculations on some sort of assumed investment return (typically a mixture of stock and bond returns) that parallels how individuals might invest a defined contribution retirement account, like an IRA or 401(k). By contrast, actuaries and economists more typically discount the employee and employer contributions to present value using government bond interest rates. In any case, depending upon the assumed rates used, the value of (3) can overwhelm (1) and (2) amounts.
I can illustrate this with an example from my personal life. Before we married, my wife worked for a nonprofit which had no retirement plan. At the recommendation of her tax preparer she annually made traditional (deductible) IRA contributions to reduce her tax liability. These amounted to $8,000 of contributions, which she invested in bank CDs and earned a few hundred dollars of interest. When we married in 1990, the IRA was worth less than $8,400. To get better returns, I recommended that she move the money to a mixture of stock and bond mutual funds. In the 29 years since, the $8,400 has grown to over $240,000. (No Warrant Buffet was investing the money – just me in plain vanilla mutual funds.) The ratio of contributions to current value is 3.5%. This simply illustrates how important the assumed return or discounting to present value calculations are, given the long periods of time involved, and why 15% is not as low a ratio as it may initially seem.
A second factor to consider is that, as noted above, Social Security also functions as a quasi-life insurance policy by providing dependent benefits if the worker dies and as a disability policy, if the worker becomes disabled. Effectively, one could argue, some sort of annual deduction from the contributions should be made to pay for premiums on that insurance. The estimates typically do not do that (particularly those making the case for private accounts), so they overstate how much is being contributed to the pure retirement/pension-like portion of the program.
The Actuary’s estimate of 15%. By contrast, the Social Security Actuary’s estimate of a pension equivalent recovery ratio (i.e., the 15%) looks at only one of the three components – i.e., employee FICA or the nondeductible part of SECA – and, then, must determine how much those raw amounts contribute actuarially to the payment of old age benefits. These are the only amounts on which tax has already been paid. If one has a background in finance, you can intuitively see why these amounts would be so low relative to benefits paid. The important point is that the Actuary, according to the description (p. 16) by the Congressional Research Service, made the estimates for workers entering the work force in 1993 (the year the estimate was made). Thus, the estimate takes into account the declining “returns” documented in the Money’s Worth Ratios. And, as I noted in my original post, there have been no material changes in social security’s tax or benefit rules that would suggest that these calculations are no longer valid. Longevity has increased and interest rates (and inflation) have declined, offsetting factors. It would be good to update the estimates, but in the absence of that, I see no reason to conclude that the 1993 one is still not valid.
Bottom line: The declining “returns” on social security participation for younger workers says nothing about the validity of the Social Security Actuary’s estimate of the 15% ratio. It continues to be the best estimate of the appropriate amount to allow recovery of after-tax employee contributions.
Since the mid-1980s Minnesota has taxed social security benefits following the federal income tax rules. The 2017 and 2019 legislatures began chipping away at the practice by allowing a state subtraction in addition to the roughly 60% to 70% of benefits that are exempt from federal tax.
In a January press conference, the Senate GOP caucus said that a centerpiece of their 2020 tax proposal is to exempt all social security benefits from Minnesota tax (covered by the Star Tribune here). I will begin collecting benefits when I turn 70 in December, so my instinctive, selfish reaction was “Yay” – I’ll be able to use my tax cut and take that long-dreamed spring training trip. (A quick calculation indicated I would save $3,000 in tax on my $45,000 in annual benefits. To provide some context, the Department of Revenue estimates that the subtraction will yield average savings of $1,070 for the 358,000 taxpayers who would benefit. Since the Trump tax cut limited my deduction for state and local taxes to $10,000, a lot less than I pay, that $3,000 would be pure savings to me, no loss of federal tax savings by paying less Minnesota tax!) But my better angels told me to slow down, put aside my selfish interests and think about whether the exemption is good for the rest of the state. After a little thought, it was obvious to me that it was not, and I decided a better option is to ask the Senate Republicans to come up with a better use for the money.
This blog post describes my thinking. It consists of three parts:
The first describes the three reasons why the proposed exemption does not make sense to me: it’s unnecessary, unaffordable, and unfair.
The second part discusses why I think the typical arguments for the exemption do not hold water.
The third part discusses the elephant in the room, politics.
PART ONE: WHAT’S WRONG WITH THE EXEMPTION
Unnecessary. Most social security benefits (about 60% to 70% according to the Congressional Budget Office) are already exempt under the federal rules that Minnesota uses in starting its tax calculations. The federal rules that the determine how much, if any, benefits are taxable are a complicated mess of income thresholds and phase-outs that only an accountant could like and explain. But they result in recipients with the lower incomes paying no tax on their social security, middle income recipients paying some, and high-income recipients paying the most. No one pays tax on more than 85% of their benefits (more on that later).
