Minnesota’s state tax collections have consistently exceeded the February MMB forecast amounts. The most recent update reports that revenues for the last quarter (July through September) were up over $200 million or 4.4%. Corporate franchise tax revenues were an exception; they were down $47 million or 9.5% below forecast.
A similar, but longer-term and more striking, trend in federal corporate tax revenues has been occurring. An August 2019 CBO letter on the 2017 federal tax act’s (TCJA’s) effects on federal revenues reports much larger drops in federal corporate tax revenues, compared to forecast (18% for FY2018 and 20% for FY2019). The letter says it cannot attribute that to the effects of TCJA, rather than other causes such as Trump’s ongoing trade wars.
The federal trends are ominous, if you are (like I am) concerned about the federal budget deficit which is running at record levels for an economic expansion. This TPC blog post by Benjamin Page speculates that it may be a sign that TCJA’s big corporate tax cut was even bigger than Treasury and congressional estimators thought it would be. If so, that may also be a bad sign for Minnesota’s corporate tax revenues, since Minnesota conformed to most of TCJA’s corporate and business base expansions, other than its international or foreign provisions.
If the federal estimates of the TCJA are off, Minnesota’s corporate revenues will be adversely affected. Minnesota uses federal taxable income as the starting point for its corporate franchise tax calculations, so dips in federal revenues tend to be mirrored in Minnesota corporate revenues. The relationship is not perfect because the Minnesota economy does not exactly mirror the national economy and Minnesota corporate tax law modestly deviates from the federal calculations. But the relationship is close. Even more important, when Minnesota conformed to most of TCJA’s corporate and business changes, the state used the federal estimates to score its revenue gain that would also result.
Minnesota choose not to conform to TCJA’s foreign provisions – specifically by not taxing GILTI or the deemed repatriation of deferred foreign income. The estimates of those provisions (as discussed below) were probably among the most difficult for federal government economists to estimate. Shouldn’t that immunize the state from the risk those federal estimates are off? On the surface one would think so, but that ain’t necessarily so. Even though the legislature rejected adopting these federal provisions and, thus, the tax bill’s score did not reflect any increase (or decrease) in state tax revenue as a result, there still might be a not insignificant effect on Minnesota’s corporate revenues.
To understand why that could happen, one needs a rudimentary understanding of the federal estimates and how Minnesota uses them in making its own estimates. TCJA dramatically restructured how the federal government taxes the foreign operations of US multinationals, as well as the US operations of foreign businesses. These changes may be the most consequential changes since those made during the Kennedy administration when subpart F imposing federal tax on US multinational’s foreign operations. As with any estimate but especially here because of the scope of the changes, the federal economists needed to forecast two different things:
- The direct effects of changes in law, i.e., the changes in tax base and rates made by the law. Doing this is more or less a matter of complicated arithmetic mainly using data from tax returns. That grossly understates the complexity, since in many cases the necessary data is not on the tax returns and must be inferred or imputed, if it cannot be matched from other records. That requires economic modeling and making assumptions.
- How those law changes will affect the behavior or decisions made by the owners and managers of the multinational businesses. These behavioral changes have two basic components – i.e., the extent to which corporations:
- Artificially shift or recharacterize income to reduce their taxes (these are largely accounting and legal maneuvers that change where or how income is reported and taxed, but not where the corporation’s actual operations are, in response the legal changes); and
- Make substantive operational changes (e.g., by moving factories or other facilities into or out of the US).
TCJA’s changes in the taxation of foreign operations were intended by its architects to reduce the artificial income shifting to low tax foreign jurisdiction, while not affecting actual operations or to encourage multinational businesses to increase their US activities. The effects of both #1 and #2 above are typically embedded in a single number in the federal estimates for each federal change or in some cases for multiple changes.
Making these estimates in an accurate manner is extraordinarily challenging, even with regard to #1 because of the data challenges, but especially #2 because of the unpredictability of how the key deciders will act when there is no precedent for changes in the law. That was the case particularly for TCJA’s foreign changes. Economists, of course, rely on the assumption that individuals and businesses act as profit maximizers in forecasting behavior – i.e., they will do whatever minimizes their taxes and increases their after-tax incomes. But given the complexity of the provisions and the myriad of different business situations, that doesn’t always provide clear guidance. Moreover, having read just a small sample of the analyzes of the TCJA international provisions, there is little consensus among the experts on what their effects will be. Thus, there is a lot of uncertainty about what will happen – there’s a good chance the estimators could be wrong by more than the typical expected error. (Estimates always miss somewhat; one just hopes by not too much.)
