Choose Your Very Own Tax and Spending Reality

As Election Day approaches and another biennial budget session looms, facts about inflation-adjusted government revenue and spending trends are flying fast and furious.  A look at why these claims deserve a highly jaundiced eye.

Consider the following two statements:

  • Inflation adjusted or “real” per pupil state aid to schools has increased 1.7% since 2003
  • Inflation adjusted or “real” per pupil state aid to schools has decreased 8.5% since 2003

Now pick the statement you like best, tell your friends, and blog about it if you are so inclined. They are both true.

Mark Twain said “facts are stubborn things, but statistics are pliable.” Price indexes have long proven to be some of the best and most reliable statistical tools to shape perception of a set of facts. That is clearly demonstrated by the statements above which are based on data reported by the Minnesota Department of Education. How you perceive state support for K-12 education will be influenced by the discount rate employed on nominal aid growth.

So what’s the best way to assess real revenue and spending trends in a state – the omnipresent, jack-of-all-trades Consumer Price Index (CPI), the specialized government input price index called the Implicit Price Deflator for State and Local Government (IPD), or something else?

A few years ago, Dr. Robert Tannenwald, former Director of the New England Public Policy Center at the Federal Reserve Bank of Boston and former senior fellow at Citizens for Budget and Policy Priorities, offered this perspective to some Massachusetts policy analysts grappling with the same question:

“Ideally the policy analyst needs a precise deflator with the following characteristics: 1) it should be based on a basket of inputs typically purchased by state and local governments within his/her state (emphasis his, not ours) … and 2) the price of each input should reflect the average prevailing throughout his or her state (again, his emphasis). Unfortunately no such deflator is readily available.”

This leads to the common use of a second best measure – the IPD, which is based on national data. He continued:

“Thus the analyst of Massachusetts spending is faced with a difficult choice under uncertainty: a price index with the wrong market basket but disaggregated to geographic areas more characteristic of the state than the nation as a whole… or a nationwide index with a more appropriate market basket.”

Dr. Tannenwald’s commentary is spot-on except for one thing: for a lot of analysts and editorialists this is not a difficult choice at all. The right measure is the one that spins the desired political narrative.

The advantage of the IPD is that it is a weighted price index based on the things that government actually purchases. At first glance this looks like an absolute slam dunk – and there are plenty of policy commentators around who will tell you just that. However, there are three reasons why the IPD has its own conceptual baggage, and why selecting an appropriate index to deflate government spending remains a “difficult choice” despite insistence to the contrary.

Differences in government price input changes between states are not only significant but likely to increase in future years.

Our recent blog post highlighted several large state efforts to solve their pension crises and discussed the impact these types of efforts will have on the IPD going forward. We also showed that after decades more or less tracking with consumer inflation, government inflation over the past decade or so featured several spikes causing the CPI and IPD trend lines to diverge. The table below suggests where investigations into causes should focus:

Relative Contributions to the Growth in IPD for State and Local Government: 1999-2013

 Fixed Capital6.5%
 Durable Goods0.0%
 Nondurable Goods16.6%
 Own-Account Investment/Sales to Other Sectors-21.6%
 Investment in Structures19.3%
 Investment in Equipment & Intellectual Property-0.6%


Since 1999 employee compensation had over three times the net impact on the growth in the IPD than investment in structures and purchases of professional services, police cars, fuel, and all other goods commonly cited as unique and distinguishing characteristics of government purchasing activity. The point is this: by using the IPD, perceptions about things like the adequacy of our own state support for K-12 education are being and will continue to be affected by the fiscal malpractice of states like California. Which points to the second reason:

Government has the unique ability to manage and exercise significant control over the price trends of its primary input.

Cost trends in wages, employee health care, pensions, health care benefits for retirees, and other forms of compensation aren’t uncontrollable factors imposed on government by the market. Rather, they are heavily influenced by state and local policy choices and often directly established by governments themselves through collective bargaining agreements. Which suggests a third reason:

The IPD is highly self-referential.

By using an input price index with the cost of oil as a discount rate it would be easy to demonstrate that the “real” cost of gasoline today is near an all-time low. As preposterous as that notion is, that same sense of circularity is present in using an internal government input price index to communicate inflation-adjusted spending or revenues. Governments’ own actions can trigger higher rates of inflation that in turn are used to assert and communicate the continuing inadequacy of governments’ own resources.

That is why many economists see internal budgeting, forecasting, or project management as the preferred and appropriate application of input price indexes. (Although some would likely express reservations about applying a national inflation measure even for this use for the geographic reasons cited above and on the grounds that a mechanical calculation of the cost of maintaining the status quo does not legitimize the status quo.)

None of this will change reporting on government fiscal policy. But as we approach another biennial budget year when revenue and spending facts start flying fast and furious, we do suggest any claims about inflation-adjusted or ”real” revenue and spending trends deserve to be viewed with a jaundiced eye and the knowledge that there is a lot more to the story.

You are receiving a perspective on reality, not reality itself.