The 2017 and 2019 Minnesota tax bills expanded the exemption Minnesota carries over from federal law to cover more benefits received by higher income recipients; maybe about 70% to 75% of benefits are now exempt from Minnesota tax. So, the effect of a total exemption would be to cut taxes for recipients with higher incomes (like me!) who don’t need the help. If Warren Buffet or Donald Trump were Minnesota residents, they would qualify. Is that really what we want to do? I don’t think so.
A natural reaction is to support government policies targeted to help the elderly. This sentiment is captured eloquently in a classic Hubert Humphrey quote: “The moral test of government is how that government treats those who are in the dawn of life, the children; those who are in the twilight of life, the elderly; those who are in the shadows of life, the sick, the needy and the handicapped.”
Exempting social security benefits will help many seniors; about 95% of them receive social security although less than half now pay any Minnesota tax on their benefits. So, can a total exemption be justified as general help for the elderly? Even a cursory review of the data and the effects of the exemption make it obvious that is not a good justification for expanding the exemption.
First, national policy has succeeded admirably in pulling seniors out of poverty. The Table compares the poverty rates for seniors, the general population, and children.
It’s clear which of the two groups could most deserve help and it’s not seniors. Seniors’ poverty rates are more than two percentage points lower than for those for the general population (almost 20% lower). By contrast the child poverty rate is 36% higher than for that for the total population. What may be even more telling, Census Bureau researchers have concluded that seniors’ poverty rates may be even lower than the regularly published statistics suggest. See Adam Bee and Joshua Mitchell, Do Older Americans Have More Income Than We Think? (July 2017) (poverty rates for the elderly drop by 2.2 percentage points using more accurate measures of their income). As table indicates, Minnesota is doing better overall (we’re a high-income state) and much better for seniors and children.
Second, expanding the exemption would target its help to better off seniors, not those in poverty or on the cusp of poverty. And the most help (the biggest dollar tax cuts) would go to those higher up the income scale. That follows directly from the structure of the tax rules, which subjects more benefits to tax as one’s income rises. If the goal is to help lower income seniors, a better tax cut could easily be designed at a much lower revenue loss. Many low-income seniors don’t even pay income tax.
Third, if the purpose is to help seniors who are above the poverty line, I question the wisdom of such a policy. But in any case, data from the Pew Research Center indicates that seniors as a group are doing okay. The American Middle Class is Losing Ground (2015)(“The biggest winners since 1971 are people 65 and older. This age group was the only one that had a smaller share in the lower-income tier in 2015 than in 1971.”) The government does not need to send untargeted help to older Minnesotans.
Bottom line: most social security benefits are not taxed now; seniors as a group are doing as well as or better than other age groups; helping well-off seniors is unnecessary. A better strategy is to focus state resources on the younger generation, on whom the state’s future rests, such as expanding Minnesota’s new tax credit for student loan payments.
Unaffordable. DOR estimates that exempting all social security benefits would reduce revenues by more than $400 million per year. But the current cost is only half the story. As more of the baby boom retirement tsunami washes across Minnesota’s fiscal shores, the tab will grow rapidly. The Social Security Actuary (2019 Annual Report of the Trustees) estimates benefit payments will increase by more than 6% annually through 2028 (intermediate alternative, Table IV.A1, p. 40). If we assume that the DOR estimate grows at that rate, the FY 2028 cost rises to over $600 million, illustrating the magic (or horror) of 6%+ compounding, even over just a few years. By the end of the decade, the revenue reduction will likely exceed $1.2 billion per biennium.
Even for the state budget, a billion dollars a biennium is a lot of money. Affordability is a household or personal budget concept that is misleading to use in the context of government budgets. (I used it to get an alliteration for my three points!) For a high priority item, other spending can be reduced or taxes increased. But to accommodate the exemption will require doing either or both – by a lot. Put simply, enacting the exemption will impose a big opportunity cost. Surely, legislators can find a higher priority use for the money – either a better designed tax cut (if you’re a Republican) or needed spending/investments (if you’re a Democrat who sees much more for government to do).
Bottom line: enacting the exemption will create big challenges for budget setting by either party.
Unfair. The exemption favors people with one type of income over others. All else equal, two people with the same income should pay the same tax. Why should retired parents pay less tax than their working children with the same income? Why should someone whose income comes from pensions, 401(k)s, and social security pay less than another senior who has less social security and retirement income and, thus, must work as a Walmart greeter to make ends meet? To pose the question, answers it. Horizontal equity (equal treatment of equals) is a cardinal rule of tax policy; the exemption plainly violates it. Moreover, because the social security of lower-income recipients is already exempt, the beneficiaries will be middle- and upper-income recipients. The exemption is not progressive, if that is your fairness norm. Certainly, we can come up with a fairer way to cut taxes.