Minnesota’s estimators use these federal estimates to score how much adopting or conforming to the federal change will affect Minnesota revenue. This is done by a variety of formulas that convert the federal estimates to approximate the effects of the federal law change on the Minnesota tax base if the state adopts or conforms to them. In most cases, as noted, all the state estimators have is a federal number that reflects both direct computational and behavioral effects. But the behavioral changes – because they are caused by enactment of the federal law – will occur whether or not Minnesota adopts or conforms to the federal change. These changes are automatic and don’t depend on the actions of the legislature to adopt or conform to the changes. Ideally, the Minnesota state government estimators would like to use estimates that reflect only amounts attributable to #1 above, but they typically are stuck with one number (per fiscal year) for the federal changes. In some cases, the federal estimators may not even know what the separate effects are because of the models they use to estimate data parameters and even when they do, they are often reluctant to disclose to state officials what the breakdown is between the two components.
The bottom line, even though Minnesota did not conform to TCJA’s foreign provisions (which would have increased revenues – varying based on how that was done), Minnesota may still experience increased or decreased corporate tax revenues as a result of those provisions. If the behavioral effects of TCJA’s federal changes are to increase or decrease the amount of income reported by US based corporations that will flow through to Minnesota.
So, where does that leave us in judging what effects of the deviations – both federal and Minnesota – from the estimates are? I think it is largely good news, but because it is so preliminary, I would not put much stock in it. The reason I think it may be glimmer of good news is that if the federal estimates are wrong because they misjudged the behavioral effects (what Minnesota picks up without conforming to the federal foreign changes), then Minnesota’s revenues should be off by more than the federal estimates. The exact opposite appears to be the case, suggesting that the big drop in federal revenues must be due to other factors – other parts of the estimates (e.g., the direct computational effects) or the trade war, which are not affecting Minnesota as much as the nation.
As an aside and to provide context, there is yet a third part of the federal estimates – the macro-economic effects of the tax cut, which typically are not done but were for the TCJA because of coercion by the Republicans in Congress. This element of the estimates garnered a lot of media attention and public debate. These are different than the micro economic behavioral assumptions to individual tax parameters, which I described above, and which are routinely made when the federal estimators have a good empirical basis for doing so. These macro-economic feedback or overall effects of tax cuts or increases are very problematic and uncertain. To be reductive simplistic, they essentially must be premised on the idea that government spending, when financed by borrowing, is a smaller drag on economic growth than when financed with taxes – ignoring the details of the tax cut.
The estimates of the revenues from GILTI and the repatriation tax, neither of which Minnesota adopted, may have been too high or the cost of the federal export incentive (foreign derived intangible income) may turned out to be more costly than expected. A big word of caution, though, needs to be expressed: behavioral effects generally take a while to manifest themselves because businesses take time to assess the new law, see how other countries will react, and then develop appropriate strategies. The federal estimates undoubtedly took that into account and behavioral effects in the first couple of years are likely small. As time goes by, the behavioral effects will be picked up by the state’s economic forecasting firm’s macroeconomic forecast, reducing the state’s exposure.
Addendum
After writing this post, I spoke with a former colleague who confirmed (I had heard a rumor) that a drafting error in the conformity bill had unintentionally exempted the cash dividends paid under the repatriation tax. (I personally am not sure whether this should be called a drafting error, an estimating error, or some combination. The Senate bill when I was still working full time – i.e., in 2018 – proposed exempting the cash dividend portion but apparently that was not the intention. This same language was incorporated in the 2019 Senate bill and was included in the final version that became law. The revenue estimates for both 2018 and 2019 Senate bills must not have reflected that intention, but rather only the cost of exempting deemed dividends. Had the estimates reflected both effects, the bill drafts probably would have been clarified.) That error may explain part of the drop in state revenues, which would be even more good news from the perspective that is not the underlying tax base that has declined.
SALT Speak - Joel Michael's Blog 