Unfairness is an intrinsic problem; it is simply wrong to arbitrarily favor one class of taxpayers. But it gets worse; it will lead to an inevitable queue of others seeking similar treatment. For example, a natural group to line up for their dollop of tax cuts are public safety retirees – police, firefighters, corrections workers, FBI agents, and similar. These high stress jobs provide government pensions that are not covered by social security, because they provide for earlier retirement than social security. But they are subject to income taxation under the standard pension tax rules, no favoritism for these retirees under either federal or Minnesota law. Naturally, these recipients will not consider that fair, if all social security is exempt. They will undoubtedly seek similar or better treatment, after all they put their lives and limbs on the line for the public. Because the social security exemption is fundamentally arbitrary and unfair, legislators will be hard pressed to explain why public safety pensions should not also be exempt. The almost inevitable result will be for the legislature to exempt these pensions sooner or later, narrowing the tax base, increasing the generational unfairness, and expanding the budget challenges. Recipients with private retirement benefits, then, will likely claim they also should be exempt.
A few states have already slid down this slippery slope and exempt all or virtually all retirement income from state taxation, putting the burden of their income taxes squarely on the shoulders of working people and business owners. And it is not as if seniors do not consume much in the way of services funded by the state. Although I don’t have data to cite, a large share of state medical assistance spending goes to pay for long term care costs of indigent elderly. To my lights, it’s reasonable to expect more affluent seniors to bear some of that burden, not just the young.
Bottom line: the exemption violates basic principle of horizontal equity (equal treatment of equals) by favoring those who receive one type of income.
Because a full exemption is unnecessary, unaffordable and unfair, the policy case against it (indeed, even expanding the current Minnesota subtraction) is overwhelming.
PART TWO: “FOR” ARGUMENTS DON’T HOLD WATER
Proponents of the exemption typically make four general points or arguments for it or least they did when I worked at the legislature. As is often the case, each has some basis in fact, but none of them support the exemption when considered carefully. Here’s why:
Point 1: I paid tax when I paid into social security; taxing the benefits is double taxation. This “double taxation” argument is regularly made. Senator Gazelka in supporting the Senate’s proposal said, according to the Strib, “Let’s be one of those states that doesn’t tax our seniors, when they’ve paid all the way through, and then now they have to pay again.” That’s not exactly the double tax argument if you parse his words carefully but it implies it. Double taxation would be good argument if it were true; it’s not. Debunking this persistent urban legend requires digging into the details of the tax rules that apply to defined benefit pensions.
Social security works like a pension plan in which the employer and the employee contribute equal amounts with the employees’ contributions being subject to income tax when they are made, while the employers’ contributions are not. Employee FICA (FICA is the acronym for social security “contributions” – tax to me) is equivalent of the employee pension contribution and the employer FICA to the employer contribution. (Self-employed individuals pay both parts but can deduct an amount equal to the employer share so the income tax treatment is the same; these payments are referred to as SECA.) When the employee retires and receives standard pension benefits, most of the benefits are taxable but a portion are exempt, allowing recovery of the employee contributions without paying tax again. Each year recipients receive Form 1099Rs from the administrator of the pension plan listing the taxable and exempt portions of their benefits.
When Congress was considering imposing income tax on social security benefits, the Social Security Administration concluded it was not administratively feasible to calculate these pension equivalent amounts for each beneficiary. Social Security benefits relative to employee FICA contributions vary widely, depending the beneficiary’s earnings history (the formula is progressive with low earners getting a more generous returns relative to their contributions, for example), marital status, age of retirement, number of dependents, and many more factors. (It’s an gross understatement to say the social security program structure is more complicated than the typical defined benefit pension plan.) So, the percentage of benefits that represents a “return” over and above recovery of employee FICA payments will vary depending upon the recipient’s personal situation. When Congress set the federal tax rules in 1993, the Social Security Actuary estimated for various groups of beneficiaries the average percentages of their benefits that consisted of recovery of employee FICA (the part that would be exempt under the pension tax rules to avoid double taxation). The highest of these percentages that represents return of employee FICA was 15% for single, high earning males (the average for the entire population was 7%). See this Congressional Research Service 2019 publication (page 16) for a little more detailed description. Congress provided that 15% of benefits would always be exempt. That 15% exemption addresses the double taxation argument.
On the surface, 15% may seem very low since employee taxes generate half of benefits, right? But most pension benefits do not consist of the nominal amount contributed but rather are provided by investment return over the long period of time between when contributions are made and benefits are distributed on retirement (typically 20 to 30 years). That’s why investment professionals advise people to start saving for retirement early, to harness the effect of compounding investment returns over long periods of time. Social security is a pay-as-you-go plan, so it doesn’t have “investment return” like a 401(k) or defined plan does. But benefits are still wildly higher than the nominal employee FICA amounts. That money is coming from FICA paid by current workers and employers, as well a trust fund that is credited with amounts of FICA that exceed benefits payments and assumed interest on those amounts, using the rates on treasury securities (note that amounts in trust fund do not affect benefit entitlements unless Congress allows the trust fund to run out of the money and the automatic benefit cuts take effect). In my personal case, I will recover all of my employee FICA tax in less than four years of benefit payments and I am in a “lower return” cohort that does not benefit from the progressivity of the benefit structure. That 4-year payback ignores the value of the disability coverage social security provided to me during my working career, as well as its life insurance-like benefits. You get the idea.
The SSA Actuary estimates are now more than 25 years old. Are they still accurate? I cannot say for sure. To my knowledge they have not been redone. But I have no reason to think they are not still accurate. The social security benefits and FICA rules have changed very little since 1993. Longevity has increased, which should increase the “return on employee FICA” (i.e., the amount of benefits relative to FICA paid) and push down the percentage. But inflation and inflation expectations have both declined and inflation is a factor in the generosity of social security benefits, since they adjusted for inflation. If Congress were to cut Social Security benefits beyond making the tweaks it typically has done or to increase FICA, the 15% assumption should definitely be re-estimated. It probably would be good practice to do a re-estimate and to revise the 15% (it could down as well as up). But, of course, that would require an act of Congress and we all know how unlikely that is.
Bottom line: The 15% exemption is probably still a generous approximation for nearly all recipients and in any case, it addresses the double taxation argument in the only feasible way for the states.
Point 2: Exempting social security benefits from taxation will help prevent seniors from moving out of Minnesota. This is a frequently made argument for the exemption. (Snarky aside: it has almost become the default argument used by GOP legislators to oppose any tax change that they don’t like, particularly an increase. In my last sessions working at the legislature, the Ground Hog Day element of its repeated use devolved into a running joke that caused wise cracks and snickers by tax committee legislators on both sides of the aisle. It was almost as if a GOP member failed to make the argument when a DFL bill was up, it wasn’t a real tax committee hearing.)
Reliable and neutral evidence for this effect simply does not exist. Careful, sophisticated econometric studies of migration by the elderly that control for as many of the relevant factors as possible simply do not find any effect of state income taxation of pensions and social security. The best study in my opinion is Karen Smith Conway and Jonathan C. Rork, “No Country for Old Men (or Women)—Do State Tax Polices Drive away the Elderly?” National Tax Journal (June 2012). Although subject to all of the usual caveats that apply any econometric research analyzing complex human behavior, such as migration decisions, the authors conclude “The results from all analyses overwhelmingly find no credible effect of state income tax breaks on elderly migration.” (p. 351).
Of course, various other sources or “studies” suggest otherwise. The recent column in the Strib by the Center of the American Experiment’s (CAE) John Phelan is a good example. The underlying “study” essentially looks at two variable using IRS data – the incomes of people who moved and state taxes. When there is the expected correlation between the two (i.e., more people/income move to low tax states), the results confirm the hypothesis. Of course, a myriad of factors, other than state taxes, affect where people chose to move or whether they do at all – family ties, access to jobs or schools, weather, amenities, and many more. Failing to control for those factors and using a robust dataset (variables that capture these other factors, across time and many states) calls into serious question whether the results mean anything.
A little reflection lends intuitive support to the econometric findings that Conway and Rork (and other neutral academic economists doing other studies) reached. Consider my personal situation from two different frames of reference – whether another state’s income tax exemption for social security would be sufficient to attract me to move or whether Minnesota’s exempting of social security would cause me to stay if I were inclined to move to a low tax state. In both cases, an income tax exemption for social security seems close to irrelevant:
#1. Luring me to Wisconsin: Wisconsin fully exempts social security and although its tax system has some material differences from Minnesota’s, the two states’ overall levels of taxation are close. Thus, all I need to do to exempt my social security from taxation is to move across the border; I wouldn’t need to abandon friends, family, and community etc. to do so because Wisconsin is close. (Please tell I don’t need to become a Packer fan!!) Let’s assume (falsely!) for the sake of argument that that is the only difference in my tax treatment. Would a $3,000/year tax savings pay for me to move to Wisconsin? A little thought says no – it would take many years of $3,000 in savings just to recover the transaction costs – moving expenses, costs related to selling and buying a house (realtors fees etc.), and so forth. The numbers will not pencil out any time soon, especially if I value my time. But you say, what if I were planning to downsize in retirement and move to a new home, thus incurring those transaction costs anyway. Then, the financial calculations become more complex. Now, I must judge whether moving to a more remote, exurban location farther away from entertainment (e.g., longer tips to the Guthrie, Ordway, and Twins games), family, and friends is worth it. Ultimately, it will come down to whether those differences in lifestyle and inconvenience are worth $3,000/year. To me, it isn’t even close, but for others who prefer an exurban or rural lifestyle, it might be enough to tip the balance from rural or exurban Minnesota locations (or it might be an unneeded windfall, if they would have picked Wisconsin locations without regard to taxes).
In any case, it’s one small factor that (to my intuition) is unlikely to affect many people – just those who are looking to move and who have the correct set of preferences to make the tax benefit a decisive difference. But from the state’s perspective, it must pay everybody with social security and for most of them it is irrelevant (me!) – because they are not seriously considering moving, because the tax benefit is not big enough to be decisive, or because they were going to move out-of-state for other reasons. To yield a fiscal benefit to the state seems almost impossible. Recall that the revenue loss is over $400 million per year.
#2. Keeping me from moving to Florida: But what if I am seriously thinking about moving to a low tax state. This is the situation that outfits like CAE (and other small government, anti-taxers) think a large proportion of the population falls into, probably because that reflects their own views and preferences. Again, consider my situation. Florida is a low tax state. My 96-year old mother-in-law lives in Florida and my wife spends a week each month with her to provide care and help take care of her affairs. It would be convenient for us to move down there and we’d avoid Minnesota’s cold winters as part of the bargain. Because Florida has no income tax at all (okay, I’d pay more in sales and property tax but still a lot less than we pay in Minnesota in total) my income tax savings would be more than 4X what the savings would be from exempting my social security from Minnesota tax. So, if I’m inclined to move to Florida would exempting my social security be enough to make a difference? Probably not. If I were the type of person CAE posits most people are, a social security exemption would be thin gruel to say the least. It’s like offering a free hamburger hoping to top or equal an offer of a free steak dinner with all the fixings.
Finally and perhaps most important, how much should we be concerned about skewing Minnesota tax policy to prevent migration of seniors out of Minnesota, if their preferences are for a low-tax, low-public service type of state and local government? My response is very little or not at all for two reasons.
First, if the state is going to get in the business of explicitly competing for residents based on tax policy (seems like a bad idea to me), I think it makes much more sense to focus efforts on young people, the base on which future growth and viability the state relies, not on seniors. Forecasts of workforce shortages won’t be met by keeping seniors here. (If retaining them as workers is the goal, we should exempt their wages, not their social security benefits.) Tax reductions targeted to keep or attract new workers would seem to be the better approach. As an aside, the young are typically much more willing to move than seniors, so they are more likely to be subject to persuasion.
Second, tax policy and tax levels are only half of the fiscal equation in the competition among states to attract residents and businesses. Public services and amenities matter as well. Using $400 million per year in state revenues for either a better designed tax cut or for more public services and amenities would probably be more successful in competing with other states.
Bottom line: High quality economic research and common sense suggest that exempting social security from income taxation will not be much of a factor in keeping seniors in Minnesota. If losing or failing to attract new residents is a concern, there are better ways to deploy over $400 million per year in state resources than exempting social security from income tax – particularly focusing on young people instead.
Point 3: The federal tax rules include fixed dollar amounts, set in 1983 and 1993, and have never been adjusted for inflation. As a result, inflation has caused more and more taxpayers to have their social security benefits taxed, which Congress never intended. The legislature should exempt all benefits or at least exempt most of them to hold tax rules constant with 1983 and/or 1993. There are two responses to this argument.
First, I think that Congress likely intended to allow the dollar thresholds to slowly erode with inflation, providing a gentle transition to taxing social security more like defined benefit pensions, the correct tax policy treatment. Here’s my thinking. When Congress first subjected benefits to tax in 1983, it had already indexed many dollar amounts used in the most important tax parameters for inflation (bracket widths, personal and dependent exemptions, standard deduction, and so forth). So, it clearly knew how to do that and chose not to. By 1993 when the latest change in the social security benefits tax rules were enacted, Congress had indexed more tax parameters for inflation, but chose not to do so for the social security tax thresholds. (The IRS annually issues two revenue procedures announcing the annual inflation adjustments, one for the general provisions and one for retirement plan provisions. The 2019 basic rev proc is 28 pages long for an impression of how many inflation adjustments there are; the 2019 retirement plan adjustment rev proc adds another 4 pages.) Congress ignored calls by AARP and others to increase the dollar amounts or to index them for inflation when it passed several big tax cut bills since 1993 (two rounds of Bush tax cuts and the latest Trump tax cut), as well as many smaller ones. All of this provides good evidence that the failure to index was a feature, not a bug. That speaks for Minnesota not being concerned either.
Second, the Minnesota legislature specifically responded to this argument (made by AARP and others) in enacting the additional Minnesota subtraction for social security benefits in 2017 and expanding it in 2019. The Minnesota subtraction is indexed for inflation. See Minn. Stat. § 290.0132, subd. 26, paragraph (f).
Bottom line: The 2017 and 2019 legislatures have effectively taken care of the indexing problem, if it even is a concern.
Point 4: Minnesota is one of only a few states that tax social security; it should align its tax system with other states. It is correct that 30 of the 43 states with income taxes fully exempt social security benefits. (West Virginia is phasing in a full exemption by 2022, which will increase the number of states with full exemptions to 31.) Two states (New Mexico and Utah) follow the federal rules; the rest (11 including Minnesota) exempt some more benefits than the federal rules, but not all. The practices of other states, however, provide little basis for Minnesota exempting social security. As my mother responded when I whined “but everyone is doing it” – would you jump off a bridge if everyone else was doing it? Do you want to be a lemming? That response is almost as appropriate here; just because most other states succumb to foolish political whims does not mean Minnesota should.
Bottom line: Minnesota prides itself on having smarter policies and making better decisions than other states. This is perfect chance to prove that is really the case.
PART THREE: THE ELEPHANT IN THE ROOM; POLITICS
Point #4 leads to the natural question: if fully exempting social security from state taxation is such bad policy, why do so many state taxes do it and why hasn’t Minnesota? The answer has three elements: Minnesota’s tax policy history, how states legally tie their income taxes to federal law, and politics (the big factor).
Minnesota’s Tax Policy History. When Congress enacted Social Security in 1935, the law did not resolve whether or how benefits would be taxed. As the program began paying benefits, it fell to the IRS to resolve the issue. The IRS analogized to public assistance benefits, paid by the government to the indigent, and ruled that benefits were not taxable under an administratively created “general welfare” exemption. That decision was questionable, particularly as social security developed into a basic building block of the retirement system paying increasingly generous benefits to nearly all seniors based on their wages and self-employment income. But the exemption granted administratively by the IRS persisted until Congress opted to tax up to half of benefits in 1983 as part of a comprehensive fix of social security’s finances. (Revenues from the federal taxation of social security benefits are used to fund social security and Medicare.)
That federal practice led to Minnesota (and I suspect all other states) initially to exempt benefits from its state income tax. When Congress began taxing social security in 1983, Minnesota would have needed to enact a law, adopting the new federal rule. The 1984 legislature failed to do so, continuing the longstanding exemption. I was working for the House as a tax staffer then. As I remember it, no one publicly proposed Minnesota conform to federal taxation. The early 1980s were difficult fiscal times for the state with the legislature raising taxes several times to close budget deficits. But when the potential to conform to the federal taxation of social security came before the legislature in 1984, the severe state budget crises had mostly passed. That permitted the legislature to ignore the issue and allow the exemption to continue.
Governor Perpich in 1983 had appointed a blue ribbon tax study commission (commonly known by its chair as the Latimer Commission), which made recommendations to the 1985 legislature. Its income tax recommendations generally were to increase conformity to federal rules and expand the base of the tax. These recommendations and a growing tax reform push in Washington by the Reagan Administration provided the backdrop for the 1985 legislative session.
The Republicans had taken control of the House by a narrow majority in the 1984 election after having been in the minority for well over a decade. By 1985, the double-dip recession was well in the rear view mirror and economic good times had returned. Enacting a big income tax cut was the new Republican majority’s signature legislative priority; they billed their overall tax proposal as a billion-dollar tax cut (fudging the numbers a bit to get to a cool billion). The Senate and governor’s office were still controlled by DFLers who had higher priorities than a big income tax cut. Contentious negotiations during a long special session resulted in a broad-based income tax cut that implemented many changes, including subjecting social security to Minnesota tax for the first time. Taxing social security was a minor factor in the discussions; the major sticking point was whether to retain Minnesota’s deduction for federal income tax. The ungainly compromise bill allowed taxpayers to chose between two different tax structures – with or without the federal tax deduction. But both structures taxed social security benefits.
Although taxing social security was a low profile issue in the 1985 tax debates, I believe the legislature agreed to it for three reasons: (1) The resulting expansion of the tax base helped offset the revenue loss from the cut in the rates (a plus for the House Republicans, since a rate cut was their main goal). (2) The Latimer commission and national tax reform discussions created a political climate or consensus accepting it as an improvement in policy. (3) Minnesota then had an exclusion for pension income, which had a strong constituency pushing for its expansion as part of the tax cut. The most vocal and active proponents of expanding the exclusion were retired government employees who were not covered by social security. A prime argument they made was that private sector retirees did not pay any tax on their social security benefits. Adopting the federal rules helped indirectly to counter that argument, but more importantly proponents of a tax cut wanted to maximize the rate cut, rather than providing targeted benefits to narrow groups of recipients, such as pensioners.
1987 was the first legislative session after Congress’s enactment of the landmark 1986 Reagan tax reform. In response, Governor Perpich and legislature (both houses were controlled by Democrats in 1987) opted to dramatically simplify the state income tax by moving into very close conformity with the new federal tax rules. One of the most difficult decisions in doing so was to repeal the old pension exclusion. Amendments to strip its repeal from the bill failed by very narrow votes. The 1985 decision to tax social security benefits was one factor that enabled the legislature to repeal the preference for pension-source income. Government pensioners without social security coverage could no longer cleanly argue that they were being treated unfairly compared to social security recipients, even though only a very small portion of social security benefits were then subject to tax. Although votes on repeal of the pension exclusion were heavily along party lines (with most Republicans voting to keep the pension exclusion), the Republicans did not make this a big campaign issue in the 1988 election to my knowledge.
The 1993 federal change that expanded social security taxation to the current treatment set off a difficult discussion in the 1994 Minnesota legislature. Governor Arne Carlson did not include conformity in his tax proposal. Probably not coincidentally, he was up for election in the fall. By contrast, no senators in the DFL-controlled Senate were scheduled to be on the ballot and the Senate tax bill did provide for conforming. After much hand wringing and discussions in conference committee, the final tax bill included conformity to the federal social security tax, but it was phased in over three years (full conformity for 1997) and was paired with expanding the elderly exclusion for lower income seniors, particularly those who derive less of their income from exempt social security. The phase-in and permanent enhancement of the elderly exclusion were included at the insistence of the governor, probably to help neutralize the change as potential fall campaign issue (it never was made one as I remember it).
Legal structure linking state law to the federal income tax. States link their income taxes to federal law in two different ways. Static conformity states, like Minnesota, tie their taxes to a version of federal law in effect at a specific time. That requires a static conformity states’s legislature to enact laws to adopt later federal changes. By contrast, rolling conformity states have laws that automatically adopt federal changes without the need for any action. According to the Tax Foundation, there are 18 rolling conformity states. Rolling conformity likely represents a commitment or preference by the state to follow federal law for reasons of simplicity or similar. (Minnesota could only become a rolling conformity state by amending its constitution, since the Minnesota Supreme Court held a 1960s era law that did so unconstitutionally delegated the legislature’s power to Congress.)
The effect that this structure has on the likelihood of adopting federal changes, including those of problematic political popularity, should be obvious. The fact that it does not require action by the legislature makes it much easier to maintain conformity. Inertia is powerful force in politics, as well as physics. Some rolling conformity states likely allowed the 1983 and 1993 federal decisions to tax social security to apply to their state laws. One would expect more of these states to tax social security. The table below bears that out – 10 of the 13 states that tax some social security are rolling conformity states.
State
Rolling conformity
Politics
Colorado
Yes
Purple
Connecticut
Yes
Blue
Kansas
Yes
Red
Minnesota
No
Purple
Missouri
Yes
Red
Montana
Yes
Red
Nebraska
Yes
Red
New Mexico
Yes
Blue
North Dakota
Yes
Red
Rhode Island
Yes
Blue
Utah
Yes
Red
Vermont
No
Blue
West Virginia
No
Red
Compiled from various sources by autthor
Politics. The obvious reason, though, that most states do not tax social security is simple and obvious. The politics are compelling. Nobody likes to pay taxes. Seniors are more likely to vote and actively participate in politics than younger voters. They have time to lobby legislators and they are well organized.
Although the policy case for taxing benefits is compelling (at least to me), the countervailing arguments, as described above, are easy to make. Each of them are superficially plausible and easy to state. The responses are complex and require some thought and understanding of pretty complex tax rules; they are cannot be easy for a legislator to quickly put across while door knocking or at a community meeting. Moreover, human nature inherently seeks to rationalize or find justifications for positions that you already hold and/or are in your self-interest. It’s basic human psychology to look for and find confirmation of what one already believes is true and/or to screen out or ignore countervailing facts, especially if it is in your own interest. That puts a putative debunker at a decided disadvantage.
Given the politics and the previous decades of exemption, it seems likely that many static conformity states, unlike Minnesota, never taxed social security. Two of our bordering states, Iowa and Wisconsin, are both static conformity states that for a time did follow the federal rules. But within the last decade or so both exempted all social security benefits from taxation (Iowa still imposes its minimum tax on them in some instances). Last year, North Dakota followed Minnesota and enacted a subtraction for some benefits for lower income filers. Similar trends have occurred nationally, as states increasingly abandon taxing social security. Last year, in addition to North Dakota’s changes, the West Virginia legislature enacted a phase-in of a full exemption. Campaigns by senior groups to expand the exemptions are likely going on in all the states that still tax benefits (based on my quick Googling of a few states).
Minnesota has managed to tax social security for so long, I believe, largely because both political parties (but especially the GOP, the party of small government and tax cutting) have restrained themselves from overtly politicizing the issue. Elected Republicans agreed to imposition of the tax in 1985 and its extension in 1994. The 1985 House was controlled by Republicans and Arne Carlson, a Republican (back then, anyway), was governor in 1994. The GOP House majority from 1999 through 2006, despite supporting and enacting several large tax cuts, and Governor Pawlenty, an inveterate tax cutter, both refrained from proposing social security tax exemptions. I assume that is because they recognized doing so for the poor policy that it is: there simply are many better ways to cut taxes, if you think state government is too big.
Indeed, this should not be a partisan issue. As is evident from the table above, the states that still tax social security do not have any one partisan profile. Some are red states, others blue and purple. (I assigned my partisan leaning judgments based on the 2016 presidential election results. If the contest was reasonably close, like Minnesota’s was, I made it a purple state with no clear partisan dominance.)
More recent experience – based on positions taken by the House GOP, Senate GOP, and Jeff Johnson, the GOP candidate for governor in 2012 and 2016 – appears to reflect a decision to abandon that quiet bipartisan consensus and make it a high profile partisan issue. I don’t expect the 2020 legislature to seriously consider enacting a full exemption, since Governor Walz is on record opposing it and given its fiscal effects will likely hold fast to that. But if the GOP persists in making it a headline partisan issue for the long run, it is very likely that Minnesota will expand and ultimately fully exempt social security. The political dynamics are too compelling, based on what has happened in other states. If I’m correct, fiscal reality will require some combination of higher taxes on others and lower levels of government services, a result that is good for nobody other than higher income recipients of social security like me.
Although I don’t expect Senator Gazelka to ever read this, I hope that he and his GOP Senate colleagues rethink their position and recognize that taxing my social security (and most others as well) makes perfect sense. As nice as an extra $3,000/year would be in my pocket, there are many better ways to deploy that money, either as tax cuts or spending.
If one feels compelled to cut taxes on seniors, expanding the elderly exclusion so that all lower-income seniors, regardless of their sources of income, are treated more equally would be better. I personally would counsel channeling attention to helping the younger generations. My generation has saddled many of them with crushing student loan debt and the economy seems to be providing diminished opportunities for the average person to do even as well as their parents. If ever there was a proposal for which the response “Okay, Boomer” is deserved, this is it in my mind.
Postscript: a senior wag might respond to my plea to tax my benefits: “If you feel that way, don’t claim your savings but let me get mine!” Implicitly, I guess, the idea is that’ll eliminate my supposed ethical qualms. Non sequitur responses like that are frequent responses to similar commentaries – e.g., Warren Buffet’s pointing out that lower tax rates apply to his income than to that earned by his secretary often yields comments that he should, then, write a check to the government – and are snarky and off point. This is about what is good tax policy, not how I or anyone else should use our money.
A new report from the Center for Retirement Research at Boston College has some interesting information about Social Security’s actuarial adjustments, confirming some of my suspicions and blowing up others. The report is short (7 pages with graphs and tables taking up much of the space) and worth reading. It lays out the history of allowing SS benefits to be claimed early (i.e., as early as age 62) and providing increased benefits for those who delay retirement (now up until age 70, previously age 72) and the actuarial adjustments that go with those choices.
Key points:
Changes in interest rates and longevity make the SS actuarial adjustments out-of-date and inaccurate. Interest rates are now lower and people are living longer than when the adjustments were set more than three decades or more ago. Both are key factors used to set the adjustments. That is what I suspected.
The downward adjustments for early retirement are too large. This surprised me but makes sense if you think through the math and the finances.
The amount is substantial – about 6% larger than it should be, relative to waiting until the full retirement age or FRA.
By contrast, the delayed retirement credit is not off by much – about 1%. This also surprised me; I had expected it was much bigger than that. Apparently, the 8% annual adjustment was too small when it was set in 1983; 9% would have been closer to accurate. Changes in longevity now make it close to accurate.
These outdated actuarial assumptions benefit the well-off, who tend to be the ones who wait until FRA or further delay retirement. That counters the progressive tilt of the SS benefit formulas.
Some of my quick observations:
The early retirement discount must help the SS trust fund by a not insignificant amount. But it does so at the expense of low-income folks who are least able to afford it and more typically take early retirement, if only because it is harder to work manual labor jobs at older ages or because of their health status.
I thought that it was a no-brainer for someone with reasonably good health to delay claiming benefits as long as possible (but not beyond 70, obviously). I thought that was the case because interest rates are so much lower now than in the early 1980s. But the case for that is not as strong as I thought because, as noted above, the original adjustment was actually too small as confirmed by the then SS actuary.
The recent news about declining longevity, which appears to be disproportionately affecting lower income folks with less education, must exacerbate this situation – again, indirectly undercutting the progressive tilt of the SS benefit structure. That probably justifies providing modest benefit increases for lower income participants, especially those who take early retirement, funded with tax increases or benefit reductions on those with higher incomes. My reaction, anyway.
Since I posted this yesterday, I discovered that the LA Times published an op-ed by Michael Hiltzik on the BC Report on December 3rd with a somewhat more inflammatory take than mine. Here’s a link. I’m not sure about his apparent conclusion that this totally uncuts the progressivity of SS; in fact, I doubt it. But I do think that more sophisticated research needs to be done on the extent to which longevity differences by income and other elements of social status affect the distribution of benefits. The more simplistic research that has traditionally been done is misleading because those effects are not taken into account.
SALT Speak - Joel Michael's Blog 